Bump

New Relay Methods + Conntrack Control + Cleanup Methods for Memory + Buffer Pool Trim + Shrink Session Vec + ME2DC Fast for unstoppable init + Config Fallback + Working Directory Setup + Logging fixes with --syslog

.github/codeql/codeql-config.yml

@@ -7,7 +7,16 @@ queries:
   - uses: security-and-quality
   - uses: ./.github/codeql/queries
 
+paths-ignore:
+  - "**/tests/**"
+  - "**/test/**"
+  - "**/*_test.rs"
+  - "**/*/tests.rs"
 query-filters:
+  - exclude:
+      tags:
+        - test
+
   - exclude:
       id:
         - rust/unwrap-on-option

.github/instructions/Architecture.instructions.md

@@ -0,0 +1,126 @@
+# Architecture Directives
+
+> Companion to `Agents.md`. These are **activation directives**, not tutorials.
+> You already know these patterns — apply them. When making any structural or
+> design decision, run the relevant section below as a checklist.
+
+---
+
+## 1. Active Principles (always on)
+
+Apply these on every non-trivial change. No exceptions.
+
+- **SRP** — one reason to change per component. If you can't name the responsibility in one noun phrase, split it.
+- **OCP** — extend by adding, not by modifying. New variants/impls over patching existing logic.
+- **ISP** — traits stay minimal. More than ~5 methods is a split signal.
+- **DIP** — high-level modules depend on traits, not concrete types. Infrastructure implements domain traits; it does not own domain logic.
+- **DRY** — one authoritative source per piece of knowledge. Copies are bugs that haven't diverged yet.
+- **YAGNI** — generic parameters, extension hooks, and pluggable strategies require an *existing* concrete use case, not a hypothetical one.
+- **KISS** — two equivalent designs: choose the one with fewer concepts. Justify complexity; never assume it.
+
+---
+
+## 2. Layered Architecture
+
+Dependencies point **inward only**: `Presentation → Application → Domain ← Infrastructure`.
+
+- Domain layer: zero I/O. No network, no filesystem, no async runtime imports.
+- Infrastructure: implements domain traits at the boundary. Never leaks SDK/wire types inward.
+- Anti-Corruption Layer (ACL): all third-party and external-protocol types are translated here. If the external format changes, only the ACL changes.
+- Presentation: translates wire/HTTP representations to domain types and back. Nothing else.
+
+---
+
+## 3. Design Pattern Selection
+
+Apply the right pattern. Do not invent a new abstraction when a named pattern fits.
+
+| Situation | Pattern to apply |
+|---|---|
+| Struct with 3+ optional/dependent fields | **Builder** — `build()` returns `Result`, never panics |
+| Cross-cutting behavior (logging, retry, metrics) on a trait impl | **Decorator** — implements same trait, delegates all calls |
+| Subsystem with multiple internal components | **Façade** — single public entry point, internals are `pub(crate)` |
+| Swappable algorithm or policy | **Strategy** — trait injection; generics for compile-time, `dyn` for runtime |
+| Component notifying decoupled consumers | **Observer** — typed channels (`broadcast`, `watch`), not callback `Vec<Box<dyn Fn>>` |
+| Exclusive mutable state serving concurrent callers | **Actor** — `mpsc` command channel + `oneshot` reply; no lock needed on state |
+| Finite state with invalid transition prevention | **Typestate** — distinct types per state; invalid ops are compile errors |
+| Fixed process skeleton with overridable steps | **Template Method** — defaulted trait method calls required hooks |
+| Request pipeline with independent handlers | **Chain/Middleware** — generic compile-time chain for hot paths, `dyn` for runtime assembly |
+| Hiding a concrete type behind a trait | **Factory Function** — returns `Box<dyn Trait>` or `impl Trait` |
+
+---
+
+## 4. Data Modeling Rules
+
+- **Make illegal states unrepresentable.** Type system enforces invariants; runtime validation is a second line, not the first.
+- **Newtype every primitive** that carries domain meaning. `SessionId(u64)` ≠ `UserId(u64)` — the compiler enforces it.
+- **Enums over booleans** for any parameter or field with two or more named states.
+- **Typed error enums** with named variants carrying full diagnostic context. `anyhow` is application-layer only; never in library code.
+- **Domain types carry no I/O concerns.** No `serde`, no codec, no DB derives on domain structs. Conversions via `From`/`TryFrom` at layer boundaries.
+
+---
+
+## 5. Concurrency Rules
+
+- Prefer message-passing over shared memory. Shared state is a fallback.
+- All channels must be **bounded**. Document the bound's rationale inline.
+- Never hold a lock across an `await` unless atomicity explicitly requires it — document why.
+- Document lock acquisition order wherever two locks are taken together.
+- Every `async fn` is cancellation-safe unless explicitly documented otherwise. Mutate shared state *after* the `await` that may be cancelled, not before.
+- High-read/low-write state: use `arc-swap` or `watch` for lock-free reads.
+
+---
+
+## 6. Error Handling Rules
+
+- Errors translated at every layer boundary — low-level errors never surface unmodified.
+- Add context at the propagation site: what operation failed and where.
+- No `unwrap()`/`expect()` in production paths without a comment proving `None`/`Err` is impossible.
+- Panics are only permitted in: tests, startup/init unrecoverable failure, and `unreachable!()` with an invariant comment.
+
+---
+
+## 7. API Design Rules
+
+- **CQS**: functions that return data must not mutate; functions that mutate return only `Result`.
+- **Least surprise**: a function does exactly what its name implies. Side effects are documented.
+- **Idempotency**: `close()`, `shutdown()`, `unregister()` called twice must not panic or error.
+- **Fallibility at the type level**: failure → `Result<T, E>`. No sentinel values.
+- **Minimal public surface**: default to `pub(crate)`. Mark `pub` only deliberate API. Re-export through a single surface in `mod.rs`.
+
+---
+
+## 8. Performance Rules (hot paths)
+
+- Annotate hot-path functions with `// HOT PATH: <throughput requirement>`.
+- Zero allocations per operation in hot paths after initialization. Preallocate in constructors, reuse buffers.
+- Pass `&[u8]` / `Bytes` slices — not `Vec<u8>`. Use `BytesMut` for reusable mutable buffers.
+- No `String` formatting in hot paths. No logging without a rate-limit or sampling gate.
+- Any allocation in a hot path gets a comment: `// ALLOC: <reason and size>`.
+
+---
+
+## 9. Testing Rules
+
+- Bug fixes require a regression test that is **red before the fix, green after**. Name it after the bug.
+- Property tests for: codec round-trips, state machine invariants, cryptographic protocol correctness.
+- No shared mutable state between tests. Each test constructs its own environment.
+- Test doubles hierarchy (simplest first): Fake → Stub → Spy → Mock. Mocks couple to implementation, not behavior — use sparingly.
+
+---
+
+## 10. Pre-Change Checklist
+
+Run this before proposing or implementing any structural decision:
+
+- [ ] Responsibility nameable in one noun phrase?
+- [ ] Layer dependencies point inward only?
+- [ ] Invalid states unrepresentable in the type system?
+- [ ] State transitions gated through a single interface?
+- [ ] All channels bounded?
+- [ ] No locks held across `await` (or documented)?
+- [ ] Errors typed and translated at layer boundaries?
+- [ ] No panics in production paths without invariant proof?
+- [ ] Hot paths annotated and allocation-free?
+- [ ] Public surface minimal — only deliberate API marked `pub`?
+- [ ] Correct pattern chosen from Section 3 table?

.github/workflows/build.yml

@@ -0,0 +1,39 @@
+name: Build
+
+on:
+  push:
+    branches: [ "*" ]
+  pull_request:
+    branches: [ "*" ]
+
+env:
+  CARGO_TERM_COLOR: always
+
+jobs:
+  build:
+    name: Build
+    runs-on: ubuntu-latest
+
+    permissions:
+      contents: read
+
+    steps:
+      - name: Checkout repository
+        uses: actions/checkout@v4
+
+      - name: Install latest stable Rust toolchain
+        uses: dtolnay/rust-toolchain@stable
+
+      - name: Cache cargo registry & build artifacts
+        uses: actions/cache@v4
+        with:
+          path: |
+            ~/.cargo/registry
+            ~/.cargo/git
+            target
+          key: ${{ runner.os }}-cargo-${{ hashFiles('**/Cargo.lock') }}
+          restore-keys: |
+            ${{ runner.os }}-cargo-
+
+      - name: Build Release
+        run: cargo build --release --verbose

.github/workflows/release.yml

@@ -5,35 +5,87 @@ on:
     tags:
       - '[0-9]+.[0-9]+.[0-9]+'
   workflow_dispatch:
+    inputs:
+      tag:
+        description: 'Release tag (example: 3.3.15)'
+        required: true
+        type: string
 
 concurrency:
-  group: release-${{ github.ref }}
+  group: release-${{ github.ref_name }}-${{ github.event.inputs.tag || 'auto' }}
   cancel-in-progress: true
 
 permissions:
   contents: read
-  packages: write
 
 env:
   CARGO_TERM_COLOR: always
   BINARY_NAME: telemt
 
 jobs:
-# ==========================
-# GNU / glibc
-# ==========================
-  build-gnu:
-    name: GNU ${{ matrix.target }}
+  prepare:
+    name: Prepare
     runs-on: ubuntu-latest
 
+    outputs:
+      version: ${{ steps.vars.outputs.version }}
+      prerelease: ${{ steps.vars.outputs.prerelease }}
+
+    steps:
+      - name: Resolve version
+        id: vars
+        shell: bash
+        run: |
+          set -euo pipefail
+
+          if [ "${GITHUB_EVENT_NAME}" = "workflow_dispatch" ]; then
+            VERSION="${{ github.event.inputs.tag }}"
+          else
+            VERSION="${GITHUB_REF#refs/tags/}"
+          fi
+
+          VERSION="${VERSION#refs/tags/}"
+
+          if [ -z "${VERSION}" ]; then
+            echo "Release version is empty" >&2
+            exit 1
+          fi
+
+          if [[ "${VERSION}" == *-* ]]; then
+            PRERELEASE=true
+          else
+            PRERELEASE=false
+          fi
+
+          echo "version=${VERSION}" >> "${GITHUB_OUTPUT}"
+          echo "prerelease=${PRERELEASE}" >> "${GITHUB_OUTPUT}"
+
+  # ==========================
+  # GNU / glibc
+  # ==========================
+  build-gnu:
+    name: GNU ${{ matrix.asset }}
+    runs-on: ubuntu-latest
+    needs: prepare
+
+    container:
+      image: rust:slim-bookworm
+
     strategy:
       fail-fast: false
       matrix:
         include:
           - target: x86_64-unknown-linux-gnu
             asset: telemt-x86_64-linux-gnu
+            cpu: baseline
+
+          - target: x86_64-unknown-linux-gnu
+            asset: telemt-x86_64-v3-linux-gnu
+            cpu: v3
+
           - target: aarch64-unknown-linux-gnu
             asset: telemt-aarch64-linux-gnu
+            cpu: generic
 
     steps:
       - uses: actions/checkout@v4
@@ -47,8 +99,8 @@ jobs:
 
       - name: Install deps
         run: |
-          sudo apt-get update
-          sudo apt-get install -y \
+          apt-get update
+          apt-get install -y \
             build-essential \
             clang \
             lld \
@@ -59,53 +111,65 @@ jobs:
       - uses: actions/cache@v4
         with:
           path: |
-            ~/.cargo/registry
-            ~/.cargo/git
+            /usr/local/cargo/registry
+            /usr/local/cargo/git
             target
-          key: gnu-${{ matrix.target }}-${{ hashFiles('**/Cargo.lock') }}
+          key: gnu-${{ matrix.asset }}-${{ hashFiles('**/Cargo.lock') }}
+          restore-keys: |
+            gnu-${{ matrix.asset }}-
+            gnu-
 
       - name: Build
+        shell: bash
         run: |
+          set -euo pipefail
+
           if [ "${{ matrix.target }}" = "aarch64-unknown-linux-gnu" ]; then
             export CC=aarch64-linux-gnu-gcc
             export CXX=aarch64-linux-gnu-g++
-            export CC_aarch64_unknown_linux_gnu=aarch64-linux-gnu-gcc
-            export CXX_aarch64_unknown_linux_gnu=aarch64-linux-gnu-g++
-            export RUSTFLAGS="-C linker=aarch64-linux-gnu-gcc"
+            export RUSTFLAGS="-C linker=aarch64-linux-gnu-gcc -C lto=fat -C panic=abort"
           else
             export CC=clang
             export CXX=clang++
-            export CC_x86_64_unknown_linux_gnu=clang
-            export CXX_x86_64_unknown_linux_gnu=clang++
-            export RUSTFLAGS="-C linker=clang -C link-arg=-fuse-ld=lld"
+
+            if [ "${{ matrix.cpu }}" = "v3" ]; then
+              CPU_FLAGS="-C target-cpu=x86-64-v3"
+            else
+              CPU_FLAGS="-C target-cpu=x86-64"
+            fi
+
+            export RUSTFLAGS="-C linker=clang -C link-arg=-fuse-ld=lld -C lto=fat -C panic=abort ${CPU_FLAGS}"
           fi
 
-          cargo build --release --target ${{ matrix.target }}
+          cargo build --release --target ${{ matrix.target }} -j "$(nproc)"
 
       - name: Package
+        shell: bash
         run: |
-          mkdir -p dist
-          BIN=target/${{ matrix.target }}/release/${{ env.BINARY_NAME }}
+          set -euo pipefail
 
-          cp "$BIN" dist/${{ env.BINARY_NAME }}-${{ matrix.target }}
+          mkdir -p dist
+          cp "target/${{ matrix.target }}/release/${{ env.BINARY_NAME }}" dist/telemt
 
           cd dist
-          tar -czf ${{ matrix.asset }}.tar.gz ${{ env.BINARY_NAME }}-${{ matrix.target }}
-          sha256sum ${{ matrix.asset }}.tar.gz > ${{ matrix.asset }}.sha256
+          tar -czf "${{ matrix.asset }}.tar.gz" \
+            --owner=0 --group=0 --numeric-owner \
+            telemt
+
+          sha256sum "${{ matrix.asset }}.tar.gz" > "${{ matrix.asset }}.tar.gz.sha256"
 
       - uses: actions/upload-artifact@v4
         with:
           name: ${{ matrix.asset }}
-          path: |
-            dist/${{ matrix.asset }}.tar.gz
-            dist/${{ matrix.asset }}.sha256
+          path: dist/*
 
-# ==========================
-# MUSL
-# ==========================
+  # ==========================
+  # MUSL
+  # ==========================
   build-musl:
-    name: MUSL ${{ matrix.target }}
+    name: MUSL ${{ matrix.asset }}
     runs-on: ubuntu-latest
+    needs: prepare
 
     container:
       image: rust:slim-bookworm
@@ -116,8 +180,15 @@ jobs:
         include:
           - target: x86_64-unknown-linux-musl
             asset: telemt-x86_64-linux-musl
+            cpu: baseline
+
+          - target: x86_64-unknown-linux-musl
+            asset: telemt-x86_64-v3-linux-musl
+            cpu: v3
+
           - target: aarch64-unknown-linux-musl
             asset: telemt-aarch64-linux-musl
+            cpu: generic
 
     steps:
       - uses: actions/checkout@v4
@@ -138,30 +209,29 @@ jobs:
 
       - name: Install aarch64 musl toolchain
         if: matrix.target == 'aarch64-unknown-linux-musl'
+        shell: bash
         run: |
-          set -e
+          set -euo pipefail
 
           TOOLCHAIN_DIR="$HOME/.musl-aarch64"
           ARCHIVE="aarch64-linux-musl-cross.tgz"
-          URL="https://github.com/telemt/telemt/releases/download/toolchains/$ARCHIVE"
+          URL="https://github.com/telemt/telemt/releases/download/toolchains/${ARCHIVE}"
 
-          if [ -x "$TOOLCHAIN_DIR/bin/aarch64-linux-musl-gcc" ]; then
-            echo "✅ MUSL toolchain already installed"
+          if [ -x "${TOOLCHAIN_DIR}/bin/aarch64-linux-musl-gcc" ]; then
+            echo "MUSL toolchain cached"
           else
-            echo "⬇️ Downloading musl toolchain from Telemt GitHub Releases..."
-
             curl -fL \
               --retry 5 \
               --retry-delay 3 \
               --connect-timeout 10 \
               --max-time 120 \
-              -o "$ARCHIVE" "$URL"
+              -o "${ARCHIVE}" "${URL}"
 
-            mkdir -p "$TOOLCHAIN_DIR"
-            tar -xzf "$ARCHIVE" --strip-components=1 -C "$TOOLCHAIN_DIR"
+            mkdir -p "${TOOLCHAIN_DIR}"
+            tar -xzf "${ARCHIVE}" --strip-components=1 -C "${TOOLCHAIN_DIR}"
           fi
 
-          echo "$TOOLCHAIN_DIR/bin" >> $GITHUB_PATH
+          echo "${TOOLCHAIN_DIR}/bin" >> "${GITHUB_PATH}"
 
       - name: Add rust target
         run: rustup target add ${{ matrix.target }}
@@ -172,96 +242,62 @@ jobs:
             /usr/local/cargo/registry
             /usr/local/cargo/git
             target
-          key: musl-${{ matrix.target }}-${{ hashFiles('**/Cargo.lock') }}
+          key: musl-${{ matrix.asset }}-${{ hashFiles('**/Cargo.lock') }}
+          restore-keys: |
+            musl-${{ matrix.asset }}-
+            musl-
 
       - name: Build
+        shell: bash
         run: |
+          set -euo pipefail
+
           if [ "${{ matrix.target }}" = "aarch64-unknown-linux-musl" ]; then
             export CC=aarch64-linux-musl-gcc
             export CC_aarch64_unknown_linux_musl=aarch64-linux-musl-gcc
-            export RUSTFLAGS="-C target-feature=+crt-static -C linker=aarch64-linux-musl-gcc"
+            export RUSTFLAGS="-C target-feature=+crt-static -C linker=aarch64-linux-musl-gcc -C lto=fat -C panic=abort"
           else
             export CC=musl-gcc
             export CC_x86_64_unknown_linux_musl=musl-gcc
-            export RUSTFLAGS="-C target-feature=+crt-static"
+
+            if [ "${{ matrix.cpu }}" = "v3" ]; then
+              CPU_FLAGS="-C target-cpu=x86-64-v3"
+            else
+              CPU_FLAGS="-C target-cpu=x86-64"
+            fi
+
+            export RUSTFLAGS="-C target-feature=+crt-static -C lto=fat -C panic=abort ${CPU_FLAGS}"
           fi
 
-          cargo build --release --target ${{ matrix.target }}
+          cargo build --release --target ${{ matrix.target }} -j "$(nproc)"
 
       - name: Package
+        shell: bash
         run: |
-          mkdir -p dist
-          BIN=target/${{ matrix.target }}/release/${{ env.BINARY_NAME }}
+          set -euo pipefail
 
-          cp "$BIN" dist/${{ env.BINARY_NAME }}-${{ matrix.target }}
+          mkdir -p dist
+          cp "target/${{ matrix.target }}/release/${{ env.BINARY_NAME }}" dist/telemt
 
           cd dist
-          tar -czf ${{ matrix.asset }}.tar.gz ${{ env.BINARY_NAME }}-${{ matrix.target }}
-          sha256sum ${{ matrix.asset }}.tar.gz > ${{ matrix.asset }}.sha256
+          tar -czf "${{ matrix.asset }}.tar.gz" \
+            --owner=0 --group=0 --numeric-owner \
+            telemt
+
+          sha256sum "${{ matrix.asset }}.tar.gz" > "${{ matrix.asset }}.tar.gz.sha256"
 
       - uses: actions/upload-artifact@v4
         with:
           name: ${{ matrix.asset }}
-          path: |
-            dist/${{ matrix.asset }}.tar.gz
-            dist/${{ matrix.asset }}.sha256
+          path: dist/*
 
-# ==========================
-# Docker
-# ==========================
-  docker:
-    name: Docker
-    runs-on: ubuntu-latest
-    needs: [build-gnu, build-musl]
-    continue-on-error: true
-
-    steps:
-      - uses: actions/checkout@v4
-
-      - uses: actions/download-artifact@v4
-        with:
-          path: artifacts
-
-      - name: Extract binaries
-        run: |
-          mkdir dist
-          find artifacts -name "*.tar.gz" -exec tar -xzf {} -C dist \;
-
-          cp dist/telemt-x86_64-unknown-linux-musl dist/telemt || true
-
-      - uses: docker/setup-qemu-action@v3
-      - uses: docker/setup-buildx-action@v3
-
-      - name: Login to GHCR
-        uses: docker/login-action@v3
-        with:
-          registry: ghcr.io
-          username: ${{ github.actor }}
-          password: ${{ secrets.GITHUB_TOKEN }}
-
-      - name: Extract version
-        id: vars
-        run: echo "VERSION=${GITHUB_REF#refs/tags/}" >> $GITHUB_OUTPUT
-
-      - name: Build & Push
-        uses: docker/build-push-action@v6
-        with:
-          context: .
-          push: true
-          platforms: linux/amd64,linux/arm64
-          tags: |
-            ghcr.io/${{ github.repository }}:${{ steps.vars.outputs.VERSION }}
-            ghcr.io/${{ github.repository }}:latest
-          build-args: |
-            BINARY=dist/telemt
-
-# ==========================
-# Release
-# ==========================
+  # ==========================
+  # Release
+  # ==========================
   release:
     name: Release
     runs-on: ubuntu-latest
-    needs: [build-gnu, build-musl]
+    needs: [prepare, build-gnu, build-musl]
 
     permissions:
       contents: write
@@ -272,14 +308,98 @@ jobs:
           path: artifacts
 
       - name: Flatten artifacts
+        shell: bash
         run: |
-          mkdir dist
+          set -euo pipefail
+          mkdir -p dist
           find artifacts -type f -exec cp {} dist/ \;
 
-      - name: Create Release
+      - name: Create GitHub Release
         uses: softprops/action-gh-release@v2
         with:
+          tag_name: ${{ needs.prepare.outputs.version }}
+          target_commitish: ${{ github.sha }}
           files: dist/*
           generate_release_notes: true
-          draft: false
-          prerelease: ${{ contains(github.ref, '-rc') || contains(github.ref, '-beta') || contains(github.ref, '-alpha') }}
+          prerelease: ${{ needs.prepare.outputs.prerelease == 'true' }}
+          overwrite_files: true
+
+  # ==========================
+  # Docker
+  # ==========================
+  docker:
+    name: Docker
+    runs-on: ubuntu-latest
+    needs: [prepare, release]
+
+    permissions:
+      contents: read
+      packages: write
+
+    steps:
+      - uses: actions/checkout@v4
+
+      - uses: docker/setup-qemu-action@v3
+
+      - uses: docker/setup-buildx-action@v3
+
+      - uses: docker/login-action@v3
+        with:
+          registry: ghcr.io
+          username: ${{ github.actor }}
+          password: ${{ secrets.GITHUB_TOKEN }}
+
+      - name: Probe release assets
+        shell: bash
+        env:
+          VERSION: ${{ needs.prepare.outputs.version }}
+        run: |
+          set -euo pipefail
+
+          for asset in \
+            telemt-x86_64-linux-musl.tar.gz \
+            telemt-x86_64-linux-musl.tar.gz.sha256 \
+            telemt-aarch64-linux-musl.tar.gz \
+            telemt-aarch64-linux-musl.tar.gz.sha256
+          do
+            curl -fsIL \
+              --retry 10 \
+              --retry-delay 3 \
+              "https://github.com/${GITHUB_REPOSITORY}/releases/download/${VERSION}/${asset}" \
+              > /dev/null
+          done
+
+      - name: Compute image tags
+        id: meta
+        shell: bash
+        env:
+          VERSION: ${{ needs.prepare.outputs.version }}
+        run: |
+          set -euo pipefail
+
+          IMAGE="$(echo "ghcr.io/${GITHUB_REPOSITORY}" | tr '[:upper:]' '[:lower:]')"
+          TAGS="${IMAGE}:${VERSION}"
+
+          if [[ "${VERSION}" != *-* ]]; then
+            TAGS="${TAGS}"$'\n'"${IMAGE}:latest"
+          fi
+
+          {
+            echo "tags<<EOF"
+            printf '%s\n' "${TAGS}"
+            echo "EOF"
+          } >> "${GITHUB_OUTPUT}"
+
+      - name: Build & Push
+        uses: docker/build-push-action@v6
+        with:
+          context: .
+          push: true
+          pull: true
+          platforms: linux/amd64,linux/arm64
+          tags: ${{ steps.meta.outputs.tags }}
+          build-args: |
+            TELEMT_REPOSITORY=${{ github.repository }}
+            TELEMT_VERSION=${{ needs.prepare.outputs.version }}
+          cache-from: type=gha
+          cache-to: type=gha,mode=max

.github/workflows/rust.yml

@@ -1,66 +0,0 @@
-name: Rust
-
-on:
-  push:
-    branches: [ "*" ]
-  pull_request:
-    branches: [ "*" ]
-
-env:
-  CARGO_TERM_COLOR: always
-
-jobs:
-  build:
-    name: Build
-    runs-on: ubuntu-latest
-
-    permissions:
-      contents: read
-      actions: write
-      checks: write
-
-    steps:
-    - name: Checkout repository
-      uses: actions/checkout@v4
-
-    - name: Install latest stable Rust toolchain
-      uses: dtolnay/rust-toolchain@stable
-      with:
-        components: rustfmt, clippy
-
-    - name: Cache cargo registry & build artifacts
-      uses: actions/cache@v4
-      with:
-        path: |
-          ~/.cargo/registry
-          ~/.cargo/git
-          target
-        key: ${{ runner.os }}-cargo-${{ hashFiles('**/Cargo.lock') }}
-        restore-keys: |
-          ${{ runner.os }}-cargo-
-
-    - name: Build Release
-      run: cargo build --release --verbose
-
-    - name: Run tests
-      run: cargo test --verbose
-
-    - name: Stress quota-lock suites (PR only)
-      if: github.event_name == 'pull_request'
-      env:
-        RUST_TEST_THREADS: 16
-      run: |
-        set -euo pipefail
-        for i in $(seq 1 12); do
-          echo "[quota-lock-stress] iteration ${i}/12"
-          cargo test quota_lock_ --bin telemt -- --nocapture --test-threads 16
-          cargo test relay_quota_wake --bin telemt -- --nocapture --test-threads 16
-        done
-
-# clippy dont fail on warnings because of active development of telemt 
-# and many warnings
-    - name: Run clippy
-      run: cargo clippy -- --cap-lints warn
-
-    - name: Check for unused dependencies
-      run: cargo udeps || true

.github/workflows/test.yml

@@ -0,0 +1,139 @@
+name: Check
+
+on:
+  push:
+    branches: [ "*" ]
+  pull_request:
+    branches: [ "*" ]
+
+env:
+  CARGO_TERM_COLOR: always
+
+concurrency:
+  group: test-${{ github.ref }}
+  cancel-in-progress: true
+
+jobs:
+# ==========================
+# Formatting
+# ==========================
+  fmt:
+    name: Fmt
+    runs-on: ubuntu-latest
+
+    permissions:
+      contents: read
+
+    steps:
+      - uses: actions/checkout@v4
+
+      - uses: dtolnay/rust-toolchain@stable
+        with:
+          components: rustfmt
+
+      - run: cargo fmt -- --check
+
+# ==========================
+# Tests
+# ==========================
+  test:
+    name: Test
+    runs-on: ubuntu-latest
+
+    permissions:
+      contents: read
+      actions: write
+      checks: write
+
+    steps:
+      - uses: actions/checkout@v4
+
+      - uses: dtolnay/rust-toolchain@stable
+
+      - name: Cache cargo
+        uses: actions/cache@v4
+        with:
+          path: |
+            ~/.cargo/bin
+            ~/.cargo/registry
+            ~/.cargo/git
+            target
+          key: ${{ runner.os }}-cargo-nextest-${{ hashFiles('**/Cargo.lock') }}
+          restore-keys: |
+            ${{ runner.os }}-cargo-nextest-
+            ${{ runner.os }}-cargo-
+
+      - name: Install cargo-nextest
+        run: cargo install --locked cargo-nextest || true
+
+      - name: Run tests with nextest
+        run: cargo nextest run -j "$(nproc)"
+
+# ==========================
+# Clippy
+# ==========================
+  clippy:
+    name: Clippy
+    runs-on: ubuntu-latest
+
+    permissions:
+      contents: read
+      checks: write
+
+    steps:
+      - uses: actions/checkout@v4
+
+      - uses: dtolnay/rust-toolchain@stable
+        with:
+          components: clippy
+
+      - name: Cache cargo
+        uses: actions/cache@v4
+        with:
+          path: |
+            ~/.cargo/registry
+            ~/.cargo/git
+            target
+          key: ${{ runner.os }}-cargo-clippy-${{ hashFiles('**/Cargo.lock') }}
+          restore-keys: |
+            ${{ runner.os }}-cargo-clippy-
+            ${{ runner.os }}-cargo-
+
+      - name: Run clippy
+        run: cargo clippy -j "$(nproc)" -- --cap-lints warn
+
+# ==========================
+# Udeps
+# ==========================
+  udeps:
+    name: Udeps
+    runs-on: ubuntu-latest
+
+    permissions:
+      contents: read
+
+    steps:
+      - uses: actions/checkout@v4
+
+      - uses: dtolnay/rust-toolchain@stable
+        with:
+          components: rust-src
+
+      - name: Cache cargo
+        uses: actions/cache@v4
+        with:
+          path: |
+            ~/.cargo/bin
+            ~/.cargo/registry
+            ~/.cargo/git
+            target
+          key: ${{ runner.os }}-cargo-udeps-${{ hashFiles('**/Cargo.lock') }}
+          restore-keys: |
+            ${{ runner.os }}-cargo-udeps-
+            ${{ runner.os }}-cargo-
+
+      - name: Install cargo-udeps
+        run: cargo install --locked cargo-udeps || true
+
+      - name: Run udeps
+        run: cargo udeps -j "$(nproc)" || true

.kilocode/rules-architect/AGENTS.md

@@ -1,58 +0,0 @@
-# Architect Mode Rules for Telemt
-
-## Architecture Overview
-
-```mermaid
-graph TB
-    subgraph Entry
-        Client[Clients] --> Listener[TCP/Unix Listener]
-    end
-    
-    subgraph Proxy Layer
-        Listener --> ClientHandler[ClientHandler]
-        ClientHandler --> Handshake[Handshake Validator]
-        Handshake --> |Valid| Relay[Relay Layer]
-        Handshake --> |Invalid| Masking[Masking/TLS Fronting]
-    end
-    
-    subgraph Transport
-        Relay --> MiddleProxy[Middle-End Proxy Pool]
-        Relay --> DirectRelay[Direct DC Relay]
-        MiddleProxy --> TelegramDC[Telegram DCs]
-        DirectRelay --> TelegramDC
-    end
-```
-
-## Module Dependencies
-- [`src/main.rs`](src/main.rs) - Entry point, spawns all async tasks
-- [`src/config/`](src/config/) - Configuration loading with auto-migration
-- [`src/error.rs`](src/error.rs) - Error types, must be used by all modules
-- [`src/crypto/`](src/crypto/) - AES, SHA, random number generation
-- [`src/protocol/`](src/protocol/) - MTProto constants, frame encoding, obfuscation
-- [`src/stream/`](src/stream/) - Stream wrappers, buffer pool, frame codecs
-- [`src/proxy/`](src/proxy/) - Client handling, handshake, relay logic
-- [`src/transport/`](src/transport/) - Upstream management, middle-proxy, SOCKS support
-- [`src/stats/`](src/stats/) - Statistics and replay protection
-- [`src/ip_tracker.rs`](src/ip_tracker.rs) - Per-user IP tracking
-
-## Key Architectural Constraints
-
-### Middle-End Proxy Mode
-- Requires public IP on interface OR 1:1 NAT with STUN probing
-- Uses separate `proxy-secret` from Telegram (NOT user secrets)
-- Falls back to direct mode automatically on STUN mismatch
-
-### TLS Fronting
-- Invalid handshakes are transparently proxied to `mask_host`
-- This is critical for DPI evasion - do not change this behavior
-- `mask_unix_sock` and `mask_host` are mutually exclusive
-
-### Stream Architecture
-- Buffer pool is shared globally via Arc - prevents allocation storms
-- Frame codecs implement tokio-util Encoder/Decoder traits
-- State machine in [`src/stream/state.rs`](src/stream/state.rs) manages stream transitions
-
-### Configuration Migration
-- [`ProxyConfig::load()`](src/config/mod.rs:641) mutates config in-place
-- New fields must have sensible defaults
-- DC203 override is auto-injected for CDN/media support

.kilocode/rules-code/AGENTS.md

@@ -1,23 +0,0 @@
-# Code Mode Rules for Telemt
-
-## Error Handling
-- Always use [`ProxyError`](src/error.rs:168) from [`src/error.rs`](src/error.rs) for proxy operations
-- [`HandshakeResult<T,R,W>`](src/error.rs:292) returns streams on bad client - these MUST be returned for masking, never dropped
-- Use [`Recoverable`](src/error.rs:110) trait to check if errors are retryable
-
-## Configuration Changes
-- [`ProxyConfig::load()`](src/config/mod.rs:641) auto-mutates config - new fields should have defaults
-- DC203 override is auto-injected if missing - do not remove this behavior
-- When adding config fields, add migration logic in [`ProxyConfig::load()`](src/config/mod.rs:641)
-
-## Crypto Code
-- [`SecureRandom`](src/crypto/random.rs) from [`src/crypto/random.rs`](src/crypto/random.rs) must be used for all crypto operations
-- Never use `rand::thread_rng()` directly - use the shared `Arc<SecureRandom>`
-
-## Stream Handling
-- Buffer pool [`BufferPool`](src/stream/buffer_pool.rs) is shared via Arc - always use it instead of allocating
-- Frame codecs in [`src/stream/frame_codec.rs`](src/stream/frame_codec.rs) implement tokio-util's Encoder/Decoder traits
-
-## Testing
-- Tests are inline in modules using `#[cfg(test)]`
-- Use `cargo test --lib <module_name>` to run tests for specific modules

.kilocode/rules-debug/AGENTS.md

@@ -1,27 +0,0 @@
-# Debug Mode Rules for Telemt
-
-## Logging
-- `RUST_LOG` environment variable takes absolute priority over all config log levels
-- Log levels: `trace`, `debug`, `info`, `warn`, `error`
-- Use `RUST_LOG=debug cargo run` for detailed operational logs
-- Use `RUST_LOG=trace cargo run` for full protocol-level debugging
-
-## Middle-End Proxy Debugging
-- Set `ME_DIAG=1` environment variable for high-precision cryptography diagnostics
-- STUN probe results are logged at startup - check for mismatch between local and reflected IP
-- If Middle-End fails, check `proxy_secret_path` points to valid file from https://core.telegram.org/getProxySecret
-
-## Connection Issues
-- DC connectivity is logged at startup with RTT measurements
-- If DC ping fails, check `dc_overrides` for custom addresses
-- Use `prefer_ipv6=false` in config if IPv6 is unreliable
-
-## TLS Fronting Issues
-- Invalid handshakes are proxied to `mask_host` - check this host is reachable
-- `mask_unix_sock` and `mask_host` are mutually exclusive - only one can be set
-- If `mask_unix_sock` is set, socket must exist before connections arrive
-
-## Common Errors
-- `ReplayAttack` - client replayed a handshake nonce, potential attack
-- `TimeSkew` - client clock is off, can disable with `ignore_time_skew=true`
-- `TgHandshakeTimeout` - upstream DC connection failed, check network

CODE_OF_CONDUCT.md

@@ -1,8 +1,8 @@
 # Code of Conduct
 
-## 1. Purpose
+## Purpose
 
-Telemt exists to solve technical problems.
+**Telemt exists to solve technical problems.**
 
 Telemt is open to contributors who want to learn, improve and build meaningful systems together.
 
@@ -18,27 +18,34 @@ Technology has consequences. Responsibility is inherent.
 
 ---
 
-## 2. Principles
+## Principles
 
 * **Technical over emotional**
+
   Arguments are grounded in data, logs, reproducible cases, or clear reasoning.
 
 * **Clarity over noise**
+
   Communication is structured, concise, and relevant.
 
 * **Openness with standards**
+
   Participation is open. The work remains disciplined.
 
 * **Independence of judgment**
+
   Claims are evaluated on technical merit, not affiliation or posture.
 
 * **Responsibility over capability**
+
   Capability does not justify careless use.
 
 * **Cooperation over friction**
+
   Progress depends on coordination, mutual support, and honest review.
 
 * **Good intent, rigorous method**
+
   Assume good intent, but require rigor.
 
 > **Aussagen gelten nach ihrer Begründung.**
@@ -47,7 +54,7 @@ Technology has consequences. Responsibility is inherent.
 
 ---
 
-## 3. Expected Behavior
+## Expected Behavior
 
 Participants are expected to:
 
@@ -69,7 +76,7 @@ New contributors are welcome. They are expected to grow into these standards. Ex
 
 ---
 
-## 4. Unacceptable Behavior
+## Unacceptable Behavior
 
 The following is not allowed:
 
@@ -89,7 +96,7 @@ Such discussions may be closed, removed, or redirected.
 
 ---
 
-## 5. Security and Misuse
+## Security and Misuse
 
 Telemt is intended for responsible use.
 
@@ -109,15 +116,13 @@ Security is both technical and behavioral.
 
 Telemt is open to contributors of different backgrounds, experience levels, and working styles.
 
-Standards are public, legible, and applied to the work itself.
-
-Questions are welcome. Careful disagreement is welcome. Honest correction is welcome.
-
-Gatekeeping by obscurity, status signaling, or hostility is not.
+- Standards are public, legible, and applied to the work itself.
+- Questions are welcome. Careful disagreement is welcome. Honest correction is welcome.
+- Gatekeeping by obscurity, status signaling, or hostility is not.
 
 ---
 
-## 7. Scope
+## Scope
 
 This Code of Conduct applies to all official spaces:
 
@@ -127,16 +132,19 @@ This Code of Conduct applies to all official spaces:
 
 ---
 
-## 8. Maintainer Stewardship
+## Maintainer Stewardship
 
 Maintainers are responsible for final decisions in matters of conduct, scope, and direction.
 
-This responsibility is stewardship: preserving continuity, protecting signal, maintaining standards, and keeping Telemt workable for others.
+This responsibility is stewardship: 
+- preserving continuity, 
+- protecting signal, 
+- maintaining standards, 
+- keeping Telemt workable for others.
 
 Judgment should be exercised with restraint, consistency, and institutional responsibility.
-
-Not every decision requires extended debate.
-Not every intervention requires public explanation.
+- Not every decision requires extended debate.
+- Not every intervention requires public explanation.
 
 All decisions are expected to serve the durability, clarity, and integrity of Telemt.
 
@@ -146,7 +154,7 @@ All decisions are expected to serve the durability, clarity, and integrity of Te
 
 ---
 
-## 9. Enforcement
+## Enforcement
 
 Maintainers may act to preserve the integrity of Telemt, including by:
 
@@ -156,44 +164,40 @@ Maintainers may act to preserve the integrity of Telemt, including by:
 * Restricting or banning participants
 
 Actions are taken to maintain function, continuity, and signal quality.
-
-Where possible, correction is preferred to exclusion.
-
-Where necessary, exclusion is preferred to decay.
+- Where possible, correction is preferred to exclusion.
+- Where necessary, exclusion is preferred to decay.
 
 ---
 
-## 10. Final
+## Final
 
 Telemt is built on discipline, structure, and shared intent.
+- Signal over noise.
+- Facts over opinion.
+- Systems over rhetoric.
 
-Signal over noise.
-Facts over opinion.
-Systems over rhetoric.
+- Work is collective.
+- Outcomes are shared.
+- Responsibility is distributed.
 
-Work is collective.
-Outcomes are shared.
-Responsibility is distributed.
-
-Precision is learned.
-Rigor is expected.
-Help is part of the work.
+- Precision is learned.
+- Rigor is expected.
+- Help is part of the work.
 
 > **Ordnung ist Voraussetzung der Freiheit.**
 
-If you contribute — contribute with care.
-If you speak — speak with substance.
-If you engage — engage constructively.
+- If you contribute — contribute with care.
+- If you speak — speak with substance.
+- If you engage — engage constructively.
 
 ---
 
-## 11. After All
+## After All
 
 Systems outlive intentions.
-
-What is built will be used.
-What is released will propagate.
-What is maintained will define the future state.
+- What is built will be used.
+- What is released will propagate.
+- What is maintained will define the future state.
 
 There is no neutral infrastructure, only infrastructure shaped well or poorly.
 
@@ -201,8 +205,8 @@ There is no neutral infrastructure, only infrastructure shaped well or poorly.
 
 > Every system carries responsibility.
 
-Stability requires discipline.
-Freedom requires structure.
-Trust requires honesty.
+- Stability requires discipline.
+- Freedom requires structure.
+- Trust requires honesty.
 
-In the end, the system reflects its contributors.
+In the end: the system reflects its contributors.

CONTRIBUTING.md

@@ -1,19 +1,82 @@
-# Issues - Rules
+# Issues
+## Warnung
+Before opening Issue, if it is more question than problem or bug - ask about that [in our chat](https://t.me/telemtrs)
+
 ## What it is not
 - NOT Question and Answer
 - NOT Helpdesk
 
-# Pull Requests - Rules
+***Each of your Issues triggers attempts to reproduce problems and analyze them, which are done manually by people***
+
+---
+
+# Pull Requests
+
 ## General
 - ONLY signed and verified commits
 - ONLY from your name
-- DO NOT commit with `codex` or `claude` as author/commiter
+- DO NOT commit with `codex`, `claude`, or other AI tools as author/committer
 - PREFER `flow` branch for development, not `main`
 
-## AI
-We are not against modern tools, like AI, where you act as a principal or architect, but we consider it important:
+---
 
-- you really understand what you're doing
-- you understand the relationships and dependencies of the components being modified
-- you understand the architecture of Telegram MTProto, MTProxy, Middle-End KDF at least generically
-- you DO NOT commit for the sake of commits, but to help the community, core-developers and ordinary users
+## Definition of Ready (MANDATORY)
+
+A Pull Request WILL be ignored or closed if:
+
+- it does NOT build
+- it does NOT pass tests
+- it does NOT follow formatting rules
+- it contains unrelated or excessive changes
+- the author cannot clearly explain the change
+
+---
+
+## Blessed Principles
+- PR must build
+- PR must pass tests
+- PR must be understood by author
+
+---
+
+## AI Usage Policy
+
+AI tools (Claude, ChatGPT, Codex, DeepSeek, etc.) are allowed as **assistants**, NOT as decision-makers.
+
+By submitting a PR, you confirm that:
+
+- you fully understand the code you submit
+- you verified correctness manually
+- you reviewed architecture and dependencies
+- you take full responsibility for the change
+
+AI-generated code is treated as **draft** and must be validated like any other external contribution.
+
+PRs that look like unverified AI dumps WILL be closed
+
+---
+
+## Maintainer Policy
+
+Maintainers reserve the right to:
+
+- close PRs that do not meet basic quality requirements
+- request explanations before review
+- ignore low-effort contributions
+
+Respect the reviewers time
+
+---
+
+## Enforcement
+
+Pull Requests that violate project standards may be closed without review.
+
+This includes (but is not limited to):
+
+- non-building code
+- failing tests
+- unverified or low-effort changes
+- inability to explain the change
+
+These actions follow the Code of Conduct and are intended to preserve signal, quality, and Telemt's integrity

Cargo.lock

@@ -183,9 +183,9 @@ dependencies = [
 
 [[package]]
 name = "aws-lc-sys"
-version = "0.39.0"
+version = "0.39.1"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "1fa7e52a4c5c547c741610a2c6f123f3881e409b714cd27e6798ef020c514f0a"
+checksum = "83a25cf98105baa966497416dbd42565ce3a8cf8dbfd59803ec9ad46f3126399"
 dependencies = [
  "cc",
  "cmake",
@@ -234,16 +234,16 @@ checksum = "843867be96c8daad0d758b57df9392b6d8d271134fce549de6ce169ff98a92af"
 
 [[package]]
 name = "blake3"
-version = "1.8.3"
+version = "1.8.4"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "2468ef7d57b3fb7e16b576e8377cdbde2320c60e1491e961d11da40fc4f02a2d"
+checksum = "4d2d5991425dfd0785aed03aedcf0b321d61975c9b5b3689c774a2610ae0b51e"
 dependencies = [
  "arrayref",
  "arrayvec",
  "cc",
  "cfg-if",
  "constant_time_eq",
- "cpufeatures 0.2.17",
+ "cpufeatures 0.3.0",
 ]
 
 [[package]]
@@ -299,9 +299,9 @@ dependencies = [
 
 [[package]]
 name = "cc"
-version = "1.2.57"
+version = "1.2.58"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "7a0dd1ca384932ff3641c8718a02769f1698e7563dc6974ffd03346116310423"
+checksum = "e1e928d4b69e3077709075a938a05ffbedfa53a84c8f766efbf8220bb1ff60e1"
 dependencies = [
  "find-msvc-tools",
  "jobserver",
@@ -441,9 +441,9 @@ checksum = "c8d4a3bb8b1e0c1050499d1815f5ab16d04f0959b233085fb31653fbfc9d98f9"
 
 [[package]]
 name = "cmake"
-version = "0.1.57"
+version = "0.1.58"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "75443c44cd6b379beb8c5b45d85d0773baf31cce901fe7bb252f4eff3008ef7d"
+checksum = "c0f78a02292a74a88ac736019ab962ece0bc380e3f977bf72e376c5d78ff0678"
 dependencies = [
  "cc",
 ]
@@ -1191,9 +1191,9 @@ checksum = "df3b46402a9d5adb4c86a0cf463f42e19994e3ee891101b1841f30a545cb49a9"
 
 [[package]]
 name = "hyper"
-version = "1.8.1"
+version = "1.9.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "2ab2d4f250c3d7b1c9fcdff1cece94ea4e2dfbec68614f7b87cb205f24ca9d11"
+checksum = "6299f016b246a94207e63da54dbe807655bf9e00044f73ded42c3ac5305fbcca"
 dependencies = [
  "atomic-waker",
  "bytes",
@@ -1206,7 +1206,6 @@ dependencies = [
  "httpdate",
  "itoa",
  "pin-project-lite",
- "pin-utils",
  "smallvec",
  "tokio",
  "want",
@@ -1245,7 +1244,7 @@ dependencies = [
  "libc",
  "percent-encoding",
  "pin-project-lite",
- "socket2 0.6.3",
+ "socket2",
  "tokio",
  "tower-service",
  "tracing",
@@ -1277,12 +1276,13 @@ dependencies = [
 
 [[package]]
 name = "icu_collections"
-version = "2.1.1"
+version = "2.2.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "4c6b649701667bbe825c3b7e6388cb521c23d88644678e83c0c4d0a621a34b43"
+checksum = "2984d1cd16c883d7935b9e07e44071dca8d917fd52ecc02c04d5fa0b5a3f191c"
 dependencies = [
  "displaydoc",
  "potential_utf",
+ "utf8_iter",
  "yoke",
  "zerofrom",
  "zerovec",
@@ -1290,9 +1290,9 @@ dependencies = [
 
 [[package]]
 name = "icu_locale_core"
-version = "2.1.1"
+version = "2.2.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "edba7861004dd3714265b4db54a3c390e880ab658fec5f7db895fae2046b5bb6"
+checksum = "92219b62b3e2b4d88ac5119f8904c10f8f61bf7e95b640d25ba3075e6cac2c29"
 dependencies = [
  "displaydoc",
  "litemap",
@@ -1303,9 +1303,9 @@ dependencies = [
 
 [[package]]
 name = "icu_normalizer"
-version = "2.1.1"
+version = "2.2.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "5f6c8828b67bf8908d82127b2054ea1b4427ff0230ee9141c54251934ab1b599"
+checksum = "c56e5ee99d6e3d33bd91c5d85458b6005a22140021cc324cea84dd0e72cff3b4"
 dependencies = [
  "icu_collections",
  "icu_normalizer_data",
@@ -1317,15 +1317,15 @@ dependencies = [
 
 [[package]]
 name = "icu_normalizer_data"
-version = "2.1.1"
+version = "2.2.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "7aedcccd01fc5fe81e6b489c15b247b8b0690feb23304303a9e560f37efc560a"
+checksum = "da3be0ae77ea334f4da67c12f149704f19f81d1adf7c51cf482943e84a2bad38"
 
 [[package]]
 name = "icu_properties"
-version = "2.1.2"
+version = "2.2.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "020bfc02fe870ec3a66d93e677ccca0562506e5872c650f893269e08615d74ec"
+checksum = "bee3b67d0ea5c2cca5003417989af8996f8604e34fb9ddf96208a033901e70de"
 dependencies = [
  "icu_collections",
  "icu_locale_core",
@@ -1337,15 +1337,15 @@ dependencies = [
 
 [[package]]
 name = "icu_properties_data"
-version = "2.1.2"
+version = "2.2.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "616c294cf8d725c6afcd8f55abc17c56464ef6211f9ed59cccffe534129c77af"
+checksum = "8e2bbb201e0c04f7b4b3e14382af113e17ba4f63e2c9d2ee626b720cbce54a14"
 
 [[package]]
 name = "icu_provider"
-version = "2.1.1"
+version = "2.2.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "85962cf0ce02e1e0a629cc34e7ca3e373ce20dda4c4d7294bbd0bf1fdb59e614"
+checksum = "139c4cf31c8b5f33d7e199446eff9c1e02decfc2f0eec2c8d71f65befa45b421"
 dependencies = [
  "displaydoc",
  "icu_locale_core",
@@ -1427,14 +1427,15 @@ dependencies = [
 
 [[package]]
 name = "ipconfig"
-version = "0.3.2"
+version = "0.3.4"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "b58db92f96b720de98181bbbe63c831e87005ab460c1bf306eb2622b4707997f"
+checksum = "4d40460c0ce33d6ce4b0630ad68ff63d6661961c48b6dba35e5a4d81cfb48222"
 dependencies = [
- "socket2 0.5.10",
+ "socket2",
  "widestring",
- "windows-sys 0.48.0",
- "winreg",
+ "windows-registry",
+ "windows-result",
+ "windows-sys 0.61.2",
 ]
 
 [[package]]
@@ -1454,9 +1455,9 @@ dependencies = [
 
 [[package]]
 name = "iri-string"
-version = "0.7.10"
+version = "0.7.12"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "c91338f0783edbd6195decb37bae672fd3b165faffb89bf7b9e6942f8b1a731a"
+checksum = "25e659a4bb38e810ebc252e53b5814ff908a8c58c2a9ce2fae1bbec24cbf4e20"
 dependencies = [
  "memchr",
  "serde",
@@ -1486,7 +1487,7 @@ dependencies = [
  "cesu8",
  "cfg-if",
  "combine",
- "jni-sys",
+ "jni-sys 0.3.1",
  "log",
  "thiserror 1.0.69",
  "walkdir",
@@ -1495,9 +1496,31 @@ dependencies = [
 
 [[package]]
 name = "jni-sys"
-version = "0.3.0"
+version = "0.3.1"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "8eaf4bc02d17cbdd7ff4c7438cafcdf7fb9a4613313ad11b4f8fefe7d3fa0130"
+checksum = "41a652e1f9b6e0275df1f15b32661cf0d4b78d4d87ddec5e0c3c20f097433258"
+dependencies = [
+ "jni-sys 0.4.1",
+]
+
+[[package]]
+name = "jni-sys"
+version = "0.4.1"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "c6377a88cb3910bee9b0fa88d4f42e1d2da8e79915598f65fb0c7ee14c878af2"
+dependencies = [
+ "jni-sys-macros",
+]
+
+[[package]]
+name = "jni-sys-macros"
+version = "0.4.1"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "38c0b942f458fe50cdac086d2f946512305e5631e720728f2a61aabcd47a6264"
+dependencies = [
+ "quote",
+ "syn",
+]
 
 [[package]]
 name = "jobserver"
@@ -1511,10 +1534,12 @@ dependencies = [
 
 [[package]]
 name = "js-sys"
-version = "0.3.91"
+version = "0.3.94"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "b49715b7073f385ba4bc528e5747d02e66cb39c6146efb66b781f131f0fb399c"
+checksum = "2e04e2ef80ce82e13552136fabeef8a5ed1f985a96805761cbb9a2c34e7664d9"
 dependencies = [
+ "cfg-if",
+ "futures-util",
  "once_cell",
  "wasm-bindgen",
 ]
@@ -1553,9 +1578,9 @@ checksum = "09edd9e8b54e49e587e4f6295a7d29c3ea94d469cb40ab8ca70b288248a81db2"
 
 [[package]]
 name = "libc"
-version = "0.2.183"
+version = "0.2.184"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "b5b646652bf6661599e1da8901b3b9522896f01e736bad5f723fe7a3a27f899d"
+checksum = "48f5d2a454e16a5ea0f4ced81bd44e4cfc7bd3a507b61887c99fd3538b28e4af"
 
 [[package]]
 name = "linux-raw-sys"
@@ -1565,9 +1590,9 @@ checksum = "32a66949e030da00e8c7d4434b251670a91556f4144941d37452769c25d58a53"
 
 [[package]]
 name = "litemap"
-version = "0.8.1"
+version = "0.8.2"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "6373607a59f0be73a39b6fe456b8192fcc3585f602af20751600e974dd455e77"
+checksum = "92daf443525c4cce67b150400bc2316076100ce0b3686209eb8cf3c31612e6f0"
 
 [[package]]
 name = "lock_api"
@@ -1647,9 +1672,9 @@ checksum = "68354c5c6bd36d73ff3feceb05efa59b6acb7626617f4962be322a825e61f79a"
 
 [[package]]
 name = "mio"
-version = "1.1.1"
+version = "1.2.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "a69bcab0ad47271a0234d9422b131806bf3968021e5dc9328caf2d4cd58557fc"
+checksum = "50b7e5b27aa02a74bac8c3f23f448f8d87ff11f92d3aac1a6ed369ee08cc56c1"
 dependencies = [
  "libc",
  "log",
@@ -1659,9 +1684,9 @@ dependencies = [
 
 [[package]]
 name = "moka"
-version = "0.12.14"
+version = "0.12.15"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "85f8024e1c8e71c778968af91d43700ce1d11b219d127d79fb2934153b82b42b"
+checksum = "957228ad12042ee839f93c8f257b62b4c0ab5eaae1d4fa60de53b27c9d7c5046"
 dependencies = [
  "crossbeam-channel",
  "crossbeam-epoch",
@@ -1745,9 +1770,9 @@ dependencies = [
 
 [[package]]
 name = "num-conv"
-version = "0.2.0"
+version = "0.2.1"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "cf97ec579c3c42f953ef76dbf8d55ac91fb219dde70e49aa4a6b7d74e9919050"
+checksum = "c6673768db2d862beb9b39a78fdcb1a69439615d5794a1be50caa9bc92c81967"
 
 [[package]]
 name = "num-integer"
@@ -1869,12 +1894,6 @@ version = "0.2.17"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "a89322df9ebe1c1578d689c92318e070967d1042b512afbe49518723f4e6d5cd"
 
-[[package]]
-name = "pin-utils"
-version = "0.1.0"
-source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "8b870d8c151b6f2fb93e84a13146138f05d02ed11c7e7c54f8826aaaf7c9f184"
-
 [[package]]
 name = "pkcs8"
 version = "0.10.2"
@@ -1944,9 +1963,9 @@ checksum = "c33a9471896f1c69cecef8d20cbe2f7accd12527ce60845ff44c153bb2a21b49"
 
 [[package]]
 name = "potential_utf"
-version = "0.1.4"
+version = "0.1.5"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "b73949432f5e2a09657003c25bca5e19a0e9c84f8058ca374f49e0ebe605af77"
+checksum = "0103b1cef7ec0cf76490e969665504990193874ea05c85ff9bab8b911d0a0564"
 dependencies = [
  "zerovec",
 ]
@@ -1987,9 +2006,9 @@ dependencies = [
 
 [[package]]
 name = "proptest"
-version = "1.10.0"
+version = "1.11.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "37566cb3fdacef14c0737f9546df7cfeadbfbc9fef10991038bf5015d0c80532"
+checksum = "4b45fcc2344c680f5025fe57779faef368840d0bd1f42f216291f0dc4ace4744"
 dependencies = [
  "bit-set",
  "bit-vec",
@@ -2023,7 +2042,7 @@ dependencies = [
  "quinn-udp",
  "rustc-hash",
  "rustls",
- "socket2 0.6.3",
+ "socket2",
  "thiserror 2.0.18",
  "tokio",
  "tracing",
@@ -2061,7 +2080,7 @@ dependencies = [
  "cfg_aliases",
  "libc",
  "once_cell",
- "socket2 0.6.3",
+ "socket2",
  "tracing",
  "windows-sys 0.60.2",
 ]
@@ -2279,9 +2298,9 @@ dependencies = [
 
 [[package]]
 name = "rustc-hash"
-version = "2.1.1"
+version = "2.1.2"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "357703d41365b4b27c590e3ed91eabb1b663f07c4c084095e60cbed4362dff0d"
+checksum = "94300abf3f1ae2e2b8ffb7b58043de3d399c73fa6f4b73826402a5c457614dbe"
 
 [[package]]
 name = "rustc_version"
@@ -2533,9 +2552,9 @@ dependencies = [
 
 [[package]]
 name = "serde_spanned"
-version = "1.0.4"
+version = "1.1.1"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "f8bbf91e5a4d6315eee45e704372590b30e260ee83af6639d64557f51b067776"
+checksum = "6662b5879511e06e8999a8a235d848113e942c9124f211511b16466ee2995f26"
 dependencies = [
  "serde_core",
 ]
@@ -2603,7 +2622,7 @@ dependencies = [
  "serde_json",
  "serde_urlencoded",
  "shadowsocks-crypto",
- "socket2 0.6.3",
+ "socket2",
  "spin",
  "thiserror 2.0.18",
  "tokio",
@@ -2675,16 +2694,6 @@ version = "1.15.1"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "67b1b7a3b5fe4f1376887184045fcf45c69e92af734b7aaddc05fb777b6fbd03"
 
-[[package]]
-name = "socket2"
-version = "0.5.10"
-source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "e22376abed350d73dd1cd119b57ffccad95b4e585a7cda43e286245ce23c0678"
-dependencies = [
- "libc",
- "windows-sys 0.52.0",
-]
-
 [[package]]
 name = "socket2"
 version = "0.6.3"
@@ -2771,7 +2780,7 @@ checksum = "7b2093cf4c8eb1e67749a6762251bc9cd836b6fc171623bd0a9d324d37af2417"
 
 [[package]]
 name = "telemt"
-version = "3.3.29"
+version = "3.3.38"
 dependencies = [
  "aes",
  "anyhow",
@@ -2812,7 +2821,7 @@ dependencies = [
  "sha1",
  "sha2",
  "shadowsocks",
- "socket2 0.6.3",
+ "socket2",
  "static_assertions",
  "subtle",
  "thiserror 2.0.18",
@@ -2822,6 +2831,7 @@ dependencies = [
  "tokio-util",
  "toml",
  "tracing",
+ "tracing-appender",
  "tracing-subscriber",
  "url",
  "webpki-roots",
@@ -2925,9 +2935,9 @@ dependencies = [
 
 [[package]]
 name = "tinystr"
-version = "0.8.2"
+version = "0.8.3"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "42d3e9c45c09de15d06dd8acf5f4e0e399e85927b7f00711024eb7ae10fa4869"
+checksum = "c8323304221c2a851516f22236c5722a72eaa19749016521d6dff0824447d96d"
 dependencies = [
  "displaydoc",
  "zerovec",
@@ -2970,7 +2980,7 @@ dependencies = [
  "parking_lot",
  "pin-project-lite",
  "signal-hook-registry",
- "socket2 0.6.3",
+ "socket2",
  "tokio-macros",
  "tracing",
  "windows-sys 0.61.2",
@@ -3031,7 +3041,7 @@ dependencies = [
  "log",
  "once_cell",
  "pin-project",
- "socket2 0.6.3",
+ "socket2",
  "tokio",
  "windows-sys 0.60.2",
 ]
@@ -3055,9 +3065,9 @@ dependencies = [
 
 [[package]]
 name = "toml"
-version = "1.0.7+spec-1.1.0"
+version = "1.1.2+spec-1.1.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "dd28d57d8a6f6e458bc0b8784f8fdcc4b99a437936056fa122cb234f18656a96"
+checksum = "81f3d15e84cbcd896376e6730314d59fb5a87f31e4b038454184435cd57defee"
 dependencies = [
  "indexmap",
  "serde_core",
@@ -3070,27 +3080,27 @@ dependencies = [
 
 [[package]]
 name = "toml_datetime"
-version = "1.0.1+spec-1.1.0"
+version = "1.1.1+spec-1.1.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "9b320e741db58cac564e26c607d3cc1fdc4a88fd36c879568c07856ed83ff3e9"
+checksum = "3165f65f62e28e0115a00b2ebdd37eb6f3b641855f9d636d3cd4103767159ad7"
 dependencies = [
  "serde_core",
 ]
 
 [[package]]
 name = "toml_parser"
-version = "1.0.10+spec-1.1.0"
+version = "1.1.2+spec-1.1.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "7df25b4befd31c4816df190124375d5a20c6b6921e2cad937316de3fccd63420"
+checksum = "a2abe9b86193656635d2411dc43050282ca48aa31c2451210f4202550afb7526"
 dependencies = [
  "winnow",
 ]
 
 [[package]]
 name = "toml_writer"
-version = "1.0.7+spec-1.1.0"
+version = "1.1.1+spec-1.1.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "f17aaa1c6e3dc22b1da4b6bba97d066e354c7945cac2f7852d4e4e7ca7a6b56d"
+checksum = "756daf9b1013ebe47a8776667b466417e2d4c5679d441c26230efd9ef78692db"
 
 [[package]]
 name = "tower"
@@ -3148,6 +3158,18 @@ dependencies = [
  "tracing-core",
 ]
 
+[[package]]
+name = "tracing-appender"
+version = "0.2.4"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "786d480bce6247ab75f005b14ae1624ad978d3029d9113f0a22fa1ac773faeaf"
+dependencies = [
+ "crossbeam-channel",
+ "thiserror 2.0.18",
+ "time",
+ "tracing-subscriber",
+]
+
 [[package]]
 name = "tracing-attributes"
 version = "0.1.31"
@@ -3275,9 +3297,9 @@ checksum = "b6c140620e7ffbb22c2dee59cafe6084a59b5ffc27a8859a5f0d494b5d52b6be"
 
 [[package]]
 name = "uuid"
-version = "1.22.0"
+version = "1.23.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "a68d3c8f01c0cfa54a75291d83601161799e4a89a39e0929f4b0354d88757a37"
+checksum = "5ac8b6f42ead25368cf5b098aeb3dc8a1a2c05a3eee8a9a1a68c640edbfc79d9"
 dependencies = [
  "getrandom 0.4.2",
  "js-sys",
@@ -3350,9 +3372,9 @@ dependencies = [
 
 [[package]]
 name = "wasm-bindgen"
-version = "0.2.114"
+version = "0.2.117"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "6532f9a5c1ece3798cb1c2cfdba640b9b3ba884f5db45973a6f442510a87d38e"
+checksum = "0551fc1bb415591e3372d0bc4780db7e587d84e2a7e79da121051c5c4b89d0b0"
 dependencies = [
  "cfg-if",
  "once_cell",
@@ -3363,23 +3385,19 @@ dependencies = [
 
 [[package]]
 name = "wasm-bindgen-futures"
-version = "0.4.64"
+version = "0.4.67"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "e9c5522b3a28661442748e09d40924dfb9ca614b21c00d3fd135720e48b67db8"
+checksum = "03623de6905b7206edd0a75f69f747f134b7f0a2323392d664448bf2d3c5d87e"
 dependencies = [
- "cfg-if",
- "futures-util",
  "js-sys",
- "once_cell",
  "wasm-bindgen",
- "web-sys",
 ]
 
 [[package]]
 name = "wasm-bindgen-macro"
-version = "0.2.114"
+version = "0.2.117"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "18a2d50fcf105fb33bb15f00e7a77b772945a2ee45dcf454961fd843e74c18e6"
+checksum = "7fbdf9a35adf44786aecd5ff89b4563a90325f9da0923236f6104e603c7e86be"
 dependencies = [
  "quote",
  "wasm-bindgen-macro-support",
@@ -3387,9 +3405,9 @@ dependencies = [
 
 [[package]]
 name = "wasm-bindgen-macro-support"
-version = "0.2.114"
+version = "0.2.117"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "03ce4caeaac547cdf713d280eda22a730824dd11e6b8c3ca9e42247b25c631e3"
+checksum = "dca9693ef2bab6d4e6707234500350d8dad079eb508dca05530c85dc3a529ff2"
 dependencies = [
  "bumpalo",
  "proc-macro2",
@@ -3400,9 +3418,9 @@ dependencies = [
 
 [[package]]
 name = "wasm-bindgen-shared"
-version = "0.2.114"
+version = "0.2.117"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "75a326b8c223ee17883a4251907455a2431acc2791c98c26279376490c378c16"
+checksum = "39129a682a6d2d841b6c429d0c51e5cb0ed1a03829d8b3d1e69a011e62cb3d3b"
 dependencies = [
  "unicode-ident",
 ]
@@ -3443,9 +3461,9 @@ dependencies = [
 
 [[package]]
 name = "web-sys"
-version = "0.3.91"
+version = "0.3.94"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "854ba17bb104abfb26ba36da9729addc7ce7f06f5c0f90f3c391f8461cca21f9"
+checksum = "cd70027e39b12f0849461e08ffc50b9cd7688d942c1c8e3c7b22273236b4dd0a"
 dependencies = [
  "js-sys",
  "wasm-bindgen",
@@ -3557,6 +3575,17 @@ version = "0.2.1"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "f0805222e57f7521d6a62e36fa9163bc891acd422f971defe97d64e70d0a4fe5"
 
+[[package]]
+name = "windows-registry"
+version = "0.6.1"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "02752bf7fbdcce7f2a27a742f798510f3e5ad88dbe84871e5168e2120c3d5720"
+dependencies = [
+ "windows-link",
+ "windows-result",
+ "windows-strings",
+]
+
 [[package]]
 name = "windows-result"
 version = "0.4.1"
@@ -3584,15 +3613,6 @@ dependencies = [
  "windows-targets 0.42.2",
 ]
 
-[[package]]
-name = "windows-sys"
-version = "0.48.0"
-source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "677d2418bec65e3338edb076e806bc1ec15693c5d0104683f2efe857f61056a9"
-dependencies = [
- "windows-targets 0.48.5",
-]
-
 [[package]]
 name = "windows-sys"
 version = "0.52.0"
@@ -3635,21 +3655,6 @@ dependencies = [
  "windows_x86_64_msvc 0.42.2",
 ]
 
-[[package]]
-name = "windows-targets"
-version = "0.48.5"
-source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "9a2fa6e2155d7247be68c096456083145c183cbbbc2764150dda45a87197940c"
-dependencies = [
- "windows_aarch64_gnullvm 0.48.5",
- "windows_aarch64_msvc 0.48.5",
- "windows_i686_gnu 0.48.5",
- "windows_i686_msvc 0.48.5",
- "windows_x86_64_gnu 0.48.5",
- "windows_x86_64_gnullvm 0.48.5",
- "windows_x86_64_msvc 0.48.5",
-]
-
 [[package]]
 name = "windows-targets"
 version = "0.52.6"
@@ -3689,12 +3694,6 @@ version = "0.42.2"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "597a5118570b68bc08d8d59125332c54f1ba9d9adeedeef5b99b02ba2b0698f8"
 
-[[package]]
-name = "windows_aarch64_gnullvm"
-version = "0.48.5"
-source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "2b38e32f0abccf9987a4e3079dfb67dcd799fb61361e53e2882c3cbaf0d905d8"
-
 [[package]]
 name = "windows_aarch64_gnullvm"
 version = "0.52.6"
@@ -3713,12 +3712,6 @@ version = "0.42.2"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "e08e8864a60f06ef0d0ff4ba04124db8b0fb3be5776a5cd47641e942e58c4d43"
 
-[[package]]
-name = "windows_aarch64_msvc"
-version = "0.48.5"
-source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "dc35310971f3b2dbbf3f0690a219f40e2d9afcf64f9ab7cc1be722937c26b4bc"
-
 [[package]]
 name = "windows_aarch64_msvc"
 version = "0.52.6"
@@ -3737,12 +3730,6 @@ version = "0.42.2"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "c61d927d8da41da96a81f029489353e68739737d3beca43145c8afec9a31a84f"
 
-[[package]]
-name = "windows_i686_gnu"
-version = "0.48.5"
-source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "a75915e7def60c94dcef72200b9a8e58e5091744960da64ec734a6c6e9b3743e"
-
 [[package]]
 name = "windows_i686_gnu"
 version = "0.52.6"
@@ -3773,12 +3760,6 @@ version = "0.42.2"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "44d840b6ec649f480a41c8d80f9c65108b92d89345dd94027bfe06ac444d1060"
 
-[[package]]
-name = "windows_i686_msvc"
-version = "0.48.5"
-source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "8f55c233f70c4b27f66c523580f78f1004e8b5a8b659e05a4eb49d4166cca406"
-
 [[package]]
 name = "windows_i686_msvc"
 version = "0.52.6"
@@ -3797,12 +3778,6 @@ version = "0.42.2"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "8de912b8b8feb55c064867cf047dda097f92d51efad5b491dfb98f6bbb70cb36"
 
-[[package]]
-name = "windows_x86_64_gnu"
-version = "0.48.5"
-source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "53d40abd2583d23e4718fddf1ebec84dbff8381c07cae67ff7768bbf19c6718e"
-
 [[package]]
 name = "windows_x86_64_gnu"
 version = "0.52.6"
@@ -3821,12 +3796,6 @@ version = "0.42.2"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "26d41b46a36d453748aedef1486d5c7a85db22e56aff34643984ea85514e94a3"
 
-[[package]]
-name = "windows_x86_64_gnullvm"
-version = "0.48.5"
-source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "0b7b52767868a23d5bab768e390dc5f5c55825b6d30b86c844ff2dc7414044cc"
-
 [[package]]
 name = "windows_x86_64_gnullvm"
 version = "0.52.6"
@@ -3845,12 +3814,6 @@ version = "0.42.2"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "9aec5da331524158c6d1a4ac0ab1541149c0b9505fde06423b02f5ef0106b9f0"
 
-[[package]]
-name = "windows_x86_64_msvc"
-version = "0.48.5"
-source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "ed94fce61571a4006852b7389a063ab983c02eb1bb37b47f8272ce92d06d9538"
-
 [[package]]
 name = "windows_x86_64_msvc"
 version = "0.52.6"
@@ -3865,19 +3828,9 @@ checksum = "d6bbff5f0aada427a1e5a6da5f1f98158182f26556f345ac9e04d36d0ebed650"
 
 [[package]]
 name = "winnow"
-version = "1.0.0"
+version = "1.0.1"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "a90e88e4667264a994d34e6d1ab2d26d398dcdca8b7f52bec8668957517fc7d8"
-
-[[package]]
-name = "winreg"
-version = "0.50.0"
-source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "524e57b2c537c0f9b1e69f1965311ec12182b4122e45035b1508cd24d2adadb1"
-dependencies = [
- "cfg-if",
- "windows-sys 0.48.0",
-]
+checksum = "09dac053f1cd375980747450bfc7250c264eaae0583872e845c0c7cd578872b5"
 
 [[package]]
 name = "wit-bindgen"
@@ -4004,9 +3957,9 @@ dependencies = [
 
 [[package]]
 name = "yoke"
-version = "0.8.1"
+version = "0.8.2"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "72d6e5c6afb84d73944e5cedb052c4680d5657337201555f9f2a16b7406d4954"
+checksum = "abe8c5fda708d9ca3df187cae8bfb9ceda00dd96231bed36e445a1a48e66f9ca"
 dependencies = [
  "stable_deref_trait",
  "yoke-derive",
@@ -4015,9 +3968,9 @@ dependencies = [
 
 [[package]]
 name = "yoke-derive"
-version = "0.8.1"
+version = "0.8.2"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "b659052874eb698efe5b9e8cf382204678a0086ebf46982b79d6ca3182927e5d"
+checksum = "de844c262c8848816172cef550288e7dc6c7b7814b4ee56b3e1553f275f1858e"
 dependencies = [
  "proc-macro2",
  "quote",
@@ -4027,18 +3980,18 @@ dependencies = [
 
 [[package]]
 name = "zerocopy"
-version = "0.8.47"
+version = "0.8.48"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "efbb2a062be311f2ba113ce66f697a4dc589f85e78a4aea276200804cea0ed87"
+checksum = "eed437bf9d6692032087e337407a86f04cd8d6a16a37199ed57949d415bd68e9"
 dependencies = [
  "zerocopy-derive",
 ]
 
 [[package]]
 name = "zerocopy-derive"
-version = "0.8.47"
+version = "0.8.48"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "0e8bc7269b54418e7aeeef514aa68f8690b8c0489a06b0136e5f57c4c5ccab89"
+checksum = "70e3cd084b1788766f53af483dd21f93881ff30d7320490ec3ef7526d203bad4"
 dependencies = [
  "proc-macro2",
  "quote",
@@ -4047,18 +4000,18 @@ dependencies = [
 
 [[package]]
 name = "zerofrom"
-version = "0.1.6"
+version = "0.1.7"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "50cc42e0333e05660c3587f3bf9d0478688e15d870fab3346451ce7f8c9fbea5"
+checksum = "69faa1f2a1ea75661980b013019ed6687ed0e83d069bc1114e2cc74c6c04c4df"
 dependencies = [
  "zerofrom-derive",
 ]
 
 [[package]]
 name = "zerofrom-derive"
-version = "0.1.6"
+version = "0.1.7"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "d71e5d6e06ab090c67b5e44993ec16b72dcbaabc526db883a360057678b48502"
+checksum = "11532158c46691caf0f2593ea8358fed6bbf68a0315e80aae9bd41fbade684a1"
 dependencies = [
  "proc-macro2",
  "quote",
@@ -4088,9 +4041,9 @@ dependencies = [
 
 [[package]]
 name = "zerotrie"
-version = "0.2.3"
+version = "0.2.4"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "2a59c17a5562d507e4b54960e8569ebee33bee890c70aa3fe7b97e85a9fd7851"
+checksum = "0f9152d31db0792fa83f70fb2f83148effb5c1f5b8c7686c3459e361d9bc20bf"
 dependencies = [
  "displaydoc",
  "yoke",
@@ -4099,9 +4052,9 @@ dependencies = [
 
 [[package]]
 name = "zerovec"
-version = "0.11.5"
+version = "0.11.6"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "6c28719294829477f525be0186d13efa9a3c602f7ec202ca9e353d310fb9a002"
+checksum = "90f911cbc359ab6af17377d242225f4d75119aec87ea711a880987b18cd7b239"
 dependencies = [
  "yoke",
  "zerofrom",
@@ -4110,9 +4063,9 @@ dependencies = [
 
 [[package]]
 name = "zerovec-derive"
-version = "0.11.2"
+version = "0.11.3"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "eadce39539ca5cb3985590102671f2567e659fca9666581ad3411d59207951f3"
+checksum = "625dc425cab0dca6dc3c3319506e6593dcb08a9f387ea3b284dbd52a92c40555"
 dependencies = [
  "proc-macro2",
  "quote",

Cargo.toml

@@ -1,8 +1,11 @@
 [package]
 name = "telemt"
-version = "3.3.29"
+version = "3.3.38"
 edition = "2024"
 
+[features]
+redteam_offline_expected_fail = []
+
 [dependencies]
 # C
 libc = "0.2"
@@ -27,7 +30,13 @@ static_assertions = "1.1"
 
 # Network
 socket2 = { version = "0.6", features = ["all"] }
-nix = { version = "0.31", default-features = false, features = ["net", "fs"] }
+nix = { version = "0.31", default-features = false, features = [
+  "net",
+  "user",
+  "process",
+  "fs",
+  "signal",
+] }
 shadowsocks = { version = "1.24", features = ["aead-cipher-2022"] }
 
 # Serialization
@@ -41,6 +50,7 @@ bytes = "1.9"
 thiserror = "2.0"
 tracing = "0.1"
 tracing-subscriber = { version = "0.3", features = ["env-filter"] }
+tracing-appender = "0.2"
 parking_lot = "0.12"
 dashmap = "6.1"
 arc-swap = "1.7"
@@ -65,8 +75,14 @@ hyper = { version = "1", features = ["server", "http1"] }
 hyper-util = { version = "0.1", features = ["tokio", "server-auto"] }
 http-body-util = "0.1"
 httpdate = "1.0"
-tokio-rustls = { version = "0.26", default-features = false, features = ["tls12"] }
-rustls = { version = "0.23", default-features = false, features = ["std", "tls12", "ring"] }
+tokio-rustls = { version = "0.26", default-features = false, features = [
+  "tls12",
+] }
+rustls = { version = "0.23", default-features = false, features = [
+  "std",
+  "tls12",
+  "ring",
+] }
 webpki-roots = "1.0"
 
 [dev-dependencies]
@@ -80,4 +96,6 @@ name = "crypto_bench"
 harness = false
 
 [profile.release]
-lto = "thin"
+lto = "fat"
+codegen-units = 1
+

Dockerfile

@@ -1,97 +1,98 @@
 # syntax=docker/dockerfile:1
 
-# ==========================
-# Stage 1: Build
-# ==========================
-FROM rust:1.88-slim-bookworm AS builder
-
-RUN apt-get update && apt-get install -y --no-install-recommends \
-    pkg-config \
-    ca-certificates \
-    && rm -rf /var/lib/apt/lists/*
-
-WORKDIR /build
-
-# Depcache
-COPY Cargo.toml Cargo.lock* ./
-RUN mkdir src && echo 'fn main() {}' > src/main.rs && \
-    cargo build --release 2>/dev/null || true && \
-    rm -rf src
-
-# Build
-COPY . .
-RUN cargo build --release && strip target/release/telemt
+ARG TELEMT_REPOSITORY=telemt/telemt
+ARG TELEMT_VERSION=latest
 
 # ==========================
-# Stage 2: Compress (strip + UPX)
+# Minimal Image
 # ==========================
 FROM debian:12-slim AS minimal
 
-RUN apt-get update && apt-get install -y --no-install-recommends \
-    upx \
-    binutils \
-    && rm -rf /var/lib/apt/lists/*
+ARG TARGETARCH
+ARG TELEMT_REPOSITORY
+ARG TELEMT_VERSION
 
-COPY --from=builder /build/target/release/telemt /telemt
+RUN set -eux; \
+    apt-get update; \
+    apt-get install -y --no-install-recommends \
+        binutils \
+        ca-certificates \
+        curl \
+        tar; \
+    rm -rf /var/lib/apt/lists/*
 
-RUN strip /telemt || true
-RUN upx --best --lzma /telemt || true
+RUN set -eux; \
+    case "${TARGETARCH}" in \
+        amd64) ASSET="telemt-x86_64-linux-musl.tar.gz" ;; \
+        arm64) ASSET="telemt-aarch64-linux-musl.tar.gz" ;; \
+        *) echo "Unsupported TARGETARCH: ${TARGETARCH}" >&2; exit 1 ;; \
+    esac; \
+    VERSION="${TELEMT_VERSION#refs/tags/}"; \
+    if [ -z "${VERSION}" ] || [ "${VERSION}" = "latest" ]; then \
+        BASE_URL="https://github.com/${TELEMT_REPOSITORY}/releases/latest/download"; \
+    else \
+        BASE_URL="https://github.com/${TELEMT_REPOSITORY}/releases/download/${VERSION}"; \
+    fi; \
+    curl -fL \
+        --retry 5 \
+        --retry-delay 3 \
+        --connect-timeout 10 \
+        --max-time 120 \
+        -o "/tmp/${ASSET}" \
+        "${BASE_URL}/${ASSET}"; \
+    curl -fL \
+        --retry 5 \
+        --retry-delay 3 \
+        --connect-timeout 10 \
+        --max-time 120 \
+        -o "/tmp/${ASSET}.sha256" \
+        "${BASE_URL}/${ASSET}.sha256"; \
+    cd /tmp; \
+    sha256sum -c "${ASSET}.sha256"; \
+    tar -xzf "${ASSET}" -C /tmp; \
+    test -f /tmp/telemt; \
+    install -m 0755 /tmp/telemt /telemt; \
+    strip --strip-unneeded /telemt || true; \
+    rm -f "/tmp/${ASSET}" "/tmp/${ASSET}.sha256" /tmp/telemt
 
 # ==========================
-# Stage 3: Debug base
+# Debug Image
 # ==========================
-FROM debian:12-slim AS debug-base
+FROM debian:12-slim AS debug
 
-RUN apt-get update && apt-get install -y --no-install-recommends \
-    ca-certificates \
-    tzdata \
-    curl \
-    iproute2 \
-    busybox \
-    && rm -rf /var/lib/apt/lists/*
-
-# ==========================
-# Stage 4: Debug image
-# ==========================
-FROM debug-base AS debug
+RUN set -eux; \
+    apt-get update; \
+    apt-get install -y --no-install-recommends \
+        ca-certificates \
+        tzdata \
+        curl \
+        iproute2 \
+        busybox; \
+    rm -rf /var/lib/apt/lists/*
 
 WORKDIR /app
 
 COPY --from=minimal /telemt /app/telemt
 COPY config.toml /app/config.toml
 
-USER root
-
-EXPOSE 443
-EXPOSE 9090
-EXPOSE 9091
+EXPOSE 443 9090 9091
 
 ENTRYPOINT ["/app/telemt"]
 CMD ["config.toml"]
 
 # ==========================
-# Stage 5: Production (distroless)
+# Production Distroless on MUSL
 # ==========================
-FROM gcr.io/distroless/base-debian12 AS prod
+FROM gcr.io/distroless/static-debian12 AS prod
 
 WORKDIR /app
 
 COPY --from=minimal /telemt /app/telemt
 COPY config.toml /app/config.toml
 
-# TLS + timezone + shell
-COPY --from=debug-base /etc/ssl/certs /etc/ssl/certs
-COPY --from=debug-base /usr/share/zoneinfo /usr/share/zoneinfo
-COPY --from=debug-base /bin/busybox /bin/busybox
-
-RUN ["/bin/busybox", "--install", "-s", "/bin"]
-
-# distroless user
 USER nonroot:nonroot
 
-EXPOSE 443
-EXPOSE 9090
-EXPOSE 9091
+EXPOSE 443 9090 9091
 
 ENTRYPOINT ["/app/telemt"]
 CMD ["config.toml"]

README.md

@@ -2,6 +2,11 @@
 
 ***Löst Probleme, bevor andere überhaupt wissen, dass sie existieren*** / ***It solves problems before others even realize they exist***
 
+### [**Telemt Chat in Telegram**](https://t.me/telemtrs)
+#### Fixed TLS ClientHello is now available in Telegram Desktop starting from version 6.7.2: to work with EE-MTProxy, please update your client;
+#### Fixed TLS ClientHello for Telegram Android Client is available in [our chat](https://t.me/telemtrs/30234/36441); official releases for Android and iOS are "work in progress";
+
+
 **Telemt** is a fast, secure, and feature-rich server written in Rust: it fully implements the official Telegram proxy algo and adds many production-ready improvements such as:
 - [ME Pool + Reader/Writer + Registry + Refill + Adaptive Floor + Trio-State + Generation Lifecycle](https://github.com/telemt/telemt/blob/main/docs/model/MODEL.en.md)
 - [Full-covered API w/ management](https://github.com/telemt/telemt/blob/main/docs/API.md)
@@ -9,60 +14,6 @@
 - Prometheus-format Metrics
 - TLS-Fronting and TCP-Splicing for masking from "prying" eyes
 
-[**Telemt Chat in Telegram**](https://t.me/telemtrs)
-
-## NEWS and EMERGENCY
-### ✈️ Telemt 3 is released!
-<table>
-<tr>
-<td width="50%" valign="top">
-
-### 🇷🇺 RU
-
-#### О релизах
-
-[3.3.27](https://github.com/telemt/telemt/releases/tag/3.3.27) даёт баланс стабильности и передового функционала, а так же последние исправления по безопасности и багам
-
-Будем рады вашему фидбеку и предложениям по улучшению — особенно в части **API**, **статистики**, **UX**
-
----
-
-Если у вас есть компетенции в:
-
-- Асинхронных сетевых приложениях  
-- Анализе трафика  
-- Реверс-инжиниринге  
-- Сетевых расследованиях  
-
-Мы открыты к архитектурным предложениям, идеям и pull requests
-</td>
-<td width="50%" valign="top">
-
-### 🇬🇧 EN
-
-#### About releases
-
-[3.3.27](https://github.com/telemt/telemt/releases/tag/3.3.27) provides a balance of stability and advanced functionality, as well as the latest security and bug fixes
-
-We are looking forward to your feedback and improvement proposals — especially regarding **API**, **statistics**, **UX**
-
----
-
-If you have expertise in:
-
-- Asynchronous network applications  
-- Traffic analysis  
-- Reverse engineering  
-- Network forensics  
-
-We welcome ideas, architectural feedback, and pull requests.
-</td>
-</tr>
-</table>
-
-# Features
-💥 The configuration structure has changed since version 1.1.0.0. change it in your environment!
-
 ⚓ Our implementation of **TLS-fronting** is one of the most deeply debugged, focused, advanced and *almost* **"behaviorally consistent to real"**:  we are confident we have it right - [see evidence on our validation and traces](#recognizability-for-dpi-and-crawler)
 
 ⚓ Our ***Middle-End Pool*** is fastest by design in standard scenarios, compared to other implementations of connecting to the Middle-End Proxy: non dramatically, but usual
@@ -103,8 +54,12 @@ We welcome ideas, architectural feedback, and pull requests.
 - [FAQ EN](docs/FAQ.en.md)
 
 ### Recognizability for DPI and crawler
-Since version 1.1.0.0, we have debugged masking perfectly: for all clients without "presenting" a key, 
-we transparently direct traffic to the target host!
+
+On April 1, 2026, we became aware of a method for detecting MTProxy Fake-TLS, 
+based on the ECH extension and the ordering of cipher suites, 
+as well as an overall unique JA3/JA4 fingerprint 
+that does not occur in modern browsers: 
+we have already submitted initial changes to the Telegram Desktop developers and are working on updates for other clients.
 
 - We consider this a breakthrough aspect, which has no stable analogues today
 - Based on this: if `telemt` configured correctly, **TLS mode is completely identical to real-life handshake + communication** with a specified host

docs/FAQ.en.md

@@ -1,107 +1,122 @@
-## How to set up "proxy sponsor" channel and statistics via @MTProxybot bot
+## How to set up a "proxy sponsor" channel and statistics via the @MTProxybot
 
-1. Go to @MTProxybot bot.
-2. Enter the command `/newproxy`
-3. Send the server IP and port. For example: 1.2.3.4:443
-4. Open the config `nano /etc/telemt.toml`.
-5. Copy and send the user secret from the [access.users] section to the bot.
-6. Copy the tag received from the bot. For example 1234567890abcdef1234567890abcdef.
+1. Go to the @MTProxybot.
+2. Enter the `/newproxy` command.
+3. Send your server's IP address and port. For example: `1.2.3.4:443`.
+4. Open the configuration file: `nano /etc/telemt/telemt.toml`.
+5. Copy and send the user secret from the `[access.users]` section to the bot.
+6. Copy the tag provided by the bot. For example: `1234567890abcdef1234567890abcdef`.
 > [!WARNING]
 > The link provided by the bot will not work. Do not copy or use it!
-7. Uncomment the ad_tag parameter and enter the tag received from the bot.
-8. Uncomment/add the parameter `use_middle_proxy = true`.
+7. Uncomment the `ad_tag` parameter and enter the tag received from the bot.
+8. Uncomment or add the `use_middle_proxy = true` parameter.
 
-Config example:
+Configuration example:
 ```toml
 [general]
 ad_tag = "1234567890abcdef1234567890abcdef"
 use_middle_proxy = true
 ```
-9. Save the config. Ctrl+S -> Ctrl+X.
-10. Restart telemt `systemctl restart telemt`.
-11. In the bot, send the command /myproxies and select the added server.
+9. Save the changes (in nano: Ctrl+S -> Ctrl+X).
+10. Restart the telemt service: `systemctl restart telemt`.
+11. Send the `/myproxies` command to the bot and select the added server.
 12. Click the "Set promotion" button.
 13. Send a **public link** to the channel. Private channels cannot be added!
-14. Wait approximately 1 hour for the information to update on Telegram servers.
+14. Wait for about 1 hour for the information to update on Telegram servers.
 > [!WARNING]
-> You will not see the "proxy sponsor" if you are already subscribed to the channel.
+> The sponsored channel will not be displayed to you if you are already subscribed to it.
 
-**You can also set up different channels for different users.**
+**You can also configure different sponsored channels for different users:**
 ```toml
 [access.user_ad_tags]
 hello = "ad_tag"
 hello2 = "ad_tag2"
 ```
 
-## How many people can use 1 link
+## Why do you need a middle proxy (ME)
+https://github.com/telemt/telemt/discussions/167
 
-By default, 1 link can be used by any number of people.  
-You can limit the number of IPs using the proxy.
+
+## How many people can use one link
+
+By default, an unlimited number of people can use a single link.  
+However, you can limit the number of unique IP addresses for each user:
 ```toml
 [access.user_max_unique_ips]
 hello = 1
 ```
-This parameter limits how many unique IPs can use 1 link simultaneously. If one user disconnects, a second user can connect. Also, multiple users can sit behind the same IP.
+This parameter sets the maximum number of unique IP addresses from which a single link can be used simultaneously. If the first user disconnects, a second one can connect. At the same time, multiple users can connect from a single IP address simultaneously (for example, devices on the same Wi-Fi network).
 
 ## How to create multiple different links
 
-1. Generate the required number of secrets `openssl rand -hex 16`
-2. Open the config `nano /etc/telemt.toml`
-3. Add new users.
+1. Generate the required number of secrets using the command: `openssl rand -hex 16`.
+2. Open the configuration file: `nano /etc/telemt/telemt.toml`.
+3. Add new users to the `[access.users]` section:
 ```toml
 [access.users]
 user1 = "00000000000000000000000000000001"
 user2 = "00000000000000000000000000000002"
 user3 = "00000000000000000000000000000003"
 ```
-4. Save the config. Ctrl+S -> Ctrl+X. You don't need to restart telemt.
-5. Get the links via
+4. Save the configuration (Ctrl+S -> Ctrl+X). There is no need to restart the telemt service.
+5. Get the ready-to-use links using the command:
 ```bash
 curl -s http://127.0.0.1:9091/v1/users | jq
 ```
 
+## "Unknown TLS SNI" error
+Usually, this error occurs if you have changed the `tls_domain` parameter, but users continue to connect using old links with the previous domain.
+
+If you need to allow connections with any domains (ignoring SNI mismatches), add the following parameters:
+```toml
+[censorship]
+unknown_sni_action = "mask"
+```
+
 ## How to view metrics
 
-1. Open the config `nano /etc/telemt.toml`
-2. Add the following parameters
+1. Open the configuration file: `nano /etc/telemt/telemt.toml`.
+2. Add the following parameters:
 ```toml
 [server]
 metrics_port = 9090
 metrics_whitelist = ["127.0.0.1/32", "::1/128", "0.0.0.0/0"]
 ```
-3. Save the config. Ctrl+S -> Ctrl+X.
-4. Metrics are available at SERVER_IP:9090/metrics.
+3. Save the changes (Ctrl+S -> Ctrl+X).
+4. After that, metrics will be available at: `SERVER_IP:9090/metrics`. 
 > [!WARNING]
-> "0.0.0.0/0" in metrics_whitelist opens access from any IP. Replace with your own IP. For example "1.2.3.4"
+> The value `"0.0.0.0/0"` in `metrics_whitelist` opens access to metrics from any IP address. It is recommended to replace it with your personal IP, for example: `"1.2.3.4/32"`.
 
 ## Additional parameters
 
-### Domain in link instead of IP
-To specify a domain in the links, add to the `[general.links]` section of the config file.
+### Domain in the link instead of IP
+To display a domain instead of an IP address in the connection links, add the following lines to the configuration file:
 ```toml
 [general.links]
 public_host = "proxy.example.com"
 ```
 
-### Server connection limit
-Limits the total number of open connections to the server:
+### Total server connection limit
+This parameter limits the total number of active connections to the server:
 ```toml
 [server]
 max_connections = 10000    # 0 - unlimited, 10000 - default
 ```
 
 ### Upstream Manager
-To specify an upstream, add to the `[[upstreams]]` section of the config.toml file:
-#### Binding to IP
+To configure outbound connections (upstreams), add the corresponding parameters to the `[[upstreams]]` section of the configuration file:
+
+#### Binding to an outbound IP address
 ```toml
 [[upstreams]]
 type = "direct"
 weight = 1
 enabled = true
-interface = "192.168.1.100" # Change to your outgoing IP
+interface = "192.168.1.100" # Replace with your outbound IP
 ```
-#### SOCKS4/5 as Upstream
-- Without authentication:
+
+#### Using SOCKS4/5 as an Upstream
+- Without authorization:
 ```toml
 [[upstreams]]
 type = "socks5"            # Specify SOCKS4 or SOCKS5
@@ -110,7 +125,7 @@ weight = 1                 # Set Weight for Scenarios
 enabled = true
 ```
 
-- With authentication:
+- With authorization:
 ```toml
 [[upstreams]]
 type = "socks5"            # Specify SOCKS4 or SOCKS5
@@ -121,8 +136,8 @@ weight = 1                 # Set Weight for Scenarios
 enabled = true
 ```
 
-#### Shadowsocks as Upstream
-Requires `use_middle_proxy = false`.
+#### Using Shadowsocks as an Upstream
+For this method to work, the `use_middle_proxy = false` parameter must be set.
 
 ```toml
 [general]

docs/FAQ.ru.md

@@ -1,106 +1,121 @@
 ## Как настроить канал "спонсор прокси" и статистику через бота @MTProxybot
 
-1. Зайти в бота @MTProxybot.
-2. Ввести команду `/newproxy`
-3. Отправить IP и порт сервера. Например: 1.2.3.4:443
-4. Открыть конфиг `nano /etc/telemt.toml`.
-5. Скопировать и отправить боту секрет пользователя из раздела [access.users].
-6. Скопировать полученный tag у бота. Например 1234567890abcdef1234567890abcdef.
+1. Зайдите в бота @MTProxybot.
+2. Введите команду `/newproxy`.
+3. Отправьте IP-адрес и порт сервера. Например: `1.2.3.4:443`.
+4. Откройте файл конфигурации: `nano /etc/telemt/telemt.toml`.
+5. Скопируйте и отправьте боту секрет пользователя из раздела `[access.users]`.
+6. Скопируйте тег (tag), который выдаст бот. Например: `1234567890abcdef1234567890abcdef`.
 > [!WARNING]
-> Ссылка, которую выдает бот, не будет работать. Не копируйте и не используйте её!
-7. Раскомментировать параметр ad_tag и вписать tag, полученный у бота.
-8. Раскомментировать/добавить параметр use_middle_proxy = true.
+> Ссылка, которую выдает бот, работать не будет. Не копируйте и не используйте её!
+7. Раскомментируйте параметр `ad_tag` и впишите тег, полученный от бота.
+8. Раскомментируйте или добавьте параметр `use_middle_proxy = true`.
 
-Пример конфига:
+Пример конфигурации:
 ```toml
 [general]
 ad_tag = "1234567890abcdef1234567890abcdef"
 use_middle_proxy = true
 ```
-9.  Сохранить конфиг. Ctrl+S -> Ctrl+X.
-10. Перезапустить telemt `systemctl restart telemt`.
-11. В боте отправить команду /myproxies и выбрать добавленный сервер.
-12. Нажать кнопку "Set promotion".
-13. Отправить **публичную ссылку** на канал. Приватный канал добавить нельзя!
-14. Подождать примерно 1 час, пока информация обновится на серверах Telegram.
+9. Сохраните изменения (в nano: Ctrl+S -> Ctrl+X).
+10. Перезапустите службу telemt: `systemctl restart telemt`.
+11. В боте отправьте команду `/myproxies` и выберите добавленный сервер.
+12. Нажмите кнопку «Set promotion».
+13. Отправьте **публичную ссылку** на канал. Приватные каналы добавлять нельзя!
+14. Подождите примерно 1 час, пока информация обновится на серверах Telegram.
 > [!WARNING]
-> У вас не будет отображаться "спонсор прокси" если вы уже подписаны на канал.
+> Спонсорский канал не будет у вас отображаться, если вы уже на него подписаны.
 
-**Также вы можете настроить разные каналы для разных пользователей.**
+**Вы также можете настроить разные спонсорские каналы для разных пользователей:**
 ```toml
 [access.user_ad_tags]
 hello = "ad_tag"
 hello2 = "ad_tag2"
 ```
 
-## Сколько человек может пользоваться 1 ссылкой
+## Зачем нужен middle proxy (ME)
+https://github.com/telemt/telemt/discussions/167
 
-По умолчанию 1 ссылкой может пользоваться сколько угодно человек.  
-Вы можете ограничить число IP, использующих прокси.
+
+## Сколько человек может пользоваться одной ссылкой
+
+По умолчанию одной ссылкой может пользоваться неограниченное число людей.  
+Однако вы можете ограничить количество уникальных IP-адресов для каждого пользователя:
 ```toml
 [access.user_max_unique_ips]
 hello = 1
 ```
-Этот параметр ограничивает, сколько уникальных IP может использовать 1 ссылку одновременно. Если один пользователь отключится, второй сможет подключиться. Также с одного IP может сидеть несколько пользователей.
+Этот параметр задает максимальное количество уникальных IP-адресов, с которых можно одновременно использовать одну ссылку. Если первый пользователь отключится, второй сможет подключиться. При этом с одного IP-адреса могут подключаться несколько пользователей одновременно (например, устройства в одной Wi-Fi сети).
 
-## Как сделать несколько разных ссылок
+## Как создать несколько разных ссылок
 
-1. Сгенерируйте нужное число секретов `openssl rand -hex 16`
-2. Открыть конфиг `nano /etc/telemt.toml`
-3. Добавить новых пользователей.
+1. Сгенерируйте необходимое количество секретов с помощью команды: `openssl rand -hex 16`.
+2. Откройте файл конфигурации: `nano /etc/telemt/telemt.toml`.
+3. Добавьте новых пользователей в секцию `[access.users]`:
 ```toml
 [access.users]
 user1 = "00000000000000000000000000000001"
 user2 = "00000000000000000000000000000002"
 user3 = "00000000000000000000000000000003"
 ```
-4. Сохранить конфиг. Ctrl+S -> Ctrl+X. Перезапускать telemt не нужно.
-5. Получить ссылки через
+4. Сохраните конфигурацию (Ctrl+S -> Ctrl+X). Перезапускать службу telemt не нужно.
+5. Получите готовые ссылки с помощью команды:
 ```bash
 curl -s http://127.0.0.1:9091/v1/users | jq
 ```
 
+## Ошибка "Unknown TLS SNI"
+Обычно эта ошибка возникает, если вы изменили параметр `tls_domain`, но пользователи продолжают подключаться по старым ссылкам с прежним доменом.
+
+Если необходимо разрешить подключение с любыми доменами (игнорируя несовпадения SNI), добавьте следующие параметры:
+```toml
+[censorship]
+unknown_sni_action = "mask"
+```
+
 ## Как посмотреть метрики
 
-1. Открыть конфиг `nano /etc/telemt.toml`
-2. Добавить следующие параметры
+1. Откройте файл конфигурации: `nano /etc/telemt/telemt.toml`.
+2. Добавьте следующие параметры:
 ```toml
 [server]
 metrics_port = 9090
 metrics_whitelist = ["127.0.0.1/32", "::1/128", "0.0.0.0/0"]
 ```
-3. Сохранить конфиг. Ctrl+S -> Ctrl+X.
-4. Метрики доступны по адресу SERVER_IP:9090/metrics. 
+3. Сохраните изменения (Ctrl+S -> Ctrl+X).
+4. После этого метрики будут доступны по адресу: `SERVER_IP:9090/metrics`. 
 > [!WARNING]
-> "0.0.0.0/0" в metrics_whitelist открывает доступ с любого IP. Замените на свой ip. Например "1.2.3.4"
+> Значение `"0.0.0.0/0"` в `metrics_whitelist` открывает доступ к метрикам с любого IP-адреса. Рекомендуется заменить его на ваш личный IP, например: `"1.2.3.4/32"`.
 
 ## Дополнительные параметры
 
 ### Домен в ссылке вместо IP
-Чтобы указать домен в ссылках, добавьте в секцию `[general.links]` файла config.
+Чтобы в ссылках для подключения отображался домен вместо IP-адреса, добавьте следующие строки в файл конфигурации:
 ```toml
 [general.links]
 public_host = "proxy.example.com"
 ```
 
 ### Общий лимит подключений к серверу
-Ограничивает общее число открытых подключений к серверу:
+Этот параметр ограничивает общее количество активных подключений к серверу:
 ```toml
 [server]
-max_connections = 10000    # 0 - unlimited, 10000 - default
+max_connections = 10000    # 0 - без ограничений, 10000 - по умолчанию
 ```
 
 ### Upstream Manager
-Чтобы указать апстрим, добавьте в секцию `[[upstreams]]` файла config.toml:
-#### Привязка к IP
+Для настройки исходящих подключений (апстримов) добавьте соответствующие параметры в секцию `[[upstreams]]` файла конфигурации:
+
+#### Привязка к исходящему IP-адресу
 ```toml
 [[upstreams]]
 type = "direct"
 weight = 1
 enabled = true
-interface = "192.168.1.100" # Change to your outgoing IP
+interface = "192.168.1.100" # Замените на ваш исходящий IP
 ```
-#### SOCKS4/5 как Upstream
+
+#### Использование SOCKS4/5 в качестве Upstream
 - Без авторизации:
 ```toml
 [[upstreams]]
@@ -121,8 +136,8 @@ weight = 1                 # Set Weight for Scenarios
 enabled = true
 ```
 
-#### Shadowsocks как Upstream
-Требует `use_middle_proxy = false`.
+#### Использование Shadowsocks в качестве Upstream
+Для работы этого метода требуется установить параметр `use_middle_proxy = false`.
 
 ```toml
 [general]

docs/QUICK_START_GUIDE.en.md

@@ -27,12 +27,12 @@ chmod +x /bin/telemt
 
 **0. Check port and generate secrets**
 
-The port you have selected for use should be MISSING from the list, when:
+The port you have selected for use should not be in the list:
 ```bash
 netstat -lnp
 ```
 
-Generate 16 bytes/32 characters HEX with OpenSSL or another way:
+Generate 16 bytes/32 characters in HEX format with OpenSSL or another way:
 ```bash
 openssl rand -hex 16
 ```
@@ -50,7 +50,7 @@ Save the obtained result somewhere. You will need it later!
 
 **1. Place your config to /etc/telemt/telemt.toml**
 
-Create config directory:
+Create the config directory:
 ```bash
 mkdir /etc/telemt
 ```
@@ -59,7 +59,7 @@ Open nano
 ```bash
 nano /etc/telemt/telemt.toml
 ```
-paste your config
+Insert your configuration:
 
 ```toml
 # === General Settings ===
@@ -93,8 +93,9 @@ hello = "00000000000000000000000000000000"
 then Ctrl+S -> Ctrl+X to save
 
 > [!WARNING]
-> Replace the value of the hello parameter with the value you obtained in step 0. 
-> Replace the value of the tls_domain parameter with another website.
+> Replace the value of the hello parameter with the value you obtained in step 0.  
+> Additionally, change the value of the tls_domain parameter to a different website.
+> Changing the tls_domain parameter will break all links that use the old domain!
 
 ---
 
@@ -105,14 +106,14 @@ useradd -d /opt/telemt -m -r -U telemt
 chown -R telemt:telemt /etc/telemt
 ```
 
-**3. Create service on /etc/systemd/system/telemt.service**
+**3. Create service in /etc/systemd/system/telemt.service**
 
 Open nano
 ```bash
 nano /etc/systemd/system/telemt.service
 ```
 
-paste this Systemd Module
+Insert this Systemd module:
 ```bash
 [Unit]
 Description=Telemt
@@ -147,13 +148,16 @@ systemctl daemon-reload
 
 **6.** For automatic startup at system boot, enter `systemctl enable telemt`
 
-**7.** To get the link(s), enter
+**7.** To get the link(s), enter:
 ```bash
 curl -s http://127.0.0.1:9091/v1/users | jq
 ```
 
 > Any number of people can use one link.
 
+> [!WARNING]
+> Only the command from step 7 can provide a working link. Do not try to create it yourself or copy it from anywhere if you are not sure what you are doing!
+
 ---
 
 # Telemt via Docker Compose

docs/QUICK_START_GUIDE.ru.md

@@ -95,6 +95,7 @@ hello = "00000000000000000000000000000000"
 > [!WARNING]
 > Замените значение параметра hello на значение, которое вы получили в пункте 0.  
 > Так же замените значение параметра tls_domain на другой сайт.
+> Изменение параметра tls_domain сделает нерабочими все ссылки, использующие старый домен!
 
 ---
 

docs/VPS_DOUBLE_HOP.en.md

@@ -0,0 +1,287 @@
+<img src="https://gist.githubusercontent.com/avbor/1f8a128e628f47249aae6e058a57610b/raw/19013276c035e91058e0a9799ab145f8e70e3ff5/scheme.svg">
+
+## Concept
+- **Server A** (__conditionally Russian Federation_):\
+  Entry point, receives Telegram proxy user traffic via **HAProxy** (port `443`)\
+  and sends it to the tunnel to Server **B**.\
+  Internal IP in the tunnel — `10.10.10.2`\
+  Port for HAProxy clients — `443\tcp`
+- **Server B** (_conditionally Netherlands_):\
+  Exit point, runs **telemt** and accepts client connections through Server **A**.\
+  The server must have unrestricted access to Telegram servers.\
+  Internal IP in the tunnel — `10.10.10.1`\
+  AmneziaWG port — `8443\udp`\
+  Port for telemt clients — `443\tcp`
+
+---
+
+## Step 1. Setting up the AmneziaWG tunnel (A <-> B)
+[AmneziaWG](https://github.com/amnezia-vpn/amneziawg-linux-kernel-module) must be installed on all servers.\
+All following commands are given for **Ubuntu 24.04**.\
+For RHEL-based distributions, installation instructions are available at the link above.
+
+### Installing AmneziaWG (Servers A and B)
+The following steps must be performed on each server:
+
+#### 1. Adding the AmneziaWG repository and installing required packages:
+```bash
+sudo apt install -y software-properties-common python3-launchpadlib gnupg2 linux-headers-$(uname -r) && \
+sudo add-apt-repository ppa:amnezia/ppa && \
+sudo apt-get install -y amneziawg
+```
+
+#### 2. Generating a unique key pair:
+```bash
+cd /etc/amnezia/amneziawg && \
+awg genkey | tee private.key | awg pubkey > public.key
+```
+
+As a result, you will get two files in the `/etc/amnezia/amneziawg` folder:\
+`private.key` - private, and\
+`public.key` - public server keys
+
+#### 3. Configuring network interfaces:
+Obfuscation parameters `S1`, `S2`, `H1`, `H2`, `H3`, `H4` must be strictly identical on both servers.\
+Parameters `Jc`, `Jmin` and `Jmax` can differ.\
+Parameters `I1-I5` ([Custom Protocol Signature](https://docs.amnezia.org/documentation/amnezia-wg/)) must be specified on the client side (Server **A**).
+
+Recommendations for choosing values:
+
+```text
+Jc — 1 ≤ Jc ≤ 128; from 4 to 12 inclusive
+Jmin — Jmax > Jmin < 1280*; recommended 8
+Jmax — Jmin < Jmax ≤ 1280*; recommended 80
+S1 — S1 ≤ 1132* (1280* - 148 = 1132); S1 + 56 ≠ S2;
+recommended range from 15 to 150 inclusive
+S2 — S2 ≤ 1188* (1280* - 92 = 1188);
+recommended range from 15 to 150 inclusive
+H1/H2/H3/H4 — must be unique and differ from each other;
+recommended range from 5 to 2147483647 inclusive
+
+* It is assumed that the Internet connection has an MTU of 1280.
+```
+
+> [!IMPORTANT]
+> It is recommended to use your own, unique values.\
+> You can use the [generator](https://htmlpreview.github.io/?https://gist.githubusercontent.com/avbor/955782b5c37b06240b243aa375baeac5/raw/13f5517ca473b47c412b9a99407066de973732bd/awg-gen.html) to select parameters.
+
+#### Server B Configuration (Netherlands):
+
+Create the interface configuration file (`awg0`)
+```bash
+nano /etc/amnezia/amneziawg/awg0.conf
+```
+
+File content
+```ini
+[Interface]
+Address = 10.10.10.1/24
+ListenPort = 8443
+PrivateKey = <PRIVATE_KEY_SERVER_B>
+SaveConfig = true
+Jc = 4
+Jmin = 8
+Jmax = 80
+S1 = 29
+S2 = 15
+S3 = 18
+S4 = 0
+H1 = 2087563914
+H2 = 188817757
+H3 = 101784570
+H4 = 432174303
+
+[Peer]
+PublicKey = <PUBLIC_KEY_SERVER_A>
+AllowedIPs = 10.10.10.2/32
+```
+`ListenPort` - the port on which the server will wait for connections, you can choose any free one.\
+`<PRIVATE_KEY_SERVER_B>` - the content of the `private.key` file from Server **B**.\
+`<PUBLIC_KEY_SERVER_A>` - the content of the `public.key` file from Server **A**.
+
+Open the port on the firewall (if enabled):
+```bash
+sudo ufw allow from <PUBLIC_IP_SERVER_A> to any port 8443 proto udp
+```
+
+`<PUBLIC_IP_SERVER_A>` - the external IP address of Server **A**.
+
+#### Server A Configuration (Russian Federation):
+Create the interface configuration file (awg0)
+
+```bash
+nano /etc/amnezia/amneziawg/awg0.conf
+```
+
+File content
+```ini
+[Interface]
+Address = 10.10.10.2/24
+PrivateKey = <PRIVATE_KEY_SERVER_A>
+Jc = 4
+Jmin = 8
+Jmax = 80
+S1 = 29
+S2 = 15
+S3 = 18
+S4 = 0
+H1 = 2087563914
+H2 = 188817757
+H3 = 101784570
+H4 = 432174303
+I1 = <b 0xc10000000108981eba846e21f74e00>
+I2 = <b 0xc20000000108981eba846e21f74e00>
+I3 = <b 0xc30000000108981eba846e21f74e00>
+I4 = <b 0x43981eba846e21f74e>
+I5 = <b 0x43981eba846e21f74e>
+
+[Peer]
+PublicKey = <PUBLIC_KEY_SERVER_B>
+Endpoint = <PUBLIC_IP_SERVER_B>:8443
+AllowedIPs = 10.10.10.1/32
+PersistentKeepalive = 25
+```
+
+`<PRIVATE_KEY_SERVER_A>` - the content of the `private.key` file from Server **A**.\
+`<PUBLIC_KEY_SERVER_B>` - the content of the `public.key` file from Server **B**.\
+`<PUBLIC_IP_SERVER_B>` - the public IP address of Server **B**.
+
+Enable the tunnel on both servers:
+```bash
+sudo systemctl enable --now awg-quick@awg0
+```
+
+Make sure Server B is accessible from Server A through the tunnel.
+```bash
+ping 10.10.10.1
+PING 10.10.10.1 (10.10.10.1) 56(84) bytes of data.
+64 bytes from 10.10.10.1: icmp_seq=1 ttl=64 time=35.1 ms
+64 bytes from 10.10.10.1: icmp_seq=2 ttl=64 time=35.0 ms
+64 bytes from 10.10.10.1: icmp_seq=3 ttl=64 time=35.1 ms
+^C
+```
+---
+
+## Step 2. Installing telemt on Server B (conditionally Netherlands)
+Installation and configuration are described [here](https://github.com/telemt/telemt/blob/main/docs/QUICK_START_GUIDE.ru.md) or [here](https://gitlab.com/An0nX/telemt-docker#-quick-start-docker-compose).\
+It is assumed that telemt expects connections on port `443\tcp`.
+
+In the telemt config, you must enable the `Proxy` protocol and restrict connections to it only through the tunnel.
+```toml
+[server]
+port = 443
+listen_addr_ipv4 = "10.10.10.1"
+proxy_protocol = true
+```
+
+Also, for correct link generation, specify the FQDN or IP address and port of Server `A`
+```toml
+[general.links]
+show = "*"
+public_host = "<FQDN_OR_IP_SERVER_A>"
+public_port = 443
+```
+
+Open the port on the firewall (if enabled):
+```bash
+sudo ufw allow from 10.10.10.2 to any port 443 proto tcp
+```
+
+---
+
+## Step 3. Configuring HAProxy on Server A (Russian Federation)
+Since the version in the standard Ubuntu repository is relatively old, it makes sense to use the official Docker image.\
+[Instructions](https://docs.docker.com/engine/install/ubuntu/) for installing Docker on Ubuntu.
+
+> [!WARNING]
+> By default, regular users do not have rights to use ports < 1024.
+> Attempts to run HAProxy on port 443 can lead to errors:
+> ```
+> [ALERT] (8) : Binding [/usr/local/etc/haproxy/haproxy.cfg:17] for frontend tcp_in_443: 
+> protocol tcpv4: cannot bind socket (Permission denied) for [0.0.0.0:443].
+> ```
+> There are two simple ways to bypass this restriction, choose one:
+> 1. At the OS level, change the net.ipv4.ip_unprivileged_port_start setting to allow users to use all ports:
+> ```
+> echo "net.ipv4.ip_unprivileged_port_start = 0" | sudo tee -a /etc/sysctl.conf && sudo sysctl -p
+> ```
+> or
+> 
+> 2. Run HAProxy as root:
+> Uncomment the `user: "root"` parameter in docker-compose.yaml.
+
+#### Create a folder for HAProxy:
+```bash
+mkdir -p /opt/docker-compose/haproxy && cd $_
+```
+
+#### Create the docker-compose.yaml file
+`nano docker-compose.yaml`
+
+File content
+```yaml
+services:
+  haproxy:
+    image: haproxy:latest
+    container_name: haproxy
+    restart: unless-stopped
+    # user: "root"
+    network_mode: "host"
+    volumes:
+      - ./haproxy.cfg:/usr/local/etc/haproxy/haproxy.cfg:ro
+    logging:
+      driver: "json-file"
+      options:
+        max-size: "1m"
+        max-file: "1"
+```
+
+#### Create the haproxy.cfg config file
+Accept connections on port 443\tcp and send them through the tunnel to Server `B` 10.10.10.1:443
+
+`nano haproxy.cfg`
+
+File content
+
+```haproxy
+global
+    log stdout format raw local0
+    maxconn 10000
+
+defaults
+    log     global
+    mode    tcp
+    option  tcplog
+    option  clitcpka
+    option  srvtcpka
+    timeout connect 5s
+    timeout client  2h
+    timeout server  2h
+    timeout check   5s
+
+frontend tcp_in_443
+    bind *:443
+    maxconn 8000
+    option tcp-smart-accept
+    default_backend telemt_nodes
+
+backend telemt_nodes
+    option tcp-smart-connect
+    server server_a 10.10.10.1:443 check inter 5s rise 2 fall 3 send-proxy-v2
+
+
+```
+> [!WARNING]
+> **The file must end with an empty line, otherwise HAProxy will not start!**
+
+#### Allow port 443\tcp in the firewall (if enabled)
+```bash
+sudo ufw allow 443/tcp
+```
+
+#### Start the HAProxy container
+```bash
+docker compose up -d
+```
+
+If everything is configured correctly, you can now try connecting Telegram clients using links from the telemt log\api.

docs/VPS_DOUBLE_HOP.ru.md

@@ -0,0 +1,291 @@
+<img src="https://gist.githubusercontent.com/avbor/1f8a128e628f47249aae6e058a57610b/raw/19013276c035e91058e0a9799ab145f8e70e3ff5/scheme.svg">
+
+## Концепция
+- **Сервер A** (_РФ_):\
+  Точка входа, принимает трафик пользователей Telegram-прокси через **HAProxy** (порт `443`)\
+  и отправляет в туннель на Сервер **B**.\
+  Внутренний IP в туннеле — `10.10.10.2`\
+  Порт для клиентов HAProxy — `443\tcp`
+- **Сервер B** (_условно Нидерланды_):\
+  Точка выхода, на нем работает **telemt** и принимает подключения клиентов через Сервер **A**.\
+  На сервере должен быть неограниченный доступ до серверов Telegram.\
+  Внутренний IP в туннеле — `10.10.10.1`\
+  Порт AmneziaWG — `8443\udp`\
+  Порт для клиентов telemt — `443\tcp`
+
+---
+
+## Шаг 1. Настройка туннеля AmneziaWG (A <-> B)
+
+На всех серверах необходимо установить [amneziawg](https://github.com/amnezia-vpn/amneziawg-linux-kernel-module).\
+Далее все команды даны для **Ununtu 24.04**.\
+Для RHEL-based дистрибутивов инструкция по установке есть по ссылке выше.
+
+### Установка AmneziaWG (Сервера A и B)
+На каждом из серверов необходимо выполнить следующие шаги:
+
+#### 1. Добавление репозитория AmneziaWG и установка необходимых пакетов:
+```bash
+sudo apt install -y software-properties-common python3-launchpadlib gnupg2 linux-headers-$(uname -r) && \
+sudo add-apt-repository ppa:amnezia/ppa && \
+sudo apt-get install -y amneziawg
+```
+ 
+#### 2. Генерация уникальной пары ключей:
+```bash
+cd /etc/amnezia/amneziawg && \
+awg genkey | tee private.key | awg pubkey > public.key
+```
+В результате вы получите в папке `/etc/amnezia/amneziawg` два файла:\
+`private.key` - приватный и\
+`public.key` - публичный ключи сервера
+
+#### 3. Настройка сетевых интерфейсов:
+
+Параметры обфускации `S1`, `S2`, `H1`, `H2`, `H3`, `H4` должны быть строго идентичными на обоих серверах.\
+Параметры `Jc`, `Jmin` и `Jmax` могут отличатся.\
+Параметры `I1-I5` ([Custom Protocol Signature](https://docs.amnezia.org/documentation/amnezia-wg/)) нужно указывать на стороне _клиента_ (Сервер **А**).
+
+Рекомендации по выбору значений:
+```text
+Jc — 1 ≤ Jc ≤ 128; от 4 до 12 включительно
+Jmin — Jmax > Jmin < 1280*; рекомендовано 8
+Jmax — Jmin < Jmax ≤ 1280*; рекомендовано 80
+S1 — S1 ≤ 1132* (1280* - 148 = 1132); S1 + 56 ≠ S2;
+рекомендованный диапазон от 15 до 150 включительно
+S2 — S2 ≤ 1188* (1280* - 92 = 1188);
+рекомендованный диапазон от 15 до 150 включительно
+H1/H2/H3/H4 — должны быть уникальны и отличаться друг от друга;
+рекомендованный диапазон от 5 до 2147483647 включительно
+
+* Предполагается, что подключение к Интернету имеет MTU 1280.
+```
+> [!IMPORTANT]
+> Рекомендуется использовать собственные, уникальные значения.\
+> Для выбора параметров можете воспользоваться [генератором](https://htmlpreview.github.io/?https://gist.githubusercontent.com/avbor/955782b5c37b06240b243aa375baeac5/raw/13f5517ca473b47c412b9a99407066de973732bd/awg-gen.html).
+
+#### Конфигурация Сервера B (_Нидерланды_):
+
+Создаем файл конфигурации интерфейса (`awg0`)
+```bash
+nano /etc/amnezia/amneziawg/awg0.conf
+```
+
+Содержимое файла
+```ini
+[Interface]
+Address = 10.10.10.1/24
+ListenPort = 8443
+PrivateKey = <PRIVATE_KEY_SERVER_B>
+SaveConfig = true
+Jc = 4
+Jmin = 8
+Jmax = 80
+S1 = 29
+S2 = 15
+S3 = 18
+S4 = 0
+H1 = 2087563914
+H2 = 188817757
+H3 = 101784570
+H4 = 432174303
+
+[Peer]
+PublicKey = <PUBLIC_KEY_SERVER_A>
+AllowedIPs = 10.10.10.2/32
+```
+
+`ListenPort` - порт, на котором сервер будет ждать подключения, можете выбрать любой свободный.\
+`<PRIVATE_KEY_SERVER_B>` - содержимое файла `private.key` с сервера **B**.\
+`<PUBLIC_KEY_SERVER_A>` - содержимое файла `public.key` с сервера **A**.
+
+Открываем порт на фаерволе (если включен):
+```bash
+sudo ufw allow from <PUBLIC_IP_SERVER_A> to any port 8443 proto udp
+```
+
+`<PUBLIC_IP_SERVER_A>` - внешний IP адрес Сервера **A**.
+
+#### Конфигурация Сервера A (_РФ_):
+
+Создаем файл конфигурации интерфейса (`awg0`)
+```bash
+nano /etc/amnezia/amneziawg/awg0.conf
+```
+
+Содержимое файла
+```ini
+[Interface]
+Address = 10.10.10.2/24
+PrivateKey = <PRIVATE_KEY_SERVER_A>
+Jc = 4
+Jmin = 8
+Jmax = 80
+S1 = 29
+S2 = 15
+S3 = 18
+S4 = 0
+H1 = 2087563914
+H2 = 188817757
+H3 = 101784570
+H4 = 432174303
+I1 = <b 0xc10000000108981eba846e21f74e00>
+I2 = <b 0xc20000000108981eba846e21f74e00>
+I3 = <b 0xc30000000108981eba846e21f74e00>
+I4 = <b 0x43981eba846e21f74e>
+I5 = <b 0x43981eba846e21f74e>
+
+[Peer]
+PublicKey = <PUBLIC_KEY_SERVER_B>
+Endpoint = <PUBLIC_IP_SERVER_B>:8443
+AllowedIPs = 10.10.10.1/32
+PersistentKeepalive = 25
+```
+
+`<PRIVATE_KEY_SERVER_A>` - содержимое файла `private.key` с сервера **A**.\
+`<PUBLIC_KEY_SERVER_B>` - содержимое файла `public.key` с сервера **B**.\
+`<PUBLIC_IP_SERVER_B>` - публичный IP адресс сервера **B**.
+
+#### Включаем туннель на обоих серверах:
+```bash
+sudo systemctl enable --now awg-quick@awg0 
+```
+
+Убедитесь, что с Сервера `A` доступен Сервер `B` через туннель.
+```bash
+ping 10.10.10.1
+PING 10.10.10.1 (10.10.10.1) 56(84) bytes of data.
+64 bytes from 10.10.10.1: icmp_seq=1 ttl=64 time=35.1 ms
+64 bytes from 10.10.10.1: icmp_seq=2 ttl=64 time=35.0 ms
+64 bytes from 10.10.10.1: icmp_seq=3 ttl=64 time=35.1 ms
+^C
+
+```
+
+---
+
+## Шаг 2. Установка telemt на Сервере B (_условно Нидерланды_)
+
+Установка и настройка описаны [здесь](https://github.com/telemt/telemt/blob/main/docs/QUICK_START_GUIDE.ru.md) или [здесь](https://gitlab.com/An0nX/telemt-docker#-quick-start-docker-compose).\
+Подразумевается что telemt ожидает подключения на порту `443\tcp`.
+
+В конфиге telemt необходимо включить протокол `Proxy` и ограничить подключения к нему только через туннель.
+
+```toml
+[server]
+port = 443
+listen_addr_ipv4 = "10.10.10.1"
+proxy_protocol = true
+```
+
+А также, для правильной генерации ссылок, указать FQDN или IP адрес и порт Сервера `A` 
+
+```toml
+[general.links]
+show = "*"
+public_host = "<FQDN_OR_IP_SERVER_A>"
+public_port = 443
+```
+
+Открываем порт на фаерволе (если включен):
+```bash
+sudo ufw allow from 10.10.10.2 to any port 443 proto tcp
+```
+
+---
+
+### Шаг 3. Настройка HAProxy на Сервере A (_РФ_)
+
+Т.к. в стандартном репозитории Ubuntu версия относительно старая, имеет смысл воспользоваться официальным образом Docker.\
+[Инструкция](https://docs.docker.com/engine/install/ubuntu/) по установке Docker на Ubuntu.
+
+> [!WARNING]
+> По умолчанию у обычных пользователей нет прав на использование портов < 1024.\
+> Попытки запустить HAProxy на 443 порту могут приводить к ошибкам:
+> ```
+> [ALERT] (8) : Binding [/usr/local/etc/haproxy/haproxy.cfg:17] for frontend tcp_in_443: 
+> protocol tcpv4: cannot bind socket (Permission denied) for [0.0.0.0:443].
+> ```
+> Есть два простых способа обойти это ограничение, выберите что-то одно:
+> 1. На уровне ОС изменить настройку net.ipv4.ip_unprivileged_port_start, разрешив пользователям использовать все порты:
+> ```
+> echo "net.ipv4.ip_unprivileged_port_start = 0" | sudo tee -a /etc/sysctl.conf && sudo sysctl -p
+> ```
+> или
+> 
+> 2. Запустить HAProxy под root:\
+> Раскомментируйте в docker-compose.yaml параметр `user: "root"`.
+
+#### Создаем папку для HAProxy:
+```bash
+mkdir -p /opt/docker-compose/haproxy && cd $_
+```
+#### Создаем файл docker-compose.yaml
+
+`nano docker-compose.yaml`
+
+Содержимое файла
+```yaml
+services:
+  haproxy:
+    image: haproxy:latest
+    container_name: haproxy
+    restart: unless-stopped
+    # user: "root"
+    network_mode: "host"
+    volumes:
+      - ./haproxy.cfg:/usr/local/etc/haproxy/haproxy.cfg:ro
+    logging:
+      driver: "json-file"
+      options:
+        max-size: "1m"
+        max-file: "1"
+```
+#### Создаем файл конфига haproxy.cfg
+Принимаем подключения на порту 443\tcp и отправляем их через туннель на Сервер `B` 10.10.10.1:443
+
+`nano haproxy.cfg`
+
+Содержимое файла
+```haproxy
+global
+    log stdout format raw local0
+    maxconn 10000
+
+defaults
+    log     global
+    mode    tcp
+    option  tcplog
+    option  clitcpka
+    option  srvtcpka
+    timeout connect 5s
+    timeout client  2h
+    timeout server  2h
+    timeout check   5s
+
+frontend tcp_in_443
+    bind *:443
+    maxconn 8000
+    option tcp-smart-accept
+    default_backend telemt_nodes
+
+backend telemt_nodes
+    option tcp-smart-connect
+    server server_a 10.10.10.1:443 check inter 5s rise 2 fall 3 send-proxy-v2
+
+
+```
+>[!WARNING]
+>**Файл должен заканчиваться пустой строкой, иначе HAProxy не запустится!**
+
+#### Разрешаем порт 443\tcp в фаерволе (если включен)
+```bash
+sudo ufw allow 443/tcp
+```
+
+#### Запускаем контейнер HAProxy
+```bash
+docker compose up -d
+```
+
+Если все настроено верно, то теперь можно пробовать подключить клиентов Telegram с использованием ссылок из лога\api telemt.

install.sh

@@ -20,6 +20,13 @@ TARGET_VERSION="${VERSION:-latest}"
 while [ $# -gt 0 ]; do
     case "$1" in
         -h|--help) ACTION="help"; shift ;;
+        -d|--domain)
+            if [ "$#" -lt 2 ] || [ -z "$2" ]; then
+                printf '[ERROR] %s requires a domain argument.\n' "$1" >&2
+                exit 1
+            fi
+            TLS_DOMAIN="$2"
+            shift 2 ;;
         uninstall|--uninstall)
             if [ "$ACTION" != "purge" ]; then ACTION="uninstall"; fi
             shift ;;
@@ -52,11 +59,12 @@ cleanup() {
 trap cleanup EXIT INT TERM
 
 show_help() {
-    say "Usage: $0 [ <version> | install | uninstall | purge | --help ]"
+    say "Usage: $0 [ <version> | install | uninstall | purge ] [ -d <domain> ] [ --help ]"
     say "  <version>    Install specific version (e.g. 3.3.15, default: latest)"
     say "  install      Install the latest version"
     say "  uninstall    Remove the binary and service (keeps config and user)"
     say "  purge        Remove everything including configuration, data, and user"
+    say "  -d, --domain Set TLS domain (default: petrovich.ru)"
     exit 0
 }
 
@@ -192,7 +200,13 @@ verify_install_deps() {
 detect_arch() {
     sys_arch="$(uname -m)"
     case "$sys_arch" in
-        x86_64|amd64) echo "x86_64" ;;
+        x86_64|amd64)
+            if [ -r /proc/cpuinfo ] && grep -q "avx2" /proc/cpuinfo 2>/dev/null && grep -q "bmi2" /proc/cpuinfo 2>/dev/null; then
+                echo "x86_64-v3"
+            else
+                echo "x86_64"
+            fi
+            ;;
         aarch64|arm64) echo "aarch64" ;;
         *) die "Unsupported architecture: $sys_arch" ;;
     esac
@@ -500,7 +514,21 @@ case "$ACTION" in
             die "Temp directory is invalid or was not created"
         fi
 
-        fetch_file "$DL_URL" "${TEMP_DIR}/${FILE_NAME}" || die "Download failed"
+        if ! fetch_file "$DL_URL" "${TEMP_DIR}/${FILE_NAME}"; then
+            if [ "$ARCH" = "x86_64-v3" ]; then
+                say "  -> x86_64-v3 build not found, falling back to standard x86_64..."
+                ARCH="x86_64"
+                FILE_NAME="${BIN_NAME}-${ARCH}-linux-${LIBC}.tar.gz"
+                if [ "$TARGET_VERSION" = "latest" ]; then
+                    DL_URL="https://github.com/${REPO}/releases/latest/download/${FILE_NAME}"
+                else 
+                    DL_URL="https://github.com/${REPO}/releases/download/${TARGET_VERSION}/${FILE_NAME}"
+                fi
+                fetch_file "$DL_URL" "${TEMP_DIR}/${FILE_NAME}" || die "Download failed"
+            else
+                die "Download failed"
+            fi
+        fi
 
         say ">>> Stage 3: Extracting archive"
         if ! gzip -dc "${TEMP_DIR}/${FILE_NAME}" | tar -xf - -C "$TEMP_DIR" 2>/dev/null; then

src/api/mod.rs

@@ -37,11 +37,12 @@ mod runtime_watch;
 mod runtime_zero;
 mod users;
 
-use config_store::{current_revision, parse_if_match};
+use config_store::{current_revision, load_config_from_disk, parse_if_match};
 use events::ApiEventStore;
 use http_utils::{error_response, read_json, read_optional_json, success_response};
 use model::{
-    ApiFailure, CreateUserRequest, HealthData, PatchUserRequest, RotateSecretRequest, SummaryData,
+    ApiFailure, CreateUserRequest, DeleteUserResponse, HealthData, PatchUserRequest,
+    RotateSecretRequest, SummaryData, UserActiveIps,
 };
 use runtime_edge::{
     EdgeConnectionsCacheEntry, build_runtime_connections_summary_data,
@@ -362,15 +363,33 @@ async fn handle(
                 );
                 Ok(success_response(StatusCode::OK, data, revision))
             }
+            ("GET", "/v1/stats/users/active-ips") => {
+                let revision = current_revision(&shared.config_path).await?;
+                let usernames: Vec<_> = cfg.access.users.keys().cloned().collect();
+                let active_ips_map = shared.ip_tracker.get_active_ips_for_users(&usernames).await;
+                let mut data: Vec<UserActiveIps> = active_ips_map
+                    .into_iter()
+                    .filter(|(_, ips)| !ips.is_empty())
+                    .map(|(username, active_ips)| UserActiveIps {
+                        username,
+                        active_ips,
+                    })
+                    .collect();
+                data.sort_by(|a, b| a.username.cmp(&b.username));
+                Ok(success_response(StatusCode::OK, data, revision))
+            }
             ("GET", "/v1/stats/users") | ("GET", "/v1/users") => {
                 let revision = current_revision(&shared.config_path).await?;
+                let disk_cfg = load_config_from_disk(&shared.config_path).await?;
+                let runtime_cfg = config_rx.borrow().clone();
                 let (detected_ip_v4, detected_ip_v6) = shared.detected_link_ips();
                 let users = users_from_config(
-                    &cfg,
+                    &disk_cfg,
                     &shared.stats,
                     &shared.ip_tracker,
                     detected_ip_v4,
                     detected_ip_v6,
+                    Some(runtime_cfg.as_ref()),
                 )
                 .await;
                 Ok(success_response(StatusCode::OK, users, revision))
@@ -389,7 +408,7 @@ async fn handle(
                 let expected_revision = parse_if_match(req.headers());
                 let body = read_json::<CreateUserRequest>(req.into_body(), body_limit).await?;
                 let result = create_user(body, expected_revision, &shared).await;
-                let (data, revision) = match result {
+                let (mut data, revision) = match result {
                     Ok(ok) => ok,
                     Err(error) => {
                         shared
@@ -398,11 +417,18 @@ async fn handle(
                         return Err(error);
                     }
                 };
+                let runtime_cfg = config_rx.borrow().clone();
+                data.user.in_runtime = runtime_cfg.access.users.contains_key(&data.user.username);
                 shared.runtime_events.record(
                     "api.user.create.ok",
                     format!("username={}", data.user.username),
                 );
-                Ok(success_response(StatusCode::CREATED, data, revision))
+                let status = if data.user.in_runtime {
+                    StatusCode::CREATED
+                } else {
+                    StatusCode::ACCEPTED
+                };
+                Ok(success_response(status, data, revision))
             }
             _ => {
                 if let Some(user) = path.strip_prefix("/v1/users/")
@@ -411,13 +437,16 @@ async fn handle(
                 {
                     if method == Method::GET {
                         let revision = current_revision(&shared.config_path).await?;
+                        let disk_cfg = load_config_from_disk(&shared.config_path).await?;
+                        let runtime_cfg = config_rx.borrow().clone();
                         let (detected_ip_v4, detected_ip_v6) = shared.detected_link_ips();
                         let users = users_from_config(
-                            &cfg,
+                            &disk_cfg,
                             &shared.stats,
                             &shared.ip_tracker,
                             detected_ip_v4,
                             detected_ip_v6,
+                            Some(runtime_cfg.as_ref()),
                         )
                         .await;
                         if let Some(user_info) =
@@ -445,7 +474,7 @@ async fn handle(
                         let body =
                             read_json::<PatchUserRequest>(req.into_body(), body_limit).await?;
                         let result = patch_user(user, body, expected_revision, &shared).await;
-                        let (data, revision) = match result {
+                        let (mut data, revision) = match result {
                             Ok(ok) => ok,
                             Err(error) => {
                                 shared.runtime_events.record(
@@ -455,10 +484,17 @@ async fn handle(
                                 return Err(error);
                             }
                         };
+                        let runtime_cfg = config_rx.borrow().clone();
+                        data.in_runtime = runtime_cfg.access.users.contains_key(&data.username);
                         shared
                             .runtime_events
                             .record("api.user.patch.ok", format!("username={}", data.username));
-                        return Ok(success_response(StatusCode::OK, data, revision));
+                        let status = if data.in_runtime {
+                            StatusCode::OK
+                        } else {
+                            StatusCode::ACCEPTED
+                        };
+                        return Ok(success_response(status, data, revision));
                     }
                     if method == Method::DELETE {
                         if api_cfg.read_only {
@@ -486,7 +522,18 @@ async fn handle(
                         shared
                             .runtime_events
                             .record("api.user.delete.ok", format!("username={}", deleted_user));
-                        return Ok(success_response(StatusCode::OK, deleted_user, revision));
+                        let runtime_cfg = config_rx.borrow().clone();
+                        let in_runtime = runtime_cfg.access.users.contains_key(&deleted_user);
+                        let response = DeleteUserResponse {
+                            username: deleted_user,
+                            in_runtime,
+                        };
+                        let status = if response.in_runtime {
+                            StatusCode::ACCEPTED
+                        } else {
+                            StatusCode::OK
+                        };
+                        return Ok(success_response(status, response, revision));
                     }
                     if method == Method::POST
                         && let Some(base_user) = user.strip_suffix("/rotate-secret")
@@ -514,7 +561,7 @@ async fn handle(
                             &shared,
                         )
                         .await;
-                        let (data, revision) = match result {
+                        let (mut data, revision) = match result {
                             Ok(ok) => ok,
                             Err(error) => {
                                 shared.runtime_events.record(
@@ -524,11 +571,19 @@ async fn handle(
                                 return Err(error);
                             }
                         };
+                        let runtime_cfg = config_rx.borrow().clone();
+                        data.user.in_runtime =
+                            runtime_cfg.access.users.contains_key(&data.user.username);
                         shared.runtime_events.record(
                             "api.user.rotate_secret.ok",
                             format!("username={}", base_user),
                         );
-                        return Ok(success_response(StatusCode::OK, data, revision));
+                        let status = if data.user.in_runtime {
+                            StatusCode::OK
+                        } else {
+                            StatusCode::ACCEPTED
+                        };
+                        return Ok(success_response(status, data, revision));
                     }
                     if method == Method::POST {
                         return Ok(error_response(

src/api/model.rs

@@ -81,10 +81,21 @@ pub(super) struct ZeroCoreData {
     pub(super) connections_total: u64,
     pub(super) connections_bad_total: u64,
     pub(super) handshake_timeouts_total: u64,
+    pub(super) accept_permit_timeout_total: u64,
     pub(super) configured_users: usize,
     pub(super) telemetry_core_enabled: bool,
     pub(super) telemetry_user_enabled: bool,
     pub(super) telemetry_me_level: String,
+    pub(super) conntrack_control_enabled: bool,
+    pub(super) conntrack_control_available: bool,
+    pub(super) conntrack_pressure_active: bool,
+    pub(super) conntrack_event_queue_depth: u64,
+    pub(super) conntrack_rule_apply_ok: bool,
+    pub(super) conntrack_delete_attempt_total: u64,
+    pub(super) conntrack_delete_success_total: u64,
+    pub(super) conntrack_delete_not_found_total: u64,
+    pub(super) conntrack_delete_error_total: u64,
+    pub(super) conntrack_close_event_drop_total: u64,
 }
 
 #[derive(Serialize, Clone)]
@@ -174,6 +185,24 @@ pub(super) struct ZeroMiddleProxyData {
     pub(super) route_drop_queue_full_total: u64,
     pub(super) route_drop_queue_full_base_total: u64,
     pub(super) route_drop_queue_full_high_total: u64,
+    pub(super) d2c_batches_total: u64,
+    pub(super) d2c_batch_frames_total: u64,
+    pub(super) d2c_batch_bytes_total: u64,
+    pub(super) d2c_flush_reason_queue_drain_total: u64,
+    pub(super) d2c_flush_reason_batch_frames_total: u64,
+    pub(super) d2c_flush_reason_batch_bytes_total: u64,
+    pub(super) d2c_flush_reason_max_delay_total: u64,
+    pub(super) d2c_flush_reason_ack_immediate_total: u64,
+    pub(super) d2c_flush_reason_close_total: u64,
+    pub(super) d2c_data_frames_total: u64,
+    pub(super) d2c_ack_frames_total: u64,
+    pub(super) d2c_payload_bytes_total: u64,
+    pub(super) d2c_write_mode_coalesced_total: u64,
+    pub(super) d2c_write_mode_split_total: u64,
+    pub(super) d2c_quota_reject_pre_write_total: u64,
+    pub(super) d2c_quota_reject_post_write_total: u64,
+    pub(super) d2c_frame_buf_shrink_total: u64,
+    pub(super) d2c_frame_buf_shrink_bytes_total: u64,
     pub(super) socks_kdf_strict_reject_total: u64,
     pub(super) socks_kdf_compat_fallback_total: u64,
     pub(super) endpoint_quarantine_total: u64,
@@ -410,6 +439,7 @@ pub(super) struct UserLinks {
 #[derive(Serialize)]
 pub(super) struct UserInfo {
     pub(super) username: String,
+    pub(super) in_runtime: bool,
     pub(super) user_ad_tag: Option<String>,
     pub(super) max_tcp_conns: Option<usize>,
     pub(super) expiration_rfc3339: Option<String>,
@@ -424,12 +454,24 @@ pub(super) struct UserInfo {
     pub(super) links: UserLinks,
 }
 
+#[derive(Serialize)]
+pub(super) struct UserActiveIps {
+    pub(super) username: String,
+    pub(super) active_ips: Vec<IpAddr>,
+}
+
 #[derive(Serialize)]
 pub(super) struct CreateUserResponse {
     pub(super) user: UserInfo,
     pub(super) secret: String,
 }
 
+#[derive(Serialize)]
+pub(super) struct DeleteUserResponse {
+    pub(super) username: String,
+    pub(super) in_runtime: bool,
+}
+
 #[derive(Deserialize)]
 pub(super) struct CreateUserRequest {
     pub(super) username: String,

src/api/runtime_stats.rs

@@ -39,10 +39,21 @@ pub(super) fn build_zero_all_data(stats: &Stats, configured_users: usize) -> Zer
             connections_total: stats.get_connects_all(),
             connections_bad_total: stats.get_connects_bad(),
             handshake_timeouts_total: stats.get_handshake_timeouts(),
+            accept_permit_timeout_total: stats.get_accept_permit_timeout_total(),
             configured_users,
             telemetry_core_enabled: telemetry.core_enabled,
             telemetry_user_enabled: telemetry.user_enabled,
             telemetry_me_level: telemetry.me_level.to_string(),
+            conntrack_control_enabled: stats.get_conntrack_control_enabled(),
+            conntrack_control_available: stats.get_conntrack_control_available(),
+            conntrack_pressure_active: stats.get_conntrack_pressure_active(),
+            conntrack_event_queue_depth: stats.get_conntrack_event_queue_depth(),
+            conntrack_rule_apply_ok: stats.get_conntrack_rule_apply_ok(),
+            conntrack_delete_attempt_total: stats.get_conntrack_delete_attempt_total(),
+            conntrack_delete_success_total: stats.get_conntrack_delete_success_total(),
+            conntrack_delete_not_found_total: stats.get_conntrack_delete_not_found_total(),
+            conntrack_delete_error_total: stats.get_conntrack_delete_error_total(),
+            conntrack_close_event_drop_total: stats.get_conntrack_close_event_drop_total(),
         },
         upstream: build_zero_upstream_data(stats),
         middle_proxy: ZeroMiddleProxyData {
@@ -68,6 +79,25 @@ pub(super) fn build_zero_all_data(stats: &Stats, configured_users: usize) -> Zer
             route_drop_queue_full_total: stats.get_me_route_drop_queue_full(),
             route_drop_queue_full_base_total: stats.get_me_route_drop_queue_full_base(),
             route_drop_queue_full_high_total: stats.get_me_route_drop_queue_full_high(),
+            d2c_batches_total: stats.get_me_d2c_batches_total(),
+            d2c_batch_frames_total: stats.get_me_d2c_batch_frames_total(),
+            d2c_batch_bytes_total: stats.get_me_d2c_batch_bytes_total(),
+            d2c_flush_reason_queue_drain_total: stats.get_me_d2c_flush_reason_queue_drain_total(),
+            d2c_flush_reason_batch_frames_total: stats.get_me_d2c_flush_reason_batch_frames_total(),
+            d2c_flush_reason_batch_bytes_total: stats.get_me_d2c_flush_reason_batch_bytes_total(),
+            d2c_flush_reason_max_delay_total: stats.get_me_d2c_flush_reason_max_delay_total(),
+            d2c_flush_reason_ack_immediate_total: stats
+                .get_me_d2c_flush_reason_ack_immediate_total(),
+            d2c_flush_reason_close_total: stats.get_me_d2c_flush_reason_close_total(),
+            d2c_data_frames_total: stats.get_me_d2c_data_frames_total(),
+            d2c_ack_frames_total: stats.get_me_d2c_ack_frames_total(),
+            d2c_payload_bytes_total: stats.get_me_d2c_payload_bytes_total(),
+            d2c_write_mode_coalesced_total: stats.get_me_d2c_write_mode_coalesced_total(),
+            d2c_write_mode_split_total: stats.get_me_d2c_write_mode_split_total(),
+            d2c_quota_reject_pre_write_total: stats.get_me_d2c_quota_reject_pre_write_total(),
+            d2c_quota_reject_post_write_total: stats.get_me_d2c_quota_reject_post_write_total(),
+            d2c_frame_buf_shrink_total: stats.get_me_d2c_frame_buf_shrink_total(),
+            d2c_frame_buf_shrink_bytes_total: stats.get_me_d2c_frame_buf_shrink_bytes_total(),
             socks_kdf_strict_reject_total: stats.get_me_socks_kdf_strict_reject(),
             socks_kdf_compat_fallback_total: stats.get_me_socks_kdf_compat_fallback(),
             endpoint_quarantine_total: stats.get_me_endpoint_quarantine_total(),

src/api/runtime_zero.rs

@@ -35,11 +35,14 @@ pub(super) struct RuntimeGatesData {
     pub(super) conditional_cast_enabled: bool,
     pub(super) me_runtime_ready: bool,
     pub(super) me2dc_fallback_enabled: bool,
+    pub(super) me2dc_fast_enabled: bool,
     pub(super) use_middle_proxy: bool,
     pub(super) route_mode: &'static str,
     pub(super) reroute_active: bool,
     #[serde(skip_serializing_if = "Option::is_none")]
     pub(super) reroute_to_direct_at_epoch_secs: Option<u64>,
+    #[serde(skip_serializing_if = "Option::is_none")]
+    pub(super) reroute_reason: Option<&'static str>,
     pub(super) startup_status: &'static str,
     pub(super) startup_stage: String,
     pub(super) startup_progress_pct: f64,
@@ -47,6 +50,7 @@ pub(super) struct RuntimeGatesData {
 
 #[derive(Serialize)]
 pub(super) struct EffectiveTimeoutLimits {
+    pub(super) client_first_byte_idle_secs: u64,
     pub(super) client_handshake_secs: u64,
     pub(super) tg_connect_secs: u64,
     pub(super) client_keepalive_secs: u64,
@@ -86,6 +90,7 @@ pub(super) struct EffectiveMiddleProxyLimits {
     pub(super) writer_pick_mode: &'static str,
     pub(super) writer_pick_sample_size: u8,
     pub(super) me2dc_fallback: bool,
+    pub(super) me2dc_fast: bool,
 }
 
 #[derive(Serialize)]
@@ -95,6 +100,11 @@ pub(super) struct EffectiveUserIpPolicyLimits {
     pub(super) window_secs: u64,
 }
 
+#[derive(Serialize)]
+pub(super) struct EffectiveUserTcpPolicyLimits {
+    pub(super) global_each: usize,
+}
+
 #[derive(Serialize)]
 pub(super) struct EffectiveLimitsData {
     pub(super) update_every_secs: u64,
@@ -104,6 +114,7 @@ pub(super) struct EffectiveLimitsData {
     pub(super) upstream: EffectiveUpstreamLimits,
     pub(super) middle_proxy: EffectiveMiddleProxyLimits,
     pub(super) user_ip_policy: EffectiveUserIpPolicyLimits,
+    pub(super) user_tcp_policy: EffectiveUserTcpPolicyLimits,
 }
 
 #[derive(Serialize)]
@@ -169,6 +180,8 @@ pub(super) async fn build_runtime_gates_data(
     let startup_summary = build_runtime_startup_summary(shared).await;
     let route_state = shared.route_runtime.snapshot();
     let route_mode = route_state.mode.as_str();
+    let fast_fallback_enabled =
+        cfg.general.use_middle_proxy && cfg.general.me2dc_fallback && cfg.general.me2dc_fast;
     let reroute_active = cfg.general.use_middle_proxy
         && cfg.general.me2dc_fallback
         && matches!(route_state.mode, RelayRouteMode::Direct);
@@ -177,6 +190,15 @@ pub(super) async fn build_runtime_gates_data(
     } else {
         None
     };
+    let reroute_reason = if reroute_active {
+        if fast_fallback_enabled {
+            Some("fast_not_ready_fallback")
+        } else {
+            Some("strict_grace_fallback")
+        }
+    } else {
+        None
+    };
     let me_runtime_ready = if !cfg.general.use_middle_proxy {
         true
     } else {
@@ -194,10 +216,12 @@ pub(super) async fn build_runtime_gates_data(
         conditional_cast_enabled: cfg.general.use_middle_proxy,
         me_runtime_ready,
         me2dc_fallback_enabled: cfg.general.me2dc_fallback,
+        me2dc_fast_enabled: fast_fallback_enabled,
         use_middle_proxy: cfg.general.use_middle_proxy,
         route_mode,
         reroute_active,
         reroute_to_direct_at_epoch_secs,
+        reroute_reason,
         startup_status: startup_summary.status,
         startup_stage: startup_summary.stage,
         startup_progress_pct: startup_summary.progress_pct,
@@ -210,8 +234,9 @@ pub(super) fn build_limits_effective_data(cfg: &ProxyConfig) -> EffectiveLimitsD
         me_reinit_every_secs: cfg.general.effective_me_reinit_every_secs(),
         me_pool_force_close_secs: cfg.general.effective_me_pool_force_close_secs(),
         timeouts: EffectiveTimeoutLimits {
+            client_first_byte_idle_secs: cfg.timeouts.client_first_byte_idle_secs,
             client_handshake_secs: cfg.timeouts.client_handshake,
-            tg_connect_secs: cfg.timeouts.tg_connect,
+            tg_connect_secs: cfg.general.tg_connect,
             client_keepalive_secs: cfg.timeouts.client_keepalive,
             client_ack_secs: cfg.timeouts.client_ack,
             me_one_retry: cfg.timeouts.me_one_retry,
@@ -263,12 +288,16 @@ pub(super) fn build_limits_effective_data(cfg: &ProxyConfig) -> EffectiveLimitsD
             writer_pick_mode: me_writer_pick_mode_label(cfg.general.me_writer_pick_mode),
             writer_pick_sample_size: cfg.general.me_writer_pick_sample_size,
             me2dc_fallback: cfg.general.me2dc_fallback,
+            me2dc_fast: cfg.general.me2dc_fast,
         },
         user_ip_policy: EffectiveUserIpPolicyLimits {
             global_each: cfg.access.user_max_unique_ips_global_each,
             mode: user_max_unique_ips_mode_label(cfg.access.user_max_unique_ips_mode),
             window_secs: cfg.access.user_max_unique_ips_window_secs,
         },
+        user_tcp_policy: EffectiveUserTcpPolicyLimits {
+            global_each: cfg.access.user_max_tcp_conns_global_each,
+        },
     }
 }
 

src/api/users.rs

@@ -136,6 +136,7 @@ pub(super) async fn create_user(
         &shared.ip_tracker,
         detected_ip_v4,
         detected_ip_v6,
+        None,
     )
     .await;
     let user = users
@@ -143,8 +144,16 @@ pub(super) async fn create_user(
         .find(|entry| entry.username == body.username)
         .unwrap_or(UserInfo {
             username: body.username.clone(),
+            in_runtime: false,
             user_ad_tag: None,
-            max_tcp_conns: None,
+            max_tcp_conns: cfg
+                .access
+                .user_max_tcp_conns
+                .get(&body.username)
+                .copied()
+                .filter(|limit| *limit > 0)
+                .or((cfg.access.user_max_tcp_conns_global_each > 0)
+                    .then_some(cfg.access.user_max_tcp_conns_global_each)),
             expiration_rfc3339: None,
             data_quota_bytes: None,
             max_unique_ips: updated_limit,
@@ -236,6 +245,7 @@ pub(super) async fn patch_user(
         &shared.ip_tracker,
         detected_ip_v4,
         detected_ip_v6,
+        None,
     )
     .await;
     let user_info = users
@@ -293,6 +303,7 @@ pub(super) async fn rotate_secret(
         &shared.ip_tracker,
         detected_ip_v4,
         detected_ip_v6,
+        None,
     )
     .await;
     let user_info = users
@@ -365,6 +376,7 @@ pub(super) async fn users_from_config(
     ip_tracker: &UserIpTracker,
     startup_detected_ip_v4: Option<IpAddr>,
     startup_detected_ip_v6: Option<IpAddr>,
+    runtime_cfg: Option<&ProxyConfig>,
 ) -> Vec<UserInfo> {
     let mut names = cfg.access.users.keys().cloned().collect::<Vec<_>>();
     names.sort();
@@ -394,8 +406,18 @@ pub(super) async fn users_from_config(
                 tls: Vec::new(),
             });
         users.push(UserInfo {
+            in_runtime: runtime_cfg
+                .map(|runtime| runtime.access.users.contains_key(&username))
+                .unwrap_or(false),
             user_ad_tag: cfg.access.user_ad_tags.get(&username).cloned(),
-            max_tcp_conns: cfg.access.user_max_tcp_conns.get(&username).copied(),
+            max_tcp_conns: cfg
+                .access
+                .user_max_tcp_conns
+                .get(&username)
+                .copied()
+                .filter(|limit| *limit > 0)
+                .or((cfg.access.user_max_tcp_conns_global_each > 0)
+                    .then_some(cfg.access.user_max_tcp_conns_global_each)),
             expiration_rfc3339: cfg
                 .access
                 .user_expirations
@@ -572,3 +594,94 @@ fn resolve_tls_domains(cfg: &ProxyConfig) -> Vec<&str> {
     }
     domains
 }
+
+#[cfg(test)]
+mod tests {
+    use super::*;
+    use crate::ip_tracker::UserIpTracker;
+    use crate::stats::Stats;
+
+    #[tokio::test]
+    async fn users_from_config_reports_effective_tcp_limit_with_global_fallback() {
+        let mut cfg = ProxyConfig::default();
+        cfg.access.users.insert(
+            "alice".to_string(),
+            "0123456789abcdef0123456789abcdef".to_string(),
+        );
+        cfg.access.user_max_tcp_conns_global_each = 7;
+
+        let stats = Stats::new();
+        let tracker = UserIpTracker::new();
+
+        let users = users_from_config(&cfg, &stats, &tracker, None, None, None).await;
+        let alice = users
+            .iter()
+            .find(|entry| entry.username == "alice")
+            .expect("alice must be present");
+        assert!(!alice.in_runtime);
+        assert_eq!(alice.max_tcp_conns, Some(7));
+
+        cfg.access.user_max_tcp_conns.insert("alice".to_string(), 5);
+        let users = users_from_config(&cfg, &stats, &tracker, None, None, None).await;
+        let alice = users
+            .iter()
+            .find(|entry| entry.username == "alice")
+            .expect("alice must be present");
+        assert!(!alice.in_runtime);
+        assert_eq!(alice.max_tcp_conns, Some(5));
+
+        cfg.access.user_max_tcp_conns.insert("alice".to_string(), 0);
+        let users = users_from_config(&cfg, &stats, &tracker, None, None, None).await;
+        let alice = users
+            .iter()
+            .find(|entry| entry.username == "alice")
+            .expect("alice must be present");
+        assert!(!alice.in_runtime);
+        assert_eq!(alice.max_tcp_conns, Some(7));
+
+        cfg.access.user_max_tcp_conns_global_each = 0;
+        let users = users_from_config(&cfg, &stats, &tracker, None, None, None).await;
+        let alice = users
+            .iter()
+            .find(|entry| entry.username == "alice")
+            .expect("alice must be present");
+        assert!(!alice.in_runtime);
+        assert_eq!(alice.max_tcp_conns, None);
+    }
+
+    #[tokio::test]
+    async fn users_from_config_marks_runtime_membership_when_snapshot_is_provided() {
+        let mut disk_cfg = ProxyConfig::default();
+        disk_cfg.access.users.insert(
+            "alice".to_string(),
+            "0123456789abcdef0123456789abcdef".to_string(),
+        );
+        disk_cfg.access.users.insert(
+            "bob".to_string(),
+            "fedcba9876543210fedcba9876543210".to_string(),
+        );
+
+        let mut runtime_cfg = ProxyConfig::default();
+        runtime_cfg.access.users.insert(
+            "alice".to_string(),
+            "0123456789abcdef0123456789abcdef".to_string(),
+        );
+
+        let stats = Stats::new();
+        let tracker = UserIpTracker::new();
+        let users =
+            users_from_config(&disk_cfg, &stats, &tracker, None, None, Some(&runtime_cfg)).await;
+
+        let alice = users
+            .iter()
+            .find(|entry| entry.username == "alice")
+            .expect("alice must be present");
+        let bob = users
+            .iter()
+            .find(|entry| entry.username == "bob")
+            .expect("bob must be present");
+
+        assert!(alice.in_runtime);
+        assert!(!bob.in_runtime);
+    }
+}

src/cli.rs

@@ -1,11 +1,270 @@
-//! CLI commands: --init (fire-and-forget setup)
+//! CLI commands: --init (fire-and-forget setup), daemon options, subcommands
+//!
+//! Subcommands:
+//! - `start [OPTIONS] [config.toml]` - Start the daemon
+//! - `stop [--pid-file PATH]` - Stop a running daemon
+//! - `reload [--pid-file PATH]` - Reload configuration (SIGHUP)
+//! - `status [--pid-file PATH]` - Check daemon status
+//! - `run [OPTIONS] [config.toml]` - Run in foreground (default behavior)
 
 use rand::RngExt;
 use std::fs;
 use std::path::{Path, PathBuf};
 use std::process::Command;
 
+#[cfg(unix)]
+use crate::daemon::{self, DEFAULT_PID_FILE, DaemonOptions};
+
+/// CLI subcommand to execute.
+#[derive(Debug, Clone, PartialEq, Eq)]
+pub enum Subcommand {
+    /// Run the proxy (default, or explicit `run` subcommand).
+    Run,
+    /// Start as daemon (`start` subcommand).
+    Start,
+    /// Stop a running daemon (`stop` subcommand).
+    Stop,
+    /// Reload configuration (`reload` subcommand).
+    Reload,
+    /// Check daemon status (`status` subcommand).
+    Status,
+    /// Fire-and-forget setup (`--init`).
+    Init,
+}
+
+/// Parsed subcommand with its options.
+#[derive(Debug)]
+pub struct ParsedCommand {
+    pub subcommand: Subcommand,
+    pub pid_file: PathBuf,
+    pub config_path: String,
+    #[cfg(unix)]
+    pub daemon_opts: DaemonOptions,
+    pub init_opts: Option<InitOptions>,
+}
+
+impl Default for ParsedCommand {
+    fn default() -> Self {
+        Self {
+            subcommand: Subcommand::Run,
+            #[cfg(unix)]
+            pid_file: PathBuf::from(DEFAULT_PID_FILE),
+            #[cfg(not(unix))]
+            pid_file: PathBuf::from("/var/run/telemt.pid"),
+            config_path: "config.toml".to_string(),
+            #[cfg(unix)]
+            daemon_opts: DaemonOptions::default(),
+            init_opts: None,
+        }
+    }
+}
+
+/// Parse CLI arguments into a command structure.
+pub fn parse_command(args: &[String]) -> ParsedCommand {
+    let mut cmd = ParsedCommand::default();
+
+    // Check for --init first (legacy form)
+    if args.iter().any(|a| a == "--init") {
+        cmd.subcommand = Subcommand::Init;
+        cmd.init_opts = parse_init_args(args);
+        return cmd;
+    }
+
+    // Check for subcommand as first argument
+    if let Some(first) = args.first() {
+        match first.as_str() {
+            "start" => {
+                cmd.subcommand = Subcommand::Start;
+                #[cfg(unix)]
+                {
+                    cmd.daemon_opts = parse_daemon_args(args);
+                    // Force daemonize for start command
+                    cmd.daemon_opts.daemonize = true;
+                }
+            }
+            "stop" => {
+                cmd.subcommand = Subcommand::Stop;
+            }
+            "reload" => {
+                cmd.subcommand = Subcommand::Reload;
+            }
+            "status" => {
+                cmd.subcommand = Subcommand::Status;
+            }
+            "run" => {
+                cmd.subcommand = Subcommand::Run;
+                #[cfg(unix)]
+                {
+                    cmd.daemon_opts = parse_daemon_args(args);
+                }
+            }
+            _ => {
+                // No subcommand, default to Run
+                #[cfg(unix)]
+                {
+                    cmd.daemon_opts = parse_daemon_args(args);
+                }
+            }
+        }
+    }
+
+    // Parse remaining options
+    let mut i = 0;
+    while i < args.len() {
+        match args[i].as_str() {
+            // Skip subcommand names
+            "start" | "stop" | "reload" | "status" | "run" => {}
+            // PID file option (for stop/reload/status)
+            "--pid-file" => {
+                i += 1;
+                if i < args.len() {
+                    cmd.pid_file = PathBuf::from(&args[i]);
+                    #[cfg(unix)]
+                    {
+                        cmd.daemon_opts.pid_file = Some(cmd.pid_file.clone());
+                    }
+                }
+            }
+            s if s.starts_with("--pid-file=") => {
+                cmd.pid_file = PathBuf::from(s.trim_start_matches("--pid-file="));
+                #[cfg(unix)]
+                {
+                    cmd.daemon_opts.pid_file = Some(cmd.pid_file.clone());
+                }
+            }
+            // Config path (positional, non-flag argument)
+            s if !s.starts_with('-') => {
+                cmd.config_path = s.to_string();
+            }
+            _ => {}
+        }
+        i += 1;
+    }
+
+    cmd
+}
+
+/// Execute a subcommand that doesn't require starting the server.
+/// Returns `Some(exit_code)` if the command was handled, `None` if server should start.
+#[cfg(unix)]
+pub fn execute_subcommand(cmd: &ParsedCommand) -> Option<i32> {
+    match cmd.subcommand {
+        Subcommand::Stop => Some(cmd_stop(&cmd.pid_file)),
+        Subcommand::Reload => Some(cmd_reload(&cmd.pid_file)),
+        Subcommand::Status => Some(cmd_status(&cmd.pid_file)),
+        Subcommand::Init => {
+            if let Some(opts) = cmd.init_opts.clone() {
+                match run_init(opts) {
+                    Ok(()) => Some(0),
+                    Err(e) => {
+                        eprintln!("[telemt] Init failed: {}", e);
+                        Some(1)
+                    }
+                }
+            } else {
+                Some(1)
+            }
+        }
+        // Run and Start need the server
+        Subcommand::Run | Subcommand::Start => None,
+    }
+}
+
+#[cfg(not(unix))]
+pub fn execute_subcommand(cmd: &ParsedCommand) -> Option<i32> {
+    match cmd.subcommand {
+        Subcommand::Stop | Subcommand::Reload | Subcommand::Status => {
+            eprintln!("[telemt] Subcommand not supported on this platform");
+            Some(1)
+        }
+        Subcommand::Init => {
+            if let Some(opts) = cmd.init_opts.clone() {
+                match run_init(opts) {
+                    Ok(()) => Some(0),
+                    Err(e) => {
+                        eprintln!("[telemt] Init failed: {}", e);
+                        Some(1)
+                    }
+                }
+            } else {
+                Some(1)
+            }
+        }
+        Subcommand::Run | Subcommand::Start => None,
+    }
+}
+
+/// Stop command: send SIGTERM to the running daemon.
+#[cfg(unix)]
+fn cmd_stop(pid_file: &Path) -> i32 {
+    use nix::sys::signal::Signal;
+
+    println!("Stopping telemt daemon...");
+
+    match daemon::signal_pid_file(pid_file, Signal::SIGTERM) {
+        Ok(()) => {
+            println!("Stop signal sent successfully");
+
+            // Wait for process to exit (up to 10 seconds)
+            for _ in 0..20 {
+                std::thread::sleep(std::time::Duration::from_millis(500));
+                if let daemon::DaemonStatus::NotRunning = daemon::check_status(pid_file) {
+                    println!("Daemon stopped");
+                    return 0;
+                }
+            }
+            println!("Daemon may still be shutting down");
+            0
+        }
+        Err(e) => {
+            eprintln!("Failed to stop daemon: {}", e);
+            1
+        }
+    }
+}
+
+/// Reload command: send SIGHUP to trigger config reload.
+#[cfg(unix)]
+fn cmd_reload(pid_file: &Path) -> i32 {
+    use nix::sys::signal::Signal;
+
+    println!("Reloading telemt configuration...");
+
+    match daemon::signal_pid_file(pid_file, Signal::SIGHUP) {
+        Ok(()) => {
+            println!("Reload signal sent successfully");
+            0
+        }
+        Err(e) => {
+            eprintln!("Failed to reload daemon: {}", e);
+            1
+        }
+    }
+}
+
+/// Status command: check if daemon is running.
+#[cfg(unix)]
+fn cmd_status(pid_file: &Path) -> i32 {
+    match daemon::check_status(pid_file) {
+        daemon::DaemonStatus::Running(pid) => {
+            println!("telemt is running (pid {})", pid);
+            0
+        }
+        daemon::DaemonStatus::Stale(pid) => {
+            println!("telemt is not running (stale pid file, was pid {})", pid);
+            // Clean up stale PID file
+            let _ = std::fs::remove_file(pid_file);
+            1
+        }
+        daemon::DaemonStatus::NotRunning => {
+            println!("telemt is not running");
+            1
+        }
+    }
+}
+
 /// Options for the init command
+#[derive(Debug, Clone)]
 pub struct InitOptions {
     pub port: u16,
     pub domain: String,
@@ -15,6 +274,64 @@ pub struct InitOptions {
     pub no_start: bool,
 }
 
+/// Parse daemon-related options from CLI args.
+#[cfg(unix)]
+pub fn parse_daemon_args(args: &[String]) -> DaemonOptions {
+    let mut opts = DaemonOptions::default();
+    let mut i = 0;
+
+    while i < args.len() {
+        match args[i].as_str() {
+            "--daemon" | "-d" => {
+                opts.daemonize = true;
+            }
+            "--foreground" | "-f" => {
+                opts.foreground = true;
+            }
+            "--pid-file" => {
+                i += 1;
+                if i < args.len() {
+                    opts.pid_file = Some(PathBuf::from(&args[i]));
+                }
+            }
+            s if s.starts_with("--pid-file=") => {
+                opts.pid_file = Some(PathBuf::from(s.trim_start_matches("--pid-file=")));
+            }
+            "--run-as-user" => {
+                i += 1;
+                if i < args.len() {
+                    opts.user = Some(args[i].clone());
+                }
+            }
+            s if s.starts_with("--run-as-user=") => {
+                opts.user = Some(s.trim_start_matches("--run-as-user=").to_string());
+            }
+            "--run-as-group" => {
+                i += 1;
+                if i < args.len() {
+                    opts.group = Some(args[i].clone());
+                }
+            }
+            s if s.starts_with("--run-as-group=") => {
+                opts.group = Some(s.trim_start_matches("--run-as-group=").to_string());
+            }
+            "--working-dir" => {
+                i += 1;
+                if i < args.len() {
+                    opts.working_dir = Some(PathBuf::from(&args[i]));
+                }
+            }
+            s if s.starts_with("--working-dir=") => {
+                opts.working_dir = Some(PathBuf::from(s.trim_start_matches("--working-dir=")));
+            }
+            _ => {}
+        }
+        i += 1;
+    }
+
+    opts
+}
+
 impl Default for InitOptions {
     fn default() -> Self {
         Self {
@@ -84,10 +401,16 @@ pub fn parse_init_args(args: &[String]) -> Option<InitOptions> {
 
 /// Run the fire-and-forget setup.
 pub fn run_init(opts: InitOptions) -> Result<(), Box<dyn std::error::Error>> {
+    use crate::service::{self, InitSystem, ServiceOptions};
+
     eprintln!("[telemt] Fire-and-forget setup");
     eprintln!();
 
-    // 1. Generate or validate secret
+    // 1. Detect init system
+    let init_system = service::detect_init_system();
+    eprintln!("[+] Detected init system: {}", init_system);
+
+    // 2. Generate or validate secret
     let secret = match opts.secret {
         Some(s) => {
             if s.len() != 32 || !s.chars().all(|c| c.is_ascii_hexdigit()) {
@@ -104,72 +427,126 @@ pub fn run_init(opts: InitOptions) -> Result<(), Box<dyn std::error::Error>> {
     eprintln!("[+] Port:   {}", opts.port);
     eprintln!("[+] Domain: {}", opts.domain);
 
-    // 2. Create config directory
+    // 3. Create config directory
     fs::create_dir_all(&opts.config_dir)?;
     let config_path = opts.config_dir.join("config.toml");
 
-    // 3. Write config
+    // 4. Write config
     let config_content = generate_config(&opts.username, &secret, opts.port, &opts.domain);
     fs::write(&config_path, &config_content)?;
     eprintln!("[+] Config written to {}", config_path.display());
 
-    // 4. Write systemd unit
+    // 5. Generate and write service file
     let exe_path =
         std::env::current_exe().unwrap_or_else(|_| PathBuf::from("/usr/local/bin/telemt"));
 
-    let unit_path = Path::new("/etc/systemd/system/telemt.service");
-    let unit_content = generate_systemd_unit(&exe_path, &config_path);
+    let service_opts = ServiceOptions {
+        exe_path: &exe_path,
+        config_path: &config_path,
+        user: None, // Let systemd/init handle user
+        group: None,
+        pid_file: "/var/run/telemt.pid",
+        working_dir: Some("/var/lib/telemt"),
+        description: "Telemt MTProxy - Telegram MTProto Proxy",
+    };
 
-    match fs::write(unit_path, &unit_content) {
+    let service_path = service::service_file_path(init_system);
+    let service_content = service::generate_service_file(init_system, &service_opts);
+
+    // Ensure parent directory exists
+    if let Some(parent) = Path::new(service_path).parent() {
+        let _ = fs::create_dir_all(parent);
+    }
+
+    match fs::write(service_path, &service_content) {
         Ok(()) => {
-            eprintln!("[+] Systemd unit written to {}", unit_path.display());
+            eprintln!("[+] Service file written to {}", service_path);
+
+            // Make script executable for OpenRC/FreeBSD
+            #[cfg(unix)]
+            if init_system == InitSystem::OpenRC || init_system == InitSystem::FreeBSDRc {
+                use std::os::unix::fs::PermissionsExt;
+                let mut perms = fs::metadata(service_path)?.permissions();
+                perms.set_mode(0o755);
+                fs::set_permissions(service_path, perms)?;
+            }
         }
         Err(e) => {
-            eprintln!("[!] Cannot write systemd unit (run as root?): {}", e);
-            eprintln!("[!] Manual unit file content:");
-            eprintln!("{}", unit_content);
+            eprintln!("[!] Cannot write service file (run as root?): {}", e);
+            eprintln!("[!] Manual service file content:");
+            eprintln!("{}", service_content);
 
-            // Still print links and config
+            // Still print links and installation instructions
+            eprintln!();
+            eprintln!("{}", service::installation_instructions(init_system));
             print_links(&opts.username, &secret, opts.port, &opts.domain);
             return Ok(());
         }
     }
 
-    // 5. Reload systemd
-    run_cmd("systemctl", &["daemon-reload"]);
+    // 6. Install and enable service based on init system
+    match init_system {
+        InitSystem::Systemd => {
+            run_cmd("systemctl", &["daemon-reload"]);
+            run_cmd("systemctl", &["enable", "telemt.service"]);
+            eprintln!("[+] Service enabled");
 
-    // 6. Enable service
-    run_cmd("systemctl", &["enable", "telemt.service"]);
-    eprintln!("[+] Service enabled");
+            if !opts.no_start {
+                run_cmd("systemctl", &["start", "telemt.service"]);
+                eprintln!("[+] Service started");
 
-    // 7. Start service (unless --no-start)
-    if !opts.no_start {
-        run_cmd("systemctl", &["start", "telemt.service"]);
-        eprintln!("[+] Service started");
+                std::thread::sleep(std::time::Duration::from_secs(1));
+                let status = Command::new("systemctl")
+                    .args(["is-active", "telemt.service"])
+                    .output();
 
-        // Brief delay then check status
-        std::thread::sleep(std::time::Duration::from_secs(1));
-        let status = Command::new("systemctl")
-            .args(["is-active", "telemt.service"])
-            .output();
-
-        match status {
-            Ok(out) if out.status.success() => {
-                eprintln!("[+] Service is running");
-            }
-            _ => {
-                eprintln!("[!] Service may not have started correctly");
-                eprintln!("[!] Check: journalctl -u telemt.service -n 20");
+                match status {
+                    Ok(out) if out.status.success() => {
+                        eprintln!("[+] Service is running");
+                    }
+                    _ => {
+                        eprintln!("[!] Service may not have started correctly");
+                        eprintln!("[!] Check: journalctl -u telemt.service -n 20");
+                    }
+                }
+            } else {
+                eprintln!("[+] Service not started (--no-start)");
+                eprintln!("[+] Start manually: systemctl start telemt.service");
             }
         }
-    } else {
-        eprintln!("[+] Service not started (--no-start)");
-        eprintln!("[+] Start manually: systemctl start telemt.service");
+        InitSystem::OpenRC => {
+            run_cmd("rc-update", &["add", "telemt", "default"]);
+            eprintln!("[+] Service enabled");
+
+            if !opts.no_start {
+                run_cmd("rc-service", &["telemt", "start"]);
+                eprintln!("[+] Service started");
+            } else {
+                eprintln!("[+] Service not started (--no-start)");
+                eprintln!("[+] Start manually: rc-service telemt start");
+            }
+        }
+        InitSystem::FreeBSDRc => {
+            run_cmd("sysrc", &["telemt_enable=YES"]);
+            eprintln!("[+] Service enabled");
+
+            if !opts.no_start {
+                run_cmd("service", &["telemt", "start"]);
+                eprintln!("[+] Service started");
+            } else {
+                eprintln!("[+] Service not started (--no-start)");
+                eprintln!("[+] Start manually: service telemt start");
+            }
+        }
+        InitSystem::Unknown => {
+            eprintln!("[!] Unknown init system - service file written but not installed");
+            eprintln!("[!] You may need to install it manually");
+        }
     }
 
     eprintln!();
 
-    // 8. Print links
+    // 7. Print links
     print_links(&opts.username, &secret, opts.port, &opts.domain);
 
     Ok(())
@@ -207,6 +584,7 @@ me_pool_drain_soft_evict_cooldown_ms = 1000
 me_bind_stale_mode = "never"
 me_pool_min_fresh_ratio = 0.8
 me_reinit_drain_timeout_secs = 90
+tg_connect = 10
 
 [network]
 ipv4 = true
@@ -232,8 +610,8 @@ ip = "0.0.0.0"
 ip = "::"
 
 [timeouts]
-client_handshake = 15
-tg_connect = 10
+client_first_byte_idle_secs = 300
+client_handshake = 60
 client_keepalive = 60
 client_ack = 300
 
@@ -245,6 +623,7 @@ fake_cert_len = 2048
 tls_full_cert_ttl_secs = 90
 
 [access]
+user_max_tcp_conns_global_each = 0
 replay_check_len = 65536
 replay_window_secs = 120
 ignore_time_skew = false
@@ -264,35 +643,6 @@ weight = 10
     )
 }
 
-fn generate_systemd_unit(exe_path: &Path, config_path: &Path) -> String {
-    format!(
-        r#"[Unit]
-Description=Telemt MTProxy
-Documentation=https://github.com/telemt/telemt
-After=network-online.target
-Wants=network-online.target
-
-[Service]
-Type=simple
-ExecStart={exe} {config}
-Restart=always
-RestartSec=5
-LimitNOFILE=65535
-# Security hardening
-NoNewPrivileges=true
-ProtectSystem=strict
-ProtectHome=true
-ReadWritePaths=/etc/telemt
-PrivateTmp=true
-
-[Install]
-WantedBy=multi-user.target
-"#,
-        exe = exe_path.display(),
-        config = config_path.display(),
-    )
-}
-
 fn run_cmd(cmd: &str, args: &[&str]) {
     match Command::new(cmd).args(args).output() {
         Ok(output) => {

src/config/defaults.rs

@@ -29,6 +29,8 @@ const DEFAULT_ME_D2C_FLUSH_BATCH_MAX_FRAMES: usize = 32;
 const DEFAULT_ME_D2C_FLUSH_BATCH_MAX_BYTES: usize = 128 * 1024;
 const DEFAULT_ME_D2C_FLUSH_BATCH_MAX_DELAY_US: u64 = 500;
 const DEFAULT_ME_D2C_ACK_FLUSH_IMMEDIATE: bool = true;
+const DEFAULT_ME_QUOTA_SOFT_OVERSHOOT_BYTES: u64 = 64 * 1024;
+const DEFAULT_ME_D2C_FRAME_BUF_SHRINK_THRESHOLD_BYTES: usize = 256 * 1024;
 const DEFAULT_DIRECT_RELAY_COPY_BUF_C2S_BYTES: usize = 64 * 1024;
 const DEFAULT_DIRECT_RELAY_COPY_BUF_S2C_BYTES: usize = 256 * 1024;
 const DEFAULT_ME_WRITER_PICK_SAMPLE_SIZE: u8 = 3;
@@ -46,6 +48,10 @@ const DEFAULT_ME_POOL_DRAIN_SOFT_EVICT_BUDGET_PER_CORE: u16 = 16;
 const DEFAULT_ME_POOL_DRAIN_SOFT_EVICT_COOLDOWN_MS: u64 = 1000;
 const DEFAULT_USER_MAX_UNIQUE_IPS_WINDOW_SECS: u64 = 30;
 const DEFAULT_ACCEPT_PERMIT_TIMEOUT_MS: u64 = 250;
+const DEFAULT_CONNTRACK_CONTROL_ENABLED: bool = true;
+const DEFAULT_CONNTRACK_PRESSURE_HIGH_WATERMARK_PCT: u8 = 85;
+const DEFAULT_CONNTRACK_PRESSURE_LOW_WATERMARK_PCT: u8 = 70;
+const DEFAULT_CONNTRACK_DELETE_BUDGET_PER_SEC: u64 = 4096;
 const DEFAULT_UPSTREAM_CONNECT_RETRY_ATTEMPTS: u32 = 2;
 const DEFAULT_UPSTREAM_UNHEALTHY_FAIL_THRESHOLD: u32 = 5;
 const DEFAULT_UPSTREAM_CONNECT_BUDGET_MS: u64 = 3000;
@@ -69,6 +75,22 @@ pub(crate) fn default_tls_fetch_scope() -> String {
     String::new()
 }
 
+pub(crate) fn default_tls_fetch_attempt_timeout_ms() -> u64 {
+    5_000
+}
+
+pub(crate) fn default_tls_fetch_total_budget_ms() -> u64 {
+    15_000
+}
+
+pub(crate) fn default_tls_fetch_strict_route() -> bool {
+    true
+}
+
+pub(crate) fn default_tls_fetch_profile_cache_ttl_secs() -> u64 {
+    600
+}
+
 pub(crate) fn default_mask_port() -> u16 {
     443
 }
@@ -78,7 +100,7 @@ pub(crate) fn default_fake_cert_len() -> usize {
 }
 
 pub(crate) fn default_tls_front_dir() -> String {
-    "tlsfront".to_string()
+    "/etc/telemt/tlsfront".to_string()
 }
 
 pub(crate) fn default_replay_check_len() -> usize {
@@ -92,7 +114,11 @@ pub(crate) fn default_replay_window_secs() -> u64 {
 }
 
 pub(crate) fn default_handshake_timeout() -> u64 {
-    30
+    60
+}
+
+pub(crate) fn default_client_first_byte_idle_secs() -> u64 {
+    300
 }
 
 pub(crate) fn default_relay_idle_policy_v2_enabled() -> bool {
@@ -183,14 +209,38 @@ pub(crate) fn default_proxy_protocol_header_timeout_ms() -> u64 {
     500
 }
 
+pub(crate) fn default_proxy_protocol_trusted_cidrs() -> Vec<IpNetwork> {
+    vec!["0.0.0.0/0".parse().unwrap(), "::/0".parse().unwrap()]
+}
+
 pub(crate) fn default_server_max_connections() -> u32 {
     10_000
 }
 
+pub(crate) fn default_listen_backlog() -> u32 {
+    1024
+}
+
 pub(crate) fn default_accept_permit_timeout_ms() -> u64 {
     DEFAULT_ACCEPT_PERMIT_TIMEOUT_MS
 }
 
+pub(crate) fn default_conntrack_control_enabled() -> bool {
+    DEFAULT_CONNTRACK_CONTROL_ENABLED
+}
+
+pub(crate) fn default_conntrack_pressure_high_watermark_pct() -> u8 {
+    DEFAULT_CONNTRACK_PRESSURE_HIGH_WATERMARK_PCT
+}
+
+pub(crate) fn default_conntrack_pressure_low_watermark_pct() -> u8 {
+    DEFAULT_CONNTRACK_PRESSURE_LOW_WATERMARK_PCT
+}
+
+pub(crate) fn default_conntrack_delete_budget_per_sec() -> u64 {
+    DEFAULT_CONNTRACK_DELETE_BUDGET_PER_SEC
+}
+
 pub(crate) fn default_prefer_4() -> u8 {
     4
 }
@@ -251,6 +301,10 @@ pub(crate) fn default_me2dc_fallback() -> bool {
     true
 }
 
+pub(crate) fn default_me2dc_fast() -> bool {
+    true
+}
+
 pub(crate) fn default_keepalive_interval() -> u64 {
     8
 }
@@ -387,6 +441,14 @@ pub(crate) fn default_me_d2c_ack_flush_immediate() -> bool {
     DEFAULT_ME_D2C_ACK_FLUSH_IMMEDIATE
 }
 
+pub(crate) fn default_me_quota_soft_overshoot_bytes() -> u64 {
+    DEFAULT_ME_QUOTA_SOFT_OVERSHOOT_BYTES
+}
+
+pub(crate) fn default_me_d2c_frame_buf_shrink_threshold_bytes() -> usize {
+    DEFAULT_ME_D2C_FRAME_BUF_SHRINK_THRESHOLD_BYTES
+}
+
 pub(crate) fn default_direct_relay_copy_buf_c2s_bytes() -> usize {
     DEFAULT_DIRECT_RELAY_COPY_BUF_C2S_BYTES
 }
@@ -496,7 +558,7 @@ pub(crate) fn default_beobachten_flush_secs() -> u64 {
 }
 
 pub(crate) fn default_beobachten_file() -> String {
-    "cache/beobachten.txt".to_string()
+    "/etc/telemt/beobachten.txt".to_string()
 }
 
 pub(crate) fn default_tls_new_session_tickets() -> u8 {
@@ -543,6 +605,20 @@ pub(crate) fn default_mask_shape_above_cap_blur_max_bytes() -> usize {
     512
 }
 
+#[cfg(not(test))]
+pub(crate) fn default_mask_relay_max_bytes() -> usize {
+    5 * 1024 * 1024
+}
+
+#[cfg(test)]
+pub(crate) fn default_mask_relay_max_bytes() -> usize {
+    32 * 1024
+}
+
+pub(crate) fn default_mask_classifier_prefetch_timeout_ms() -> u64 {
+    5
+}
+
 pub(crate) fn default_mask_timing_normalization_enabled() -> bool {
     false
 }
@@ -755,6 +831,10 @@ pub(crate) fn default_user_max_unique_ips_window_secs() -> u64 {
     DEFAULT_USER_MAX_UNIQUE_IPS_WINDOW_SECS
 }
 
+pub(crate) fn default_user_max_tcp_conns_global_each() -> usize {
+    0
+}
+
 pub(crate) fn default_user_max_unique_ips_global_each() -> usize {
     0
 }

src/config/hot_reload.rs

@@ -106,6 +106,8 @@ pub struct HotFields {
     pub me_d2c_flush_batch_max_bytes: usize,
     pub me_d2c_flush_batch_max_delay_us: u64,
     pub me_d2c_ack_flush_immediate: bool,
+    pub me_quota_soft_overshoot_bytes: u64,
+    pub me_d2c_frame_buf_shrink_threshold_bytes: usize,
     pub direct_relay_copy_buf_c2s_bytes: usize,
     pub direct_relay_copy_buf_s2c_bytes: usize,
     pub me_health_interval_ms_unhealthy: u64,
@@ -115,6 +117,7 @@ pub struct HotFields {
     pub users: std::collections::HashMap<String, String>,
     pub user_ad_tags: std::collections::HashMap<String, String>,
     pub user_max_tcp_conns: std::collections::HashMap<String, usize>,
+    pub user_max_tcp_conns_global_each: usize,
     pub user_expirations: std::collections::HashMap<String, chrono::DateTime<chrono::Utc>>,
     pub user_data_quota: std::collections::HashMap<String, u64>,
     pub user_max_unique_ips: std::collections::HashMap<String, usize>,
@@ -225,6 +228,10 @@ impl HotFields {
             me_d2c_flush_batch_max_bytes: cfg.general.me_d2c_flush_batch_max_bytes,
             me_d2c_flush_batch_max_delay_us: cfg.general.me_d2c_flush_batch_max_delay_us,
             me_d2c_ack_flush_immediate: cfg.general.me_d2c_ack_flush_immediate,
+            me_quota_soft_overshoot_bytes: cfg.general.me_quota_soft_overshoot_bytes,
+            me_d2c_frame_buf_shrink_threshold_bytes: cfg
+                .general
+                .me_d2c_frame_buf_shrink_threshold_bytes,
             direct_relay_copy_buf_c2s_bytes: cfg.general.direct_relay_copy_buf_c2s_bytes,
             direct_relay_copy_buf_s2c_bytes: cfg.general.direct_relay_copy_buf_s2c_bytes,
             me_health_interval_ms_unhealthy: cfg.general.me_health_interval_ms_unhealthy,
@@ -234,6 +241,7 @@ impl HotFields {
             users: cfg.access.users.clone(),
             user_ad_tags: cfg.access.user_ad_tags.clone(),
             user_max_tcp_conns: cfg.access.user_max_tcp_conns.clone(),
+            user_max_tcp_conns_global_each: cfg.access.user_max_tcp_conns_global_each,
             user_expirations: cfg.access.user_expirations.clone(),
             user_data_quota: cfg.access.user_data_quota.clone(),
             user_max_unique_ips: cfg.access.user_max_unique_ips.clone(),
@@ -511,6 +519,9 @@ fn overlay_hot_fields(old: &ProxyConfig, new: &ProxyConfig) -> ProxyConfig {
     cfg.general.me_d2c_flush_batch_max_bytes = new.general.me_d2c_flush_batch_max_bytes;
     cfg.general.me_d2c_flush_batch_max_delay_us = new.general.me_d2c_flush_batch_max_delay_us;
     cfg.general.me_d2c_ack_flush_immediate = new.general.me_d2c_ack_flush_immediate;
+    cfg.general.me_quota_soft_overshoot_bytes = new.general.me_quota_soft_overshoot_bytes;
+    cfg.general.me_d2c_frame_buf_shrink_threshold_bytes =
+        new.general.me_d2c_frame_buf_shrink_threshold_bytes;
     cfg.general.direct_relay_copy_buf_c2s_bytes = new.general.direct_relay_copy_buf_c2s_bytes;
     cfg.general.direct_relay_copy_buf_s2c_bytes = new.general.direct_relay_copy_buf_s2c_bytes;
     cfg.general.me_health_interval_ms_unhealthy = new.general.me_health_interval_ms_unhealthy;
@@ -521,6 +532,7 @@ fn overlay_hot_fields(old: &ProxyConfig, new: &ProxyConfig) -> ProxyConfig {
     cfg.access.users = new.access.users.clone();
     cfg.access.user_ad_tags = new.access.user_ad_tags.clone();
     cfg.access.user_max_tcp_conns = new.access.user_max_tcp_conns.clone();
+    cfg.access.user_max_tcp_conns_global_each = new.access.user_max_tcp_conns_global_each;
     cfg.access.user_expirations = new.access.user_expirations.clone();
     cfg.access.user_data_quota = new.access.user_data_quota.clone();
     cfg.access.user_max_unique_ips = new.access.user_max_unique_ips.clone();
@@ -561,6 +573,7 @@ fn warn_non_hot_changes(old: &ProxyConfig, new: &ProxyConfig, non_hot_changed: b
     }
     if old.server.proxy_protocol != new.server.proxy_protocol
         || !listeners_equal(&old.server.listeners, &new.server.listeners)
+        || old.server.listen_backlog != new.server.listen_backlog
         || old.server.listen_addr_ipv4 != new.server.listen_addr_ipv4
         || old.server.listen_addr_ipv6 != new.server.listen_addr_ipv6
         || old.server.listen_tcp != new.server.listen_tcp
@@ -593,6 +606,9 @@ fn warn_non_hot_changes(old: &ProxyConfig, new: &ProxyConfig, non_hot_changed: b
         || old.censorship.mask_shape_above_cap_blur != new.censorship.mask_shape_above_cap_blur
         || old.censorship.mask_shape_above_cap_blur_max_bytes
             != new.censorship.mask_shape_above_cap_blur_max_bytes
+        || old.censorship.mask_relay_max_bytes != new.censorship.mask_relay_max_bytes
+        || old.censorship.mask_classifier_prefetch_timeout_ms
+            != new.censorship.mask_classifier_prefetch_timeout_ms
         || old.censorship.mask_timing_normalization_enabled
             != new.censorship.mask_timing_normalization_enabled
         || old.censorship.mask_timing_normalization_floor_ms
@@ -639,6 +655,9 @@ fn warn_non_hot_changes(old: &ProxyConfig, new: &ProxyConfig, non_hot_changed: b
     }
     if old.general.me_route_no_writer_mode != new.general.me_route_no_writer_mode
         || old.general.me_route_no_writer_wait_ms != new.general.me_route_no_writer_wait_ms
+        || old.general.me_route_hybrid_max_wait_ms != new.general.me_route_hybrid_max_wait_ms
+        || old.general.me_route_blocking_send_timeout_ms
+            != new.general.me_route_blocking_send_timeout_ms
         || old.general.me_route_inline_recovery_attempts
             != new.general.me_route_inline_recovery_attempts
         || old.general.me_route_inline_recovery_wait_ms
@@ -657,9 +676,11 @@ fn warn_non_hot_changes(old: &ProxyConfig, new: &ProxyConfig, non_hot_changed: b
         warned = true;
         warn!("config reload: general.me_init_retry_attempts changed; restart required");
     }
-    if old.general.me2dc_fallback != new.general.me2dc_fallback {
+    if old.general.me2dc_fallback != new.general.me2dc_fallback
+        || old.general.me2dc_fast != new.general.me2dc_fast
+    {
         warned = true;
-        warn!("config reload: general.me2dc_fallback changed; restart required");
+        warn!("config reload: general.me2dc_fallback/me2dc_fast changed; restart required");
     }
     if old.general.proxy_config_v4_cache_path != new.general.proxy_config_v4_cache_path
         || old.general.proxy_config_v6_cache_path != new.general.proxy_config_v6_cache_path
@@ -678,6 +699,7 @@ fn warn_non_hot_changes(old: &ProxyConfig, new: &ProxyConfig, non_hot_changed: b
     if old.general.upstream_connect_retry_attempts != new.general.upstream_connect_retry_attempts
         || old.general.upstream_connect_retry_backoff_ms
             != new.general.upstream_connect_retry_backoff_ms
+        || old.general.tg_connect != new.general.tg_connect
         || old.general.upstream_unhealthy_fail_threshold
             != new.general.upstream_unhealthy_fail_threshold
         || old.general.upstream_connect_failfast_hard_errors
@@ -1030,15 +1052,20 @@ fn log_changes(
         || old_hot.me_d2c_flush_batch_max_bytes != new_hot.me_d2c_flush_batch_max_bytes
         || old_hot.me_d2c_flush_batch_max_delay_us != new_hot.me_d2c_flush_batch_max_delay_us
         || old_hot.me_d2c_ack_flush_immediate != new_hot.me_d2c_ack_flush_immediate
+        || old_hot.me_quota_soft_overshoot_bytes != new_hot.me_quota_soft_overshoot_bytes
+        || old_hot.me_d2c_frame_buf_shrink_threshold_bytes
+            != new_hot.me_d2c_frame_buf_shrink_threshold_bytes
         || old_hot.direct_relay_copy_buf_c2s_bytes != new_hot.direct_relay_copy_buf_c2s_bytes
         || old_hot.direct_relay_copy_buf_s2c_bytes != new_hot.direct_relay_copy_buf_s2c_bytes
     {
         info!(
-            "config reload: relay_tuning: me_d2c_frames={} me_d2c_bytes={} me_d2c_delay_us={} me_ack_flush_immediate={} direct_buf_c2s={} direct_buf_s2c={}",
+            "config reload: relay_tuning: me_d2c_frames={} me_d2c_bytes={} me_d2c_delay_us={} me_ack_flush_immediate={} me_quota_soft_overshoot_bytes={} me_d2c_frame_buf_shrink_threshold_bytes={} direct_buf_c2s={} direct_buf_s2c={}",
             new_hot.me_d2c_flush_batch_max_frames,
             new_hot.me_d2c_flush_batch_max_bytes,
             new_hot.me_d2c_flush_batch_max_delay_us,
             new_hot.me_d2c_ack_flush_immediate,
+            new_hot.me_quota_soft_overshoot_bytes,
+            new_hot.me_d2c_frame_buf_shrink_threshold_bytes,
             new_hot.direct_relay_copy_buf_c2s_bytes,
             new_hot.direct_relay_copy_buf_s2c_bytes,
         );
@@ -1121,6 +1148,12 @@ fn log_changes(
             new_hot.user_max_tcp_conns.len()
         );
     }
+    if old_hot.user_max_tcp_conns_global_each != new_hot.user_max_tcp_conns_global_each {
+        info!(
+            "config reload: user_max_tcp_conns policy global_each={}",
+            new_hot.user_max_tcp_conns_global_each
+        );
+    }
     if old_hot.user_expirations != new_hot.user_expirations {
         info!(
             "config reload: user_expirations updated ({} entries)",

src/config/load.rs

@@ -1,6 +1,6 @@
 #![allow(deprecated)]
 
-use std::collections::{BTreeSet, HashMap};
+use std::collections::{BTreeSet, HashMap, HashSet};
 use std::hash::{DefaultHasher, Hash, Hasher};
 use std::net::{IpAddr, SocketAddr};
 use std::path::{Path, PathBuf};
@@ -346,6 +346,12 @@ impl ProxyConfig {
             ));
         }
 
+        if config.general.tg_connect == 0 {
+            return Err(ProxyError::Config(
+                "general.tg_connect must be > 0".to_string(),
+            ));
+        }
+
         if config.general.upstream_unhealthy_fail_threshold == 0 {
             return Err(ProxyError::Config(
                 "general.upstream_unhealthy_fail_threshold must be > 0".to_string(),
@@ -430,6 +436,24 @@ impl ProxyConfig {
             ));
         }
 
+        if config.censorship.mask_relay_max_bytes == 0 {
+            return Err(ProxyError::Config(
+                "censorship.mask_relay_max_bytes must be > 0".to_string(),
+            ));
+        }
+
+        if config.censorship.mask_relay_max_bytes > 67_108_864 {
+            return Err(ProxyError::Config(
+                "censorship.mask_relay_max_bytes must be <= 67108864".to_string(),
+            ));
+        }
+
+        if !(5..=50).contains(&config.censorship.mask_classifier_prefetch_timeout_ms) {
+            return Err(ProxyError::Config(
+                "censorship.mask_classifier_prefetch_timeout_ms must be within [5, 50]".to_string(),
+            ));
+        }
+
         if config.censorship.mask_timing_normalization_ceiling_ms
             < config.censorship.mask_timing_normalization_floor_ms
         {
@@ -533,6 +557,21 @@ impl ProxyConfig {
             ));
         }
 
+        if config.general.me_quota_soft_overshoot_bytes > 16 * 1024 * 1024 {
+            return Err(ProxyError::Config(
+                "general.me_quota_soft_overshoot_bytes must be within [0, 16777216]".to_string(),
+            ));
+        }
+
+        if !(4096..=16 * 1024 * 1024)
+            .contains(&config.general.me_d2c_frame_buf_shrink_threshold_bytes)
+        {
+            return Err(ProxyError::Config(
+                "general.me_d2c_frame_buf_shrink_threshold_bytes must be within [4096, 16777216]"
+                    .to_string(),
+            ));
+        }
+
         if !(4096..=1024 * 1024).contains(&config.general.direct_relay_copy_buf_c2s_bytes) {
             return Err(ProxyError::Config(
                 "general.direct_relay_copy_buf_c2s_bytes must be within [4096, 1048576]"
@@ -883,6 +922,43 @@ impl ProxyConfig {
             ));
         }
 
+        if config.server.conntrack_control.pressure_high_watermark_pct == 0
+            || config.server.conntrack_control.pressure_high_watermark_pct > 100
+        {
+            return Err(ProxyError::Config(
+                "server.conntrack_control.pressure_high_watermark_pct must be within [1, 100]"
+                    .to_string(),
+            ));
+        }
+
+        if config.server.conntrack_control.pressure_low_watermark_pct
+            >= config.server.conntrack_control.pressure_high_watermark_pct
+        {
+            return Err(ProxyError::Config(
+                "server.conntrack_control.pressure_low_watermark_pct must be < pressure_high_watermark_pct"
+                    .to_string(),
+            ));
+        }
+
+        if config.server.conntrack_control.delete_budget_per_sec == 0 {
+            return Err(ProxyError::Config(
+                "server.conntrack_control.delete_budget_per_sec must be > 0".to_string(),
+            ));
+        }
+
+        if matches!(config.server.conntrack_control.mode, ConntrackMode::Hybrid)
+            && config
+                .server
+                .conntrack_control
+                .hybrid_listener_ips
+                .is_empty()
+        {
+            return Err(ProxyError::Config(
+                "server.conntrack_control.hybrid_listener_ips must be non-empty in mode=hybrid"
+                    .to_string(),
+            ));
+        }
+
         if config.general.effective_me_pool_force_close_secs() > 0
             && config.general.effective_me_pool_force_close_secs()
                 < config.general.me_pool_drain_ttl_secs
@@ -944,6 +1020,28 @@ impl ProxyConfig {
         // Normalize optional TLS fetch scope: whitespace-only values disable scoped routing.
         config.censorship.tls_fetch_scope = config.censorship.tls_fetch_scope.trim().to_string();
 
+        if config.censorship.tls_fetch.profiles.is_empty() {
+            config.censorship.tls_fetch.profiles = TlsFetchConfig::default().profiles;
+        } else {
+            let mut seen = HashSet::new();
+            config
+                .censorship
+                .tls_fetch
+                .profiles
+                .retain(|profile| seen.insert(*profile));
+        }
+
+        if config.censorship.tls_fetch.attempt_timeout_ms == 0 {
+            return Err(ProxyError::Config(
+                "censorship.tls_fetch.attempt_timeout_ms must be > 0".to_string(),
+            ));
+        }
+        if config.censorship.tls_fetch.total_budget_ms == 0 {
+            return Err(ProxyError::Config(
+                "censorship.tls_fetch.total_budget_ms must be > 0".to_string(),
+            ));
+        }
+
         // Merge primary + extra TLS domains, deduplicate (primary always first).
         if !config.censorship.tls_domains.is_empty() {
             let mut all = Vec::with_capacity(1 + config.censorship.tls_domains.len());
@@ -1121,6 +1219,10 @@ mod load_security_tests;
 #[path = "tests/load_mask_shape_security_tests.rs"]
 mod load_mask_shape_security_tests;
 
+#[cfg(test)]
+#[path = "tests/load_mask_classifier_prefetch_timeout_security_tests.rs"]
+mod load_mask_classifier_prefetch_timeout_security_tests;
+
 #[cfg(test)]
 mod tests {
     use super::*;
@@ -1158,6 +1260,7 @@ mod tests {
             default_me_init_retry_attempts()
         );
         assert_eq!(cfg.general.me2dc_fallback, default_me2dc_fallback());
+        assert_eq!(cfg.general.me2dc_fast, default_me2dc_fast());
         assert_eq!(
             cfg.general.proxy_config_v4_cache_path,
             default_proxy_config_v4_cache_path()
@@ -1226,6 +1329,11 @@ mod tests {
         assert_eq!(cfg.general.update_every, default_update_every());
         assert_eq!(cfg.server.listen_addr_ipv4, default_listen_addr_ipv4());
         assert_eq!(cfg.server.listen_addr_ipv6, default_listen_addr_ipv6_opt());
+        assert_eq!(
+            cfg.server.proxy_protocol_trusted_cidrs,
+            default_proxy_protocol_trusted_cidrs()
+        );
+        assert_eq!(cfg.censorship.unknown_sni_action, UnknownSniAction::Drop);
         assert_eq!(cfg.server.api.listen, default_api_listen());
         assert_eq!(cfg.server.api.whitelist, default_api_whitelist());
         assert_eq!(
@@ -1256,7 +1364,36 @@ mod tests {
             cfg.server.api.runtime_edge_events_capacity,
             default_api_runtime_edge_events_capacity()
         );
+        assert_eq!(
+            cfg.server.conntrack_control.inline_conntrack_control,
+            default_conntrack_control_enabled()
+        );
+        assert_eq!(cfg.server.conntrack_control.mode, ConntrackMode::default());
+        assert_eq!(
+            cfg.server.conntrack_control.backend,
+            ConntrackBackend::default()
+        );
+        assert_eq!(
+            cfg.server.conntrack_control.profile,
+            ConntrackPressureProfile::default()
+        );
+        assert_eq!(
+            cfg.server.conntrack_control.pressure_high_watermark_pct,
+            default_conntrack_pressure_high_watermark_pct()
+        );
+        assert_eq!(
+            cfg.server.conntrack_control.pressure_low_watermark_pct,
+            default_conntrack_pressure_low_watermark_pct()
+        );
+        assert_eq!(
+            cfg.server.conntrack_control.delete_budget_per_sec,
+            default_conntrack_delete_budget_per_sec()
+        );
         assert_eq!(cfg.access.users, default_access_users());
+        assert_eq!(
+            cfg.access.user_max_tcp_conns_global_each,
+            default_user_max_tcp_conns_global_each()
+        );
         assert_eq!(
             cfg.access.user_max_unique_ips_mode,
             UserMaxUniqueIpsMode::default()
@@ -1292,6 +1429,7 @@ mod tests {
             default_me_init_retry_attempts()
         );
         assert_eq!(general.me2dc_fallback, default_me2dc_fallback());
+        assert_eq!(general.me2dc_fast, default_me2dc_fast());
         assert_eq!(
             general.proxy_config_v4_cache_path,
             default_proxy_config_v4_cache_path()
@@ -1358,6 +1496,14 @@ mod tests {
 
         let server = ServerConfig::default();
         assert_eq!(server.listen_addr_ipv6, Some(default_listen_addr_ipv6()));
+        assert_eq!(
+            server.proxy_protocol_trusted_cidrs,
+            default_proxy_protocol_trusted_cidrs()
+        );
+        assert_eq!(
+            AntiCensorshipConfig::default().unknown_sni_action,
+            UnknownSniAction::Drop
+        );
         assert_eq!(server.api.listen, default_api_listen());
         assert_eq!(server.api.whitelist, default_api_whitelist());
         assert_eq!(
@@ -1388,9 +1534,107 @@ mod tests {
             server.api.runtime_edge_events_capacity,
             default_api_runtime_edge_events_capacity()
         );
+        assert_eq!(
+            server.conntrack_control.inline_conntrack_control,
+            default_conntrack_control_enabled()
+        );
+        assert_eq!(server.conntrack_control.mode, ConntrackMode::default());
+        assert_eq!(
+            server.conntrack_control.backend,
+            ConntrackBackend::default()
+        );
+        assert_eq!(
+            server.conntrack_control.profile,
+            ConntrackPressureProfile::default()
+        );
+        assert_eq!(
+            server.conntrack_control.pressure_high_watermark_pct,
+            default_conntrack_pressure_high_watermark_pct()
+        );
+        assert_eq!(
+            server.conntrack_control.pressure_low_watermark_pct,
+            default_conntrack_pressure_low_watermark_pct()
+        );
+        assert_eq!(
+            server.conntrack_control.delete_budget_per_sec,
+            default_conntrack_delete_budget_per_sec()
+        );
 
         let access = AccessConfig::default();
         assert_eq!(access.users, default_access_users());
+        assert_eq!(
+            access.user_max_tcp_conns_global_each,
+            default_user_max_tcp_conns_global_each()
+        );
+    }
+
+    #[test]
+    fn proxy_protocol_trusted_cidrs_missing_uses_trust_all_but_explicit_empty_stays_empty() {
+        let cfg_missing: ProxyConfig = toml::from_str(
+            r#"
+            [server]
+            [general]
+            [network]
+            [access]
+            "#,
+        )
+        .unwrap();
+        assert_eq!(
+            cfg_missing.server.proxy_protocol_trusted_cidrs,
+            default_proxy_protocol_trusted_cidrs()
+        );
+
+        let cfg_explicit_empty: ProxyConfig = toml::from_str(
+            r#"
+            [server]
+            proxy_protocol_trusted_cidrs = []
+
+            [general]
+            [network]
+            [access]
+            "#,
+        )
+        .unwrap();
+        assert!(
+            cfg_explicit_empty
+                .server
+                .proxy_protocol_trusted_cidrs
+                .is_empty()
+        );
+    }
+
+    #[test]
+    fn unknown_sni_action_parses_and_defaults_to_drop() {
+        let cfg_default: ProxyConfig = toml::from_str(
+            r#"
+            [server]
+            [general]
+            [network]
+            [access]
+            [censorship]
+            "#,
+        )
+        .unwrap();
+        assert_eq!(
+            cfg_default.censorship.unknown_sni_action,
+            UnknownSniAction::Drop
+        );
+
+        let cfg_mask: ProxyConfig = toml::from_str(
+            r#"
+            [server]
+            [general]
+            [network]
+            [access]
+            [censorship]
+            unknown_sni_action = "mask"
+            "#,
+        )
+        .unwrap();
+        assert_eq!(
+            cfg_mask.censorship.unknown_sni_action,
+            UnknownSniAction::Mask
+        );
     }
 
     #[test]
@@ -1764,6 +2008,26 @@ mod tests {
         let _ = std::fs::remove_file(path);
     }
 
+    #[test]
+    fn tg_connect_zero_is_rejected() {
+        let toml = r#"
+            [general]
+            tg_connect = 0
+
+            [censorship]
+            tls_domain = "example.com"
+
+            [access.users]
+            user = "00000000000000000000000000000000"
+        "#;
+        let dir = std::env::temp_dir();
+        let path = dir.join("telemt_tg_connect_zero_test.toml");
+        std::fs::write(&path, toml).unwrap();
+        let err = ProxyConfig::load(&path).unwrap_err().to_string();
+        assert!(err.contains("general.tg_connect must be > 0"));
+        let _ = std::fs::remove_file(path);
+    }
+
     #[test]
     fn rpc_proxy_req_every_out_of_range_is_rejected() {
         let toml = r#"
@@ -2227,6 +2491,118 @@ mod tests {
         let _ = std::fs::remove_file(path);
     }
 
+    #[test]
+    fn conntrack_pressure_high_watermark_out_of_range_is_rejected() {
+        let toml = r#"
+            [server.conntrack_control]
+            pressure_high_watermark_pct = 0
+
+            [censorship]
+            tls_domain = "example.com"
+
+            [access.users]
+            user = "00000000000000000000000000000000"
+        "#;
+        let dir = std::env::temp_dir();
+        let path = dir.join("telemt_conntrack_high_watermark_invalid_test.toml");
+        std::fs::write(&path, toml).unwrap();
+        let err = ProxyConfig::load(&path).unwrap_err().to_string();
+        assert!(err.contains(
+            "server.conntrack_control.pressure_high_watermark_pct must be within [1, 100]"
+        ));
+        let _ = std::fs::remove_file(path);
+    }
+
+    #[test]
+    fn conntrack_pressure_low_watermark_must_be_below_high() {
+        let toml = r#"
+            [server.conntrack_control]
+            pressure_high_watermark_pct = 50
+            pressure_low_watermark_pct = 50
+
+            [censorship]
+            tls_domain = "example.com"
+
+            [access.users]
+            user = "00000000000000000000000000000000"
+        "#;
+        let dir = std::env::temp_dir();
+        let path = dir.join("telemt_conntrack_low_watermark_invalid_test.toml");
+        std::fs::write(&path, toml).unwrap();
+        let err = ProxyConfig::load(&path).unwrap_err().to_string();
+        assert!(
+            err.contains(
+                "server.conntrack_control.pressure_low_watermark_pct must be < pressure_high_watermark_pct"
+            )
+        );
+        let _ = std::fs::remove_file(path);
+    }
+
+    #[test]
+    fn conntrack_delete_budget_zero_is_rejected() {
+        let toml = r#"
+            [server.conntrack_control]
+            delete_budget_per_sec = 0
+
+            [censorship]
+            tls_domain = "example.com"
+
+            [access.users]
+            user = "00000000000000000000000000000000"
+        "#;
+        let dir = std::env::temp_dir();
+        let path = dir.join("telemt_conntrack_delete_budget_invalid_test.toml");
+        std::fs::write(&path, toml).unwrap();
+        let err = ProxyConfig::load(&path).unwrap_err().to_string();
+        assert!(err.contains("server.conntrack_control.delete_budget_per_sec must be > 0"));
+        let _ = std::fs::remove_file(path);
+    }
+
+    #[test]
+    fn conntrack_hybrid_mode_requires_listener_allow_list() {
+        let toml = r#"
+            [server.conntrack_control]
+            mode = "hybrid"
+
+            [censorship]
+            tls_domain = "example.com"
+
+            [access.users]
+            user = "00000000000000000000000000000000"
+        "#;
+        let dir = std::env::temp_dir();
+        let path = dir.join("telemt_conntrack_hybrid_requires_ips_test.toml");
+        std::fs::write(&path, toml).unwrap();
+        let err = ProxyConfig::load(&path).unwrap_err().to_string();
+        assert!(err.contains(
+            "server.conntrack_control.hybrid_listener_ips must be non-empty in mode=hybrid"
+        ));
+        let _ = std::fs::remove_file(path);
+    }
+
+    #[test]
+    fn conntrack_profile_is_loaded_from_config() {
+        let toml = r#"
+            [server.conntrack_control]
+            profile = "aggressive"
+
+            [censorship]
+            tls_domain = "example.com"
+
+            [access.users]
+            user = "00000000000000000000000000000000"
+        "#;
+        let dir = std::env::temp_dir();
+        let path = dir.join("telemt_conntrack_profile_parse_test.toml");
+        std::fs::write(&path, toml).unwrap();
+        let cfg = ProxyConfig::load(&path).unwrap();
+        assert_eq!(
+            cfg.server.conntrack_control.profile,
+            ConntrackPressureProfile::Aggressive
+        );
+        let _ = std::fs::remove_file(path);
+    }
+
     #[test]
     fn force_close_default_matches_drain_ttl() {
         let toml = r#"
@@ -2340,6 +2716,94 @@ mod tests {
         let _ = std::fs::remove_file(path);
     }
 
+    #[test]
+    fn tls_fetch_defaults_are_applied() {
+        let toml = r#"
+            [censorship]
+            tls_domain = "example.com"
+
+            [access.users]
+            user = "00000000000000000000000000000000"
+        "#;
+        let dir = std::env::temp_dir();
+        let path = dir.join("telemt_tls_fetch_defaults_test.toml");
+        std::fs::write(&path, toml).unwrap();
+        let cfg = ProxyConfig::load(&path).unwrap();
+        assert_eq!(
+            cfg.censorship.tls_fetch.profiles,
+            TlsFetchConfig::default().profiles
+        );
+        assert!(cfg.censorship.tls_fetch.strict_route);
+        assert_eq!(cfg.censorship.tls_fetch.attempt_timeout_ms, 5_000);
+        assert_eq!(cfg.censorship.tls_fetch.total_budget_ms, 15_000);
+        assert_eq!(cfg.censorship.tls_fetch.profile_cache_ttl_secs, 600);
+        let _ = std::fs::remove_file(path);
+    }
+
+    #[test]
+    fn tls_fetch_profiles_are_deduplicated_preserving_order() {
+        let toml = r#"
+            [censorship]
+            tls_domain = "example.com"
+            [censorship.tls_fetch]
+            profiles = ["compat_tls12", "modern_chrome_like", "compat_tls12", "legacy_minimal"]
+
+            [access.users]
+            user = "00000000000000000000000000000000"
+        "#;
+        let dir = std::env::temp_dir();
+        let path = dir.join("telemt_tls_fetch_profiles_dedup_test.toml");
+        std::fs::write(&path, toml).unwrap();
+        let cfg = ProxyConfig::load(&path).unwrap();
+        assert_eq!(
+            cfg.censorship.tls_fetch.profiles,
+            vec![
+                TlsFetchProfile::CompatTls12,
+                TlsFetchProfile::ModernChromeLike,
+                TlsFetchProfile::LegacyMinimal
+            ]
+        );
+        let _ = std::fs::remove_file(path);
+    }
+
+    #[test]
+    fn tls_fetch_attempt_timeout_zero_is_rejected() {
+        let toml = r#"
+            [censorship]
+            tls_domain = "example.com"
+            [censorship.tls_fetch]
+            attempt_timeout_ms = 0
+
+            [access.users]
+            user = "00000000000000000000000000000000"
+        "#;
+        let dir = std::env::temp_dir();
+        let path = dir.join("telemt_tls_fetch_attempt_timeout_zero_test.toml");
+        std::fs::write(&path, toml).unwrap();
+        let err = ProxyConfig::load(&path).unwrap_err().to_string();
+        assert!(err.contains("censorship.tls_fetch.attempt_timeout_ms must be > 0"));
+        let _ = std::fs::remove_file(path);
+    }
+
+    #[test]
+    fn tls_fetch_total_budget_zero_is_rejected() {
+        let toml = r#"
+            [censorship]
+            tls_domain = "example.com"
+            [censorship.tls_fetch]
+            total_budget_ms = 0
+
+            [access.users]
+            user = "00000000000000000000000000000000"
+        "#;
+        let dir = std::env::temp_dir();
+        let path = dir.join("telemt_tls_fetch_total_budget_zero_test.toml");
+        std::fs::write(&path, toml).unwrap();
+        let err = ProxyConfig::load(&path).unwrap_err().to_string();
+        assert!(err.contains("censorship.tls_fetch.total_budget_ms must be > 0"));
+        let _ = std::fs::remove_file(path);
+    }
+
     #[test]
     fn invalid_ad_tag_is_disabled_during_load() {
         let toml = r#"

src/config/tests/load_idle_policy_tests.rs

@@ -17,6 +17,28 @@ fn remove_temp_config(path: &PathBuf) {
     let _ = fs::remove_file(path);
 }
 
+#[test]
+fn default_timeouts_enable_apple_compatible_handshake_profile() {
+    let cfg = ProxyConfig::default();
+    assert_eq!(cfg.timeouts.client_first_byte_idle_secs, 300);
+    assert_eq!(cfg.timeouts.client_handshake, 60);
+}
+
+#[test]
+fn load_accepts_zero_first_byte_idle_timeout_as_legacy_opt_out() {
+    let path = write_temp_config(
+        r#"
+[timeouts]
+client_first_byte_idle_secs = 0
+"#,
+    );
+
+    let cfg = ProxyConfig::load(&path).expect("config with zero first-byte idle timeout must load");
+    assert_eq!(cfg.timeouts.client_first_byte_idle_secs, 0);
+
+    remove_temp_config(&path);
+}
+
 #[test]
 fn load_rejects_relay_hard_idle_smaller_than_soft_idle_with_clear_error() {
     let path = write_temp_config(

src/config/tests/load_mask_classifier_prefetch_timeout_security_tests.rs

@@ -0,0 +1,76 @@
+use super::*;
+use std::fs;
+use std::path::PathBuf;
+use std::time::{SystemTime, UNIX_EPOCH};
+
+fn write_temp_config(contents: &str) -> PathBuf {
+    let nonce = SystemTime::now()
+        .duration_since(UNIX_EPOCH)
+        .expect("system time must be after unix epoch")
+        .as_nanos();
+    let path = std::env::temp_dir().join(format!(
+        "telemt-load-mask-prefetch-timeout-security-{nonce}.toml"
+    ));
+    fs::write(&path, contents).expect("temp config write must succeed");
+    path
+}
+
+fn remove_temp_config(path: &PathBuf) {
+    let _ = fs::remove_file(path);
+}
+
+#[test]
+fn load_rejects_mask_classifier_prefetch_timeout_below_min_bound() {
+    let path = write_temp_config(
+        r#"
+[censorship]
+mask_classifier_prefetch_timeout_ms = 4
+"#,
+    );
+
+    let err = ProxyConfig::load(&path)
+        .expect_err("prefetch timeout below minimum security bound must be rejected");
+    let msg = err.to_string();
+    assert!(
+        msg.contains("censorship.mask_classifier_prefetch_timeout_ms must be within [5, 50]"),
+        "error must explain timeout bound invariant, got: {msg}"
+    );
+
+    remove_temp_config(&path);
+}
+
+#[test]
+fn load_rejects_mask_classifier_prefetch_timeout_above_max_bound() {
+    let path = write_temp_config(
+        r#"
+[censorship]
+mask_classifier_prefetch_timeout_ms = 51
+"#,
+    );
+
+    let err = ProxyConfig::load(&path)
+        .expect_err("prefetch timeout above max security bound must be rejected");
+    let msg = err.to_string();
+    assert!(
+        msg.contains("censorship.mask_classifier_prefetch_timeout_ms must be within [5, 50]"),
+        "error must explain timeout bound invariant, got: {msg}"
+    );
+
+    remove_temp_config(&path);
+}
+
+#[test]
+fn load_accepts_mask_classifier_prefetch_timeout_within_bounds() {
+    let path = write_temp_config(
+        r#"
+[censorship]
+mask_classifier_prefetch_timeout_ms = 20
+"#,
+    );
+
+    let cfg =
+        ProxyConfig::load(&path).expect("prefetch timeout within security bounds must be accepted");
+    assert_eq!(cfg.censorship.mask_classifier_prefetch_timeout_ms, 20);
+
+    remove_temp_config(&path);
+}

src/config/tests/load_mask_shape_security_tests.rs

@@ -236,3 +236,57 @@ mask_shape_above_cap_blur_max_bytes = 8
 
     remove_temp_config(&path);
 }
+
+#[test]
+fn load_rejects_zero_mask_relay_max_bytes() {
+    let path = write_temp_config(
+        r#"
+[censorship]
+mask_relay_max_bytes = 0
+"#,
+    );
+
+    let err = ProxyConfig::load(&path).expect_err("mask_relay_max_bytes must be > 0");
+    let msg = err.to_string();
+    assert!(
+        msg.contains("censorship.mask_relay_max_bytes must be > 0"),
+        "error must explain non-zero relay cap invariant, got: {msg}"
+    );
+
+    remove_temp_config(&path);
+}
+
+#[test]
+fn load_rejects_mask_relay_max_bytes_above_upper_bound() {
+    let path = write_temp_config(
+        r#"
+[censorship]
+mask_relay_max_bytes = 67108865
+"#,
+    );
+
+    let err =
+        ProxyConfig::load(&path).expect_err("mask_relay_max_bytes above hard cap must be rejected");
+    let msg = err.to_string();
+    assert!(
+        msg.contains("censorship.mask_relay_max_bytes must be <= 67108864"),
+        "error must explain relay cap upper bound invariant, got: {msg}"
+    );
+
+    remove_temp_config(&path);
+}
+
+#[test]
+fn load_accepts_valid_mask_relay_max_bytes() {
+    let path = write_temp_config(
+        r#"
+[censorship]
+mask_relay_max_bytes = 8388608
+"#,
+    );
+
+    let cfg = ProxyConfig::load(&path).expect("valid mask_relay_max_bytes must be accepted");
+    assert_eq!(cfg.censorship.mask_relay_max_bytes, 8_388_608);
+
+    remove_temp_config(&path);
+}

src/config/types.rs

@@ -429,6 +429,11 @@ pub struct GeneralConfig {
     #[serde(default = "default_me2dc_fallback")]
     pub me2dc_fallback: bool,
 
+    /// Fast ME->Direct fallback mode for new sessions.
+    /// Active only when both `use_middle_proxy=true` and `me2dc_fallback=true`.
+    #[serde(default = "default_me2dc_fast")]
+    pub me2dc_fast: bool,
+
     /// Enable ME keepalive padding frames.
     #[serde(default = "default_true")]
     pub me_keepalive_enabled: bool,
@@ -468,7 +473,7 @@ pub struct GeneralConfig {
     pub me_c2me_send_timeout_ms: u64,
 
     /// Bounded wait in milliseconds for routing ME DATA to per-connection queue.
-    /// `0` keeps legacy no-wait behavior.
+    /// `0` keeps non-blocking routing; values >0 enable bounded wait for compatibility.
     #[serde(default = "default_me_reader_route_data_wait_ms")]
     pub me_reader_route_data_wait_ms: u64,
 
@@ -489,6 +494,14 @@ pub struct GeneralConfig {
     #[serde(default = "default_me_d2c_ack_flush_immediate")]
     pub me_d2c_ack_flush_immediate: bool,
 
+    /// Additional bytes above strict per-user quota allowed in hot-path soft mode.
+    #[serde(default = "default_me_quota_soft_overshoot_bytes")]
+    pub me_quota_soft_overshoot_bytes: u64,
+
+    /// Shrink threshold for reusable ME->Client frame assembly buffer.
+    #[serde(default = "default_me_d2c_frame_buf_shrink_threshold_bytes")]
+    pub me_d2c_frame_buf_shrink_threshold_bytes: usize,
+
     /// Copy buffer size for client->DC direction in direct relay.
     #[serde(default = "default_direct_relay_copy_buf_c2s_bytes")]
     pub direct_relay_copy_buf_c2s_bytes: usize,
@@ -650,6 +663,10 @@ pub struct GeneralConfig {
     #[serde(default = "default_upstream_connect_budget_ms")]
     pub upstream_connect_budget_ms: u64,
 
+    /// Per-attempt TCP connect timeout to Telegram DC (seconds).
+    #[serde(default = "default_connect_timeout")]
+    pub tg_connect: u64,
+
     /// Consecutive failed requests before upstream is marked unhealthy.
     #[serde(default = "default_upstream_unhealthy_fail_threshold")]
     pub upstream_unhealthy_fail_threshold: u32,
@@ -931,6 +948,7 @@ impl Default for GeneralConfig {
             middle_proxy_warm_standby: default_middle_proxy_warm_standby(),
             me_init_retry_attempts: default_me_init_retry_attempts(),
             me2dc_fallback: default_me2dc_fallback(),
+            me2dc_fast: default_me2dc_fast(),
             me_keepalive_enabled: default_true(),
             me_keepalive_interval_secs: default_keepalive_interval(),
             me_keepalive_jitter_secs: default_keepalive_jitter(),
@@ -945,6 +963,9 @@ impl Default for GeneralConfig {
             me_d2c_flush_batch_max_bytes: default_me_d2c_flush_batch_max_bytes(),
             me_d2c_flush_batch_max_delay_us: default_me_d2c_flush_batch_max_delay_us(),
             me_d2c_ack_flush_immediate: default_me_d2c_ack_flush_immediate(),
+            me_quota_soft_overshoot_bytes: default_me_quota_soft_overshoot_bytes(),
+            me_d2c_frame_buf_shrink_threshold_bytes:
+                default_me_d2c_frame_buf_shrink_threshold_bytes(),
             direct_relay_copy_buf_c2s_bytes: default_direct_relay_copy_buf_c2s_bytes(),
             direct_relay_copy_buf_s2c_bytes: default_direct_relay_copy_buf_s2c_bytes(),
             me_warmup_stagger_enabled: default_true(),
@@ -990,6 +1011,7 @@ impl Default for GeneralConfig {
             upstream_connect_retry_attempts: default_upstream_connect_retry_attempts(),
             upstream_connect_retry_backoff_ms: default_upstream_connect_retry_backoff_ms(),
             upstream_connect_budget_ms: default_upstream_connect_budget_ms(),
+            tg_connect: default_connect_timeout(),
             upstream_unhealthy_fail_threshold: default_upstream_unhealthy_fail_threshold(),
             upstream_connect_failfast_hard_errors: default_upstream_connect_failfast_hard_errors(),
             stun_iface_mismatch_ignore: false,
@@ -1194,6 +1216,118 @@ impl Default for ApiConfig {
     }
 }
 
+#[derive(Debug, Clone, Copy, PartialEq, Eq, Serialize, Deserialize, Default)]
+#[serde(rename_all = "lowercase")]
+pub enum ConntrackMode {
+    #[default]
+    Tracked,
+    Notrack,
+    Hybrid,
+}
+
+#[derive(Debug, Clone, Copy, PartialEq, Eq, Serialize, Deserialize, Default)]
+#[serde(rename_all = "lowercase")]
+pub enum ConntrackBackend {
+    #[default]
+    Auto,
+    Nftables,
+    Iptables,
+}
+
+#[derive(Debug, Clone, Copy, PartialEq, Eq, Serialize, Deserialize, Default)]
+#[serde(rename_all = "lowercase")]
+pub enum ConntrackPressureProfile {
+    Conservative,
+    #[default]
+    Balanced,
+    Aggressive,
+}
+
+impl ConntrackPressureProfile {
+    pub fn client_first_byte_idle_cap_secs(self) -> u64 {
+        match self {
+            Self::Conservative => 30,
+            Self::Balanced => 20,
+            Self::Aggressive => 10,
+        }
+    }
+
+    pub fn direct_activity_timeout_secs(self) -> u64 {
+        match self {
+            Self::Conservative => 180,
+            Self::Balanced => 120,
+            Self::Aggressive => 60,
+        }
+    }
+
+    pub fn middle_soft_idle_cap_secs(self) -> u64 {
+        match self {
+            Self::Conservative => 60,
+            Self::Balanced => 30,
+            Self::Aggressive => 20,
+        }
+    }
+
+    pub fn middle_hard_idle_cap_secs(self) -> u64 {
+        match self {
+            Self::Conservative => 180,
+            Self::Balanced => 90,
+            Self::Aggressive => 60,
+        }
+    }
+}
+
+#[derive(Debug, Clone, Serialize, Deserialize)]
+pub struct ConntrackControlConfig {
+    /// Enables runtime conntrack-control worker for pressure mitigation.
+    #[serde(default = "default_conntrack_control_enabled")]
+    pub inline_conntrack_control: bool,
+
+    /// Conntrack mode for listener ingress traffic.
+    #[serde(default)]
+    pub mode: ConntrackMode,
+
+    /// Netfilter backend used to reconcile notrack rules.
+    #[serde(default)]
+    pub backend: ConntrackBackend,
+
+    /// Pressure profile for timeout caps under resource saturation.
+    #[serde(default)]
+    pub profile: ConntrackPressureProfile,
+
+    /// Listener IP allow-list for hybrid mode.
+    /// Ignored in tracked/notrack mode.
+    #[serde(default)]
+    pub hybrid_listener_ips: Vec<IpAddr>,
+
+    /// Pressure high watermark as percentage.
+    #[serde(default = "default_conntrack_pressure_high_watermark_pct")]
+    pub pressure_high_watermark_pct: u8,
+
+    /// Pressure low watermark as percentage.
+    #[serde(default = "default_conntrack_pressure_low_watermark_pct")]
+    pub pressure_low_watermark_pct: u8,
+
+    /// Maximum conntrack delete operations per second.
+    #[serde(default = "default_conntrack_delete_budget_per_sec")]
+    pub delete_budget_per_sec: u64,
+}
+
+impl Default for ConntrackControlConfig {
+    fn default() -> Self {
+        Self {
+            inline_conntrack_control: default_conntrack_control_enabled(),
+            mode: ConntrackMode::default(),
+            backend: ConntrackBackend::default(),
+            profile: ConntrackPressureProfile::default(),
+            hybrid_listener_ips: Vec::new(),
+            pressure_high_watermark_pct: default_conntrack_pressure_high_watermark_pct(),
+            pressure_low_watermark_pct: default_conntrack_pressure_low_watermark_pct(),
+            delete_budget_per_sec: default_conntrack_delete_budget_per_sec(),
+        }
+    }
+}
+
 #[derive(Debug, Clone, Serialize, Deserialize)]
 pub struct ServerConfig {
     #[serde(default = "default_port")]
@@ -1229,9 +1363,10 @@ pub struct ServerConfig {
 
     /// Trusted source CIDRs allowed to send incoming PROXY protocol headers.
     ///
-    /// When non-empty, connections from addresses outside this allowlist are
-    /// rejected before `src_addr` is applied.
-    #[serde(default)]
+    /// If this field is omitted in config, it defaults to trust-all CIDRs
+    /// (`0.0.0.0/0` and `::/0`). If it is explicitly set to an empty list,
+    /// all PROXY protocol headers are rejected.
+    #[serde(default = "default_proxy_protocol_trusted_cidrs")]
     pub proxy_protocol_trusted_cidrs: Vec<IpNetwork>,
 
     /// Port for the Prometheus-compatible metrics endpoint.
@@ -1254,6 +1389,11 @@ pub struct ServerConfig {
     #[serde(default)]
     pub listeners: Vec<ListenerConfig>,
 
+    /// TCP `listen(2)` backlog for client-facing sockets (also used for the metrics HTTP listener).
+    /// The effective queue is capped by the kernel (for example `somaxconn` on Linux).
+    #[serde(default = "default_listen_backlog")]
+    pub listen_backlog: u32,
+
     /// Maximum number of concurrent client connections.
     /// 0 means unlimited.
     #[serde(default = "default_server_max_connections")]
@@ -1263,6 +1403,10 @@ pub struct ServerConfig {
     /// `0` keeps legacy unbounded wait behavior.
     #[serde(default = "default_accept_permit_timeout_ms")]
     pub accept_permit_timeout_ms: u64,
+
+    /// Runtime conntrack control and pressure policy.
+    #[serde(default)]
+    pub conntrack_control: ConntrackControlConfig,
 }
 
 impl Default for ServerConfig {
@@ -1276,20 +1420,28 @@ impl Default for ServerConfig {
             listen_tcp: None,
             proxy_protocol: false,
             proxy_protocol_header_timeout_ms: default_proxy_protocol_header_timeout_ms(),
-            proxy_protocol_trusted_cidrs: Vec::new(),
+            proxy_protocol_trusted_cidrs: default_proxy_protocol_trusted_cidrs(),
             metrics_port: None,
             metrics_listen: None,
             metrics_whitelist: default_metrics_whitelist(),
             api: ApiConfig::default(),
             listeners: Vec::new(),
+            listen_backlog: default_listen_backlog(),
             max_connections: default_server_max_connections(),
             accept_permit_timeout_ms: default_accept_permit_timeout_ms(),
+            conntrack_control: ConntrackControlConfig::default(),
         }
     }
 }
 
 #[derive(Debug, Clone, Serialize, Deserialize)]
 pub struct TimeoutsConfig {
+    /// Maximum idle wait in seconds for the first client byte before handshake parsing starts.
+    /// `0` disables the separate idle phase and keeps legacy timeout behavior.
+    #[serde(default = "default_client_first_byte_idle_secs")]
+    pub client_first_byte_idle_secs: u64,
+
+    /// Maximum active handshake duration in seconds after the first client byte is received.
     #[serde(default = "default_handshake_timeout")]
     pub client_handshake: u64,
 
@@ -1311,9 +1463,6 @@ pub struct TimeoutsConfig {
     #[serde(default = "default_relay_idle_grace_after_downstream_activity_secs")]
     pub relay_idle_grace_after_downstream_activity_secs: u64,
 
-    #[serde(default = "default_connect_timeout")]
-    pub tg_connect: u64,
-
     #[serde(default = "default_keepalive")]
     pub client_keepalive: u64,
 
@@ -1332,13 +1481,13 @@ pub struct TimeoutsConfig {
 impl Default for TimeoutsConfig {
     fn default() -> Self {
         Self {
+            client_first_byte_idle_secs: default_client_first_byte_idle_secs(),
             client_handshake: default_handshake_timeout(),
             relay_idle_policy_v2_enabled: default_relay_idle_policy_v2_enabled(),
             relay_client_idle_soft_secs: default_relay_client_idle_soft_secs(),
             relay_client_idle_hard_secs: default_relay_client_idle_hard_secs(),
             relay_idle_grace_after_downstream_activity_secs:
                 default_relay_idle_grace_after_downstream_activity_secs(),
-            tg_connect: default_connect_timeout(),
             client_keepalive: default_keepalive(),
             client_ack: default_ack_timeout(),
             me_one_retry: default_me_one_retry(),
@@ -1347,6 +1496,90 @@ impl Default for TimeoutsConfig {
     }
 }
 
+#[derive(Debug, Clone, Copy, PartialEq, Eq, Serialize, Deserialize, Default)]
+#[serde(rename_all = "lowercase")]
+pub enum UnknownSniAction {
+    #[default]
+    Drop,
+    Mask,
+}
+
+#[derive(Debug, Clone, Copy, PartialEq, Eq, Hash, Serialize, Deserialize)]
+#[serde(rename_all = "snake_case")]
+pub enum TlsFetchProfile {
+    ModernChromeLike,
+    ModernFirefoxLike,
+    CompatTls12,
+    LegacyMinimal,
+}
+
+impl TlsFetchProfile {
+    pub fn as_str(self) -> &'static str {
+        match self {
+            TlsFetchProfile::ModernChromeLike => "modern_chrome_like",
+            TlsFetchProfile::ModernFirefoxLike => "modern_firefox_like",
+            TlsFetchProfile::CompatTls12 => "compat_tls12",
+            TlsFetchProfile::LegacyMinimal => "legacy_minimal",
+        }
+    }
+}
+
+fn default_tls_fetch_profiles() -> Vec<TlsFetchProfile> {
+    vec![
+        TlsFetchProfile::ModernChromeLike,
+        TlsFetchProfile::ModernFirefoxLike,
+        TlsFetchProfile::CompatTls12,
+        TlsFetchProfile::LegacyMinimal,
+    ]
+}
+
+#[derive(Debug, Clone, Serialize, Deserialize)]
+pub struct TlsFetchConfig {
+    /// Ordered list of ClientHello profiles used for adaptive fallback.
+    #[serde(default = "default_tls_fetch_profiles")]
+    pub profiles: Vec<TlsFetchProfile>,
+
+    /// When true and upstream route is configured, TLS fetch fails closed on
+    /// upstream connect errors and does not fallback to direct TCP.
+    #[serde(default = "default_tls_fetch_strict_route")]
+    pub strict_route: bool,
+
+    /// Timeout per one profile attempt in milliseconds.
+    #[serde(default = "default_tls_fetch_attempt_timeout_ms")]
+    pub attempt_timeout_ms: u64,
+
+    /// Total wall-clock budget in milliseconds across all profile attempts.
+    #[serde(default = "default_tls_fetch_total_budget_ms")]
+    pub total_budget_ms: u64,
+
+    /// Adds GREASE-style values into selected ClientHello extensions.
+    #[serde(default)]
+    pub grease_enabled: bool,
+
+    /// Produces deterministic ClientHello randomness for debugging/tests.
+    #[serde(default)]
+    pub deterministic: bool,
+
+    /// TTL for winner-profile cache entries in seconds.
+    /// Set to 0 to disable profile cache.
+    #[serde(default = "default_tls_fetch_profile_cache_ttl_secs")]
+    pub profile_cache_ttl_secs: u64,
+}
+
+impl Default for TlsFetchConfig {
+    fn default() -> Self {
+        Self {
+            profiles: default_tls_fetch_profiles(),
+            strict_route: default_tls_fetch_strict_route(),
+            attempt_timeout_ms: default_tls_fetch_attempt_timeout_ms(),
+            total_budget_ms: default_tls_fetch_total_budget_ms(),
+            grease_enabled: false,
+            deterministic: false,
+            profile_cache_ttl_secs: default_tls_fetch_profile_cache_ttl_secs(),
+        }
+    }
+}
+
 #[derive(Debug, Clone, Serialize, Deserialize)]
 pub struct AntiCensorshipConfig {
     #[serde(default = "default_tls_domain")]
@@ -1356,11 +1589,19 @@ pub struct AntiCensorshipConfig {
     #[serde(default)]
     pub tls_domains: Vec<String>,
 
+    /// Policy for TLS ClientHello with unknown (non-configured) SNI.
+    #[serde(default)]
+    pub unknown_sni_action: UnknownSniAction,
+
     /// Upstream scope used for TLS front metadata fetches.
     /// Empty value keeps default upstream routing behavior.
     #[serde(default = "default_tls_fetch_scope")]
     pub tls_fetch_scope: String,
 
+    /// Fetch strategy for TLS front metadata bootstrap and periodic refresh.
+    #[serde(default)]
+    pub tls_fetch: TlsFetchConfig,
+
     #[serde(default = "default_true")]
     pub mask: bool,
 
@@ -1440,6 +1681,14 @@ pub struct AntiCensorshipConfig {
     #[serde(default = "default_mask_shape_above_cap_blur_max_bytes")]
     pub mask_shape_above_cap_blur_max_bytes: usize,
 
+    /// Maximum bytes relayed per direction on unauthenticated masking fallback paths.
+    #[serde(default = "default_mask_relay_max_bytes")]
+    pub mask_relay_max_bytes: usize,
+
+    /// Prefetch timeout (ms) for extending fragmented masking classifier window.
+    #[serde(default = "default_mask_classifier_prefetch_timeout_ms")]
+    pub mask_classifier_prefetch_timeout_ms: u64,
+
     /// Enable outcome-time normalization envelope for masking fallback.
     #[serde(default = "default_mask_timing_normalization_enabled")]
     pub mask_timing_normalization_enabled: bool,
@@ -1458,7 +1707,9 @@ impl Default for AntiCensorshipConfig {
         Self {
             tls_domain: default_tls_domain(),
             tls_domains: Vec::new(),
+            unknown_sni_action: UnknownSniAction::Drop,
             tls_fetch_scope: default_tls_fetch_scope(),
+            tls_fetch: TlsFetchConfig::default(),
             mask: default_true(),
             mask_host: None,
             mask_port: default_mask_port(),
@@ -1478,6 +1729,8 @@ impl Default for AntiCensorshipConfig {
             mask_shape_bucket_cap_bytes: default_mask_shape_bucket_cap_bytes(),
             mask_shape_above_cap_blur: default_mask_shape_above_cap_blur(),
             mask_shape_above_cap_blur_max_bytes: default_mask_shape_above_cap_blur_max_bytes(),
+            mask_relay_max_bytes: default_mask_relay_max_bytes(),
+            mask_classifier_prefetch_timeout_ms: default_mask_classifier_prefetch_timeout_ms(),
             mask_timing_normalization_enabled: default_mask_timing_normalization_enabled(),
             mask_timing_normalization_floor_ms: default_mask_timing_normalization_floor_ms(),
             mask_timing_normalization_ceiling_ms: default_mask_timing_normalization_ceiling_ms(),
@@ -1497,6 +1750,12 @@ pub struct AccessConfig {
     #[serde(default)]
     pub user_max_tcp_conns: HashMap<String, usize>,
 
+    /// Global per-user TCP connection limit applied when a user has no
+    /// positive individual override.
+    /// `0` disables the inherited limit.
+    #[serde(default = "default_user_max_tcp_conns_global_each")]
+    pub user_max_tcp_conns_global_each: usize,
+
     #[serde(default)]
     pub user_expirations: HashMap<String, DateTime<Utc>>,
 
@@ -1533,6 +1792,7 @@ impl Default for AccessConfig {
             users: default_access_users(),
             user_ad_tags: HashMap::new(),
             user_max_tcp_conns: HashMap::new(),
+            user_max_tcp_conns_global_each: default_user_max_tcp_conns_global_each(),
             user_expirations: HashMap::new(),
             user_data_quota: HashMap::new(),
             user_max_unique_ips: HashMap::new(),

src/conntrack_control.rs

@@ -0,0 +1,755 @@
+use std::collections::BTreeSet;
+use std::net::IpAddr;
+use std::path::PathBuf;
+use std::sync::Arc;
+use std::time::Duration;
+
+use tokio::io::AsyncWriteExt;
+use tokio::process::Command;
+use tokio::sync::{mpsc, watch};
+use tracing::{debug, info, warn};
+
+use crate::config::{ConntrackBackend, ConntrackMode, ProxyConfig};
+use crate::proxy::middle_relay::note_global_relay_pressure;
+use crate::proxy::shared_state::{ConntrackCloseEvent, ConntrackCloseReason, ProxySharedState};
+use crate::stats::Stats;
+
+const CONNTRACK_EVENT_QUEUE_CAPACITY: usize = 32_768;
+const PRESSURE_RELEASE_TICKS: u8 = 3;
+const PRESSURE_SAMPLE_INTERVAL: Duration = Duration::from_secs(1);
+
+#[derive(Clone, Copy, Debug, PartialEq, Eq)]
+enum NetfilterBackend {
+    Nftables,
+    Iptables,
+}
+
+#[derive(Clone, Copy)]
+struct PressureSample {
+    conn_pct: Option<u8>,
+    fd_pct: Option<u8>,
+    accept_timeout_delta: u64,
+    me_queue_pressure_delta: u64,
+}
+
+struct PressureState {
+    active: bool,
+    low_streak: u8,
+    prev_accept_timeout_total: u64,
+    prev_me_queue_pressure_total: u64,
+}
+
+impl PressureState {
+    fn new(stats: &Stats) -> Self {
+        Self {
+            active: false,
+            low_streak: 0,
+            prev_accept_timeout_total: stats.get_accept_permit_timeout_total(),
+            prev_me_queue_pressure_total: stats.get_me_c2me_send_full_total(),
+        }
+    }
+}
+
+pub(crate) fn spawn_conntrack_controller(
+    config_rx: watch::Receiver<Arc<ProxyConfig>>,
+    stats: Arc<Stats>,
+    shared: Arc<ProxySharedState>,
+) {
+    if !cfg!(target_os = "linux") {
+        let enabled = config_rx
+            .borrow()
+            .server
+            .conntrack_control
+            .inline_conntrack_control;
+        stats.set_conntrack_control_enabled(enabled);
+        stats.set_conntrack_control_available(false);
+        stats.set_conntrack_pressure_active(false);
+        stats.set_conntrack_event_queue_depth(0);
+        stats.set_conntrack_rule_apply_ok(false);
+        shared.disable_conntrack_close_sender();
+        shared.set_conntrack_pressure_active(false);
+        if enabled {
+            warn!(
+                "conntrack control is configured but unsupported on this OS; disabling runtime worker"
+            );
+        }
+        return;
+    }
+
+    let (tx, rx) = mpsc::channel(CONNTRACK_EVENT_QUEUE_CAPACITY);
+    shared.set_conntrack_close_sender(tx);
+    tokio::spawn(async move {
+        run_conntrack_controller(config_rx, stats, shared, rx).await;
+    });
+}
+
+async fn run_conntrack_controller(
+    mut config_rx: watch::Receiver<Arc<ProxyConfig>>,
+    stats: Arc<Stats>,
+    shared: Arc<ProxySharedState>,
+    mut close_rx: mpsc::Receiver<ConntrackCloseEvent>,
+) {
+    let mut cfg = config_rx.borrow().clone();
+    let mut pressure_state = PressureState::new(stats.as_ref());
+    let mut delete_budget_tokens = cfg.server.conntrack_control.delete_budget_per_sec;
+    let mut backend = pick_backend(cfg.server.conntrack_control.backend);
+
+    apply_runtime_state(
+        stats.as_ref(),
+        shared.as_ref(),
+        &cfg,
+        backend.is_some(),
+        false,
+    );
+    reconcile_rules(&cfg, backend, stats.as_ref()).await;
+
+    loop {
+        tokio::select! {
+            changed = config_rx.changed() => {
+                if changed.is_err() {
+                    break;
+                }
+                cfg = config_rx.borrow_and_update().clone();
+                backend = pick_backend(cfg.server.conntrack_control.backend);
+                delete_budget_tokens = cfg.server.conntrack_control.delete_budget_per_sec;
+                apply_runtime_state(stats.as_ref(), shared.as_ref(), &cfg, backend.is_some(), pressure_state.active);
+                reconcile_rules(&cfg, backend, stats.as_ref()).await;
+            }
+            event = close_rx.recv() => {
+                let Some(event) = event else {
+                    break;
+                };
+                stats.set_conntrack_event_queue_depth(close_rx.len() as u64);
+                if !cfg.server.conntrack_control.inline_conntrack_control {
+                    continue;
+                }
+                if !pressure_state.active {
+                    continue;
+                }
+                if !matches!(event.reason, ConntrackCloseReason::Timeout | ConntrackCloseReason::Pressure | ConntrackCloseReason::Reset) {
+                    continue;
+                }
+                if delete_budget_tokens == 0 {
+                    continue;
+                }
+                stats.increment_conntrack_delete_attempt_total();
+                match delete_conntrack_entry(event).await {
+                    DeleteOutcome::Deleted => {
+                        delete_budget_tokens = delete_budget_tokens.saturating_sub(1);
+                        stats.increment_conntrack_delete_success_total();
+                    }
+                    DeleteOutcome::NotFound => {
+                        delete_budget_tokens = delete_budget_tokens.saturating_sub(1);
+                        stats.increment_conntrack_delete_not_found_total();
+                    }
+                    DeleteOutcome::Error => {
+                        delete_budget_tokens = delete_budget_tokens.saturating_sub(1);
+                        stats.increment_conntrack_delete_error_total();
+                    }
+                }
+            }
+            _ = tokio::time::sleep(PRESSURE_SAMPLE_INTERVAL) => {
+                delete_budget_tokens = cfg.server.conntrack_control.delete_budget_per_sec;
+                stats.set_conntrack_event_queue_depth(close_rx.len() as u64);
+                let sample = collect_pressure_sample(stats.as_ref(), &cfg, &mut pressure_state);
+                update_pressure_state(
+                    stats.as_ref(),
+                    shared.as_ref(),
+                    &cfg,
+                    &sample,
+                    &mut pressure_state,
+                );
+                if pressure_state.active {
+                    note_global_relay_pressure(shared.as_ref());
+                }
+            }
+        }
+    }
+
+    shared.disable_conntrack_close_sender();
+    shared.set_conntrack_pressure_active(false);
+    stats.set_conntrack_pressure_active(false);
+}
+
+fn apply_runtime_state(
+    stats: &Stats,
+    shared: &ProxySharedState,
+    cfg: &ProxyConfig,
+    backend_available: bool,
+    pressure_active: bool,
+) {
+    let enabled = cfg.server.conntrack_control.inline_conntrack_control;
+    let available = enabled && backend_available && has_cap_net_admin();
+    if enabled && !available {
+        warn!(
+            "conntrack control enabled but unavailable (missing CAP_NET_ADMIN or backend binaries)"
+        );
+    }
+    stats.set_conntrack_control_enabled(enabled);
+    stats.set_conntrack_control_available(available);
+    shared.set_conntrack_pressure_active(enabled && pressure_active);
+    stats.set_conntrack_pressure_active(enabled && pressure_active);
+}
+
+fn collect_pressure_sample(
+    stats: &Stats,
+    cfg: &ProxyConfig,
+    state: &mut PressureState,
+) -> PressureSample {
+    let current_connections = stats.get_current_connections_total();
+    let conn_pct = if cfg.server.max_connections == 0 {
+        None
+    } else {
+        Some(
+            ((current_connections.saturating_mul(100)) / u64::from(cfg.server.max_connections))
+                .min(100) as u8,
+        )
+    };
+
+    let fd_pct = fd_usage_pct();
+
+    let accept_total = stats.get_accept_permit_timeout_total();
+    let accept_delta = accept_total.saturating_sub(state.prev_accept_timeout_total);
+    state.prev_accept_timeout_total = accept_total;
+
+    let me_total = stats.get_me_c2me_send_full_total();
+    let me_delta = me_total.saturating_sub(state.prev_me_queue_pressure_total);
+    state.prev_me_queue_pressure_total = me_total;
+
+    PressureSample {
+        conn_pct,
+        fd_pct,
+        accept_timeout_delta: accept_delta,
+        me_queue_pressure_delta: me_delta,
+    }
+}
+
+fn update_pressure_state(
+    stats: &Stats,
+    shared: &ProxySharedState,
+    cfg: &ProxyConfig,
+    sample: &PressureSample,
+    state: &mut PressureState,
+) {
+    if !cfg.server.conntrack_control.inline_conntrack_control {
+        if state.active {
+            state.active = false;
+            state.low_streak = 0;
+            shared.set_conntrack_pressure_active(false);
+            stats.set_conntrack_pressure_active(false);
+            info!("Conntrack pressure mode deactivated (feature disabled)");
+        }
+        return;
+    }
+
+    let high = cfg.server.conntrack_control.pressure_high_watermark_pct;
+    let low = cfg.server.conntrack_control.pressure_low_watermark_pct;
+
+    let high_hit = sample.conn_pct.is_some_and(|v| v >= high)
+        || sample.fd_pct.is_some_and(|v| v >= high)
+        || sample.accept_timeout_delta > 0
+        || sample.me_queue_pressure_delta > 0;
+
+    let low_clear = sample.conn_pct.is_none_or(|v| v <= low)
+        && sample.fd_pct.is_none_or(|v| v <= low)
+        && sample.accept_timeout_delta == 0
+        && sample.me_queue_pressure_delta == 0;
+
+    if !state.active && high_hit {
+        state.active = true;
+        state.low_streak = 0;
+        shared.set_conntrack_pressure_active(true);
+        stats.set_conntrack_pressure_active(true);
+        info!(
+            conn_pct = ?sample.conn_pct,
+            fd_pct = ?sample.fd_pct,
+            accept_timeout_delta = sample.accept_timeout_delta,
+            me_queue_pressure_delta = sample.me_queue_pressure_delta,
+            "Conntrack pressure mode activated"
+        );
+        return;
+    }
+
+    if state.active && low_clear {
+        state.low_streak = state.low_streak.saturating_add(1);
+        if state.low_streak >= PRESSURE_RELEASE_TICKS {
+            state.active = false;
+            state.low_streak = 0;
+            shared.set_conntrack_pressure_active(false);
+            stats.set_conntrack_pressure_active(false);
+            info!("Conntrack pressure mode deactivated");
+        }
+        return;
+    }
+
+    state.low_streak = 0;
+}
+
+async fn reconcile_rules(cfg: &ProxyConfig, backend: Option<NetfilterBackend>, stats: &Stats) {
+    if !cfg.server.conntrack_control.inline_conntrack_control {
+        clear_notrack_rules_all_backends().await;
+        stats.set_conntrack_rule_apply_ok(true);
+        return;
+    }
+
+    if !has_cap_net_admin() {
+        stats.set_conntrack_rule_apply_ok(false);
+        return;
+    }
+
+    let Some(backend) = backend else {
+        stats.set_conntrack_rule_apply_ok(false);
+        return;
+    };
+
+    let apply_result = match backend {
+        NetfilterBackend::Nftables => apply_nft_rules(cfg).await,
+        NetfilterBackend::Iptables => apply_iptables_rules(cfg).await,
+    };
+
+    if let Err(error) = apply_result {
+        warn!(error = %error, "Failed to reconcile conntrack/notrack rules");
+        stats.set_conntrack_rule_apply_ok(false);
+    } else {
+        stats.set_conntrack_rule_apply_ok(true);
+    }
+}
+
+fn pick_backend(configured: ConntrackBackend) -> Option<NetfilterBackend> {
+    match configured {
+        ConntrackBackend::Auto => {
+            if command_exists("nft") {
+                Some(NetfilterBackend::Nftables)
+            } else if command_exists("iptables") {
+                Some(NetfilterBackend::Iptables)
+            } else {
+                None
+            }
+        }
+        ConntrackBackend::Nftables => command_exists("nft").then_some(NetfilterBackend::Nftables),
+        ConntrackBackend::Iptables => {
+            command_exists("iptables").then_some(NetfilterBackend::Iptables)
+        }
+    }
+}
+
+fn command_exists(binary: &str) -> bool {
+    let Some(path_var) = std::env::var_os("PATH") else {
+        return false;
+    };
+    std::env::split_paths(&path_var).any(|dir| {
+        let candidate: PathBuf = dir.join(binary);
+        candidate.exists() && candidate.is_file()
+    })
+}
+
+fn notrack_targets(cfg: &ProxyConfig) -> (Vec<Option<IpAddr>>, Vec<Option<IpAddr>>) {
+    let mode = cfg.server.conntrack_control.mode;
+    let mut v4_targets: BTreeSet<Option<IpAddr>> = BTreeSet::new();
+    let mut v6_targets: BTreeSet<Option<IpAddr>> = BTreeSet::new();
+
+    match mode {
+        ConntrackMode::Tracked => {}
+        ConntrackMode::Notrack => {
+            if cfg.server.listeners.is_empty() {
+                if let Some(ipv4) = cfg
+                    .server
+                    .listen_addr_ipv4
+                    .as_ref()
+                    .and_then(|s| s.parse::<IpAddr>().ok())
+                {
+                    if ipv4.is_unspecified() {
+                        v4_targets.insert(None);
+                    } else {
+                        v4_targets.insert(Some(ipv4));
+                    }
+                }
+                if let Some(ipv6) = cfg
+                    .server
+                    .listen_addr_ipv6
+                    .as_ref()
+                    .and_then(|s| s.parse::<IpAddr>().ok())
+                {
+                    if ipv6.is_unspecified() {
+                        v6_targets.insert(None);
+                    } else {
+                        v6_targets.insert(Some(ipv6));
+                    }
+                }
+            } else {
+                for listener in &cfg.server.listeners {
+                    if listener.ip.is_ipv4() {
+                        if listener.ip.is_unspecified() {
+                            v4_targets.insert(None);
+                        } else {
+                            v4_targets.insert(Some(listener.ip));
+                        }
+                    } else if listener.ip.is_unspecified() {
+                        v6_targets.insert(None);
+                    } else {
+                        v6_targets.insert(Some(listener.ip));
+                    }
+                }
+            }
+        }
+        ConntrackMode::Hybrid => {
+            for ip in &cfg.server.conntrack_control.hybrid_listener_ips {
+                if ip.is_ipv4() {
+                    v4_targets.insert(Some(*ip));
+                } else {
+                    v6_targets.insert(Some(*ip));
+                }
+            }
+        }
+    }
+
+    (
+        v4_targets.into_iter().collect(),
+        v6_targets.into_iter().collect(),
+    )
+}
+
+async fn apply_nft_rules(cfg: &ProxyConfig) -> Result<(), String> {
+    let _ = run_command(
+        "nft",
+        &["delete", "table", "inet", "telemt_conntrack"],
+        None,
+    )
+    .await;
+    if matches!(cfg.server.conntrack_control.mode, ConntrackMode::Tracked) {
+        return Ok(());
+    }
+
+    let (v4_targets, v6_targets) = notrack_targets(cfg);
+    let mut rules = Vec::new();
+    for ip in v4_targets {
+        let rule = if let Some(ip) = ip {
+            format!("tcp dport {} ip daddr {} notrack", cfg.server.port, ip)
+        } else {
+            format!("tcp dport {} notrack", cfg.server.port)
+        };
+        rules.push(rule);
+    }
+    for ip in v6_targets {
+        let rule = if let Some(ip) = ip {
+            format!("tcp dport {} ip6 daddr {} notrack", cfg.server.port, ip)
+        } else {
+            format!("tcp dport {} notrack", cfg.server.port)
+        };
+        rules.push(rule);
+    }
+
+    let rule_blob = if rules.is_empty() {
+        String::new()
+    } else {
+        format!("    {}\n", rules.join("\n    "))
+    };
+    let script = format!(
+        "table inet telemt_conntrack {{\n  chain preraw {{\n    type filter hook prerouting priority raw; policy accept;\n{rule_blob}  }}\n}}\n"
+    );
+    run_command("nft", &["-f", "-"], Some(script)).await
+}
+
+async fn apply_iptables_rules(cfg: &ProxyConfig) -> Result<(), String> {
+    apply_iptables_rules_for_binary("iptables", cfg, true).await?;
+    apply_iptables_rules_for_binary("ip6tables", cfg, false).await?;
+    Ok(())
+}
+
+async fn apply_iptables_rules_for_binary(
+    binary: &str,
+    cfg: &ProxyConfig,
+    ipv4: bool,
+) -> Result<(), String> {
+    if !command_exists(binary) {
+        return Ok(());
+    }
+    let chain = "TELEMT_NOTRACK";
+    let _ = run_command(
+        binary,
+        &["-t", "raw", "-D", "PREROUTING", "-j", chain],
+        None,
+    )
+    .await;
+    let _ = run_command(binary, &["-t", "raw", "-F", chain], None).await;
+    let _ = run_command(binary, &["-t", "raw", "-X", chain], None).await;
+
+    if matches!(cfg.server.conntrack_control.mode, ConntrackMode::Tracked) {
+        return Ok(());
+    }
+
+    run_command(binary, &["-t", "raw", "-N", chain], None).await?;
+    run_command(binary, &["-t", "raw", "-F", chain], None).await?;
+    if run_command(
+        binary,
+        &["-t", "raw", "-C", "PREROUTING", "-j", chain],
+        None,
+    )
+    .await
+    .is_err()
+    {
+        run_command(
+            binary,
+            &["-t", "raw", "-I", "PREROUTING", "1", "-j", chain],
+            None,
+        )
+        .await?;
+    }
+
+    let (v4_targets, v6_targets) = notrack_targets(cfg);
+    let selected = if ipv4 { v4_targets } else { v6_targets };
+    for ip in selected {
+        let mut args = vec![
+            "-t".to_string(),
+            "raw".to_string(),
+            "-A".to_string(),
+            chain.to_string(),
+            "-p".to_string(),
+            "tcp".to_string(),
+            "--dport".to_string(),
+            cfg.server.port.to_string(),
+        ];
+        if let Some(ip) = ip {
+            args.push("-d".to_string());
+            args.push(ip.to_string());
+        }
+        args.push("-j".to_string());
+        args.push("CT".to_string());
+        args.push("--notrack".to_string());
+        let arg_refs: Vec<&str> = args.iter().map(String::as_str).collect();
+        run_command(binary, &arg_refs, None).await?;
+    }
+    Ok(())
+}
+
+async fn clear_notrack_rules_all_backends() {
+    let _ = run_command(
+        "nft",
+        &["delete", "table", "inet", "telemt_conntrack"],
+        None,
+    )
+    .await;
+    let _ = run_command(
+        "iptables",
+        &["-t", "raw", "-D", "PREROUTING", "-j", "TELEMT_NOTRACK"],
+        None,
+    )
+    .await;
+    let _ = run_command("iptables", &["-t", "raw", "-F", "TELEMT_NOTRACK"], None).await;
+    let _ = run_command("iptables", &["-t", "raw", "-X", "TELEMT_NOTRACK"], None).await;
+    let _ = run_command(
+        "ip6tables",
+        &["-t", "raw", "-D", "PREROUTING", "-j", "TELEMT_NOTRACK"],
+        None,
+    )
+    .await;
+    let _ = run_command("ip6tables", &["-t", "raw", "-F", "TELEMT_NOTRACK"], None).await;
+    let _ = run_command("ip6tables", &["-t", "raw", "-X", "TELEMT_NOTRACK"], None).await;
+}
+
+enum DeleteOutcome {
+    Deleted,
+    NotFound,
+    Error,
+}
+
+async fn delete_conntrack_entry(event: ConntrackCloseEvent) -> DeleteOutcome {
+    if !command_exists("conntrack") {
+        return DeleteOutcome::Error;
+    }
+    let args = vec![
+        "-D".to_string(),
+        "-p".to_string(),
+        "tcp".to_string(),
+        "-s".to_string(),
+        event.src.ip().to_string(),
+        "--sport".to_string(),
+        event.src.port().to_string(),
+        "-d".to_string(),
+        event.dst.ip().to_string(),
+        "--dport".to_string(),
+        event.dst.port().to_string(),
+    ];
+    let arg_refs: Vec<&str> = args.iter().map(String::as_str).collect();
+    match run_command("conntrack", &arg_refs, None).await {
+        Ok(()) => DeleteOutcome::Deleted,
+        Err(error) => {
+            if error.contains("0 flow entries have been deleted") {
+                DeleteOutcome::NotFound
+            } else {
+                debug!(error = %error, "conntrack delete failed");
+                DeleteOutcome::Error
+            }
+        }
+    }
+}
+
+async fn run_command(binary: &str, args: &[&str], stdin: Option<String>) -> Result<(), String> {
+    if !command_exists(binary) {
+        return Err(format!("{binary} is not available"));
+    }
+    let mut command = Command::new(binary);
+    command.args(args);
+    if stdin.is_some() {
+        command.stdin(std::process::Stdio::piped());
+    }
+    command.stdout(std::process::Stdio::null());
+    command.stderr(std::process::Stdio::piped());
+    let mut child = command
+        .spawn()
+        .map_err(|e| format!("spawn {binary} failed: {e}"))?;
+    if let Some(blob) = stdin
+        && let Some(mut writer) = child.stdin.take()
+    {
+        writer
+            .write_all(blob.as_bytes())
+            .await
+            .map_err(|e| format!("stdin write {binary} failed: {e}"))?;
+    }
+    let output = child
+        .wait_with_output()
+        .await
+        .map_err(|e| format!("wait {binary} failed: {e}"))?;
+    if output.status.success() {
+        return Ok(());
+    }
+    let stderr = String::from_utf8_lossy(&output.stderr).trim().to_string();
+    Err(if stderr.is_empty() {
+        format!("{binary} exited with status {}", output.status)
+    } else {
+        stderr
+    })
+}
+
+fn fd_usage_pct() -> Option<u8> {
+    let soft_limit = nofile_soft_limit()?;
+    if soft_limit == 0 {
+        return None;
+    }
+    let fd_count = std::fs::read_dir("/proc/self/fd").ok()?.count() as u64;
+    Some(((fd_count.saturating_mul(100)) / soft_limit).min(100) as u8)
+}
+
+fn nofile_soft_limit() -> Option<u64> {
+    #[cfg(target_os = "linux")]
+    {
+        let mut lim = libc::rlimit {
+            rlim_cur: 0,
+            rlim_max: 0,
+        };
+        let rc = unsafe { libc::getrlimit(libc::RLIMIT_NOFILE, &mut lim) };
+        if rc != 0 {
+            return None;
+        }
+        return Some(lim.rlim_cur);
+    }
+    #[cfg(not(target_os = "linux"))]
+    {
+        None
+    }
+}
+
+fn has_cap_net_admin() -> bool {
+    #[cfg(target_os = "linux")]
+    {
+        let Ok(status) = std::fs::read_to_string("/proc/self/status") else {
+            return false;
+        };
+        for line in status.lines() {
+            if let Some(raw) = line.strip_prefix("CapEff:") {
+                let caps = raw.trim();
+                if let Ok(bits) = u64::from_str_radix(caps, 16) {
+                    const CAP_NET_ADMIN_BIT: u64 = 12;
+                    return (bits & (1u64 << CAP_NET_ADMIN_BIT)) != 0;
+                }
+            }
+        }
+        false
+    }
+    #[cfg(not(target_os = "linux"))]
+    {
+        false
+    }
+}
+
+#[cfg(test)]
+mod tests {
+    use super::*;
+    use crate::config::ProxyConfig;
+
+    #[test]
+    fn pressure_activates_on_accept_timeout_spike() {
+        let stats = Stats::new();
+        let shared = ProxySharedState::new();
+        let mut cfg = ProxyConfig::default();
+        cfg.server.conntrack_control.inline_conntrack_control = true;
+        let mut state = PressureState::new(&stats);
+        let sample = PressureSample {
+            conn_pct: Some(10),
+            fd_pct: Some(10),
+            accept_timeout_delta: 1,
+            me_queue_pressure_delta: 0,
+        };
+
+        update_pressure_state(&stats, shared.as_ref(), &cfg, &sample, &mut state);
+
+        assert!(state.active);
+        assert!(shared.conntrack_pressure_active());
+        assert!(stats.get_conntrack_pressure_active());
+    }
+
+    #[test]
+    fn pressure_releases_after_hysteresis_window() {
+        let stats = Stats::new();
+        let shared = ProxySharedState::new();
+        let mut cfg = ProxyConfig::default();
+        cfg.server.conntrack_control.inline_conntrack_control = true;
+        let mut state = PressureState::new(&stats);
+
+        let high_sample = PressureSample {
+            conn_pct: Some(95),
+            fd_pct: Some(95),
+            accept_timeout_delta: 0,
+            me_queue_pressure_delta: 0,
+        };
+        update_pressure_state(&stats, shared.as_ref(), &cfg, &high_sample, &mut state);
+        assert!(state.active);
+
+        let low_sample = PressureSample {
+            conn_pct: Some(10),
+            fd_pct: Some(10),
+            accept_timeout_delta: 0,
+            me_queue_pressure_delta: 0,
+        };
+        update_pressure_state(&stats, shared.as_ref(), &cfg, &low_sample, &mut state);
+        assert!(state.active);
+        update_pressure_state(&stats, shared.as_ref(), &cfg, &low_sample, &mut state);
+        assert!(state.active);
+        update_pressure_state(&stats, shared.as_ref(), &cfg, &low_sample, &mut state);
+
+        assert!(!state.active);
+        assert!(!shared.conntrack_pressure_active());
+        assert!(!stats.get_conntrack_pressure_active());
+    }
+
+    #[test]
+    fn pressure_does_not_activate_when_disabled() {
+        let stats = Stats::new();
+        let shared = ProxySharedState::new();
+        let mut cfg = ProxyConfig::default();
+        cfg.server.conntrack_control.inline_conntrack_control = false;
+        let mut state = PressureState::new(&stats);
+        let sample = PressureSample {
+            conn_pct: Some(100),
+            fd_pct: Some(100),
+            accept_timeout_delta: 10,
+            me_queue_pressure_delta: 10,
+        };
+
+        update_pressure_state(&stats, shared.as_ref(), &cfg, &sample, &mut state);
+
+        assert!(!state.active);
+        assert!(!shared.conntrack_pressure_active());
+        assert!(!stats.get_conntrack_pressure_active());
+    }
+}

src/daemon/mod.rs

@@ -0,0 +1,541 @@
+//! Unix daemon support for telemt.
+//!
+//! Provides classic Unix daemonization (double-fork), PID file management,
+//! and privilege dropping for running telemt as a background service.
+
+use std::fs::{self, File, OpenOptions};
+use std::io::{self, Read, Write};
+use std::os::unix::fs::OpenOptionsExt;
+use std::path::{Path, PathBuf};
+
+use nix::fcntl::{Flock, FlockArg};
+use nix::unistd::{self, ForkResult, Gid, Pid, Uid, chdir, close, fork, getpid, setsid};
+use tracing::{debug, info, warn};
+
+/// Default PID file location.
+pub const DEFAULT_PID_FILE: &str = "/var/run/telemt.pid";
+
+/// Daemon configuration options parsed from CLI.
+#[derive(Debug, Clone, Default)]
+pub struct DaemonOptions {
+    /// Run as daemon (fork to background).
+    pub daemonize: bool,
+    /// Path to PID file.
+    pub pid_file: Option<PathBuf>,
+    /// User to run as after binding sockets.
+    pub user: Option<String>,
+    /// Group to run as after binding sockets.
+    pub group: Option<String>,
+    /// Working directory for the daemon.
+    pub working_dir: Option<PathBuf>,
+    /// Explicit foreground mode (for systemd Type=simple).
+    pub foreground: bool,
+}
+
+impl DaemonOptions {
+    /// Returns the effective PID file path.
+    pub fn pid_file_path(&self) -> &Path {
+        self.pid_file
+            .as_deref()
+            .unwrap_or(Path::new(DEFAULT_PID_FILE))
+    }
+
+    /// Returns true if we should actually daemonize.
+    /// Foreground flag takes precedence.
+    pub fn should_daemonize(&self) -> bool {
+        self.daemonize && !self.foreground
+    }
+}
+
+/// Error types for daemon operations.
+#[derive(Debug, thiserror::Error)]
+pub enum DaemonError {
+    #[error("fork failed: {0}")]
+    ForkFailed(#[source] nix::Error),
+
+    #[error("setsid failed: {0}")]
+    SetsidFailed(#[source] nix::Error),
+
+    #[error("chdir failed: {0}")]
+    ChdirFailed(#[source] nix::Error),
+
+    #[error("failed to open /dev/null: {0}")]
+    DevNullFailed(#[source] io::Error),
+
+    #[error("failed to redirect stdio: {0}")]
+    RedirectFailed(#[source] nix::Error),
+
+    #[error("PID file error: {0}")]
+    PidFile(String),
+
+    #[error("another instance is already running (pid {0})")]
+    AlreadyRunning(i32),
+
+    #[error("user '{0}' not found")]
+    UserNotFound(String),
+
+    #[error("group '{0}' not found")]
+    GroupNotFound(String),
+
+    #[error("failed to set uid/gid: {0}")]
+    PrivilegeDrop(#[source] nix::Error),
+
+    #[error("io error: {0}")]
+    Io(#[from] io::Error),
+}
+
+/// Result of a successful daemonize() call.
+#[derive(Debug)]
+pub enum DaemonizeResult {
+    /// We are the parent process and should exit.
+    Parent,
+    /// We are the daemon child process and should continue.
+    Child,
+}
+
+/// Performs classic Unix double-fork daemonization.
+///
+/// This detaches the process from the controlling terminal:
+/// 1. First fork - parent exits, child continues
+/// 2. setsid() - become session leader
+/// 3. Second fork - ensure we can never acquire a controlling terminal
+/// 4. chdir("/") - don't hold any directory open
+/// 5. Redirect stdin/stdout/stderr to /dev/null
+///
+/// Returns `DaemonizeResult::Parent` in the original parent (which should exit),
+/// or `DaemonizeResult::Child` in the final daemon child.
+pub fn daemonize(working_dir: Option<&Path>) -> Result<DaemonizeResult, DaemonError> {
+    // First fork
+    match unsafe { fork() } {
+        Ok(ForkResult::Parent { .. }) => {
+            // Parent exits
+            return Ok(DaemonizeResult::Parent);
+        }
+        Ok(ForkResult::Child) => {
+            // Child continues
+        }
+        Err(e) => return Err(DaemonError::ForkFailed(e)),
+    }
+
+    // Create new session, become session leader
+    setsid().map_err(DaemonError::SetsidFailed)?;
+
+    // Second fork to ensure we can never acquire a controlling terminal
+    match unsafe { fork() } {
+        Ok(ForkResult::Parent { .. }) => {
+            // Intermediate parent exits
+            std::process::exit(0);
+        }
+        Ok(ForkResult::Child) => {
+            // Final daemon child continues
+        }
+        Err(e) => return Err(DaemonError::ForkFailed(e)),
+    }
+
+    // Change working directory
+    let target_dir = working_dir.unwrap_or(Path::new("/"));
+    chdir(target_dir).map_err(DaemonError::ChdirFailed)?;
+
+    // Redirect stdin, stdout, stderr to /dev/null
+    redirect_stdio_to_devnull()?;
+
+    Ok(DaemonizeResult::Child)
+}
+
+/// Redirects stdin, stdout, and stderr to /dev/null.
+fn redirect_stdio_to_devnull() -> Result<(), DaemonError> {
+    let devnull = File::options()
+        .read(true)
+        .write(true)
+        .open("/dev/null")
+        .map_err(DaemonError::DevNullFailed)?;
+
+    let devnull_fd = std::os::unix::io::AsRawFd::as_raw_fd(&devnull);
+
+    // Use libc::dup2 directly for redirecting standard file descriptors
+    // nix 0.31's dup2 requires OwnedFd which doesn't work well with stdio fds
+    unsafe {
+        // Redirect stdin (fd 0)
+        if libc::dup2(devnull_fd, 0) < 0 {
+            return Err(DaemonError::RedirectFailed(nix::errno::Errno::last()));
+        }
+        // Redirect stdout (fd 1)
+        if libc::dup2(devnull_fd, 1) < 0 {
+            return Err(DaemonError::RedirectFailed(nix::errno::Errno::last()));
+        }
+        // Redirect stderr (fd 2)
+        if libc::dup2(devnull_fd, 2) < 0 {
+            return Err(DaemonError::RedirectFailed(nix::errno::Errno::last()));
+        }
+    }
+
+    // Close original devnull fd if it's not one of the standard fds
+    if devnull_fd > 2 {
+        let _ = close(devnull_fd);
+    }
+
+    Ok(())
+}
+
+/// PID file manager with flock-based locking.
+pub struct PidFile {
+    path: PathBuf,
+    file: Option<File>,
+    locked: bool,
+}
+
+impl PidFile {
+    /// Creates a new PID file manager for the given path.
+    pub fn new<P: AsRef<Path>>(path: P) -> Self {
+        Self {
+            path: path.as_ref().to_path_buf(),
+            file: None,
+            locked: false,
+        }
+    }
+
+    /// Checks if another instance is already running.
+    ///
+    /// Returns the PID of the running instance if one exists.
+    pub fn check_running(&self) -> Result<Option<i32>, DaemonError> {
+        if !self.path.exists() {
+            return Ok(None);
+        }
+
+        // Try to read existing PID
+        let mut contents = String::new();
+        File::open(&self.path)
+            .and_then(|mut f| f.read_to_string(&mut contents))
+            .map_err(|e| {
+                DaemonError::PidFile(format!("cannot read {}: {}", self.path.display(), e))
+            })?;
+
+        let pid: i32 = contents
+            .trim()
+            .parse()
+            .map_err(|_| DaemonError::PidFile(format!("invalid PID in {}", self.path.display())))?;
+
+        // Check if process is still running
+        if is_process_running(pid) {
+            Ok(Some(pid))
+        } else {
+            // Stale PID file
+            debug!(pid, path = %self.path.display(), "Removing stale PID file");
+            let _ = fs::remove_file(&self.path);
+            Ok(None)
+        }
+    }
+
+    /// Acquires the PID file lock and writes the current PID.
+    ///
+    /// Fails if another instance is already running.
+    pub fn acquire(&mut self) -> Result<(), DaemonError> {
+        // Check for running instance first
+        if let Some(pid) = self.check_running()? {
+            return Err(DaemonError::AlreadyRunning(pid));
+        }
+
+        // Ensure parent directory exists
+        if let Some(parent) = self.path.parent() {
+            if !parent.exists() {
+                fs::create_dir_all(parent).map_err(|e| {
+                    DaemonError::PidFile(format!(
+                        "cannot create directory {}: {}",
+                        parent.display(),
+                        e
+                    ))
+                })?;
+            }
+        }
+
+        // Open/create PID file with exclusive lock
+        let file = OpenOptions::new()
+            .write(true)
+            .create(true)
+            .truncate(true)
+            .mode(0o644)
+            .open(&self.path)
+            .map_err(|e| {
+                DaemonError::PidFile(format!("cannot open {}: {}", self.path.display(), e))
+            })?;
+
+        // Try to acquire exclusive lock (non-blocking)
+        let flock = Flock::lock(file, FlockArg::LockExclusiveNonblock).map_err(|(_, errno)| {
+            // Check if another instance grabbed the lock
+            if let Some(pid) = self.check_running().ok().flatten() {
+                DaemonError::AlreadyRunning(pid)
+            } else {
+                DaemonError::PidFile(format!("cannot lock {}: {}", self.path.display(), errno))
+            }
+        })?;
+
+        // Write our PID
+        let pid = getpid();
+        let mut file = flock
+            .unlock()
+            .map_err(|(_, errno)| DaemonError::PidFile(format!("unlock failed: {}", errno)))?;
+
+        writeln!(file, "{}", pid).map_err(|e| {
+            DaemonError::PidFile(format!(
+                "cannot write PID to {}: {}",
+                self.path.display(),
+                e
+            ))
+        })?;
+
+        // Re-acquire lock and keep it
+        let flock = Flock::lock(file, FlockArg::LockExclusiveNonblock).map_err(|(_, errno)| {
+            DaemonError::PidFile(format!("cannot re-lock {}: {}", self.path.display(), errno))
+        })?;
+
+        self.file = Some(flock.unlock().map_err(|(_, errno)| {
+            DaemonError::PidFile(format!("unlock for storage failed: {}", errno))
+        })?);
+        self.locked = true;
+
+        info!(pid = pid.as_raw(), path = %self.path.display(), "PID file created");
+        Ok(())
+    }
+
+    /// Releases the PID file lock and removes the file.
+    pub fn release(&mut self) -> Result<(), DaemonError> {
+        if let Some(file) = self.file.take() {
+            drop(file);
+        }
+        self.locked = false;
+
+        if self.path.exists() {
+            fs::remove_file(&self.path).map_err(|e| {
+                DaemonError::PidFile(format!("cannot remove {}: {}", self.path.display(), e))
+            })?;
+            debug!(path = %self.path.display(), "PID file removed");
+        }
+
+        Ok(())
+    }
+
+    /// Returns the path to this PID file.
+    #[allow(dead_code)]
+    pub fn path(&self) -> &Path {
+        &self.path
+    }
+}
+
+impl Drop for PidFile {
+    fn drop(&mut self) {
+        if self.locked {
+            if let Err(e) = self.release() {
+                warn!(error = %e, "Failed to clean up PID file on drop");
+            }
+        }
+    }
+}
+
+/// Checks if a process with the given PID is running.
+fn is_process_running(pid: i32) -> bool {
+    // kill(pid, 0) checks if process exists without sending a signal
+    nix::sys::signal::kill(Pid::from_raw(pid), None).is_ok()
+}
+
+/// Drops privileges to the specified user and group.
+///
+/// This should be called after binding privileged ports but before
+/// entering the main event loop.
+pub fn drop_privileges(user: Option<&str>, group: Option<&str>) -> Result<(), DaemonError> {
+    // Look up group first (need to do this while still root)
+    let target_gid = if let Some(group_name) = group {
+        Some(lookup_group(group_name)?)
+    } else if let Some(user_name) = user {
+        // If no group specified but user is, use user's primary group
+        Some(lookup_user_primary_gid(user_name)?)
+    } else {
+        None
+    };
+
+    // Look up user
+    let target_uid = if let Some(user_name) = user {
+        Some(lookup_user(user_name)?)
+    } else {
+        None
+    };
+
+    // Drop privileges: set GID first, then UID
+    // (Setting UID first would prevent us from setting GID)
+    if let Some(gid) = target_gid {
+        unistd::setgid(gid).map_err(DaemonError::PrivilegeDrop)?;
+        // Also set supplementary groups to just this one
+        unistd::setgroups(&[gid]).map_err(DaemonError::PrivilegeDrop)?;
+        info!(gid = gid.as_raw(), "Dropped group privileges");
+    }
+
+    if let Some(uid) = target_uid {
+        unistd::setuid(uid).map_err(DaemonError::PrivilegeDrop)?;
+        info!(uid = uid.as_raw(), "Dropped user privileges");
+    }
+
+    Ok(())
+}
+
+/// Looks up a user by name and returns their UID.
+fn lookup_user(name: &str) -> Result<Uid, DaemonError> {
+    // Use libc getpwnam
+    let c_name =
+        std::ffi::CString::new(name).map_err(|_| DaemonError::UserNotFound(name.to_string()))?;
+
+    unsafe {
+        let pwd = libc::getpwnam(c_name.as_ptr());
+        if pwd.is_null() {
+            Err(DaemonError::UserNotFound(name.to_string()))
+        } else {
+            Ok(Uid::from_raw((*pwd).pw_uid))
+        }
+    }
+}
+
+/// Looks up a user's primary GID by username.
+fn lookup_user_primary_gid(name: &str) -> Result<Gid, DaemonError> {
+    let c_name =
+        std::ffi::CString::new(name).map_err(|_| DaemonError::UserNotFound(name.to_string()))?;
+
+    unsafe {
+        let pwd = libc::getpwnam(c_name.as_ptr());
+        if pwd.is_null() {
+            Err(DaemonError::UserNotFound(name.to_string()))
+        } else {
+            Ok(Gid::from_raw((*pwd).pw_gid))
+        }
+    }
+}
+
+/// Looks up a group by name and returns its GID.
+fn lookup_group(name: &str) -> Result<Gid, DaemonError> {
+    let c_name =
+        std::ffi::CString::new(name).map_err(|_| DaemonError::GroupNotFound(name.to_string()))?;
+
+    unsafe {
+        let grp = libc::getgrnam(c_name.as_ptr());
+        if grp.is_null() {
+            Err(DaemonError::GroupNotFound(name.to_string()))
+        } else {
+            Ok(Gid::from_raw((*grp).gr_gid))
+        }
+    }
+}
+
+/// Reads PID from a PID file.
+#[allow(dead_code)]
+pub fn read_pid_file<P: AsRef<Path>>(path: P) -> Result<i32, DaemonError> {
+    let path = path.as_ref();
+    let mut contents = String::new();
+    File::open(path)
+        .and_then(|mut f| f.read_to_string(&mut contents))
+        .map_err(|e| DaemonError::PidFile(format!("cannot read {}: {}", path.display(), e)))?;
+
+    contents
+        .trim()
+        .parse()
+        .map_err(|_| DaemonError::PidFile(format!("invalid PID in {}", path.display())))
+}
+
+/// Sends a signal to the process specified in a PID file.
+#[allow(dead_code)]
+pub fn signal_pid_file<P: AsRef<Path>>(
+    path: P,
+    signal: nix::sys::signal::Signal,
+) -> Result<(), DaemonError> {
+    let pid = read_pid_file(&path)?;
+
+    if !is_process_running(pid) {
+        return Err(DaemonError::PidFile(format!(
+            "process {} from {} is not running",
+            pid,
+            path.as_ref().display()
+        )));
+    }
+
+    nix::sys::signal::kill(Pid::from_raw(pid), signal)
+        .map_err(|e| DaemonError::PidFile(format!("cannot signal process {}: {}", pid, e)))?;
+
+    Ok(())
+}
+
+/// Returns the status of the daemon based on PID file.
+#[allow(dead_code)]
+#[derive(Debug, Clone, PartialEq, Eq)]
+pub enum DaemonStatus {
+    /// Daemon is running with the given PID.
+    Running(i32),
+    /// PID file exists but process is not running.
+    Stale(i32),
+    /// No PID file exists.
+    NotRunning,
+}
+
+/// Checks the daemon status from a PID file.
+#[allow(dead_code)]
+pub fn check_status<P: AsRef<Path>>(path: P) -> DaemonStatus {
+    let path = path.as_ref();
+
+    if !path.exists() {
+        return DaemonStatus::NotRunning;
+    }
+
+    match read_pid_file(path) {
+        Ok(pid) => {
+            if is_process_running(pid) {
+                DaemonStatus::Running(pid)
+            } else {
+                DaemonStatus::Stale(pid)
+            }
+        }
+        Err(_) => DaemonStatus::NotRunning,
+    }
+}
+
+#[cfg(test)]
+mod tests {
+    use super::*;
+
+    #[test]
+    fn test_daemon_options_default() {
+        let opts = DaemonOptions::default();
+        assert!(!opts.daemonize);
+        assert!(!opts.should_daemonize());
+        assert_eq!(opts.pid_file_path(), Path::new(DEFAULT_PID_FILE));
+    }
+
+    #[test]
+    fn test_daemon_options_foreground_overrides() {
+        let opts = DaemonOptions {
+            daemonize: true,
+            foreground: true,
+            ..Default::default()
+        };
+        assert!(!opts.should_daemonize());
+    }
+
+    #[test]
+    fn test_check_status_not_running() {
+        let path = "/tmp/telemt_test_nonexistent.pid";
+        assert_eq!(check_status(path), DaemonStatus::NotRunning);
+    }
+
+    #[test]
+    fn test_pid_file_basic() {
+        let path = "/tmp/telemt_test_pidfile.pid";
+        let _ = fs::remove_file(path);
+
+        let mut pf = PidFile::new(path);
+        assert!(pf.check_running().unwrap().is_none());
+
+        pf.acquire().unwrap();
+        assert!(Path::new(path).exists());
+
+        // Read it back
+        let pid = read_pid_file(path).unwrap();
+        assert_eq!(pid, std::process::id() as i32);
+
+        pf.release().unwrap();
+        assert!(!Path::new(path).exists());
+    }
+}

src/error.rs

@@ -216,6 +216,9 @@ pub enum ProxyError {
     #[error("Invalid proxy protocol header")]
     InvalidProxyProtocol,
 
+    #[error("Unknown TLS SNI")]
+    UnknownTlsSni,
+
     #[error("Proxy error: {0}")]
     Proxy(String),
 

src/ip_tracker.rs

@@ -26,6 +26,15 @@ pub struct UserIpTracker {
     cleanup_drain_lock: Arc<AsyncMutex<()>>,
 }
 
+#[derive(Debug, Clone, Copy)]
+pub struct UserIpTrackerMemoryStats {
+    pub active_users: usize,
+    pub recent_users: usize,
+    pub active_entries: usize,
+    pub recent_entries: usize,
+    pub cleanup_queue_len: usize,
+}
+
 impl UserIpTracker {
     pub fn new() -> Self {
         Self {
@@ -141,6 +150,13 @@ impl UserIpTracker {
 
         let mut active_ips = self.active_ips.write().await;
         let mut recent_ips = self.recent_ips.write().await;
+        let window = *self.limit_window.read().await;
+        let now = Instant::now();
+
+        for user_recent in recent_ips.values_mut() {
+            Self::prune_recent(user_recent, now, window);
+        }
+
         let mut users =
             Vec::<String>::with_capacity(active_ips.len().saturating_add(recent_ips.len()));
         users.extend(active_ips.keys().cloned());
@@ -166,6 +182,26 @@ impl UserIpTracker {
         }
     }
 
+    pub async fn memory_stats(&self) -> UserIpTrackerMemoryStats {
+        let cleanup_queue_len = self
+            .cleanup_queue
+            .lock()
+            .unwrap_or_else(|poisoned| poisoned.into_inner())
+            .len();
+        let active_ips = self.active_ips.read().await;
+        let recent_ips = self.recent_ips.read().await;
+        let active_entries = active_ips.values().map(HashMap::len).sum();
+        let recent_entries = recent_ips.values().map(HashMap::len).sum();
+
+        UserIpTrackerMemoryStats {
+            active_users: active_ips.len(),
+            recent_users: recent_ips.len(),
+            active_entries,
+            recent_entries,
+            cleanup_queue_len,
+        }
+    }
+
     pub async fn set_limit_policy(&self, mode: UserMaxUniqueIpsMode, window_secs: u64) {
         {
             let mut current_mode = self.limit_mode.write().await;
@@ -451,6 +487,7 @@ impl Default for UserIpTracker {
 mod tests {
     use super::*;
     use std::net::{IpAddr, Ipv4Addr, Ipv6Addr};
+    use std::sync::atomic::Ordering;
 
     fn test_ipv4(oct1: u8, oct2: u8, oct3: u8, oct4: u8) -> IpAddr {
         IpAddr::V4(Ipv4Addr::new(oct1, oct2, oct3, oct4))
@@ -764,4 +801,54 @@ mod tests {
         tokio::time::sleep(Duration::from_millis(1100)).await;
         assert!(tracker.check_and_add("test_user", ip2).await.is_ok());
     }
+
+    #[tokio::test]
+    async fn test_memory_stats_reports_queue_and_entry_counts() {
+        let tracker = UserIpTracker::new();
+        tracker.set_user_limit("test_user", 4).await;
+        let ip1 = test_ipv4(10, 2, 0, 1);
+        let ip2 = test_ipv4(10, 2, 0, 2);
+
+        tracker.check_and_add("test_user", ip1).await.unwrap();
+        tracker.check_and_add("test_user", ip2).await.unwrap();
+        tracker.enqueue_cleanup("test_user".to_string(), ip1);
+
+        let snapshot = tracker.memory_stats().await;
+        assert_eq!(snapshot.active_users, 1);
+        assert_eq!(snapshot.recent_users, 1);
+        assert_eq!(snapshot.active_entries, 2);
+        assert_eq!(snapshot.recent_entries, 2);
+        assert_eq!(snapshot.cleanup_queue_len, 1);
+    }
+
+    #[tokio::test]
+    async fn test_compact_prunes_stale_recent_entries() {
+        let tracker = UserIpTracker::new();
+        tracker
+            .set_limit_policy(UserMaxUniqueIpsMode::TimeWindow, 1)
+            .await;
+
+        let stale_user = "stale-user".to_string();
+        let stale_ip = test_ipv4(10, 3, 0, 1);
+        {
+            let mut recent_ips = tracker.recent_ips.write().await;
+            recent_ips
+                .entry(stale_user.clone())
+                .or_insert_with(HashMap::new)
+                .insert(stale_ip, Instant::now() - Duration::from_secs(5));
+        }
+
+        tracker.last_compact_epoch_secs.store(0, Ordering::Relaxed);
+        tracker
+            .check_and_add("trigger-user", test_ipv4(10, 3, 0, 2))
+            .await
+            .unwrap();
+
+        let recent_ips = tracker.recent_ips.read().await;
+        let stale_exists = recent_ips
+            .get(&stale_user)
+            .map(|ips| ips.contains_key(&stale_ip))
+            .unwrap_or(false);
+        assert!(!stale_exists);
+    }
 }

src/logging.rs

@@ -0,0 +1,343 @@
+//! Logging configuration for telemt.
+//!
+//! Supports multiple log destinations:
+//! - stderr (default, works with systemd journald)
+//! - syslog (Unix only, for traditional init systems)
+//! - file (with optional rotation)
+
+#![allow(dead_code)] // Infrastructure module - used via CLI flags
+
+use std::path::Path;
+
+use tracing_subscriber::layer::SubscriberExt;
+use tracing_subscriber::util::SubscriberInitExt;
+use tracing_subscriber::{EnvFilter, fmt, reload};
+
+/// Log destination configuration.
+#[derive(Debug, Clone, Default)]
+pub enum LogDestination {
+    /// Log to stderr (default, captured by systemd journald).
+    #[default]
+    Stderr,
+    /// Log to syslog (Unix only).
+    #[cfg(unix)]
+    Syslog,
+    /// Log to a file with optional rotation.
+    File {
+        path: String,
+        /// Rotate daily if true.
+        rotate_daily: bool,
+    },
+}
+
+/// Logging options parsed from CLI/config.
+#[derive(Debug, Clone, Default)]
+pub struct LoggingOptions {
+    /// Where to send logs.
+    pub destination: LogDestination,
+    /// Disable ANSI colors.
+    pub disable_colors: bool,
+}
+
+/// Guard that must be held to keep file logging active.
+/// When dropped, flushes and closes log files.
+pub struct LoggingGuard {
+    _guard: Option<tracing_appender::non_blocking::WorkerGuard>,
+}
+
+impl LoggingGuard {
+    fn new(guard: Option<tracing_appender::non_blocking::WorkerGuard>) -> Self {
+        Self { _guard: guard }
+    }
+
+    /// Creates a no-op guard for stderr/syslog logging.
+    pub fn noop() -> Self {
+        Self { _guard: None }
+    }
+}
+
+/// Initialize the tracing subscriber with the specified options.
+///
+/// Returns a reload handle for dynamic log level changes and a guard
+/// that must be kept alive for file logging.
+pub fn init_logging(
+    opts: &LoggingOptions,
+    initial_filter: &str,
+) -> (
+    reload::Handle<EnvFilter, impl tracing::Subscriber + Send + Sync>,
+    LoggingGuard,
+) {
+    let (filter_layer, filter_handle) = reload::Layer::new(EnvFilter::new(initial_filter));
+
+    match &opts.destination {
+        LogDestination::Stderr => {
+            let fmt_layer = fmt::Layer::default()
+                .with_ansi(!opts.disable_colors)
+                .with_target(true);
+
+            tracing_subscriber::registry()
+                .with(filter_layer)
+                .with(fmt_layer)
+                .init();
+
+            (filter_handle, LoggingGuard::noop())
+        }
+
+        #[cfg(unix)]
+        LogDestination::Syslog => {
+            // Use a custom fmt layer that writes to syslog
+            let fmt_layer = fmt::Layer::default()
+                .with_ansi(false)
+                .with_target(false)
+                .with_level(false)
+                .without_time()
+                .with_writer(SyslogMakeWriter::new());
+
+            tracing_subscriber::registry()
+                .with(filter_layer)
+                .with(fmt_layer)
+                .init();
+
+            (filter_handle, LoggingGuard::noop())
+        }
+
+        LogDestination::File { path, rotate_daily } => {
+            let (non_blocking, guard) = if *rotate_daily {
+                // Extract directory and filename prefix
+                let path = Path::new(path);
+                let dir = path.parent().unwrap_or(Path::new("/var/log"));
+                let prefix = path
+                    .file_name()
+                    .and_then(|s| s.to_str())
+                    .unwrap_or("telemt");
+
+                let file_appender = tracing_appender::rolling::daily(dir, prefix);
+                tracing_appender::non_blocking(file_appender)
+            } else {
+                let file = std::fs::OpenOptions::new()
+                    .create(true)
+                    .append(true)
+                    .open(path)
+                    .expect("Failed to open log file");
+                tracing_appender::non_blocking(file)
+            };
+
+            let fmt_layer = fmt::Layer::default()
+                .with_ansi(false)
+                .with_target(true)
+                .with_writer(non_blocking);
+
+            tracing_subscriber::registry()
+                .with(filter_layer)
+                .with(fmt_layer)
+                .init();
+
+            (filter_handle, LoggingGuard::new(Some(guard)))
+        }
+    }
+}
+
+/// Syslog writer for tracing.
+#[cfg(unix)]
+#[derive(Clone, Copy)]
+struct SyslogMakeWriter;
+
+#[cfg(unix)]
+#[derive(Clone, Copy)]
+struct SyslogWriter {
+    priority: libc::c_int,
+}
+
+#[cfg(unix)]
+impl SyslogMakeWriter {
+    fn new() -> Self {
+        // Open syslog connection on first use
+        static INIT: std::sync::Once = std::sync::Once::new();
+        INIT.call_once(|| {
+            unsafe {
+                // Open syslog with ident "telemt", LOG_PID, LOG_DAEMON facility
+                let ident = b"telemt\0".as_ptr() as *const libc::c_char;
+                libc::openlog(ident, libc::LOG_PID | libc::LOG_NDELAY, libc::LOG_DAEMON);
+            }
+        });
+        Self
+    }
+}
+
+#[cfg(unix)]
+fn syslog_priority_for_level(level: &tracing::Level) -> libc::c_int {
+    match *level {
+        tracing::Level::ERROR => libc::LOG_ERR,
+        tracing::Level::WARN => libc::LOG_WARNING,
+        tracing::Level::INFO => libc::LOG_INFO,
+        tracing::Level::DEBUG => libc::LOG_DEBUG,
+        tracing::Level::TRACE => libc::LOG_DEBUG,
+    }
+}
+
+#[cfg(unix)]
+impl std::io::Write for SyslogWriter {
+    fn write(&mut self, buf: &[u8]) -> std::io::Result<usize> {
+        // Convert to C string, stripping newlines
+        let msg = String::from_utf8_lossy(buf);
+        let msg = msg.trim_end();
+
+        if msg.is_empty() {
+            return Ok(buf.len());
+        }
+
+        // Write to syslog
+        let c_msg = std::ffi::CString::new(msg.as_bytes())
+            .unwrap_or_else(|_| std::ffi::CString::new("(invalid utf8)").unwrap());
+
+        unsafe {
+            libc::syslog(
+                self.priority,
+                b"%s\0".as_ptr() as *const libc::c_char,
+                c_msg.as_ptr(),
+            );
+        }
+
+        Ok(buf.len())
+    }
+
+    fn flush(&mut self) -> std::io::Result<()> {
+        Ok(())
+    }
+}
+
+#[cfg(unix)]
+impl<'a> tracing_subscriber::fmt::MakeWriter<'a> for SyslogMakeWriter {
+    type Writer = SyslogWriter;
+
+    fn make_writer(&'a self) -> Self::Writer {
+        SyslogWriter {
+            priority: libc::LOG_INFO,
+        }
+    }
+
+    fn make_writer_for(&'a self, meta: &tracing::Metadata<'_>) -> Self::Writer {
+        SyslogWriter {
+            priority: syslog_priority_for_level(meta.level()),
+        }
+    }
+}
+
+/// Parse log destination from CLI arguments.
+pub fn parse_log_destination(args: &[String]) -> LogDestination {
+    let mut i = 0;
+    while i < args.len() {
+        match args[i].as_str() {
+            #[cfg(unix)]
+            "--syslog" => {
+                return LogDestination::Syslog;
+            }
+            "--log-file" => {
+                i += 1;
+                if i < args.len() {
+                    return LogDestination::File {
+                        path: args[i].clone(),
+                        rotate_daily: false,
+                    };
+                }
+            }
+            s if s.starts_with("--log-file=") => {
+                return LogDestination::File {
+                    path: s.trim_start_matches("--log-file=").to_string(),
+                    rotate_daily: false,
+                };
+            }
+            "--log-file-daily" => {
+                i += 1;
+                if i < args.len() {
+                    return LogDestination::File {
+                        path: args[i].clone(),
+                        rotate_daily: true,
+                    };
+                }
+            }
+            s if s.starts_with("--log-file-daily=") => {
+                return LogDestination::File {
+                    path: s.trim_start_matches("--log-file-daily=").to_string(),
+                    rotate_daily: true,
+                };
+            }
+            _ => {}
+        }
+        i += 1;
+    }
+    LogDestination::Stderr
+}
+
+#[cfg(test)]
+mod tests {
+    use super::*;
+
+    #[test]
+    fn test_parse_log_destination_default() {
+        let args: Vec<String> = vec![];
+        assert!(matches!(
+            parse_log_destination(&args),
+            LogDestination::Stderr
+        ));
+    }
+
+    #[test]
+    fn test_parse_log_destination_file() {
+        let args = vec!["--log-file".to_string(), "/var/log/telemt.log".to_string()];
+        match parse_log_destination(&args) {
+            LogDestination::File { path, rotate_daily } => {
+                assert_eq!(path, "/var/log/telemt.log");
+                assert!(!rotate_daily);
+            }
+            _ => panic!("Expected File destination"),
+        }
+    }
+
+    #[test]
+    fn test_parse_log_destination_file_daily() {
+        let args = vec!["--log-file-daily=/var/log/telemt".to_string()];
+        match parse_log_destination(&args) {
+            LogDestination::File { path, rotate_daily } => {
+                assert_eq!(path, "/var/log/telemt");
+                assert!(rotate_daily);
+            }
+            _ => panic!("Expected File destination"),
+        }
+    }
+
+    #[cfg(unix)]
+    #[test]
+    fn test_parse_log_destination_syslog() {
+        let args = vec!["--syslog".to_string()];
+        assert!(matches!(
+            parse_log_destination(&args),
+            LogDestination::Syslog
+        ));
+    }
+
+    #[cfg(unix)]
+    #[test]
+    fn test_syslog_priority_for_level_mapping() {
+        assert_eq!(
+            syslog_priority_for_level(&tracing::Level::ERROR),
+            libc::LOG_ERR
+        );
+        assert_eq!(
+            syslog_priority_for_level(&tracing::Level::WARN),
+            libc::LOG_WARNING
+        );
+        assert_eq!(
+            syslog_priority_for_level(&tracing::Level::INFO),
+            libc::LOG_INFO
+        );
+        assert_eq!(
+            syslog_priority_for_level(&tracing::Level::DEBUG),
+            libc::LOG_DEBUG
+        );
+        assert_eq!(
+            syslog_priority_for_level(&tracing::Level::TRACE),
+            libc::LOG_DEBUG
+        );
+    }
+}

src/maestro/admission.rs

@@ -21,10 +21,29 @@ pub(crate) async fn configure_admission_gate(
     if config.general.use_middle_proxy {
         if let Some(pool) = me_pool.as_ref() {
             let initial_ready = pool.admission_ready_conditional_cast().await;
-            admission_tx.send_replace(initial_ready);
-            let _ = route_runtime.set_mode(RelayRouteMode::Middle);
+            let mut fallback_enabled = config.general.me2dc_fallback;
+            let mut fast_fallback_enabled = fallback_enabled && config.general.me2dc_fast;
+            let (initial_gate_open, initial_route_mode, initial_fallback_reason) = if initial_ready
+            {
+                (true, RelayRouteMode::Middle, None)
+            } else if fast_fallback_enabled {
+                (
+                    true,
+                    RelayRouteMode::Direct,
+                    Some("fast_not_ready_fallback"),
+                )
+            } else {
+                (false, RelayRouteMode::Middle, None)
+            };
+            admission_tx.send_replace(initial_gate_open);
+            let _ = route_runtime.set_mode(initial_route_mode);
             if initial_ready {
                 info!("Conditional-admission gate: open / ME pool READY");
+            } else if let Some(reason) = initial_fallback_reason {
+                warn!(
+                    fallback_reason = reason,
+                    "Conditional-admission gate opened in ME fast fallback mode"
+                );
             } else {
                 warn!("Conditional-admission gate: closed / ME pool is NOT ready)");
             }
@@ -34,10 +53,9 @@ pub(crate) async fn configure_admission_gate(
             let route_runtime_gate = route_runtime.clone();
             let mut config_rx_gate = config_rx.clone();
             let mut admission_poll_ms = config.general.me_admission_poll_ms.max(1);
-            let mut fallback_enabled = config.general.me2dc_fallback;
             tokio::spawn(async move {
-                let mut gate_open = initial_ready;
-                let mut route_mode = RelayRouteMode::Middle;
+                let mut gate_open = initial_gate_open;
+                let mut route_mode = initial_route_mode;
                 let mut ready_observed = initial_ready;
                 let mut not_ready_since = if initial_ready {
                     None
@@ -53,16 +71,23 @@ pub(crate) async fn configure_admission_gate(
                             let cfg = config_rx_gate.borrow_and_update().clone();
                             admission_poll_ms = cfg.general.me_admission_poll_ms.max(1);
                             fallback_enabled = cfg.general.me2dc_fallback;
+                            fast_fallback_enabled = cfg.general.me2dc_fallback && cfg.general.me2dc_fast;
                             continue;
                         }
                         _ = tokio::time::sleep(Duration::from_millis(admission_poll_ms)) => {}
                     }
                     let ready = pool_for_gate.admission_ready_conditional_cast().await;
                     let now = Instant::now();
-                    let (next_gate_open, next_route_mode, next_fallback_active) = if ready {
+                    let (next_gate_open, next_route_mode, next_fallback_reason) = if ready {
                         ready_observed = true;
                         not_ready_since = None;
-                        (true, RelayRouteMode::Middle, false)
+                        (true, RelayRouteMode::Middle, None)
+                    } else if fast_fallback_enabled {
+                        (
+                            true,
+                            RelayRouteMode::Direct,
+                            Some("fast_not_ready_fallback"),
+                        )
                     } else {
                         let not_ready_started_at = *not_ready_since.get_or_insert(now);
                         let not_ready_for = now.saturating_duration_since(not_ready_started_at);
@@ -72,11 +97,12 @@ pub(crate) async fn configure_admission_gate(
                             STARTUP_FALLBACK_AFTER
                         };
                         if fallback_enabled && not_ready_for > fallback_after {
-                            (true, RelayRouteMode::Direct, true)
+                            (true, RelayRouteMode::Direct, Some("strict_grace_fallback"))
                         } else {
-                            (false, RelayRouteMode::Middle, false)
+                            (false, RelayRouteMode::Middle, None)
                         }
                     };
+                    let next_fallback_active = next_fallback_reason.is_some();
 
                     if next_route_mode != route_mode {
                         route_mode = next_route_mode;
@@ -88,17 +114,28 @@ pub(crate) async fn configure_admission_gate(
                                     "Middle-End routing restored for new sessions"
                                 );
                             } else {
-                                let fallback_after = if ready_observed {
-                                    RUNTIME_FALLBACK_AFTER
+                                let fallback_reason = next_fallback_reason.unwrap_or("unknown");
+                                if fallback_reason == "strict_grace_fallback" {
+                                    let fallback_after = if ready_observed {
+                                        RUNTIME_FALLBACK_AFTER
+                                    } else {
+                                        STARTUP_FALLBACK_AFTER
+                                    };
+                                    warn!(
+                                        target_mode = route_mode.as_str(),
+                                        cutover_generation = snapshot.generation,
+                                        grace_secs = fallback_after.as_secs(),
+                                        fallback_reason,
+                                        "ME pool stayed not-ready beyond grace; routing new sessions via Direct-DC"
+                                    );
                                 } else {
-                                    STARTUP_FALLBACK_AFTER
-                                };
-                                warn!(
-                                    target_mode = route_mode.as_str(),
-                                    cutover_generation = snapshot.generation,
-                                    grace_secs = fallback_after.as_secs(),
-                                    "ME pool stayed not-ready beyond grace; routing new sessions via Direct-DC"
-                                );
+                                    warn!(
+                                        target_mode = route_mode.as_str(),
+                                        cutover_generation = snapshot.generation,
+                                        fallback_reason,
+                                        "ME pool not-ready; routing new sessions via Direct-DC (fast mode)"
+                                    );
+                                }
                             }
                         }
                     }
@@ -108,7 +145,10 @@ pub(crate) async fn configure_admission_gate(
                         admission_tx_gate.send_replace(gate_open);
                         if gate_open {
                             if next_fallback_active {
-                                warn!("Conditional-admission gate opened in ME fallback mode");
+                                warn!(
+                                    fallback_reason = next_fallback_reason.unwrap_or("unknown"),
+                                    "Conditional-admission gate opened in ME fallback mode"
+                                );
                             } else {
                                 info!("Conditional-admission gate opened / ME pool READY");
                             }

src/maestro/helpers.rs

@@ -8,31 +8,66 @@ use tracing::{debug, error, info, warn};
 
 use crate::cli;
 use crate::config::ProxyConfig;
+use crate::logging::LogDestination;
+use crate::transport::UpstreamManager;
 use crate::transport::middle_proxy::{
-    ProxyConfigData, fetch_proxy_config_with_raw, load_proxy_config_cache, save_proxy_config_cache,
+    ProxyConfigData, fetch_proxy_config_with_raw_via_upstream, load_proxy_config_cache,
+    save_proxy_config_cache,
 };
 
 pub(crate) fn resolve_runtime_config_path(
     config_path_cli: &str,
     startup_cwd: &std::path::Path,
+    config_path_explicit: bool,
 ) -> PathBuf {
-    let raw = PathBuf::from(config_path_cli);
-    let absolute = if raw.is_absolute() {
-        raw
-    } else {
-        startup_cwd.join(raw)
-    };
-    absolute.canonicalize().unwrap_or(absolute)
+    if config_path_explicit {
+        let raw = PathBuf::from(config_path_cli);
+        let absolute = if raw.is_absolute() {
+            raw
+        } else {
+            startup_cwd.join(raw)
+        };
+        return absolute.canonicalize().unwrap_or(absolute);
+    }
+
+    let etc_telemt = std::path::Path::new("/etc/telemt");
+    let candidates = [
+        startup_cwd.join("config.toml"),
+        startup_cwd.join("telemt.toml"),
+        etc_telemt.join("telemt.toml"),
+        etc_telemt.join("config.toml"),
+    ];
+    for candidate in candidates {
+        if candidate.is_file() {
+            return candidate.canonicalize().unwrap_or(candidate);
+        }
+    }
+
+    startup_cwd.join("config.toml")
 }
 
-pub(crate) fn parse_cli() -> (String, Option<PathBuf>, bool, Option<String>) {
+/// Parsed CLI arguments.
+pub(crate) struct CliArgs {
+    pub config_path: String,
+    pub config_path_explicit: bool,
+    pub data_path: Option<PathBuf>,
+    pub silent: bool,
+    pub log_level: Option<String>,
+    pub log_destination: LogDestination,
+}
+
+pub(crate) fn parse_cli() -> CliArgs {
     let mut config_path = "config.toml".to_string();
+    let mut config_path_explicit = false;
     let mut data_path: Option<PathBuf> = None;
     let mut silent = false;
     let mut log_level: Option<String> = None;
 
     let args: Vec<String> = std::env::args().skip(1).collect();
 
+    // Parse log destination
+    let log_destination = crate::logging::parse_log_destination(&args);
+
     // Check for --init first (handled before tokio)
     if let Some(init_opts) = cli::parse_init_args(&args) {
         if let Err(e) = cli::run_init(init_opts) {
@@ -59,6 +94,20 @@ pub(crate) fn parse_cli() -> (String, Option<PathBuf>, bool, Option<String>) {
                     s.trim_start_matches("--data-path=").to_string(),
                 ));
             }
+            "--working-dir" => {
+                i += 1;
+                if i < args.len() {
+                    data_path = Some(PathBuf::from(args[i].clone()));
+                } else {
+                    eprintln!("Missing value for --working-dir");
+                    std::process::exit(0);
+                }
+            }
+            s if s.starts_with("--working-dir=") => {
+                data_path = Some(PathBuf::from(
+                    s.trim_start_matches("--working-dir=").to_string(),
+                ));
+            }
             "--silent" | "-s" => {
                 silent = true;
             }
@@ -72,38 +121,35 @@ pub(crate) fn parse_cli() -> (String, Option<PathBuf>, bool, Option<String>) {
                 log_level = Some(s.trim_start_matches("--log-level=").to_string());
             }
             "--help" | "-h" => {
-                eprintln!("Usage: telemt [config.toml] [OPTIONS]");
-                eprintln!();
-                eprintln!("Options:");
-                eprintln!(
-                    "  --data-path <DIR>       Set data directory (absolute path; overrides config value)"
-                );
-                eprintln!("  --silent, -s            Suppress info logs");
-                eprintln!("  --log-level <LEVEL>     debug|verbose|normal|silent");
-                eprintln!("  --help, -h              Show this help");
-                eprintln!();
-                eprintln!("Setup (fire-and-forget):");
-                eprintln!(
-                    "  --init                  Generate config, install systemd service, start"
-                );
-                eprintln!("    --port <PORT>          Listen port (default: 443)");
-                eprintln!(
-                    "    --domain <DOMAIN>      TLS domain for masking (default: www.google.com)"
-                );
-                eprintln!(
-                    "    --secret <HEX>         32-char hex secret (auto-generated if omitted)"
-                );
-                eprintln!("    --user <NAME>          Username (default: user)");
-                eprintln!("    --config-dir <DIR>     Config directory (default: /etc/telemt)");
-                eprintln!("    --no-start             Don't start the service after install");
+                print_help();
                 std::process::exit(0);
             }
             "--version" | "-V" => {
                 println!("telemt {}", env!("CARGO_PKG_VERSION"));
                 std::process::exit(0);
             }
+            // Skip daemon-related flags (already parsed)
+            "--daemon" | "-d" | "--foreground" | "-f" => {}
+            s if s.starts_with("--pid-file") => {
+                if !s.contains('=') {
+                    i += 1; // skip value
+                }
+            }
+            s if s.starts_with("--run-as-user") => {
+                if !s.contains('=') {
+                    i += 1;
+                }
+            }
+            s if s.starts_with("--run-as-group") => {
+                if !s.contains('=') {
+                    i += 1;
+                }
+            }
             s if !s.starts_with('-') => {
-                config_path = s.to_string();
+                if !matches!(s, "run" | "start" | "stop" | "reload" | "status") {
+                    config_path = s.to_string();
+                    config_path_explicit = true;
+                }
             }
             other => {
                 eprintln!("Unknown option: {}", other);
@@ -112,7 +158,75 @@ pub(crate) fn parse_cli() -> (String, Option<PathBuf>, bool, Option<String>) {
         i += 1;
     }
 
-    (config_path, data_path, silent, log_level)
+    CliArgs {
+        config_path,
+        config_path_explicit,
+        data_path,
+        silent,
+        log_level,
+        log_destination,
+    }
+}
+
+fn print_help() {
+    eprintln!("Usage: telemt [COMMAND] [OPTIONS] [config.toml]");
+    eprintln!();
+    eprintln!("Commands:");
+    eprintln!("  run                     Run in foreground (default if no command given)");
+    #[cfg(unix)]
+    {
+        eprintln!("  start                   Start as background daemon");
+        eprintln!("  stop                    Stop a running daemon");
+        eprintln!("  reload                  Reload configuration (send SIGHUP)");
+        eprintln!("  status                  Check if daemon is running");
+    }
+    eprintln!();
+    eprintln!("Options:");
+    eprintln!(
+        "  --data-path <DIR>       Set data directory (absolute path; overrides config value)"
+    );
+    eprintln!("  --working-dir <DIR>     Alias for --data-path");
+    eprintln!("  --silent, -s            Suppress info logs");
+    eprintln!("  --log-level <LEVEL>     debug|verbose|normal|silent");
+    eprintln!("  --help, -h              Show this help");
+    eprintln!("  --version, -V           Show version");
+    eprintln!();
+    eprintln!("Logging options:");
+    eprintln!("  --log-file <PATH>       Log to file (default: stderr)");
+    eprintln!("  --log-file-daily <PATH> Log to file with daily rotation");
+    #[cfg(unix)]
+    eprintln!("  --syslog                Log to syslog (Unix only)");
+    eprintln!();
+    #[cfg(unix)]
+    {
+        eprintln!("Daemon options (Unix only):");
+        eprintln!("  --daemon, -d            Fork to background (daemonize)");
+        eprintln!("  --foreground, -f        Explicit foreground mode (for systemd)");
+        eprintln!("  --pid-file <PATH>       PID file path (default: /var/run/telemt.pid)");
+        eprintln!("  --run-as-user <USER>    Drop privileges to this user after binding");
+        eprintln!("  --run-as-group <GROUP>  Drop privileges to this group after binding");
+        eprintln!("  --working-dir <DIR>     Working directory for daemon mode");
+        eprintln!();
+    }
+    eprintln!("Setup (fire-and-forget):");
+    eprintln!("  --init                  Generate config, install systemd service, start");
+    eprintln!("    --port <PORT>          Listen port (default: 443)");
+    eprintln!("    --domain <DOMAIN>      TLS domain for masking (default: www.google.com)");
+    eprintln!("    --secret <HEX>         32-char hex secret (auto-generated if omitted)");
+    eprintln!("    --user <NAME>          Username (default: user)");
+    eprintln!("    --config-dir <DIR>     Config directory (default: /etc/telemt)");
+    eprintln!("    --no-start             Don't start the service after install");
+    #[cfg(unix)]
+    {
+        eprintln!();
+        eprintln!("Examples:");
+        eprintln!("  telemt config.toml                    Run in foreground");
+        eprintln!("  telemt start config.toml              Start as daemon");
+        eprintln!("  telemt start --pid-file /tmp/t.pid    Start with custom PID file");
+        eprintln!("  telemt stop                           Stop daemon");
+        eprintln!("  telemt reload                         Reload configuration");
+        eprintln!("  telemt status                         Check daemon status");
+    }
 }
 
 #[cfg(test)]
@@ -130,7 +244,7 @@ mod tests {
         let target = startup_cwd.join("config.toml");
         std::fs::write(&target, " ").unwrap();
 
-        let resolved = resolve_runtime_config_path("config.toml", &startup_cwd);
+        let resolved = resolve_runtime_config_path("config.toml", &startup_cwd, true);
         assert_eq!(resolved, target.canonicalize().unwrap());
 
         let _ = std::fs::remove_file(&target);
@@ -146,11 +260,45 @@ mod tests {
         let startup_cwd = std::env::temp_dir().join(format!("telemt_cfg_path_missing_{nonce}"));
         std::fs::create_dir_all(&startup_cwd).unwrap();
 
-        let resolved = resolve_runtime_config_path("missing.toml", &startup_cwd);
+        let resolved = resolve_runtime_config_path("missing.toml", &startup_cwd, true);
         assert_eq!(resolved, startup_cwd.join("missing.toml"));
 
         let _ = std::fs::remove_dir(&startup_cwd);
     }
+
+    #[test]
+    fn resolve_runtime_config_path_uses_startup_candidates_when_not_explicit() {
+        let nonce = std::time::SystemTime::now()
+            .duration_since(std::time::UNIX_EPOCH)
+            .unwrap()
+            .as_nanos();
+        let startup_cwd =
+            std::env::temp_dir().join(format!("telemt_cfg_startup_candidates_{nonce}"));
+        std::fs::create_dir_all(&startup_cwd).unwrap();
+        let telemt = startup_cwd.join("telemt.toml");
+        std::fs::write(&telemt, " ").unwrap();
+
+        let resolved = resolve_runtime_config_path("config.toml", &startup_cwd, false);
+        assert_eq!(resolved, telemt.canonicalize().unwrap());
+
+        let _ = std::fs::remove_file(&telemt);
+        let _ = std::fs::remove_dir(&startup_cwd);
+    }
+
+    #[test]
+    fn resolve_runtime_config_path_defaults_to_startup_config_when_none_found() {
+        let nonce = std::time::SystemTime::now()
+            .duration_since(std::time::UNIX_EPOCH)
+            .unwrap()
+            .as_nanos();
+        let startup_cwd = std::env::temp_dir().join(format!("telemt_cfg_startup_default_{nonce}"));
+        std::fs::create_dir_all(&startup_cwd).unwrap();
+
+        let resolved = resolve_runtime_config_path("config.toml", &startup_cwd, false);
+        assert_eq!(resolved, startup_cwd.join("config.toml"));
+
+        let _ = std::fs::remove_dir(&startup_cwd);
+    }
 }
 
 pub(crate) fn print_proxy_links(host: &str, port: u16, config: &ProxyConfig) {
@@ -288,9 +436,10 @@ pub(crate) async fn load_startup_proxy_config_snapshot(
     cache_path: Option<&str>,
     me2dc_fallback: bool,
     label: &'static str,
+    upstream: Option<std::sync::Arc<UpstreamManager>>,
 ) -> Option<ProxyConfigData> {
     loop {
-        match fetch_proxy_config_with_raw(url).await {
+        match fetch_proxy_config_with_raw_via_upstream(url, upstream.clone()).await {
             Ok((cfg, raw)) => {
                 if !cfg.map.is_empty() {
                     if let Some(path) = cache_path

src/maestro/listeners.rs

@@ -14,6 +14,7 @@ use crate::crypto::SecureRandom;
 use crate::ip_tracker::UserIpTracker;
 use crate::proxy::ClientHandler;
 use crate::proxy::route_mode::{ROUTE_SWITCH_ERROR_MSG, RouteRuntimeController};
+use crate::proxy::shared_state::ProxySharedState;
 use crate::startup::{COMPONENT_LISTENERS_BIND, StartupTracker};
 use crate::stats::beobachten::BeobachtenStore;
 use crate::stats::{ReplayChecker, Stats};
@@ -49,6 +50,7 @@ pub(crate) async fn bind_listeners(
     tls_cache: Option<Arc<TlsFrontCache>>,
     ip_tracker: Arc<UserIpTracker>,
     beobachten: Arc<BeobachtenStore>,
+    shared: Arc<ProxySharedState>,
     max_connections: Arc<Semaphore>,
 ) -> Result<BoundListeners, Box<dyn Error>> {
     startup_tracker
@@ -72,6 +74,7 @@ pub(crate) async fn bind_listeners(
         let options = ListenOptions {
             reuse_port: listener_conf.reuse_allow,
             ipv6_only: listener_conf.ip.is_ipv6(),
+            backlog: config.server.listen_backlog,
             ..Default::default()
         };
 
@@ -223,6 +226,7 @@ pub(crate) async fn bind_listeners(
         let tls_cache = tls_cache.clone();
         let ip_tracker = ip_tracker.clone();
         let beobachten = beobachten.clone();
+        let shared = shared.clone();
         let max_connections_unix = max_connections.clone();
 
         tokio::spawn(async move {
@@ -258,6 +262,7 @@ pub(crate) async fn bind_listeners(
                                     break;
                                 }
                                 Err(_) => {
+                                    stats.increment_accept_permit_timeout_total();
                                     debug!(
                                         timeout_ms = accept_permit_timeout_ms,
                                         "Dropping accepted unix connection: permit wait timeout"
@@ -283,11 +288,12 @@ pub(crate) async fn bind_listeners(
                         let tls_cache = tls_cache.clone();
                         let ip_tracker = ip_tracker.clone();
                         let beobachten = beobachten.clone();
+                        let shared = shared.clone();
                         let proxy_protocol_enabled = config.server.proxy_protocol;
 
                         tokio::spawn(async move {
                             let _permit = permit;
-                            if let Err(e) = crate::proxy::client::handle_client_stream(
+                            if let Err(e) = crate::proxy::client::handle_client_stream_with_shared(
                                 stream,
                                 fake_peer,
                                 config,
@@ -301,6 +307,7 @@ pub(crate) async fn bind_listeners(
                                 tls_cache,
                                 ip_tracker,
                                 beobachten,
+                                shared,
                                 proxy_protocol_enabled,
                             )
                             .await
@@ -350,6 +357,7 @@ pub(crate) fn spawn_tcp_accept_loops(
     tls_cache: Option<Arc<TlsFrontCache>>,
     ip_tracker: Arc<UserIpTracker>,
     beobachten: Arc<BeobachtenStore>,
+    shared: Arc<ProxySharedState>,
     max_connections: Arc<Semaphore>,
 ) {
     for (listener, listener_proxy_protocol) in listeners {
@@ -365,6 +373,7 @@ pub(crate) fn spawn_tcp_accept_loops(
         let tls_cache = tls_cache.clone();
         let ip_tracker = ip_tracker.clone();
         let beobachten = beobachten.clone();
+        let shared = shared.clone();
         let max_connections_tcp = max_connections.clone();
 
         tokio::spawn(async move {
@@ -399,6 +408,7 @@ pub(crate) fn spawn_tcp_accept_loops(
                                     break;
                                 }
                                 Err(_) => {
+                                    stats.increment_accept_permit_timeout_total();
                                     debug!(
                                         peer = %peer_addr,
                                         timeout_ms = accept_permit_timeout_ms,
@@ -420,13 +430,14 @@ pub(crate) fn spawn_tcp_accept_loops(
                         let tls_cache = tls_cache.clone();
                         let ip_tracker = ip_tracker.clone();
                         let beobachten = beobachten.clone();
+                        let shared = shared.clone();
                         let proxy_protocol_enabled = listener_proxy_protocol;
                         let real_peer_report = Arc::new(std::sync::Mutex::new(None));
                         let real_peer_report_for_handler = real_peer_report.clone();
 
                         tokio::spawn(async move {
                             let _permit = permit;
-                            if let Err(e) = ClientHandler::new(
+                            if let Err(e) = ClientHandler::new_with_shared(
                                 stream,
                                 peer_addr,
                                 config,
@@ -440,6 +451,7 @@ pub(crate) fn spawn_tcp_accept_loops(
                                 tls_cache,
                                 ip_tracker,
                                 beobachten,
+                                shared,
                                 proxy_protocol_enabled,
                                 real_peer_report_for_handler,
                             )

src/maestro/me_startup.rs

@@ -63,9 +63,10 @@ pub(crate) async fn initialize_me_pool(
     let proxy_secret_path = config.general.proxy_secret_path.as_deref();
     let pool_size = config.general.middle_proxy_pool_size.max(1);
     let proxy_secret = loop {
-        match crate::transport::middle_proxy::fetch_proxy_secret(
+        match crate::transport::middle_proxy::fetch_proxy_secret_with_upstream(
             proxy_secret_path,
             config.general.proxy_secret_len_max,
+            Some(upstream_manager.clone()),
         )
         .await
         {
@@ -129,6 +130,7 @@ pub(crate) async fn initialize_me_pool(
                 config.general.proxy_config_v4_cache_path.as_deref(),
                 me2dc_fallback,
                 "getProxyConfig",
+                Some(upstream_manager.clone()),
             )
             .await;
             if cfg_v4.is_some() {
@@ -160,6 +162,7 @@ pub(crate) async fn initialize_me_pool(
                 config.general.proxy_config_v6_cache_path.as_deref(),
                 me2dc_fallback,
                 "getProxyConfigV6",
+                Some(upstream_manager.clone()),
             )
             .await;
             if cfg_v6.is_some() {
@@ -274,6 +277,8 @@ pub(crate) async fn initialize_me_pool(
                     config.general.me_warn_rate_limit_ms,
                     config.general.me_route_no_writer_mode,
                     config.general.me_route_no_writer_wait_ms,
+                    config.general.me_route_hybrid_max_wait_ms,
+                    config.general.me_route_blocking_send_timeout_ms,
                     config.general.me_route_inline_recovery_attempts,
                     config.general.me_route_inline_recovery_wait_ms,
                 );

src/maestro/mod.rs

@@ -29,10 +29,12 @@ use tracing_subscriber::{EnvFilter, fmt, prelude::*, reload};
 
 use crate::api;
 use crate::config::{LogLevel, ProxyConfig};
+use crate::conntrack_control;
 use crate::crypto::SecureRandom;
 use crate::ip_tracker::UserIpTracker;
 use crate::network::probe::{decide_network_capabilities, log_probe_result, run_probe};
 use crate::proxy::route_mode::{RelayRouteMode, RouteRuntimeController};
+use crate::proxy::shared_state::ProxySharedState;
 use crate::startup::{
     COMPONENT_API_BOOTSTRAP, COMPONENT_CONFIG_LOAD, COMPONENT_ME_POOL_CONSTRUCT,
     COMPONENT_ME_POOL_INIT_STAGE1, COMPONENT_ME_PROXY_CONFIG_V4, COMPONENT_ME_PROXY_CONFIG_V6,
@@ -47,8 +49,55 @@ use crate::transport::UpstreamManager;
 use crate::transport::middle_proxy::MePool;
 use helpers::{parse_cli, resolve_runtime_config_path};
 
+#[cfg(unix)]
+use crate::daemon::{DaemonOptions, PidFile, drop_privileges};
+
 /// Runs the full telemt runtime startup pipeline and blocks until shutdown.
+///
+/// On Unix, daemon options should be handled before calling this function
+/// (daemonization must happen before tokio runtime starts).
+#[cfg(unix)]
+pub async fn run_with_daemon(
+    daemon_opts: DaemonOptions,
+) -> std::result::Result<(), Box<dyn std::error::Error>> {
+    run_inner(daemon_opts).await
+}
+
+/// Runs the full telemt runtime startup pipeline and blocks until shutdown.
+///
+/// This is the main entry point for non-daemon mode or when called as a library.
+#[allow(dead_code)]
 pub async fn run() -> std::result::Result<(), Box<dyn std::error::Error>> {
+    #[cfg(unix)]
+    {
+        // Parse CLI to get daemon options even in simple run() path
+        let args: Vec<String> = std::env::args().skip(1).collect();
+        let daemon_opts = crate::cli::parse_daemon_args(&args);
+        run_inner(daemon_opts).await
+    }
+    #[cfg(not(unix))]
+    {
+        run_inner().await
+    }
+}
+
+#[cfg(unix)]
+async fn run_inner(
+    daemon_opts: DaemonOptions,
+) -> std::result::Result<(), Box<dyn std::error::Error>> {
+    // Acquire PID file if daemonizing or if explicitly requested
+    // Keep it alive until shutdown (underscore prefix = intentionally kept for RAII cleanup)
+    let _pid_file = if daemon_opts.daemonize || daemon_opts.pid_file.is_some() {
+        let mut pf = PidFile::new(daemon_opts.pid_file_path());
+        if let Err(e) = pf.acquire() {
+            eprintln!("[telemt] {}", e);
+            std::process::exit(1);
+        }
+        Some(pf)
+    } else {
+        None
+    };
+
     let process_started_at = Instant::now();
     let process_started_at_epoch_secs = SystemTime::now()
         .duration_since(UNIX_EPOCH)
@@ -61,7 +110,13 @@ pub async fn run() -> std::result::Result<(), Box<dyn std::error::Error>> {
             Some("load and validate config".to_string()),
         )
         .await;
-    let (config_path_cli, data_path, cli_silent, cli_log_level) = parse_cli();
+    let cli_args = parse_cli();
+    let config_path_cli = cli_args.config_path;
+    let config_path_explicit = cli_args.config_path_explicit;
+    let data_path = cli_args.data_path;
+    let cli_silent = cli_args.silent;
+    let cli_log_level = cli_args.log_level;
+    let log_destination = cli_args.log_destination;
     let startup_cwd = match std::env::current_dir() {
         Ok(cwd) => cwd,
         Err(e) => {
@@ -69,7 +124,8 @@ pub async fn run() -> std::result::Result<(), Box<dyn std::error::Error>> {
             std::process::exit(1);
         }
     };
-    let config_path = resolve_runtime_config_path(&config_path_cli, &startup_cwd);
+    let mut config_path =
+        resolve_runtime_config_path(&config_path_cli, &startup_cwd, config_path_explicit);
 
     let mut config = match ProxyConfig::load(&config_path) {
         Ok(c) => c,
@@ -79,11 +135,99 @@ pub async fn run() -> std::result::Result<(), Box<dyn std::error::Error>> {
                 std::process::exit(1);
             } else {
                 let default = ProxyConfig::default();
-                std::fs::write(&config_path, toml::to_string_pretty(&default).unwrap()).unwrap();
-                eprintln!(
-                    "[telemt] Created default config at {}",
-                    config_path.display()
-                );
+
+                let serialized =
+                    match toml::to_string_pretty(&default).or_else(|_| toml::to_string(&default)) {
+                        Ok(value) => Some(value),
+                        Err(serialize_error) => {
+                            eprintln!(
+                                "[telemt] Warning: failed to serialize default config: {}",
+                                serialize_error
+                            );
+                            None
+                        }
+                    };
+
+                if config_path_explicit {
+                    if let Some(serialized) = serialized.as_ref() {
+                        if let Err(write_error) = std::fs::write(&config_path, serialized) {
+                            eprintln!(
+                                "[telemt] Error: failed to create explicit config at {}: {}",
+                                config_path.display(),
+                                write_error
+                            );
+                            std::process::exit(1);
+                        }
+                        eprintln!(
+                            "[telemt] Created default config at {}",
+                            config_path.display()
+                        );
+                    } else {
+                        eprintln!(
+                            "[telemt] Warning: running with in-memory default config without writing to disk"
+                        );
+                    }
+                } else {
+                    let system_dir = std::path::Path::new("/etc/telemt");
+                    let system_config_path = system_dir.join("telemt.toml");
+                    let startup_config_path = startup_cwd.join("config.toml");
+                    let mut persisted = false;
+
+                    if let Some(serialized) = serialized.as_ref() {
+                        match std::fs::create_dir_all(system_dir) {
+                            Ok(()) => match std::fs::write(&system_config_path, serialized) {
+                                Ok(()) => {
+                                    config_path = system_config_path;
+                                    eprintln!(
+                                        "[telemt] Created default config at {}",
+                                        config_path.display()
+                                    );
+                                    persisted = true;
+                                }
+                                Err(write_error) => {
+                                    eprintln!(
+                                        "[telemt] Warning: failed to write default config at {}: {}",
+                                        system_config_path.display(),
+                                        write_error
+                                    );
+                                }
+                            },
+                            Err(create_error) => {
+                                eprintln!(
+                                    "[telemt] Warning: failed to create {}: {}",
+                                    system_dir.display(),
+                                    create_error
+                                );
+                            }
+                        }
+
+                        if !persisted {
+                            match std::fs::write(&startup_config_path, serialized) {
+                                Ok(()) => {
+                                    config_path = startup_config_path;
+                                    eprintln!(
+                                        "[telemt] Created default config at {}",
+                                        config_path.display()
+                                    );
+                                    persisted = true;
+                                }
+                                Err(write_error) => {
+                                    eprintln!(
+                                        "[telemt] Warning: failed to write default config at {}: {}",
+                                        startup_config_path.display(),
+                                        write_error
+                                    );
+                                }
+                            }
+                        }
+                    }
+
+                    if !persisted {
+                        eprintln!(
+                            "[telemt] Warning: running with in-memory default config without writing to disk"
+                        );
+                    }
+                }
                 default
             }
         }
@@ -115,15 +259,13 @@ pub async fn run() -> std::result::Result<(), Box<dyn std::error::Error>> {
                 );
                 std::process::exit(1);
             }
-        } else {
-            if let Err(e) = std::fs::create_dir_all(data_path) {
-                eprintln!(
-                    "[telemt] Can't create data_path {}: {}",
-                    data_path.display(),
-                    e
-                );
-                std::process::exit(1);
-            }
+        } else if let Err(e) = std::fs::create_dir_all(data_path) {
+            eprintln!(
+                "[telemt] Can't create data_path {}: {}",
+                data_path.display(),
+                e
+            );
+            std::process::exit(1);
         }
 
         if let Err(e) = std::env::set_current_dir(data_path) {
@@ -161,17 +303,43 @@ pub async fn run() -> std::result::Result<(), Box<dyn std::error::Error>> {
         )
         .await;
 
-    // Configure color output based on config
-    let fmt_layer = if config.general.disable_colors {
-        fmt::Layer::default().with_ansi(false)
-    } else {
-        fmt::Layer::default().with_ansi(true)
-    };
+    // Initialize logging based on destination
+    let _logging_guard: Option<crate::logging::LoggingGuard>;
+    match log_destination {
+        crate::logging::LogDestination::Stderr => {
+            // Default: log to stderr (works with systemd journald)
+            let fmt_layer = if config.general.disable_colors {
+                fmt::Layer::default().with_ansi(false)
+            } else {
+                fmt::Layer::default().with_ansi(true)
+            };
+            tracing_subscriber::registry()
+                .with(filter_layer)
+                .with(fmt_layer)
+                .init();
+            _logging_guard = None;
+        }
+        #[cfg(unix)]
+        crate::logging::LogDestination::Syslog => {
+            // Syslog: for OpenRC/FreeBSD
+            let logging_opts = crate::logging::LoggingOptions {
+                destination: log_destination,
+                disable_colors: true,
+            };
+            let (_, guard) = crate::logging::init_logging(&logging_opts, "info");
+            _logging_guard = Some(guard);
+        }
+        crate::logging::LogDestination::File { .. } => {
+            // File logging with optional rotation
+            let logging_opts = crate::logging::LoggingOptions {
+                destination: log_destination,
+                disable_colors: true,
+            };
+            let (_, guard) = crate::logging::init_logging(&logging_opts, "info");
+            _logging_guard = Some(guard);
+        }
+    }
 
-    tracing_subscriber::registry()
-        .with(filter_layer)
-        .with(fmt_layer)
-        .init();
     startup_tracker
         .complete_component(
             COMPONENT_TRACING_INIT,
@@ -225,6 +393,7 @@ pub async fn run() -> std::result::Result<(), Box<dyn std::error::Error>> {
         config.general.upstream_connect_retry_attempts,
         config.general.upstream_connect_retry_backoff_ms,
         config.general.upstream_connect_budget_ms,
+        config.general.tg_connect,
         config.general.upstream_unhealthy_fail_threshold,
         config.general.upstream_connect_failfast_hard_errors,
         stats.clone(),
@@ -554,6 +723,12 @@ pub async fn run() -> std::result::Result<(), Box<dyn std::error::Error>> {
     )
     .await;
     let _admission_tx_hold = admission_tx;
+    let shared_state = ProxySharedState::new();
+    conntrack_control::spawn_conntrack_controller(
+        config_rx.clone(),
+        stats.clone(),
+        shared_state.clone(),
+    );
 
     let bound = listeners::bind_listeners(
         &config,
@@ -574,6 +749,7 @@ pub async fn run() -> std::result::Result<(), Box<dyn std::error::Error>> {
         tls_cache.clone(),
         ip_tracker.clone(),
         beobachten.clone(),
+        shared_state.clone(),
         max_connections.clone(),
     )
     .await?;
@@ -585,6 +761,14 @@ pub async fn run() -> std::result::Result<(), Box<dyn std::error::Error>> {
         std::process::exit(1);
     }
 
+    // Drop privileges after binding sockets (which may require root for port < 1024)
+    if daemon_opts.user.is_some() || daemon_opts.group.is_some() {
+        if let Err(e) = drop_privileges(daemon_opts.user.as_deref(), daemon_opts.group.as_deref()) {
+            error!(error = %e, "Failed to drop privileges");
+            std::process::exit(1);
+        }
+    }
+
     runtime_tasks::apply_runtime_log_filter(
         has_rust_log,
         &effective_log_level,
@@ -605,6 +789,9 @@ pub async fn run() -> std::result::Result<(), Box<dyn std::error::Error>> {
 
     runtime_tasks::mark_runtime_ready(&startup_tracker).await;
 
+    // Spawn signal handlers for SIGUSR1/SIGUSR2 (non-shutdown signals)
+    shutdown::spawn_signal_handlers(stats.clone(), process_started_at);
+
     listeners::spawn_tcp_accept_loops(
         listeners,
         config_rx.clone(),
@@ -619,10 +806,11 @@ pub async fn run() -> std::result::Result<(), Box<dyn std::error::Error>> {
         tls_cache.clone(),
         ip_tracker.clone(),
         beobachten.clone(),
+        shared_state,
         max_connections.clone(),
     );
 
-    shutdown::wait_for_shutdown(process_started_at, me_pool).await;
+    shutdown::wait_for_shutdown(process_started_at, me_pool, stats).await;
 
     Ok(())
 }

src/maestro/runtime_tasks.rs

@@ -323,10 +323,12 @@ pub(crate) async fn spawn_metrics_if_configured(
         let config_rx_metrics = config_rx.clone();
         let ip_tracker_metrics = ip_tracker.clone();
         let whitelist = config.server.metrics_whitelist.clone();
+        let listen_backlog = config.server.listen_backlog;
         tokio::spawn(async move {
             metrics::serve(
                 port,
                 listen,
+                listen_backlog,
                 stats,
                 beobachten,
                 ip_tracker_metrics,

src/maestro/shutdown.rs

@@ -1,45 +1,206 @@
+//! Shutdown and signal handling for telemt.
+//!
+//! Handles graceful shutdown on various signals:
+//! - SIGINT (Ctrl+C) / SIGTERM: Graceful shutdown
+//! - SIGQUIT: Graceful shutdown with stats dump
+//! - SIGUSR1: Reserved for log rotation (logs acknowledgment)
+//! - SIGUSR2: Dump runtime status to log
+//!
+//! SIGHUP is handled separately in config/hot_reload.rs for config reload.
+
 use std::sync::Arc;
 use std::time::{Duration, Instant};
 
+#[cfg(not(unix))]
 use tokio::signal;
-use tracing::{error, info, warn};
+#[cfg(unix)]
+use tokio::signal::unix::{SignalKind, signal};
+use tracing::{info, warn};
 
+use crate::stats::Stats;
 use crate::transport::middle_proxy::MePool;
 
 use super::helpers::{format_uptime, unit_label};
 
-pub(crate) async fn wait_for_shutdown(process_started_at: Instant, me_pool: Option<Arc<MePool>>) {
-    match signal::ctrl_c().await {
-        Ok(()) => {
-            let shutdown_started_at = Instant::now();
-            info!("Shutting down...");
-            let uptime_secs = process_started_at.elapsed().as_secs();
-            info!("Uptime: {}", format_uptime(uptime_secs));
-            if let Some(pool) = &me_pool {
-                match tokio::time::timeout(
-                    Duration::from_secs(2),
-                    pool.shutdown_send_close_conn_all(),
-                )
-                .await
-                {
-                    Ok(total) => {
-                        info!(
-                            close_conn_sent = total,
-                            "ME shutdown: RPC_CLOSE_CONN broadcast completed"
-                        );
-                    }
-                    Err(_) => {
-                        warn!("ME shutdown: RPC_CLOSE_CONN broadcast timed out");
-                    }
-                }
-            }
-            let shutdown_secs = shutdown_started_at.elapsed().as_secs();
-            info!(
-                "Shutdown completed successfully in {} {}.",
-                shutdown_secs,
-                unit_label(shutdown_secs, "second", "seconds")
-            );
+/// Signal that triggered shutdown.
+#[derive(Debug, Clone, Copy, PartialEq, Eq)]
+pub enum ShutdownSignal {
+    /// SIGINT (Ctrl+C)
+    Interrupt,
+    /// SIGTERM
+    Terminate,
+    /// SIGQUIT (with stats dump)
+    Quit,
+}
+
+impl std::fmt::Display for ShutdownSignal {
+    fn fmt(&self, f: &mut std::fmt::Formatter<'_>) -> std::fmt::Result {
+        match self {
+            ShutdownSignal::Interrupt => write!(f, "SIGINT"),
+            ShutdownSignal::Terminate => write!(f, "SIGTERM"),
+            ShutdownSignal::Quit => write!(f, "SIGQUIT"),
         }
-        Err(e) => error!("Signal error: {}", e),
     }
 }
+
+/// Waits for a shutdown signal and performs graceful shutdown.
+pub(crate) async fn wait_for_shutdown(
+    process_started_at: Instant,
+    me_pool: Option<Arc<MePool>>,
+    stats: Arc<Stats>,
+) {
+    let signal = wait_for_shutdown_signal().await;
+    perform_shutdown(signal, process_started_at, me_pool, &stats).await;
+}
+
+/// Waits for any shutdown signal (SIGINT, SIGTERM, SIGQUIT).
+#[cfg(unix)]
+async fn wait_for_shutdown_signal() -> ShutdownSignal {
+    let mut sigint = signal(SignalKind::interrupt()).expect("Failed to register SIGINT handler");
+    let mut sigterm = signal(SignalKind::terminate()).expect("Failed to register SIGTERM handler");
+    let mut sigquit = signal(SignalKind::quit()).expect("Failed to register SIGQUIT handler");
+
+    tokio::select! {
+        _ = sigint.recv() => ShutdownSignal::Interrupt,
+        _ = sigterm.recv() => ShutdownSignal::Terminate,
+        _ = sigquit.recv() => ShutdownSignal::Quit,
+    }
+}
+
+#[cfg(not(unix))]
+async fn wait_for_shutdown_signal() -> ShutdownSignal {
+    signal::ctrl_c().await.expect("Failed to listen for Ctrl+C");
+    ShutdownSignal::Interrupt
+}
+
+/// Performs graceful shutdown sequence.
+async fn perform_shutdown(
+    signal: ShutdownSignal,
+    process_started_at: Instant,
+    me_pool: Option<Arc<MePool>>,
+    stats: &Stats,
+) {
+    let shutdown_started_at = Instant::now();
+    info!(signal = %signal, "Received shutdown signal");
+
+    // Dump stats if SIGQUIT
+    if signal == ShutdownSignal::Quit {
+        dump_stats(stats, process_started_at);
+    }
+
+    info!("Shutting down...");
+    let uptime_secs = process_started_at.elapsed().as_secs();
+    info!("Uptime: {}", format_uptime(uptime_secs));
+
+    // Graceful ME pool shutdown
+    if let Some(pool) = &me_pool {
+        match tokio::time::timeout(Duration::from_secs(2), pool.shutdown_send_close_conn_all())
+            .await
+        {
+            Ok(total) => {
+                info!(
+                    close_conn_sent = total,
+                    "ME shutdown: RPC_CLOSE_CONN broadcast completed"
+                );
+            }
+            Err(_) => {
+                warn!("ME shutdown: RPC_CLOSE_CONN broadcast timed out");
+            }
+        }
+    }
+
+    let shutdown_secs = shutdown_started_at.elapsed().as_secs();
+    info!(
+        "Shutdown completed successfully in {} {}.",
+        shutdown_secs,
+        unit_label(shutdown_secs, "second", "seconds")
+    );
+}
+
+/// Dumps runtime statistics to the log.
+fn dump_stats(stats: &Stats, process_started_at: Instant) {
+    let uptime_secs = process_started_at.elapsed().as_secs();
+
+    info!("=== Runtime Statistics Dump ===");
+    info!("Uptime: {}", format_uptime(uptime_secs));
+
+    // Connection stats
+    info!(
+        "Connections: total={}, current={} (direct={}, me={}), bad={}",
+        stats.get_connects_all(),
+        stats.get_current_connections_total(),
+        stats.get_current_connections_direct(),
+        stats.get_current_connections_me(),
+        stats.get_connects_bad(),
+    );
+
+    // ME pool stats
+    info!(
+        "ME keepalive: sent={}, pong={}, failed={}, timeout={}",
+        stats.get_me_keepalive_sent(),
+        stats.get_me_keepalive_pong(),
+        stats.get_me_keepalive_failed(),
+        stats.get_me_keepalive_timeout(),
+    );
+
+    // Relay stats
+    info!(
+        "Relay idle: soft_mark={}, hard_close={}, pressure_evict={}",
+        stats.get_relay_idle_soft_mark_total(),
+        stats.get_relay_idle_hard_close_total(),
+        stats.get_relay_pressure_evict_total(),
+    );
+
+    info!("=== End Statistics Dump ===");
+}
+
+/// Spawns a background task to handle operational signals (SIGUSR1, SIGUSR2).
+///
+/// These signals don't trigger shutdown but perform specific actions:
+/// - SIGUSR1: Log rotation acknowledgment (for external log rotation tools)
+/// - SIGUSR2: Dump runtime status to log
+#[cfg(unix)]
+pub(crate) fn spawn_signal_handlers(stats: Arc<Stats>, process_started_at: Instant) {
+    tokio::spawn(async move {
+        let mut sigusr1 =
+            signal(SignalKind::user_defined1()).expect("Failed to register SIGUSR1 handler");
+        let mut sigusr2 =
+            signal(SignalKind::user_defined2()).expect("Failed to register SIGUSR2 handler");
+
+        loop {
+            tokio::select! {
+                _ = sigusr1.recv() => {
+                    handle_sigusr1();
+                }
+                _ = sigusr2.recv() => {
+                    handle_sigusr2(&stats, process_started_at);
+                }
+            }
+        }
+    });
+}
+
+/// No-op on non-Unix platforms.
+#[cfg(not(unix))]
+pub(crate) fn spawn_signal_handlers(_stats: Arc<Stats>, _process_started_at: Instant) {
+    // No SIGUSR1/SIGUSR2 on non-Unix
+}
+
+/// Handles SIGUSR1 - log rotation signal.
+///
+/// This signal is typically sent by logrotate or similar tools after
+/// rotating log files. Since tracing-subscriber doesn't natively support
+/// reopening files, we just acknowledge the signal. If file logging is
+/// added in the future, this would reopen log file handles.
+#[cfg(unix)]
+fn handle_sigusr1() {
+    info!("SIGUSR1 received - log rotation acknowledged");
+    // Future: If using file-based logging, reopen file handles here
+}
+
+/// Handles SIGUSR2 - dump runtime status.
+#[cfg(unix)]
+fn handle_sigusr2(stats: &Stats, process_started_at: Instant) {
+    info!("SIGUSR2 received - dumping runtime status");
+    dump_stats(stats, process_started_at);
+}

src/maestro/tls_bootstrap.rs

@@ -7,6 +7,7 @@ use tracing::warn;
 use crate::config::ProxyConfig;
 use crate::startup::{COMPONENT_TLS_FRONT_BOOTSTRAP, StartupTracker};
 use crate::tls_front::TlsFrontCache;
+use crate::tls_front::fetcher::TlsFetchStrategy;
 use crate::transport::UpstreamManager;
 
 pub(crate) async fn bootstrap_tls_front(
@@ -40,7 +41,17 @@ pub(crate) async fn bootstrap_tls_front(
         let mask_unix_sock = config.censorship.mask_unix_sock.clone();
         let tls_fetch_scope = (!config.censorship.tls_fetch_scope.is_empty())
             .then(|| config.censorship.tls_fetch_scope.clone());
-        let fetch_timeout = Duration::from_secs(5);
+        let tls_fetch = config.censorship.tls_fetch.clone();
+        let fetch_strategy = TlsFetchStrategy {
+            profiles: tls_fetch.profiles,
+            strict_route: tls_fetch.strict_route,
+            attempt_timeout: Duration::from_millis(tls_fetch.attempt_timeout_ms.max(1)),
+            total_budget: Duration::from_millis(tls_fetch.total_budget_ms.max(1)),
+            grease_enabled: tls_fetch.grease_enabled,
+            deterministic: tls_fetch.deterministic,
+            profile_cache_ttl: Duration::from_secs(tls_fetch.profile_cache_ttl_secs),
+        };
+        let fetch_timeout = fetch_strategy.total_budget;
 
         let cache_initial = cache.clone();
         let domains_initial = tls_domains.to_vec();
@@ -48,6 +59,7 @@ pub(crate) async fn bootstrap_tls_front(
         let unix_sock_initial = mask_unix_sock.clone();
         let scope_initial = tls_fetch_scope.clone();
         let upstream_initial = upstream_manager.clone();
+        let strategy_initial = fetch_strategy.clone();
         tokio::spawn(async move {
             let mut join = tokio::task::JoinSet::new();
             for domain in domains_initial {
@@ -56,12 +68,13 @@ pub(crate) async fn bootstrap_tls_front(
                 let unix_sock_domain = unix_sock_initial.clone();
                 let scope_domain = scope_initial.clone();
                 let upstream_domain = upstream_initial.clone();
+                let strategy_domain = strategy_initial.clone();
                 join.spawn(async move {
-                    match crate::tls_front::fetcher::fetch_real_tls(
+                    match crate::tls_front::fetcher::fetch_real_tls_with_strategy(
                         &host_domain,
                         port,
                         &domain,
-                        fetch_timeout,
+                        &strategy_domain,
                         Some(upstream_domain),
                         scope_domain.as_deref(),
                         proxy_protocol,
@@ -107,6 +120,7 @@ pub(crate) async fn bootstrap_tls_front(
         let unix_sock_refresh = mask_unix_sock.clone();
         let scope_refresh = tls_fetch_scope.clone();
         let upstream_refresh = upstream_manager.clone();
+        let strategy_refresh = fetch_strategy.clone();
         tokio::spawn(async move {
             loop {
                 let base_secs = rand::rng().random_range(4 * 3600..=6 * 3600);
@@ -120,12 +134,13 @@ pub(crate) async fn bootstrap_tls_front(
                     let unix_sock_domain = unix_sock_refresh.clone();
                     let scope_domain = scope_refresh.clone();
                     let upstream_domain = upstream_refresh.clone();
+                    let strategy_domain = strategy_refresh.clone();
                     join.spawn(async move {
-                        match crate::tls_front::fetcher::fetch_real_tls(
+                        match crate::tls_front::fetcher::fetch_real_tls_with_strategy(
                             &host_domain,
                             port,
                             &domain,
-                            fetch_timeout,
+                            &strategy_domain,
                             Some(upstream_domain),
                             scope_domain.as_deref(),
                             proxy_protocol,

src/main.rs

@@ -3,23 +3,28 @@
 mod api;
 mod cli;
 mod config;
+mod conntrack_control;
 mod crypto;
+#[cfg(unix)]
+mod daemon;
 mod error;
 mod ip_tracker;
 #[cfg(test)]
-#[path = "tests/ip_tracker_hotpath_adversarial_tests.rs"]
-mod ip_tracker_hotpath_adversarial_tests;
-#[cfg(test)]
 #[path = "tests/ip_tracker_encapsulation_adversarial_tests.rs"]
 mod ip_tracker_encapsulation_adversarial_tests;
 #[cfg(test)]
+#[path = "tests/ip_tracker_hotpath_adversarial_tests.rs"]
+mod ip_tracker_hotpath_adversarial_tests;
+#[cfg(test)]
 #[path = "tests/ip_tracker_regression_tests.rs"]
 mod ip_tracker_regression_tests;
+mod logging;
 mod maestro;
 mod metrics;
 mod network;
 mod protocol;
 mod proxy;
+mod service;
 mod startup;
 mod stats;
 mod stream;
@@ -27,7 +32,49 @@ mod tls_front;
 mod transport;
 mod util;
 
-#[tokio::main]
-async fn main() -> std::result::Result<(), Box<dyn std::error::Error>> {
-    maestro::run().await
+fn main() -> std::result::Result<(), Box<dyn std::error::Error>> {
+    // Install rustls crypto provider early
+    let _ = rustls::crypto::ring::default_provider().install_default();
+
+    let args: Vec<String> = std::env::args().skip(1).collect();
+    let cmd = cli::parse_command(&args);
+
+    // Handle subcommands that don't need the server (stop, reload, status, init)
+    if let Some(exit_code) = cli::execute_subcommand(&cmd) {
+        std::process::exit(exit_code);
+    }
+
+    #[cfg(unix)]
+    {
+        let daemon_opts = cmd.daemon_opts;
+
+        // Daemonize BEFORE runtime
+        if daemon_opts.should_daemonize() {
+            match daemon::daemonize(daemon_opts.working_dir.as_deref()) {
+                Ok(daemon::DaemonizeResult::Parent) => {
+                    std::process::exit(0);
+                }
+                Ok(daemon::DaemonizeResult::Child) => {
+                    // continue
+                }
+                Err(e) => {
+                    eprintln!("[telemt] Daemonization failed: {}", e);
+                    std::process::exit(1);
+                }
+            }
+        }
+
+        tokio::runtime::Builder::new_multi_thread()
+            .enable_all()
+            .build()?
+            .block_on(maestro::run_with_daemon(daemon_opts))
+    }
+
+    #[cfg(not(unix))]
+    {
+        tokio::runtime::Builder::new_multi_thread()
+            .enable_all()
+            .build()?
+            .block_on(maestro::run())
+    }
 }

src/metrics.rs

@@ -22,6 +22,7 @@ use crate::transport::{ListenOptions, create_listener};
 pub async fn serve(
     port: u16,
     listen: Option<String>,
+    listen_backlog: u32,
     stats: Arc<Stats>,
     beobachten: Arc<BeobachtenStore>,
     ip_tracker: Arc<UserIpTracker>,
@@ -40,7 +41,7 @@ pub async fn serve(
             }
         };
         let is_ipv6 = addr.is_ipv6();
-        match bind_metrics_listener(addr, is_ipv6) {
+        match bind_metrics_listener(addr, is_ipv6, listen_backlog) {
             Ok(listener) => {
                 info!("Metrics endpoint: http://{}/metrics and /beobachten", addr);
                 serve_listener(
@@ -60,7 +61,7 @@ pub async fn serve(
     let mut listener_v6 = None;
 
     let addr_v4 = SocketAddr::from(([0, 0, 0, 0], port));
-    match bind_metrics_listener(addr_v4, false) {
+    match bind_metrics_listener(addr_v4, false, listen_backlog) {
         Ok(listener) => {
             info!(
                 "Metrics endpoint: http://{}/metrics and /beobachten",
@@ -74,7 +75,7 @@ pub async fn serve(
     }
 
     let addr_v6 = SocketAddr::from(([0, 0, 0, 0, 0, 0, 0, 0], port));
-    match bind_metrics_listener(addr_v6, true) {
+    match bind_metrics_listener(addr_v6, true, listen_backlog) {
         Ok(listener) => {
             info!(
                 "Metrics endpoint: http://[::]:{}/metrics and /beobachten",
@@ -122,10 +123,15 @@ pub async fn serve(
     }
 }
 
-fn bind_metrics_listener(addr: SocketAddr, ipv6_only: bool) -> std::io::Result<TcpListener> {
+fn bind_metrics_listener(
+    addr: SocketAddr,
+    ipv6_only: bool,
+    listen_backlog: u32,
+) -> std::io::Result<TcpListener> {
     let options = ListenOptions {
         reuse_port: false,
         ipv6_only,
+        backlog: listen_backlog,
         ..Default::default()
     };
     let socket = create_listener(addr, &options)?;
@@ -287,6 +293,27 @@ async fn render_metrics(stats: &Stats, config: &ProxyConfig, ip_tracker: &UserIp
         }
     );
 
+    let _ = writeln!(
+        out,
+        "# HELP telemt_buffer_pool_buffers_total Snapshot of pooled and allocated buffers"
+    );
+    let _ = writeln!(out, "# TYPE telemt_buffer_pool_buffers_total gauge");
+    let _ = writeln!(
+        out,
+        "telemt_buffer_pool_buffers_total{{kind=\"pooled\"}} {}",
+        stats.get_buffer_pool_pooled_gauge()
+    );
+    let _ = writeln!(
+        out,
+        "telemt_buffer_pool_buffers_total{{kind=\"allocated\"}} {}",
+        stats.get_buffer_pool_allocated_gauge()
+    );
+    let _ = writeln!(
+        out,
+        "telemt_buffer_pool_buffers_total{{kind=\"in_use\"}} {}",
+        stats.get_buffer_pool_in_use_gauge()
+    );
+
     let _ = writeln!(
         out,
         "# HELP telemt_connections_total Total accepted connections"
@@ -332,6 +359,134 @@ async fn render_metrics(stats: &Stats, config: &ProxyConfig, ip_tracker: &UserIp
         }
     );
 
+    let _ = writeln!(
+        out,
+        "# HELP telemt_accept_permit_timeout_total Accepted connections dropped due to permit wait timeout"
+    );
+    let _ = writeln!(out, "# TYPE telemt_accept_permit_timeout_total counter");
+    let _ = writeln!(
+        out,
+        "telemt_accept_permit_timeout_total {}",
+        if core_enabled {
+            stats.get_accept_permit_timeout_total()
+        } else {
+            0
+        }
+    );
+
+    let _ = writeln!(
+        out,
+        "# HELP telemt_conntrack_control_state Runtime conntrack control state flags"
+    );
+    let _ = writeln!(out, "# TYPE telemt_conntrack_control_state gauge");
+    let _ = writeln!(
+        out,
+        "telemt_conntrack_control_state{{flag=\"enabled\"}} {}",
+        if stats.get_conntrack_control_enabled() {
+            1
+        } else {
+            0
+        }
+    );
+    let _ = writeln!(
+        out,
+        "telemt_conntrack_control_state{{flag=\"available\"}} {}",
+        if stats.get_conntrack_control_available() {
+            1
+        } else {
+            0
+        }
+    );
+    let _ = writeln!(
+        out,
+        "telemt_conntrack_control_state{{flag=\"pressure_active\"}} {}",
+        if stats.get_conntrack_pressure_active() {
+            1
+        } else {
+            0
+        }
+    );
+    let _ = writeln!(
+        out,
+        "telemt_conntrack_control_state{{flag=\"rule_apply_ok\"}} {}",
+        if stats.get_conntrack_rule_apply_ok() {
+            1
+        } else {
+            0
+        }
+    );
+
+    let _ = writeln!(
+        out,
+        "# HELP telemt_conntrack_event_queue_depth Pending close events in conntrack control queue"
+    );
+    let _ = writeln!(out, "# TYPE telemt_conntrack_event_queue_depth gauge");
+    let _ = writeln!(
+        out,
+        "telemt_conntrack_event_queue_depth {}",
+        stats.get_conntrack_event_queue_depth()
+    );
+
+    let _ = writeln!(
+        out,
+        "# HELP telemt_conntrack_delete_total Conntrack delete attempts by outcome"
+    );
+    let _ = writeln!(out, "# TYPE telemt_conntrack_delete_total counter");
+    let _ = writeln!(
+        out,
+        "telemt_conntrack_delete_total{{result=\"attempt\"}} {}",
+        if core_enabled {
+            stats.get_conntrack_delete_attempt_total()
+        } else {
+            0
+        }
+    );
+    let _ = writeln!(
+        out,
+        "telemt_conntrack_delete_total{{result=\"success\"}} {}",
+        if core_enabled {
+            stats.get_conntrack_delete_success_total()
+        } else {
+            0
+        }
+    );
+    let _ = writeln!(
+        out,
+        "telemt_conntrack_delete_total{{result=\"not_found\"}} {}",
+        if core_enabled {
+            stats.get_conntrack_delete_not_found_total()
+        } else {
+            0
+        }
+    );
+    let _ = writeln!(
+        out,
+        "telemt_conntrack_delete_total{{result=\"error\"}} {}",
+        if core_enabled {
+            stats.get_conntrack_delete_error_total()
+        } else {
+            0
+        }
+    );
+
+    let _ = writeln!(
+        out,
+        "# HELP telemt_conntrack_close_event_drop_total Dropped conntrack close events due to queue pressure or unavailable sender"
+    );
+    let _ = writeln!(
+        out,
+        "# TYPE telemt_conntrack_close_event_drop_total counter"
+    );
+    let _ = writeln!(
+        out,
+        "telemt_conntrack_close_event_drop_total {}",
+        if core_enabled {
+            stats.get_conntrack_close_event_drop_total()
+        } else {
+            0
+        }
+    );
+
     let _ = writeln!(
         out,
         "# HELP telemt_upstream_connect_attempt_total Upstream connect attempts across all requests"
@@ -935,6 +1090,492 @@ async fn render_metrics(stats: &Stats, config: &ProxyConfig, ip_tracker: &UserIp
         }
     );
 
+    let _ = writeln!(
+        out,
+        "# HELP telemt_me_c2me_enqueue_events_total ME client->ME enqueue outcomes"
+    );
+    let _ = writeln!(out, "# TYPE telemt_me_c2me_enqueue_events_total counter");
+    let _ = writeln!(
+        out,
+        "telemt_me_c2me_enqueue_events_total{{event=\"full\"}} {}",
+        if me_allows_normal {
+            stats.get_me_c2me_send_full_total()
+        } else {
+            0
+        }
+    );
+    let _ = writeln!(
+        out,
+        "telemt_me_c2me_enqueue_events_total{{event=\"high_water\"}} {}",
+        if me_allows_normal {
+            stats.get_me_c2me_send_high_water_total()
+        } else {
+            0
+        }
+    );
+    let _ = writeln!(
+        out,
+        "telemt_me_c2me_enqueue_events_total{{event=\"timeout\"}} {}",
+        if me_allows_normal {
+            stats.get_me_c2me_send_timeout_total()
+        } else {
+            0
+        }
+    );
+
+    let _ = writeln!(
+        out,
+        "# HELP telemt_me_d2c_batches_total Total DC->Client flush batches"
+    );
+    let _ = writeln!(out, "# TYPE telemt_me_d2c_batches_total counter");
+    let _ = writeln!(
+        out,
+        "telemt_me_d2c_batches_total {}",
+        if me_allows_normal {
+            stats.get_me_d2c_batches_total()
+        } else {
+            0
+        }
+    );
+
+    let _ = writeln!(
+        out,
+        "# HELP telemt_me_d2c_batch_frames_total Total DC->Client frames flushed in batches"
+    );
+    let _ = writeln!(out, "# TYPE telemt_me_d2c_batch_frames_total counter");
+    let _ = writeln!(
+        out,
+        "telemt_me_d2c_batch_frames_total {}",
+        if me_allows_normal {
+            stats.get_me_d2c_batch_frames_total()
+        } else {
+            0
+        }
+    );
+
+    let _ = writeln!(
+        out,
+        "# HELP telemt_me_d2c_batch_bytes_total Total DC->Client bytes flushed in batches"
+    );
+    let _ = writeln!(out, "# TYPE telemt_me_d2c_batch_bytes_total counter");
+    let _ = writeln!(
+        out,
+        "telemt_me_d2c_batch_bytes_total {}",
+        if me_allows_normal {
+            stats.get_me_d2c_batch_bytes_total()
+        } else {
+            0
+        }
+    );
+
+    let _ = writeln!(
+        out,
+        "# HELP telemt_me_d2c_flush_reason_total DC->Client flush reasons"
+    );
+    let _ = writeln!(out, "# TYPE telemt_me_d2c_flush_reason_total counter");
+    let _ = writeln!(
+        out,
+        "telemt_me_d2c_flush_reason_total{{reason=\"queue_drain\"}} {}",
+        if me_allows_normal {
+            stats.get_me_d2c_flush_reason_queue_drain_total()
+        } else {
+            0
+        }
+    );
+    let _ = writeln!(
+        out,
+        "telemt_me_d2c_flush_reason_total{{reason=\"batch_frames\"}} {}",
+        if me_allows_normal {
+            stats.get_me_d2c_flush_reason_batch_frames_total()
+        } else {
+            0
+        }
+    );
+    let _ = writeln!(
+        out,
+        "telemt_me_d2c_flush_reason_total{{reason=\"batch_bytes\"}} {}",
+        if me_allows_normal {
+            stats.get_me_d2c_flush_reason_batch_bytes_total()
+        } else {
+            0
+        }
+    );
+    let _ = writeln!(
+        out,
+        "telemt_me_d2c_flush_reason_total{{reason=\"max_delay\"}} {}",
+        if me_allows_normal {
+            stats.get_me_d2c_flush_reason_max_delay_total()
+        } else {
+            0
+        }
+    );
+    let _ = writeln!(
+        out,
+        "telemt_me_d2c_flush_reason_total{{reason=\"ack_immediate\"}} {}",
+        if me_allows_normal {
+            stats.get_me_d2c_flush_reason_ack_immediate_total()
+        } else {
+            0
+        }
+    );
+    let _ = writeln!(
+        out,
+        "telemt_me_d2c_flush_reason_total{{reason=\"close\"}} {}",
+        if me_allows_normal {
+            stats.get_me_d2c_flush_reason_close_total()
+        } else {
+            0
+        }
+    );
+
+    let _ = writeln!(
+        out,
+        "# HELP telemt_me_d2c_data_frames_total DC->Client data frames"
+    );
+    let _ = writeln!(out, "# TYPE telemt_me_d2c_data_frames_total counter");
+    let _ = writeln!(
+        out,
+        "telemt_me_d2c_data_frames_total {}",
+        if me_allows_normal {
+            stats.get_me_d2c_data_frames_total()
+        } else {
+            0
+        }
+    );
+
+    let _ = writeln!(
+        out,
+        "# HELP telemt_me_d2c_ack_frames_total DC->Client quick-ack frames"
+    );
+    let _ = writeln!(out, "# TYPE telemt_me_d2c_ack_frames_total counter");
+    let _ = writeln!(
+        out,
+        "telemt_me_d2c_ack_frames_total {}",
+        if me_allows_normal {
+            stats.get_me_d2c_ack_frames_total()
+        } else {
+            0
+        }
+    );
+
+    let _ = writeln!(
+        out,
+        "# HELP telemt_me_d2c_payload_bytes_total DC->Client payload bytes before transport framing"
+    );
+    let _ = writeln!(out, "# TYPE telemt_me_d2c_payload_bytes_total counter");
+    let _ = writeln!(
+        out,
+        "telemt_me_d2c_payload_bytes_total {}",
+        if me_allows_normal {
+            stats.get_me_d2c_payload_bytes_total()
+        } else {
+            0
+        }
+    );
+
+    let _ = writeln!(
+        out,
+        "# HELP telemt_me_d2c_write_mode_total DC->Client writer mode selection"
+    );
+    let _ = writeln!(out, "# TYPE telemt_me_d2c_write_mode_total counter");
+    let _ = writeln!(
+        out,
+        "telemt_me_d2c_write_mode_total{{mode=\"coalesced\"}} {}",
+        if me_allows_normal {
+            stats.get_me_d2c_write_mode_coalesced_total()
+        } else {
+            0
+        }
+    );
+    let _ = writeln!(
+        out,
+        "telemt_me_d2c_write_mode_total{{mode=\"split\"}} {}",
+        if me_allows_normal {
+            stats.get_me_d2c_write_mode_split_total()
+        } else {
+            0
+        }
+    );
+
+    let _ = writeln!(
+        out,
+        "# HELP telemt_me_d2c_quota_reject_total DC->Client quota rejects"
+    );
+    let _ = writeln!(out, "# TYPE telemt_me_d2c_quota_reject_total counter");
+    let _ = writeln!(
+        out,
+        "telemt_me_d2c_quota_reject_total{{stage=\"pre_write\"}} {}",
+        if me_allows_normal {
+            stats.get_me_d2c_quota_reject_pre_write_total()
+        } else {
+            0
+        }
+    );
+    let _ = writeln!(
+        out,
+        "telemt_me_d2c_quota_reject_total{{stage=\"post_write\"}} {}",
+        if me_allows_normal {
+            stats.get_me_d2c_quota_reject_post_write_total()
+        } else {
+            0
+        }
+    );
+
+    let _ = writeln!(
+        out,
+        "# HELP telemt_me_d2c_frame_buf_shrink_total DC->Client reusable frame buffer shrink events"
+    );
+    let _ = writeln!(out, "# TYPE telemt_me_d2c_frame_buf_shrink_total counter");
+    let _ = writeln!(
+        out,
+        "telemt_me_d2c_frame_buf_shrink_total {}",
+        if me_allows_normal {
+            stats.get_me_d2c_frame_buf_shrink_total()
+        } else {
+            0
+        }
+    );
+
+    let _ = writeln!(
+        out,
+        "# HELP telemt_me_d2c_frame_buf_shrink_bytes_total DC->Client reusable frame buffer bytes released"
+    );
+    let _ = writeln!(
+        out,
+        "# TYPE telemt_me_d2c_frame_buf_shrink_bytes_total counter"
+    );
+    let _ = writeln!(
+        out,
+        "telemt_me_d2c_frame_buf_shrink_bytes_total {}",
+        if me_allows_normal {
+            stats.get_me_d2c_frame_buf_shrink_bytes_total()
+        } else {
+            0
+        }
+    );
+
+    let _ = writeln!(
+        out,
+        "# HELP telemt_me_d2c_batch_frames_bucket_total DC->Client batch frame count buckets"
+    );
+    let _ = writeln!(
+        out,
+        "# TYPE telemt_me_d2c_batch_frames_bucket_total counter"
+    );
+    let _ = writeln!(
+        out,
+        "telemt_me_d2c_batch_frames_bucket_total{{bucket=\"1\"}} {}",
+        if me_allows_debug {
+            stats.get_me_d2c_batch_frames_bucket_1()
+        } else {
+            0
+        }
+    );
+    let _ = writeln!(
+        out,
+        "telemt_me_d2c_batch_frames_bucket_total{{bucket=\"2_4\"}} {}",
+        if me_allows_debug {
+            stats.get_me_d2c_batch_frames_bucket_2_4()
+        } else {
+            0
+        }
+    );
+    let _ = writeln!(
+        out,
+        "telemt_me_d2c_batch_frames_bucket_total{{bucket=\"5_8\"}} {}",
+        if me_allows_debug {
+            stats.get_me_d2c_batch_frames_bucket_5_8()
+        } else {
+            0
+        }
+    );
+    let _ = writeln!(
+        out,
+        "telemt_me_d2c_batch_frames_bucket_total{{bucket=\"9_16\"}} {}",
+        if me_allows_debug {
+            stats.get_me_d2c_batch_frames_bucket_9_16()
+        } else {
+            0
+        }
+    );
+    let _ = writeln!(
+        out,
+        "telemt_me_d2c_batch_frames_bucket_total{{bucket=\"17_32\"}} {}",
+        if me_allows_debug {
+            stats.get_me_d2c_batch_frames_bucket_17_32()
+        } else {
+            0
+        }
+    );
+    let _ = writeln!(
+        out,
+        "telemt_me_d2c_batch_frames_bucket_total{{bucket=\"gt_32\"}} {}",
+        if me_allows_debug {
+            stats.get_me_d2c_batch_frames_bucket_gt_32()
+        } else {
+            0
+        }
+    );
+
+    let _ = writeln!(
+        out,
+        "# HELP telemt_me_d2c_batch_bytes_bucket_total DC->Client batch byte size buckets"
+    );
+    let _ = writeln!(out, "# TYPE telemt_me_d2c_batch_bytes_bucket_total counter");
+    let _ = writeln!(
+        out,
+        "telemt_me_d2c_batch_bytes_bucket_total{{bucket=\"0_1k\"}} {}",
+        if me_allows_debug {
+            stats.get_me_d2c_batch_bytes_bucket_0_1k()
+        } else {
+            0
+        }
+    );
+    let _ = writeln!(
+        out,
+        "telemt_me_d2c_batch_bytes_bucket_total{{bucket=\"1k_4k\"}} {}",
+        if me_allows_debug {
+            stats.get_me_d2c_batch_bytes_bucket_1k_4k()
+        } else {
+            0
+        }
+    );
+    let _ = writeln!(
+        out,
+        "telemt_me_d2c_batch_bytes_bucket_total{{bucket=\"4k_16k\"}} {}",
+        if me_allows_debug {
+            stats.get_me_d2c_batch_bytes_bucket_4k_16k()
+        } else {
+            0
+        }
+    );
+    let _ = writeln!(
+        out,
+        "telemt_me_d2c_batch_bytes_bucket_total{{bucket=\"16k_64k\"}} {}",
+        if me_allows_debug {
+            stats.get_me_d2c_batch_bytes_bucket_16k_64k()
+        } else {
+            0
+        }
+    );
+    let _ = writeln!(
+        out,
+        "telemt_me_d2c_batch_bytes_bucket_total{{bucket=\"64k_128k\"}} {}",
+        if me_allows_debug {
+            stats.get_me_d2c_batch_bytes_bucket_64k_128k()
+        } else {
+            0
+        }
+    );
+    let _ = writeln!(
+        out,
+        "telemt_me_d2c_batch_bytes_bucket_total{{bucket=\"gt_128k\"}} {}",
+        if me_allows_debug {
+            stats.get_me_d2c_batch_bytes_bucket_gt_128k()
+        } else {
+            0
+        }
+    );
+
+    let _ = writeln!(
+        out,
+        "# HELP telemt_me_d2c_flush_duration_us_bucket_total DC->Client flush duration buckets"
+    );
+    let _ = writeln!(
+        out,
+        "# TYPE telemt_me_d2c_flush_duration_us_bucket_total counter"
+    );
+    let _ = writeln!(
+        out,
+        "telemt_me_d2c_flush_duration_us_bucket_total{{bucket=\"0_50\"}} {}",
+        if me_allows_debug {
+            stats.get_me_d2c_flush_duration_us_bucket_0_50()
+        } else {
+            0
+        }
+    );
+    let _ = writeln!(
+        out,
+        "telemt_me_d2c_flush_duration_us_bucket_total{{bucket=\"51_200\"}} {}",
+        if me_allows_debug {
+            stats.get_me_d2c_flush_duration_us_bucket_51_200()
+        } else {
+            0
+        }
+    );
+    let _ = writeln!(
+        out,
+        "telemt_me_d2c_flush_duration_us_bucket_total{{bucket=\"201_1000\"}} {}",
+        if me_allows_debug {
+            stats.get_me_d2c_flush_duration_us_bucket_201_1000()
+        } else {
+            0
+        }
+    );
+    let _ = writeln!(
+        out,
+        "telemt_me_d2c_flush_duration_us_bucket_total{{bucket=\"1001_5000\"}} {}",
+        if me_allows_debug {
+            stats.get_me_d2c_flush_duration_us_bucket_1001_5000()
+        } else {
+            0
+        }
+    );
+    let _ = writeln!(
+        out,
+        "telemt_me_d2c_flush_duration_us_bucket_total{{bucket=\"5001_20000\"}} {}",
+        if me_allows_debug {
+            stats.get_me_d2c_flush_duration_us_bucket_5001_20000()
+        } else {
+            0
+        }
+    );
+    let _ = writeln!(
+        out,
+        "telemt_me_d2c_flush_duration_us_bucket_total{{bucket=\"gt_20000\"}} {}",
+        if me_allows_debug {
+            stats.get_me_d2c_flush_duration_us_bucket_gt_20000()
+        } else {
+            0
+        }
+    );
+
+    let _ = writeln!(
+        out,
+        "# HELP telemt_me_d2c_batch_timeout_armed_total DC->Client max-delay timer armed events"
+    );
+    let _ = writeln!(
+        out,
+        "# TYPE telemt_me_d2c_batch_timeout_armed_total counter"
+    );
+    let _ = writeln!(
+        out,
+        "telemt_me_d2c_batch_timeout_armed_total {}",
+        if me_allows_debug {
+            stats.get_me_d2c_batch_timeout_armed_total()
+        } else {
+            0
+        }
+    );
+
+    let _ = writeln!(
+        out,
+        "# HELP telemt_me_d2c_batch_timeout_fired_total DC->Client max-delay timer fired events"
+    );
+    let _ = writeln!(
+        out,
+        "# TYPE telemt_me_d2c_batch_timeout_fired_total counter"
+    );
+    let _ = writeln!(
+        out,
+        "telemt_me_d2c_batch_timeout_fired_total {}",
+        if me_allows_debug {
+            stats.get_me_d2c_batch_timeout_fired_total()
+        } else {
+            0
+        }
+    );
+
     let _ = writeln!(
         out,
         "# HELP telemt_me_writer_pick_total ME writer-pick outcomes by mode and result"
@@ -1105,6 +1746,40 @@ async fn render_metrics(stats: &Stats, config: &ProxyConfig, ip_tracker: &UserIp
             0
         }
     );
+    let _ = writeln!(
+        out,
+        "# HELP telemt_me_endpoint_quarantine_unexpected_total ME endpoint quarantines caused by unexpected writer removals"
+    );
+    let _ = writeln!(
+        out,
+        "# TYPE telemt_me_endpoint_quarantine_unexpected_total counter"
+    );
+    let _ = writeln!(
+        out,
+        "telemt_me_endpoint_quarantine_unexpected_total {}",
+        if me_allows_normal {
+            stats.get_me_endpoint_quarantine_unexpected_total()
+        } else {
+            0
+        }
+    );
+    let _ = writeln!(
+        out,
+        "# HELP telemt_me_endpoint_quarantine_draining_suppressed_total Draining writer removals that skipped endpoint quarantine"
+    );
+    let _ = writeln!(
+        out,
+        "# TYPE telemt_me_endpoint_quarantine_draining_suppressed_total counter"
+    );
+    let _ = writeln!(
+        out,
+        "telemt_me_endpoint_quarantine_draining_suppressed_total {}",
+        if me_allows_normal {
+            stats.get_me_endpoint_quarantine_draining_suppressed_total()
+        } else {
+            0
+        }
+    );
 
     let _ = writeln!(
         out,
@@ -1865,6 +2540,20 @@ async fn render_metrics(stats: &Stats, config: &ProxyConfig, ip_tracker: &UserIp
             0
         }
     );
+    let _ = writeln!(
+        out,
+        "# HELP telemt_me_hybrid_timeout_total ME hybrid route timeouts after bounded retry window"
+    );
+    let _ = writeln!(out, "# TYPE telemt_me_hybrid_timeout_total counter");
+    let _ = writeln!(
+        out,
+        "telemt_me_hybrid_timeout_total {}",
+        if me_allows_normal {
+            stats.get_me_hybrid_timeout_total()
+        } else {
+            0
+        }
+    );
     let _ = writeln!(
         out,
         "# HELP telemt_me_async_recovery_trigger_total Async ME recovery trigger attempts from route path"
@@ -1983,6 +2672,48 @@ async fn render_metrics(stats: &Stats, config: &ProxyConfig, ip_tracker: &UserIp
         if user_enabled { 0 } else { 1 }
     );
 
+    let ip_memory = ip_tracker.memory_stats().await;
+    let _ = writeln!(
+        out,
+        "# HELP telemt_ip_tracker_users Number of users tracked by IP limiter state"
+    );
+    let _ = writeln!(out, "# TYPE telemt_ip_tracker_users gauge");
+    let _ = writeln!(
+        out,
+        "telemt_ip_tracker_users{{scope=\"active\"}} {}",
+        ip_memory.active_users
+    );
+    let _ = writeln!(
+        out,
+        "telemt_ip_tracker_users{{scope=\"recent\"}} {}",
+        ip_memory.recent_users
+    );
+    let _ = writeln!(
+        out,
+        "# HELP telemt_ip_tracker_entries Number of IP entries tracked by limiter state"
+    );
+    let _ = writeln!(out, "# TYPE telemt_ip_tracker_entries gauge");
+    let _ = writeln!(
+        out,
+        "telemt_ip_tracker_entries{{scope=\"active\"}} {}",
+        ip_memory.active_entries
+    );
+    let _ = writeln!(
+        out,
+        "telemt_ip_tracker_entries{{scope=\"recent\"}} {}",
+        ip_memory.recent_entries
+    );
+    let _ = writeln!(
+        out,
+        "# HELP telemt_ip_tracker_cleanup_queue_len Deferred disconnect cleanup queue length"
+    );
+    let _ = writeln!(out, "# TYPE telemt_ip_tracker_cleanup_queue_len gauge");
+    let _ = writeln!(
+        out,
+        "telemt_ip_tracker_cleanup_queue_len {}",
+        ip_memory.cleanup_queue_len
+    );
+
     if user_enabled {
         for entry in stats.iter_user_stats() {
             let user = entry.key();
@@ -2145,6 +2876,19 @@ mod tests {
         stats.increment_relay_idle_hard_close_total();
         stats.increment_relay_pressure_evict_total();
         stats.increment_relay_protocol_desync_close_total();
+        stats.increment_me_d2c_batches_total();
+        stats.add_me_d2c_batch_frames_total(3);
+        stats.add_me_d2c_batch_bytes_total(2048);
+        stats.increment_me_d2c_flush_reason(crate::stats::MeD2cFlushReason::AckImmediate);
+        stats.increment_me_d2c_data_frames_total();
+        stats.increment_me_d2c_ack_frames_total();
+        stats.add_me_d2c_payload_bytes_total(1800);
+        stats.increment_me_d2c_write_mode(crate::stats::MeD2cWriteMode::Coalesced);
+        stats.increment_me_d2c_quota_reject_total(crate::stats::MeD2cQuotaRejectStage::PostWrite);
+        stats.observe_me_d2c_frame_buf_shrink(4096);
+        stats.increment_me_endpoint_quarantine_total();
+        stats.increment_me_endpoint_quarantine_unexpected_total();
+        stats.increment_me_endpoint_quarantine_draining_suppressed_total();
         stats.increment_user_connects("alice");
         stats.increment_user_curr_connects("alice");
         stats.add_user_octets_from("alice", 1024);
@@ -2184,6 +2928,20 @@ mod tests {
         assert!(output.contains("telemt_relay_idle_hard_close_total 1"));
         assert!(output.contains("telemt_relay_pressure_evict_total 1"));
         assert!(output.contains("telemt_relay_protocol_desync_close_total 1"));
+        assert!(output.contains("telemt_me_d2c_batches_total 1"));
+        assert!(output.contains("telemt_me_d2c_batch_frames_total 3"));
+        assert!(output.contains("telemt_me_d2c_batch_bytes_total 2048"));
+        assert!(output.contains("telemt_me_d2c_flush_reason_total{reason=\"ack_immediate\"} 1"));
+        assert!(output.contains("telemt_me_d2c_data_frames_total 1"));
+        assert!(output.contains("telemt_me_d2c_ack_frames_total 1"));
+        assert!(output.contains("telemt_me_d2c_payload_bytes_total 1800"));
+        assert!(output.contains("telemt_me_d2c_write_mode_total{mode=\"coalesced\"} 1"));
+        assert!(output.contains("telemt_me_d2c_quota_reject_total{stage=\"post_write\"} 1"));
+        assert!(output.contains("telemt_me_d2c_frame_buf_shrink_total 1"));
+        assert!(output.contains("telemt_me_d2c_frame_buf_shrink_bytes_total 4096"));
+        assert!(output.contains("telemt_me_endpoint_quarantine_total 1"));
+        assert!(output.contains("telemt_me_endpoint_quarantine_unexpected_total 1"));
+        assert!(output.contains("telemt_me_endpoint_quarantine_draining_suppressed_total 1"));
         assert!(output.contains("telemt_user_connections_total{user=\"alice\"} 1"));
         assert!(output.contains("telemt_user_connections_current{user=\"alice\"} 1"));
         assert!(output.contains("telemt_user_octets_from_client{user=\"alice\"} 1024"));
@@ -2194,6 +2952,9 @@ mod tests {
         assert!(output.contains("telemt_user_unique_ips_recent_window{user=\"alice\"} 1"));
         assert!(output.contains("telemt_user_unique_ips_limit{user=\"alice\"} 4"));
         assert!(output.contains("telemt_user_unique_ips_utilization{user=\"alice\"} 0.250000"));
+        assert!(output.contains("telemt_ip_tracker_users{scope=\"active\"} 1"));
+        assert!(output.contains("telemt_ip_tracker_entries{scope=\"active\"} 1"));
+        assert!(output.contains("telemt_ip_tracker_cleanup_queue_len 0"));
     }
 
     #[tokio::test]
@@ -2245,6 +3006,17 @@ mod tests {
         assert!(output.contains("# TYPE telemt_relay_idle_hard_close_total counter"));
         assert!(output.contains("# TYPE telemt_relay_pressure_evict_total counter"));
         assert!(output.contains("# TYPE telemt_relay_protocol_desync_close_total counter"));
+        assert!(output.contains("# TYPE telemt_me_d2c_batches_total counter"));
+        assert!(output.contains("# TYPE telemt_me_d2c_flush_reason_total counter"));
+        assert!(output.contains("# TYPE telemt_me_d2c_write_mode_total counter"));
+        assert!(output.contains("# TYPE telemt_me_d2c_batch_frames_bucket_total counter"));
+        assert!(output.contains("# TYPE telemt_me_d2c_flush_duration_us_bucket_total counter"));
+        assert!(output.contains("# TYPE telemt_me_endpoint_quarantine_total counter"));
+        assert!(output.contains("# TYPE telemt_me_endpoint_quarantine_unexpected_total counter"));
+        assert!(
+            output
+                .contains("# TYPE telemt_me_endpoint_quarantine_draining_suppressed_total counter")
+        );
         assert!(output.contains("# TYPE telemt_me_writer_removed_total counter"));
         assert!(
             output
@@ -2254,6 +3026,9 @@ mod tests {
         assert!(output.contains("# TYPE telemt_user_unique_ips_recent_window gauge"));
         assert!(output.contains("# TYPE telemt_user_unique_ips_limit gauge"));
         assert!(output.contains("# TYPE telemt_user_unique_ips_utilization gauge"));
+        assert!(output.contains("# TYPE telemt_ip_tracker_users gauge"));
+        assert!(output.contains("# TYPE telemt_ip_tracker_entries gauge"));
+        assert!(output.contains("# TYPE telemt_ip_tracker_cleanup_queue_len gauge"));
     }
 
     #[tokio::test]

src/proxy/adaptive_buffers.rs

@@ -24,6 +24,8 @@ const DIRECT_S2C_CAP_BYTES: usize = 512 * 1024;
 const ME_FRAMES_CAP: usize = 96;
 const ME_BYTES_CAP: usize = 384 * 1024;
 const ME_DELAY_MIN_US: u64 = 150;
+const MAX_USER_PROFILES_ENTRIES: usize = 50_000;
+const MAX_USER_KEY_BYTES: usize = 512;
 
 #[derive(Debug, Clone, Copy, PartialEq, Eq, PartialOrd, Ord)]
 pub enum AdaptiveTier {
@@ -234,32 +236,50 @@ fn profiles() -> &'static DashMap<String, UserAdaptiveProfile> {
 }
 
 pub fn seed_tier_for_user(user: &str) -> AdaptiveTier {
+    if user.len() > MAX_USER_KEY_BYTES {
+        return AdaptiveTier::Base;
+    }
     let now = Instant::now();
     if let Some(entry) = profiles().get(user) {
-        let value = entry.value();
-        if now.duration_since(value.seen_at) <= PROFILE_TTL {
+        let value = *entry.value();
+        drop(entry);
+        if now.saturating_duration_since(value.seen_at) <= PROFILE_TTL {
             return value.tier;
         }
+        profiles().remove_if(user, |_, v| {
+            now.saturating_duration_since(v.seen_at) > PROFILE_TTL
+        });
     }
     AdaptiveTier::Base
 }
 
 pub fn record_user_tier(user: &str, tier: AdaptiveTier) {
-    let now = Instant::now();
-    if let Some(mut entry) = profiles().get_mut(user) {
-        let existing = *entry;
-        let effective = if now.duration_since(existing.seen_at) > PROFILE_TTL {
-            tier
-        } else {
-            max(existing.tier, tier)
-        };
-        *entry = UserAdaptiveProfile {
-            tier: effective,
-            seen_at: now,
-        };
+    if user.len() > MAX_USER_KEY_BYTES {
         return;
     }
-    profiles().insert(user.to_string(), UserAdaptiveProfile { tier, seen_at: now });
+    let now = Instant::now();
+    let mut was_vacant = false;
+    match profiles().entry(user.to_string()) {
+        dashmap::mapref::entry::Entry::Occupied(mut entry) => {
+            let existing = *entry.get();
+            let effective = if now.saturating_duration_since(existing.seen_at) > PROFILE_TTL {
+                tier
+            } else {
+                max(existing.tier, tier)
+            };
+            entry.insert(UserAdaptiveProfile {
+                tier: effective,
+                seen_at: now,
+            });
+        }
+        dashmap::mapref::entry::Entry::Vacant(slot) => {
+            slot.insert(UserAdaptiveProfile { tier, seen_at: now });
+            was_vacant = true;
+        }
+    }
+    if was_vacant && profiles().len() > MAX_USER_PROFILES_ENTRIES {
+        profiles().retain(|_, v| now.saturating_duration_since(v.seen_at) <= PROFILE_TTL);
+    }
 }
 
 pub fn direct_copy_buffers_for_tier(
@@ -310,6 +330,14 @@ fn scale(base: usize, numerator: usize, denominator: usize, cap: usize) -> usize
     scaled.min(cap).max(1)
 }
 
+#[cfg(test)]
+#[path = "tests/adaptive_buffers_security_tests.rs"]
+mod adaptive_buffers_security_tests;
+
+#[cfg(test)]
+#[path = "tests/adaptive_buffers_record_race_security_tests.rs"]
+mod adaptive_buffers_record_race_security_tests;
+
 #[cfg(test)]
 mod tests {
     use super::*;

src/proxy/client.rs

@@ -80,11 +80,16 @@ use crate::transport::middle_proxy::MePool;
 use crate::transport::socket::normalize_ip;
 use crate::transport::{UpstreamManager, configure_client_socket, parse_proxy_protocol};
 
-use crate::proxy::direct_relay::handle_via_direct;
-use crate::proxy::handshake::{HandshakeSuccess, handle_mtproto_handshake, handle_tls_handshake};
+use crate::proxy::direct_relay::handle_via_direct_with_shared;
+use crate::proxy::handshake::{
+    HandshakeSuccess, handle_mtproto_handshake_with_shared, handle_tls_handshake_with_shared,
+};
+#[cfg(test)]
+use crate::proxy::handshake::{handle_mtproto_handshake, handle_tls_handshake};
 use crate::proxy::masking::handle_bad_client;
 use crate::proxy::middle_relay::handle_via_middle_proxy;
 use crate::proxy::route_mode::{RelayRouteMode, RouteRuntimeController};
+use crate::proxy::shared_state::ProxySharedState;
 
 fn beobachten_ttl(config: &ProxyConfig) -> Duration {
     const BEOBACHTEN_TTL_MAX_MINUTES: u64 = 24 * 60;
@@ -186,6 +191,90 @@ fn handshake_timeout_with_mask_grace(config: &ProxyConfig) -> Duration {
     }
 }
 
+fn effective_client_first_byte_idle_secs(config: &ProxyConfig, shared: &ProxySharedState) -> u64 {
+    let idle_secs = config.timeouts.client_first_byte_idle_secs;
+    if idle_secs == 0 {
+        return 0;
+    }
+    if shared.conntrack_pressure_active() {
+        idle_secs.min(
+            config
+                .server
+                .conntrack_control
+                .profile
+                .client_first_byte_idle_cap_secs(),
+        )
+    } else {
+        idle_secs
+    }
+}
+
+const MASK_CLASSIFIER_PREFETCH_WINDOW: usize = 16;
+#[cfg(test)]
+const MASK_CLASSIFIER_PREFETCH_TIMEOUT: Duration = Duration::from_millis(5);
+
+fn mask_classifier_prefetch_timeout(config: &ProxyConfig) -> Duration {
+    Duration::from_millis(config.censorship.mask_classifier_prefetch_timeout_ms)
+}
+
+fn should_prefetch_mask_classifier_window(initial_data: &[u8]) -> bool {
+    if initial_data.len() >= MASK_CLASSIFIER_PREFETCH_WINDOW {
+        return false;
+    }
+
+    if initial_data.is_empty() {
+        // Empty initial_data means there is no client probe prefix to refine.
+        // Prefetching in this case can consume fallback relay payload bytes and
+        // accidentally route them through shaping heuristics.
+        return false;
+    }
+
+    if initial_data[0] == 0x16 || initial_data.starts_with(b"SSH-") {
+        return false;
+    }
+
+    initial_data
+        .iter()
+        .all(|b| b.is_ascii_alphabetic() || *b == b' ')
+}
+
+#[cfg(test)]
+async fn extend_masking_initial_window<R>(reader: &mut R, initial_data: &mut Vec<u8>)
+where
+    R: AsyncRead + Unpin,
+{
+    extend_masking_initial_window_with_timeout(
+        reader,
+        initial_data,
+        MASK_CLASSIFIER_PREFETCH_TIMEOUT,
+    )
+    .await;
+}
+
+async fn extend_masking_initial_window_with_timeout<R>(
+    reader: &mut R,
+    initial_data: &mut Vec<u8>,
+    prefetch_timeout: Duration,
+) where
+    R: AsyncRead + Unpin,
+{
+    if !should_prefetch_mask_classifier_window(initial_data) {
+        return;
+    }
+
+    let need = MASK_CLASSIFIER_PREFETCH_WINDOW.saturating_sub(initial_data.len());
+    if need == 0 {
+        return;
+    }
+
+    let mut extra = [0u8; MASK_CLASSIFIER_PREFETCH_WINDOW];
+    if let Ok(Ok(n)) = timeout(prefetch_timeout, reader.read(&mut extra[..need])).await
+        && n > 0
+    {
+        initial_data.extend_from_slice(&extra[..n]);
+    }
+}
+
 fn masking_outcome<R, W>(
     reader: R,
     writer: W,
@@ -200,6 +289,15 @@ where
     W: AsyncWrite + Unpin + Send + 'static,
 {
     HandshakeOutcome::NeedsMasking(Box::pin(async move {
+        let mut reader = reader;
+        let mut initial_data = initial_data;
+        extend_masking_initial_window_with_timeout(
+            &mut reader,
+            &mut initial_data,
+            mask_classifier_prefetch_timeout(&config),
+        )
+        .await;
+
         handle_bad_client(
             reader,
             writer,
@@ -242,13 +340,20 @@ fn record_handshake_failure_class(
     record_beobachten_class(beobachten, config, peer_ip, class);
 }
 
+#[inline]
+fn increment_bad_on_unknown_tls_sni(stats: &Stats, error: &ProxyError) {
+    if matches!(error, ProxyError::UnknownTlsSni) {
+        stats.increment_connects_bad();
+    }
+}
+
 fn is_trusted_proxy_source(peer_ip: IpAddr, trusted: &[IpNetwork]) -> bool {
     if trusted.is_empty() {
         static EMPTY_PROXY_TRUST_WARNED: OnceLock<AtomicBool> = OnceLock::new();
         let warned = EMPTY_PROXY_TRUST_WARNED.get_or_init(|| AtomicBool::new(false));
         if !warned.swap(true, Ordering::Relaxed) {
             warn!(
-                "PROXY protocol enabled but server.proxy_protocol_trusted_cidrs is empty; rejecting all PROXY headers by default"
+                "PROXY protocol enabled but server.proxy_protocol_trusted_cidrs is empty; rejecting all PROXY headers"
             );
         }
         return false;
@@ -260,7 +365,48 @@ fn synthetic_local_addr(port: u16) -> SocketAddr {
     SocketAddr::from(([0, 0, 0, 0], port))
 }
 
+#[cfg(test)]
 pub async fn handle_client_stream<S>(
+    stream: S,
+    peer: SocketAddr,
+    config: Arc<ProxyConfig>,
+    stats: Arc<Stats>,
+    upstream_manager: Arc<UpstreamManager>,
+    replay_checker: Arc<ReplayChecker>,
+    buffer_pool: Arc<BufferPool>,
+    rng: Arc<SecureRandom>,
+    me_pool: Option<Arc<MePool>>,
+    route_runtime: Arc<RouteRuntimeController>,
+    tls_cache: Option<Arc<TlsFrontCache>>,
+    ip_tracker: Arc<UserIpTracker>,
+    beobachten: Arc<BeobachtenStore>,
+    proxy_protocol_enabled: bool,
+) -> Result<()>
+where
+    S: AsyncRead + AsyncWrite + Unpin + Send + 'static,
+{
+    handle_client_stream_with_shared(
+        stream,
+        peer,
+        config,
+        stats,
+        upstream_manager,
+        replay_checker,
+        buffer_pool,
+        rng,
+        me_pool,
+        route_runtime,
+        tls_cache,
+        ip_tracker,
+        beobachten,
+        ProxySharedState::new(),
+        proxy_protocol_enabled,
+    )
+    .await
+}
+
+#[allow(clippy::too_many_arguments)]
+pub async fn handle_client_stream_with_shared<S>(
     mut stream: S,
     peer: SocketAddr,
     config: Arc<ProxyConfig>,
@@ -274,6 +420,7 @@ pub async fn handle_client_stream<S>(
     tls_cache: Option<Arc<TlsFrontCache>>,
     ip_tracker: Arc<UserIpTracker>,
     beobachten: Arc<BeobachtenStore>,
+    shared: Arc<ProxySharedState>,
     proxy_protocol_enabled: bool,
 ) -> Result<()>
 where
@@ -334,16 +481,69 @@ where
 
     debug!(peer = %real_peer, "New connection (generic stream)");
 
+    let first_byte_idle_secs = effective_client_first_byte_idle_secs(&config, shared.as_ref());
+    let first_byte = if first_byte_idle_secs == 0 {
+        None
+    } else {
+        let idle_timeout = Duration::from_secs(first_byte_idle_secs);
+        let mut first_byte = [0u8; 1];
+        match timeout(idle_timeout, stream.read(&mut first_byte)).await {
+            Ok(Ok(0)) => {
+                debug!(peer = %real_peer, "Connection closed before first client byte");
+                return Ok(());
+            }
+            Ok(Ok(_)) => Some(first_byte[0]),
+            Ok(Err(e))
+                if matches!(
+                    e.kind(),
+                    std::io::ErrorKind::UnexpectedEof
+                        | std::io::ErrorKind::ConnectionReset
+                        | std::io::ErrorKind::ConnectionAborted
+                        | std::io::ErrorKind::BrokenPipe
+                        | std::io::ErrorKind::NotConnected
+                ) =>
+            {
+                debug!(
+                    peer = %real_peer,
+                    error = %e,
+                    "Connection closed before first client byte"
+                );
+                return Ok(());
+            }
+            Ok(Err(e)) => {
+                debug!(
+                    peer = %real_peer,
+                    error = %e,
+                    "Failed while waiting for first client byte"
+                );
+                return Err(ProxyError::Io(e));
+            }
+            Err(_) => {
+                debug!(
+                    peer = %real_peer,
+                    idle_secs = first_byte_idle_secs,
+                    "Closing idle pooled connection before first client byte"
+                );
+                return Ok(());
+            }
+        }
+    };
+
     let handshake_timeout = handshake_timeout_with_mask_grace(&config);
     let stats_for_timeout = stats.clone();
     let config_for_timeout = config.clone();
     let beobachten_for_timeout = beobachten.clone();
     let peer_for_timeout = real_peer.ip();
 
-    // Phase 1: handshake (with timeout)
+    // Phase 2: active handshake (with timeout after the first client byte)
     let outcome = match timeout(handshake_timeout, async {
         let mut first_bytes = [0u8; 5];
-        stream.read_exact(&mut first_bytes).await?;
+        if let Some(first_byte) = first_byte {
+            first_bytes[0] = first_byte;
+            stream.read_exact(&mut first_bytes[1..]).await?;
+        } else {
+            stream.read_exact(&mut first_bytes).await?;
+        }
 
         let is_tls = tls::is_tls_handshake(&first_bytes[..3]);
         debug!(peer = %real_peer, is_tls = is_tls, "Handshake type detected");
@@ -416,9 +616,10 @@ where
 
             let (read_half, write_half) = tokio::io::split(stream);
 
-            let (mut tls_reader, tls_writer, tls_user) = match handle_tls_handshake(
+            let (mut tls_reader, tls_writer, tls_user) = match handle_tls_handshake_with_shared(
                 &handshake, read_half, write_half, real_peer,
                 &config, &replay_checker, &rng, tls_cache.clone(),
+                shared.as_ref(),
             ).await {
                 HandshakeResult::Success(result) => result,
                 HandshakeResult::BadClient { reader, writer } => {
@@ -433,7 +634,10 @@ where
                         beobachten.clone(),
                     ));
                 }
-                HandshakeResult::Error(e) => return Err(e),
+                HandshakeResult::Error(e) => {
+                    increment_bad_on_unknown_tls_sni(stats.as_ref(), &e);
+                    return Err(e);
+                }
             };
 
             debug!(peer = %peer, "Reading MTProto handshake through TLS");
@@ -441,9 +645,10 @@ where
             let mtproto_handshake: [u8; HANDSHAKE_LEN] = mtproto_data[..].try_into()
                 .map_err(|_| ProxyError::InvalidHandshake("Short MTProto handshake".into()))?;
 
-            let (crypto_reader, crypto_writer, success) = match handle_mtproto_handshake(
+            let (crypto_reader, crypto_writer, success) = match handle_mtproto_handshake_with_shared(
                 &mtproto_handshake, tls_reader, tls_writer, real_peer,
                 &config, &replay_checker, true, Some(tls_user.as_str()),
+                shared.as_ref(),
             ).await {
                 HandshakeResult::Success(result) => result,
                 HandshakeResult::BadClient { reader, writer } => {
@@ -477,11 +682,12 @@ where
             };
 
             Ok(HandshakeOutcome::NeedsRelay(Box::pin(
-                RunningClientHandler::handle_authenticated_static(
+                RunningClientHandler::handle_authenticated_static_with_shared(
                     crypto_reader, crypto_writer, success,
                     upstream_manager, stats, config, buffer_pool, rng, me_pool,
                     route_runtime.clone(),
                     local_addr, real_peer, ip_tracker.clone(),
+                    shared.clone(),
                 ),
             )))
         } else {
@@ -507,9 +713,10 @@ where
 
             let (read_half, write_half) = tokio::io::split(stream);
 
-            let (crypto_reader, crypto_writer, success) = match handle_mtproto_handshake(
+            let (crypto_reader, crypto_writer, success) = match handle_mtproto_handshake_with_shared(
                 &handshake, read_half, write_half, real_peer,
                 &config, &replay_checker, false, None,
+                shared.as_ref(),
             ).await {
                 HandshakeResult::Success(result) => result,
                 HandshakeResult::BadClient { reader, writer } => {
@@ -528,7 +735,7 @@ where
             };
 
             Ok(HandshakeOutcome::NeedsRelay(Box::pin(
-                RunningClientHandler::handle_authenticated_static(
+                RunningClientHandler::handle_authenticated_static_with_shared(
                     crypto_reader,
                     crypto_writer,
                     success,
@@ -542,6 +749,7 @@ where
                     local_addr,
                     real_peer,
                     ip_tracker.clone(),
+                    shared.clone(),
                 )
             )))
         }
@@ -594,10 +802,12 @@ pub struct RunningClientHandler {
     tls_cache: Option<Arc<TlsFrontCache>>,
     ip_tracker: Arc<UserIpTracker>,
     beobachten: Arc<BeobachtenStore>,
+    shared: Arc<ProxySharedState>,
     proxy_protocol_enabled: bool,
 }
 
 impl ClientHandler {
+    #[cfg(test)]
     pub fn new(
         stream: TcpStream,
         peer: SocketAddr,
@@ -614,6 +824,45 @@ impl ClientHandler {
         beobachten: Arc<BeobachtenStore>,
         proxy_protocol_enabled: bool,
         real_peer_report: Arc<std::sync::Mutex<Option<SocketAddr>>>,
+    ) -> RunningClientHandler {
+        Self::new_with_shared(
+            stream,
+            peer,
+            config,
+            stats,
+            upstream_manager,
+            replay_checker,
+            buffer_pool,
+            rng,
+            me_pool,
+            route_runtime,
+            tls_cache,
+            ip_tracker,
+            beobachten,
+            ProxySharedState::new(),
+            proxy_protocol_enabled,
+            real_peer_report,
+        )
+    }
+
+    #[allow(clippy::too_many_arguments)]
+    pub fn new_with_shared(
+        stream: TcpStream,
+        peer: SocketAddr,
+        config: Arc<ProxyConfig>,
+        stats: Arc<Stats>,
+        upstream_manager: Arc<UpstreamManager>,
+        replay_checker: Arc<ReplayChecker>,
+        buffer_pool: Arc<BufferPool>,
+        rng: Arc<SecureRandom>,
+        me_pool: Option<Arc<MePool>>,
+        route_runtime: Arc<RouteRuntimeController>,
+        tls_cache: Option<Arc<TlsFrontCache>>,
+        ip_tracker: Arc<UserIpTracker>,
+        beobachten: Arc<BeobachtenStore>,
+        shared: Arc<ProxySharedState>,
+        proxy_protocol_enabled: bool,
+        real_peer_report: Arc<std::sync::Mutex<Option<SocketAddr>>>,
     ) -> RunningClientHandler {
         let normalized_peer = normalize_ip(peer);
         RunningClientHandler {
@@ -632,6 +881,7 @@ impl ClientHandler {
             tls_cache,
             ip_tracker,
             beobachten,
+            shared,
             proxy_protocol_enabled,
         }
     }
@@ -651,36 +901,9 @@ impl RunningClientHandler {
             debug!(peer = %peer, error = %e, "Failed to configure client socket");
         }
 
-        let handshake_timeout = handshake_timeout_with_mask_grace(&self.config);
-        let stats = self.stats.clone();
-        let config_for_timeout = self.config.clone();
-        let beobachten_for_timeout = self.beobachten.clone();
-        let peer_for_timeout = peer.ip();
-
-        // Phase 1: handshake (with timeout)
-        let outcome = match timeout(handshake_timeout, self.do_handshake()).await {
-            Ok(Ok(outcome)) => outcome,
-            Ok(Err(e)) => {
-                debug!(peer = %peer, error = %e, "Handshake failed");
-                record_handshake_failure_class(
-                    &beobachten_for_timeout,
-                    &config_for_timeout,
-                    peer_for_timeout,
-                    &e,
-                );
-                return Err(e);
-            }
-            Err(_) => {
-                stats.increment_handshake_timeouts();
-                debug!(peer = %peer, "Handshake timeout");
-                record_beobachten_class(
-                    &beobachten_for_timeout,
-                    &config_for_timeout,
-                    peer_for_timeout,
-                    "other",
-                );
-                return Err(ProxyError::TgHandshakeTimeout);
-            }
+        let outcome = match self.do_handshake().await? {
+            Some(outcome) => outcome,
+            None => return Ok(()),
         };
 
         // Phase 2: relay (WITHOUT handshake timeout — relay has its own activity timeouts)
@@ -689,7 +912,7 @@ impl RunningClientHandler {
         }
     }
 
-    async fn do_handshake(mut self) -> Result<HandshakeOutcome> {
+    async fn do_handshake(mut self) -> Result<Option<HandshakeOutcome>> {
         let mut local_addr = self.stream.local_addr().map_err(ProxyError::Io)?;
 
         if self.proxy_protocol_enabled {
@@ -764,19 +987,109 @@ impl RunningClientHandler {
             }
         }
 
-        let mut first_bytes = [0u8; 5];
-        self.stream.read_exact(&mut first_bytes).await?;
-
-        let is_tls = tls::is_tls_handshake(&first_bytes[..3]);
-        let peer = self.peer;
-
-        debug!(peer = %peer, is_tls = is_tls, "Handshake type detected");
-
-        if is_tls {
-            self.handle_tls_client(first_bytes, local_addr).await
+        let first_byte_idle_secs =
+            effective_client_first_byte_idle_secs(&self.config, self.shared.as_ref());
+        let first_byte = if first_byte_idle_secs == 0 {
+            None
         } else {
-            self.handle_direct_client(first_bytes, local_addr).await
-        }
+            let idle_timeout = Duration::from_secs(first_byte_idle_secs);
+            let mut first_byte = [0u8; 1];
+            match timeout(idle_timeout, self.stream.read(&mut first_byte)).await {
+                Ok(Ok(0)) => {
+                    debug!(peer = %self.peer, "Connection closed before first client byte");
+                    return Ok(None);
+                }
+                Ok(Ok(_)) => Some(first_byte[0]),
+                Ok(Err(e))
+                    if matches!(
+                        e.kind(),
+                        std::io::ErrorKind::UnexpectedEof
+                            | std::io::ErrorKind::ConnectionReset
+                            | std::io::ErrorKind::ConnectionAborted
+                            | std::io::ErrorKind::BrokenPipe
+                            | std::io::ErrorKind::NotConnected
+                    ) =>
+                {
+                    debug!(
+                        peer = %self.peer,
+                        error = %e,
+                        "Connection closed before first client byte"
+                    );
+                    return Ok(None);
+                }
+                Ok(Err(e)) => {
+                    debug!(
+                        peer = %self.peer,
+                        error = %e,
+                        "Failed while waiting for first client byte"
+                    );
+                    return Err(ProxyError::Io(e));
+                }
+                Err(_) => {
+                    debug!(
+                        peer = %self.peer,
+                        idle_secs = first_byte_idle_secs,
+                        "Closing idle pooled connection before first client byte"
+                    );
+                    return Ok(None);
+                }
+            }
+        };
+
+        let handshake_timeout = handshake_timeout_with_mask_grace(&self.config);
+        let stats = self.stats.clone();
+        let config_for_timeout = self.config.clone();
+        let beobachten_for_timeout = self.beobachten.clone();
+        let peer_for_timeout = self.peer.ip();
+        let peer_for_log = self.peer;
+
+        let outcome = match timeout(handshake_timeout, async {
+            let mut first_bytes = [0u8; 5];
+            if let Some(first_byte) = first_byte {
+                first_bytes[0] = first_byte;
+                self.stream.read_exact(&mut first_bytes[1..]).await?;
+            } else {
+                self.stream.read_exact(&mut first_bytes).await?;
+            }
+
+            let is_tls = tls::is_tls_handshake(&first_bytes[..3]);
+            let peer = self.peer;
+
+            debug!(peer = %peer, is_tls = is_tls, "Handshake type detected");
+
+            if is_tls {
+                self.handle_tls_client(first_bytes, local_addr).await
+            } else {
+                self.handle_direct_client(first_bytes, local_addr).await
+            }
+        })
+        .await
+        {
+            Ok(Ok(outcome)) => outcome,
+            Ok(Err(e)) => {
+                debug!(peer = %peer_for_log, error = %e, "Handshake failed");
+                record_handshake_failure_class(
+                    &beobachten_for_timeout,
+                    &config_for_timeout,
+                    peer_for_timeout,
+                    &e,
+                );
+                return Err(e);
+            }
+            Err(_) => {
+                stats.increment_handshake_timeouts();
+                debug!(peer = %peer_for_log, "Handshake timeout");
+                record_beobachten_class(
+                    &beobachten_for_timeout,
+                    &config_for_timeout,
+                    peer_for_timeout,
+                    "other",
+                );
+                return Err(ProxyError::TgHandshakeTimeout);
+            }
+        };
+
+        Ok(Some(outcome))
     }
 
     async fn handle_tls_client(
@@ -859,7 +1172,7 @@ impl RunningClientHandler {
 
         let (read_half, write_half) = self.stream.into_split();
 
-        let (mut tls_reader, tls_writer, tls_user) = match handle_tls_handshake(
+        let (mut tls_reader, tls_writer, tls_user) = match handle_tls_handshake_with_shared(
             &handshake,
             read_half,
             write_half,
@@ -868,6 +1181,7 @@ impl RunningClientHandler {
             &replay_checker,
             &self.rng,
             self.tls_cache.clone(),
+            self.shared.as_ref(),
         )
         .await
         {
@@ -884,7 +1198,10 @@ impl RunningClientHandler {
                     self.beobachten.clone(),
                 ));
             }
-            HandshakeResult::Error(e) => return Err(e),
+            HandshakeResult::Error(e) => {
+                increment_bad_on_unknown_tls_sni(stats.as_ref(), &e);
+                return Err(e);
+            }
         };
 
         debug!(peer = %peer, "Reading MTProto handshake through TLS");
@@ -893,7 +1210,7 @@ impl RunningClientHandler {
             .try_into()
             .map_err(|_| ProxyError::InvalidHandshake("Short MTProto handshake".into()))?;
 
-        let (crypto_reader, crypto_writer, success) = match handle_mtproto_handshake(
+        let (crypto_reader, crypto_writer, success) = match handle_mtproto_handshake_with_shared(
             &mtproto_handshake,
             tls_reader,
             tls_writer,
@@ -902,6 +1219,7 @@ impl RunningClientHandler {
             &replay_checker,
             true,
             Some(tls_user.as_str()),
+            self.shared.as_ref(),
         )
         .await
         {
@@ -938,7 +1256,7 @@ impl RunningClientHandler {
         };
 
         Ok(HandshakeOutcome::NeedsRelay(Box::pin(
-            Self::handle_authenticated_static(
+            Self::handle_authenticated_static_with_shared(
                 crypto_reader,
                 crypto_writer,
                 success,
@@ -952,6 +1270,7 @@ impl RunningClientHandler {
                 local_addr,
                 peer,
                 self.ip_tracker,
+                self.shared,
             ),
         )))
     }
@@ -990,7 +1309,7 @@ impl RunningClientHandler {
 
         let (read_half, write_half) = self.stream.into_split();
 
-        let (crypto_reader, crypto_writer, success) = match handle_mtproto_handshake(
+        let (crypto_reader, crypto_writer, success) = match handle_mtproto_handshake_with_shared(
             &handshake,
             read_half,
             write_half,
@@ -999,6 +1318,7 @@ impl RunningClientHandler {
             &replay_checker,
             false,
             None,
+            self.shared.as_ref(),
         )
         .await
         {
@@ -1019,7 +1339,7 @@ impl RunningClientHandler {
         };
 
         Ok(HandshakeOutcome::NeedsRelay(Box::pin(
-            Self::handle_authenticated_static(
+            Self::handle_authenticated_static_with_shared(
                 crypto_reader,
                 crypto_writer,
                 success,
@@ -1033,6 +1353,7 @@ impl RunningClientHandler {
                 local_addr,
                 peer,
                 self.ip_tracker,
+                self.shared,
             ),
         )))
     }
@@ -1041,6 +1362,7 @@ impl RunningClientHandler {
     /// Two modes:
     ///   - Direct: TCP relay to TG DC (existing behavior)  
     ///   - Middle Proxy: RPC multiplex through ME pool (new — supports CDN DCs)
+    #[cfg(test)]
     async fn handle_authenticated_static<R, W>(
         client_reader: CryptoReader<R>,
         client_writer: CryptoWriter<W>,
@@ -1056,6 +1378,45 @@ impl RunningClientHandler {
         peer_addr: SocketAddr,
         ip_tracker: Arc<UserIpTracker>,
     ) -> Result<()>
+    where
+        R: AsyncRead + Unpin + Send + 'static,
+        W: AsyncWrite + Unpin + Send + 'static,
+    {
+        Self::handle_authenticated_static_with_shared(
+            client_reader,
+            client_writer,
+            success,
+            upstream_manager,
+            stats,
+            config,
+            buffer_pool,
+            rng,
+            me_pool,
+            route_runtime,
+            local_addr,
+            peer_addr,
+            ip_tracker,
+            ProxySharedState::new(),
+        )
+        .await
+    }
+
+    async fn handle_authenticated_static_with_shared<R, W>(
+        client_reader: CryptoReader<R>,
+        client_writer: CryptoWriter<W>,
+        success: HandshakeSuccess,
+        upstream_manager: Arc<UpstreamManager>,
+        stats: Arc<Stats>,
+        config: Arc<ProxyConfig>,
+        buffer_pool: Arc<BufferPool>,
+        rng: Arc<SecureRandom>,
+        me_pool: Option<Arc<MePool>>,
+        route_runtime: Arc<RouteRuntimeController>,
+        local_addr: SocketAddr,
+        peer_addr: SocketAddr,
+        ip_tracker: Arc<UserIpTracker>,
+        shared: Arc<ProxySharedState>,
+    ) -> Result<()>
     where
         R: AsyncRead + Unpin + Send + 'static,
         W: AsyncWrite + Unpin + Send + 'static,
@@ -1097,11 +1458,12 @@ impl RunningClientHandler {
                     route_runtime.subscribe(),
                     route_snapshot,
                     session_id,
+                    shared.clone(),
                 )
                 .await
             } else {
                 warn!("use_middle_proxy=true but MePool not initialized, falling back to direct");
-                handle_via_direct(
+                handle_via_direct_with_shared(
                     client_reader,
                     client_writer,
                     success,
@@ -1113,12 +1475,14 @@ impl RunningClientHandler {
                     route_runtime.subscribe(),
                     route_snapshot,
                     session_id,
+                    local_addr,
+                    shared.clone(),
                 )
                 .await
             }
         } else {
             // Direct mode (original behavior)
-            handle_via_direct(
+            handle_via_direct_with_shared(
                 client_reader,
                 client_writer,
                 success,
@@ -1130,6 +1494,8 @@ impl RunningClientHandler {
                 route_runtime.subscribe(),
                 route_snapshot,
                 session_id,
+                local_addr,
+                shared.clone(),
             )
             .await
         };
@@ -1153,7 +1519,7 @@ impl RunningClientHandler {
         }
 
         if let Some(quota) = config.access.user_data_quota.get(user)
-            && stats.get_user_total_octets(user) >= *quota
+            && stats.get_user_quota_used(user) >= *quota
         {
             return Err(ProxyError::DataQuotaExceeded {
                 user: user.to_string(),
@@ -1164,7 +1530,11 @@ impl RunningClientHandler {
             .access
             .user_max_tcp_conns
             .get(user)
-            .map(|v| *v as u64);
+            .copied()
+            .filter(|limit| *limit > 0)
+            .or((config.access.user_max_tcp_conns_global_each > 0)
+                .then_some(config.access.user_max_tcp_conns_global_each))
+            .map(|v| v as u64);
         if !stats.try_acquire_user_curr_connects(user, limit) {
             return Err(ProxyError::ConnectionLimitExceeded {
                 user: user.to_string(),
@@ -1212,7 +1582,7 @@ impl RunningClientHandler {
         }
 
         if let Some(quota) = config.access.user_data_quota.get(user)
-            && stats.get_user_total_octets(user) >= *quota
+            && stats.get_user_quota_used(user) >= *quota
         {
             return Err(ProxyError::DataQuotaExceeded {
                 user: user.to_string(),
@@ -1223,7 +1593,11 @@ impl RunningClientHandler {
             .access
             .user_max_tcp_conns
             .get(user)
-            .map(|v| *v as u64);
+            .copied()
+            .filter(|limit| *limit > 0)
+            .or((config.access.user_max_tcp_conns_global_each > 0)
+                .then_some(config.access.user_max_tcp_conns_global_each))
+            .map(|v| v as u64);
         if !stats.try_acquire_user_curr_connects(user, limit) {
             return Err(ProxyError::ConnectionLimitExceeded {
                 user: user.to_string(),
@@ -1321,6 +1695,38 @@ mod masking_shape_classifier_fuzz_redteam_expected_fail_tests;
 #[path = "tests/client_masking_probe_evasion_blackhat_tests.rs"]
 mod masking_probe_evasion_blackhat_tests;
 
+#[cfg(test)]
+#[path = "tests/client_masking_fragmented_classifier_security_tests.rs"]
+mod masking_fragmented_classifier_security_tests;
+
+#[cfg(test)]
+#[path = "tests/client_masking_replay_timing_security_tests.rs"]
+mod masking_replay_timing_security_tests;
+
+#[cfg(test)]
+#[path = "tests/client_masking_http2_fragmented_preface_security_tests.rs"]
+mod masking_http2_fragmented_preface_security_tests;
+
+#[cfg(test)]
+#[path = "tests/client_masking_prefetch_invariant_security_tests.rs"]
+mod masking_prefetch_invariant_security_tests;
+
+#[cfg(test)]
+#[path = "tests/client_masking_prefetch_timing_matrix_security_tests.rs"]
+mod masking_prefetch_timing_matrix_security_tests;
+
+#[cfg(test)]
+#[path = "tests/client_masking_prefetch_config_runtime_security_tests.rs"]
+mod masking_prefetch_config_runtime_security_tests;
+
+#[cfg(test)]
+#[path = "tests/client_masking_prefetch_config_pipeline_integration_security_tests.rs"]
+mod masking_prefetch_config_pipeline_integration_security_tests;
+
+#[cfg(test)]
+#[path = "tests/client_masking_prefetch_strict_boundary_security_tests.rs"]
+mod masking_prefetch_strict_boundary_security_tests;
+
 #[cfg(test)]
 #[path = "tests/client_beobachten_ttl_bounds_security_tests.rs"]
 mod beobachten_ttl_bounds_security_tests;
@@ -1328,3 +1734,15 @@ mod beobachten_ttl_bounds_security_tests;
 #[cfg(test)]
 #[path = "tests/client_tls_record_wrap_hardening_security_tests.rs"]
 mod tls_record_wrap_hardening_security_tests;
+
+#[cfg(test)]
+#[path = "tests/client_clever_advanced_tests.rs"]
+mod client_clever_advanced_tests;
+
+#[cfg(test)]
+#[path = "tests/client_more_advanced_tests.rs"]
+mod client_more_advanced_tests;
+
+#[cfg(test)]
+#[path = "tests/client_deep_invariants_tests.rs"]
+mod client_deep_invariants_tests;

src/proxy/direct_relay.rs

@@ -6,6 +6,7 @@ use std::net::SocketAddr;
 use std::path::{Component, Path, PathBuf};
 use std::sync::Arc;
 use std::sync::{Mutex, OnceLock};
+use std::time::Duration;
 
 use tokio::io::{AsyncRead, AsyncWrite, AsyncWriteExt, ReadHalf, WriteHalf, split};
 use tokio::sync::watch;
@@ -16,11 +17,13 @@ use crate::crypto::SecureRandom;
 use crate::error::{ProxyError, Result};
 use crate::protocol::constants::*;
 use crate::proxy::handshake::{HandshakeSuccess, encrypt_tg_nonce_with_ciphers, generate_tg_nonce};
-use crate::proxy::relay::relay_bidirectional;
 use crate::proxy::route_mode::{
     ROUTE_SWITCH_ERROR_MSG, RelayRouteMode, RouteCutoverState, affected_cutover_state,
     cutover_stagger_delay,
 };
+use crate::proxy::shared_state::{
+    ConntrackCloseEvent, ConntrackClosePublishResult, ConntrackCloseReason, ProxySharedState,
+};
 use crate::stats::Stats;
 use crate::stream::{BufferPool, CryptoReader, CryptoWriter};
 use crate::transport::UpstreamManager;
@@ -225,7 +228,43 @@ fn unknown_dc_test_lock() -> &'static Mutex<()> {
     TEST_LOCK.get_or_init(|| Mutex::new(()))
 }
 
+#[allow(dead_code)]
 pub(crate) async fn handle_via_direct<R, W>(
+    client_reader: CryptoReader<R>,
+    client_writer: CryptoWriter<W>,
+    success: HandshakeSuccess,
+    upstream_manager: Arc<UpstreamManager>,
+    stats: Arc<Stats>,
+    config: Arc<ProxyConfig>,
+    buffer_pool: Arc<BufferPool>,
+    rng: Arc<SecureRandom>,
+    route_rx: watch::Receiver<RouteCutoverState>,
+    route_snapshot: RouteCutoverState,
+    session_id: u64,
+) -> Result<()>
+where
+    R: AsyncRead + Unpin + Send + 'static,
+    W: AsyncWrite + Unpin + Send + 'static,
+{
+    handle_via_direct_with_shared(
+        client_reader,
+        client_writer,
+        success,
+        upstream_manager,
+        stats,
+        config.clone(),
+        buffer_pool,
+        rng,
+        route_rx,
+        route_snapshot,
+        session_id,
+        SocketAddr::from(([0, 0, 0, 0], config.server.port)),
+        ProxySharedState::new(),
+    )
+    .await
+}
+
+pub(crate) async fn handle_via_direct_with_shared<R, W>(
     client_reader: CryptoReader<R>,
     client_writer: CryptoWriter<W>,
     success: HandshakeSuccess,
@@ -237,6 +276,8 @@ pub(crate) async fn handle_via_direct<R, W>(
     mut route_rx: watch::Receiver<RouteCutoverState>,
     route_snapshot: RouteCutoverState,
     session_id: u64,
+    local_addr: SocketAddr,
+    shared: Arc<ProxySharedState>,
 ) -> Result<()>
 where
     R: AsyncRead + Unpin + Send + 'static,
@@ -276,7 +317,19 @@ where
     stats.increment_user_connects(user);
     let _direct_connection_lease = stats.acquire_direct_connection_lease();
 
-    let relay_result = relay_bidirectional(
+    let buffer_pool_trim = Arc::clone(&buffer_pool);
+    let relay_activity_timeout = if shared.conntrack_pressure_active() {
+        Duration::from_secs(
+            config
+                .server
+                .conntrack_control
+                .profile
+                .direct_activity_timeout_secs(),
+        )
+    } else {
+        Duration::from_secs(1800)
+    };
+    let relay_result = crate::proxy::relay::relay_bidirectional_with_activity_timeout(
         client_reader,
         client_writer,
         tg_reader,
@@ -287,6 +340,7 @@ where
         Arc::clone(&stats),
         config.access.user_data_quota.get(user).copied(),
         buffer_pool,
+        relay_activity_timeout,
     );
     tokio::pin!(relay_result);
     let relay_result = loop {
@@ -321,9 +375,59 @@ where
         Err(e) => debug!(user = %user, error = %e, "Direct relay ended with error"),
     }
 
+    buffer_pool_trim.trim_to(buffer_pool_trim.max_buffers().min(64));
+    let pool_snapshot = buffer_pool_trim.stats();
+    stats.set_buffer_pool_gauges(
+        pool_snapshot.pooled,
+        pool_snapshot.allocated,
+        pool_snapshot.allocated.saturating_sub(pool_snapshot.pooled),
+    );
+
+    let close_reason = classify_conntrack_close_reason(&relay_result);
+    let publish_result = shared.publish_conntrack_close_event(ConntrackCloseEvent {
+        src: success.peer,
+        dst: local_addr,
+        reason: close_reason,
+    });
+    if !matches!(
+        publish_result,
+        ConntrackClosePublishResult::Sent | ConntrackClosePublishResult::Disabled
+    ) {
+        stats.increment_conntrack_close_event_drop_total();
+    }
+
     relay_result
 }
 
+fn classify_conntrack_close_reason(result: &Result<()>) -> ConntrackCloseReason {
+    match result {
+        Ok(()) => ConntrackCloseReason::NormalEof,
+        Err(crate::error::ProxyError::Io(error))
+            if matches!(error.kind(), std::io::ErrorKind::TimedOut) =>
+        {
+            ConntrackCloseReason::Timeout
+        }
+        Err(crate::error::ProxyError::Io(error))
+            if matches!(
+                error.kind(),
+                std::io::ErrorKind::ConnectionReset
+                    | std::io::ErrorKind::ConnectionAborted
+                    | std::io::ErrorKind::BrokenPipe
+                    | std::io::ErrorKind::NotConnected
+                    | std::io::ErrorKind::UnexpectedEof
+            ) =>
+        {
+            ConntrackCloseReason::Reset
+        }
+        Err(crate::error::ProxyError::Proxy(message))
+            if message.contains("pressure") || message.contains("evicted") =>
+        {
+            ConntrackCloseReason::Pressure
+        }
+        Err(_) => ConntrackCloseReason::Other,
+    }
+}
+
 fn get_dc_addr_static(dc_idx: i16, config: &ProxyConfig) -> Result<SocketAddr> {
     let prefer_v6 = config.network.prefer == 6 && config.network.ipv6.unwrap_or(true);
     let datacenters = if prefer_v6 {

src/proxy/masking.rs

@@ -4,14 +4,23 @@ use crate::config::ProxyConfig;
 use crate::network::dns_overrides::resolve_socket_addr;
 use crate::stats::beobachten::BeobachtenStore;
 use crate::transport::proxy_protocol::{ProxyProtocolV1Builder, ProxyProtocolV2Builder};
-use rand::{Rng, RngExt};
-use std::net::SocketAddr;
+#[cfg(unix)]
+use nix::ifaddrs::getifaddrs;
+use rand::rngs::StdRng;
+use rand::{Rng, RngExt, SeedableRng};
+use std::net::{IpAddr, SocketAddr};
 use std::str;
-use std::time::Duration;
+#[cfg(test)]
+use std::sync::atomic::{AtomicUsize, Ordering};
+#[cfg(unix)]
+use std::sync::{Mutex, OnceLock};
+use std::time::{Duration, Instant as StdInstant};
 use tokio::io::{AsyncRead, AsyncReadExt, AsyncWrite, AsyncWriteExt};
 use tokio::net::TcpStream;
 #[cfg(unix)]
 use tokio::net::UnixStream;
+#[cfg(unix)]
+use tokio::sync::Mutex as AsyncMutex;
 use tokio::time::{Instant, timeout};
 use tracing::debug;
 
@@ -30,28 +39,55 @@ const MASK_RELAY_IDLE_TIMEOUT: Duration = Duration::from_secs(5);
 #[cfg(test)]
 const MASK_RELAY_IDLE_TIMEOUT: Duration = Duration::from_millis(100);
 const MASK_BUFFER_SIZE: usize = 8192;
+#[cfg(unix)]
+#[cfg(not(test))]
+const LOCAL_INTERFACE_CACHE_TTL: Duration = Duration::from_secs(300);
+#[cfg(all(unix, test))]
+const LOCAL_INTERFACE_CACHE_TTL: Duration = Duration::from_secs(1);
 
 struct CopyOutcome {
     total: usize,
     ended_by_eof: bool,
 }
 
-async fn copy_with_idle_timeout<R, W>(reader: &mut R, writer: &mut W) -> CopyOutcome
+async fn copy_with_idle_timeout<R, W>(
+    reader: &mut R,
+    writer: &mut W,
+    byte_cap: usize,
+    shutdown_on_eof: bool,
+) -> CopyOutcome
 where
     R: AsyncRead + Unpin,
     W: AsyncWrite + Unpin,
 {
-    let mut buf = [0u8; MASK_BUFFER_SIZE];
+    let mut buf = Box::new([0u8; MASK_BUFFER_SIZE]);
     let mut total = 0usize;
     let mut ended_by_eof = false;
+
+    if byte_cap == 0 {
+        return CopyOutcome {
+            total,
+            ended_by_eof,
+        };
+    }
+
     loop {
-        let read_res = timeout(MASK_RELAY_IDLE_TIMEOUT, reader.read(&mut buf)).await;
+        let remaining_budget = byte_cap.saturating_sub(total);
+        if remaining_budget == 0 {
+            break;
+        }
+
+        let read_len = remaining_budget.min(MASK_BUFFER_SIZE);
+        let read_res = timeout(MASK_RELAY_IDLE_TIMEOUT, reader.read(&mut buf[..read_len])).await;
         let n = match read_res {
             Ok(Ok(n)) => n,
             Ok(Err(_)) | Err(_) => break,
         };
         if n == 0 {
             ended_by_eof = true;
+            if shutdown_on_eof {
+                let _ = timeout(MASK_RELAY_IDLE_TIMEOUT, writer.shutdown()).await;
+            }
             break;
         }
         total = total.saturating_add(n);
@@ -68,6 +104,31 @@ where
     }
 }
 
+fn is_http_probe(data: &[u8]) -> bool {
+    // RFC 7540 section 3.5: HTTP/2 client preface starts with "PRI ".
+    const HTTP_METHODS: [&[u8]; 10] = [
+        b"GET ", b"POST", b"HEAD", b"PUT ", b"DELETE", b"OPTIONS", b"CONNECT", b"TRACE", b"PATCH",
+        b"PRI ",
+    ];
+
+    if data.is_empty() {
+        return false;
+    }
+
+    let window = &data[..data.len().min(16)];
+    for method in HTTP_METHODS {
+        if data.len() >= method.len() && window.starts_with(method) {
+            return true;
+        }
+
+        if (2..=3).contains(&window.len()) && method.starts_with(window) {
+            return true;
+        }
+    }
+
+    false
+}
+
 fn next_mask_shape_bucket(total: usize, floor: usize, cap: usize) -> usize {
     if total == 0 || floor == 0 || cap < floor {
         return total;
@@ -125,6 +186,11 @@ async fn maybe_write_shape_padding<W>(
     let mut remaining = target_total - total_sent;
     let mut pad_chunk = [0u8; 1024];
     let deadline = Instant::now() + MASK_TIMEOUT;
+    // Use a Send RNG so relay futures remain spawn-safe under Tokio.
+    let mut rng = {
+        let mut seed_source = rand::rng();
+        StdRng::from_rng(&mut seed_source)
+    };
 
     while remaining > 0 {
         let now = Instant::now();
@@ -133,10 +199,7 @@ async fn maybe_write_shape_padding<W>(
         }
 
         let write_len = remaining.min(pad_chunk.len());
-        {
-            let mut rng = rand::rng();
-            rng.fill_bytes(&mut pad_chunk[..write_len]);
-        }
+        rng.fill_bytes(&mut pad_chunk[..write_len]);
         let write_budget = deadline.saturating_duration_since(now);
         match timeout(write_budget, mask_write.write_all(&pad_chunk[..write_len])).await {
             Ok(Ok(())) => {}
@@ -167,11 +230,11 @@ where
     }
 }
 
-async fn consume_client_data_with_timeout<R>(reader: R)
+async fn consume_client_data_with_timeout_and_cap<R>(reader: R, byte_cap: usize)
 where
     R: AsyncRead + Unpin,
 {
-    if timeout(MASK_RELAY_TIMEOUT, consume_client_data(reader))
+    if timeout(MASK_RELAY_TIMEOUT, consume_client_data(reader, byte_cap))
         .await
         .is_err()
     {
@@ -186,15 +249,63 @@ async fn wait_mask_connect_budget(started: Instant) {
     }
 }
 
+// Log-normal sample bounded to [floor, ceiling]. Median = sqrt(floor * ceiling).
+// Implements Box-Muller transform for standard normal sampling — no external
+// dependency on rand_distr (which is incompatible with rand 0.10).
+// sigma is chosen so ~99% of raw samples land inside [floor, ceiling] before clamp.
+// When floor > ceiling (misconfiguration), returns ceiling (the smaller value).
+// When floor == ceiling, returns that value. When both are 0, returns 0.
+pub(crate) fn sample_lognormal_percentile_bounded(
+    floor: u64,
+    ceiling: u64,
+    rng: &mut impl Rng,
+) -> u64 {
+    if ceiling == 0 && floor == 0 {
+        return 0;
+    }
+    if floor > ceiling {
+        return ceiling;
+    }
+    if floor == ceiling {
+        return floor;
+    }
+    let floor_f = floor.max(1) as f64;
+    let ceiling_f = ceiling.max(1) as f64;
+    let mu = (floor_f.ln() + ceiling_f.ln()) / 2.0;
+    // 4.65 ≈ 2 * 2.326 (double-sided z-score for 99th percentile)
+    let sigma = ((ceiling_f / floor_f).ln() / 4.65).max(0.01);
+    // Box-Muller transform: two uniform samples → one standard normal sample
+    let u1: f64 = rng.random_range(f64::MIN_POSITIVE..1.0);
+    let u2: f64 = rng.random_range(0.0_f64..std::f64::consts::TAU);
+    let normal_sample = (-2.0_f64 * u1.ln()).sqrt() * u2.cos();
+    let raw = (mu + sigma * normal_sample).exp();
+    if raw.is_finite() {
+        (raw as u64).clamp(floor, ceiling)
+    } else {
+        ((floor_f * ceiling_f).sqrt()) as u64
+    }
+}
+
 fn mask_outcome_target_budget(config: &ProxyConfig) -> Duration {
     if config.censorship.mask_timing_normalization_enabled {
         let floor = config.censorship.mask_timing_normalization_floor_ms;
         let ceiling = config.censorship.mask_timing_normalization_ceiling_ms;
+        if floor == 0 {
+            if ceiling == 0 {
+                return Duration::from_millis(0);
+            }
+            // floor=0 stays uniform: log-normal cannot model distribution anchored at zero
+            let mut rng = rand::rng();
+            return Duration::from_millis(rng.random_range(0..=ceiling));
+        }
         if ceiling > floor {
             let mut rng = rand::rng();
-            return Duration::from_millis(rng.random_range(floor..=ceiling));
+            return Duration::from_millis(sample_lognormal_percentile_bounded(
+                floor, ceiling, &mut rng,
+            ));
         }
-        return Duration::from_millis(floor);
+        // ceiling <= floor: use the larger value (fail-closed: preserve longer delay)
+        return Duration::from_millis(floor.max(ceiling));
     }
 
     MASK_TIMEOUT
@@ -219,14 +330,7 @@ async fn wait_mask_outcome_budget(started: Instant, config: &ProxyConfig) {
 /// Detect client type based on initial data
 fn detect_client_type(data: &[u8]) -> &'static str {
     // Check for HTTP request
-    if data.len() > 4
-        && (data.starts_with(b"GET ")
-            || data.starts_with(b"POST")
-            || data.starts_with(b"HEAD")
-            || data.starts_with(b"PUT ")
-            || data.starts_with(b"DELETE")
-            || data.starts_with(b"OPTIONS"))
-    {
+    if is_http_probe(data) {
         return "HTTP";
     }
 
@@ -248,6 +352,247 @@ fn detect_client_type(data: &[u8]) -> &'static str {
     "unknown"
 }
 
+fn parse_mask_host_ip_literal(host: &str) -> Option<IpAddr> {
+    if host.starts_with('[') && host.ends_with(']') {
+        return host[1..host.len() - 1].parse::<IpAddr>().ok();
+    }
+    host.parse::<IpAddr>().ok()
+}
+
+fn canonical_ip(ip: IpAddr) -> IpAddr {
+    match ip {
+        IpAddr::V6(v6) => v6
+            .to_ipv4_mapped()
+            .map(IpAddr::V4)
+            .unwrap_or(IpAddr::V6(v6)),
+        IpAddr::V4(v4) => IpAddr::V4(v4),
+    }
+}
+
+#[cfg(unix)]
+fn collect_local_interface_ips() -> Vec<IpAddr> {
+    #[cfg(test)]
+    LOCAL_INTERFACE_ENUMERATIONS.fetch_add(1, Ordering::Relaxed);
+
+    let mut out = Vec::new();
+    if let Ok(addrs) = getifaddrs() {
+        for iface in addrs {
+            if let Some(address) = iface.address {
+                if let Some(v4) = address.as_sockaddr_in() {
+                    out.push(canonical_ip(IpAddr::V4(v4.ip())));
+                } else if let Some(v6) = address.as_sockaddr_in6() {
+                    out.push(canonical_ip(IpAddr::V6(v6.ip())));
+                }
+            }
+        }
+    }
+    out
+}
+
+fn choose_interface_snapshot(previous: &[IpAddr], refreshed: Vec<IpAddr>) -> Vec<IpAddr> {
+    if refreshed.is_empty() && !previous.is_empty() {
+        return previous.to_vec();
+    }
+
+    refreshed
+}
+
+#[cfg(unix)]
+#[derive(Default)]
+struct LocalInterfaceCache {
+    ips: Vec<IpAddr>,
+    refreshed_at: Option<StdInstant>,
+}
+
+#[cfg(unix)]
+static LOCAL_INTERFACE_CACHE: OnceLock<Mutex<LocalInterfaceCache>> = OnceLock::new();
+
+#[cfg(unix)]
+static LOCAL_INTERFACE_REFRESH_LOCK: OnceLock<AsyncMutex<()>> = OnceLock::new();
+
+#[cfg(all(unix, test))]
+fn local_interface_ips() -> Vec<IpAddr> {
+    let cache = LOCAL_INTERFACE_CACHE.get_or_init(|| Mutex::new(LocalInterfaceCache::default()));
+    let mut guard = cache.lock().unwrap_or_else(|poison| poison.into_inner());
+
+    let stale = guard
+        .refreshed_at
+        .is_none_or(|at| at.elapsed() >= LOCAL_INTERFACE_CACHE_TTL);
+    if stale {
+        let refreshed = collect_local_interface_ips();
+        guard.ips = choose_interface_snapshot(&guard.ips, refreshed);
+        guard.refreshed_at = Some(StdInstant::now());
+    }
+
+    guard.ips.clone()
+}
+
+#[cfg(unix)]
+async fn local_interface_ips_async() -> Vec<IpAddr> {
+    let cache = LOCAL_INTERFACE_CACHE.get_or_init(|| Mutex::new(LocalInterfaceCache::default()));
+
+    {
+        let guard = cache.lock().unwrap_or_else(|poison| poison.into_inner());
+        let stale = guard
+            .refreshed_at
+            .is_none_or(|at| at.elapsed() >= LOCAL_INTERFACE_CACHE_TTL);
+        if !stale {
+            return guard.ips.clone();
+        }
+    }
+
+    let refresh_lock = LOCAL_INTERFACE_REFRESH_LOCK.get_or_init(|| AsyncMutex::new(()));
+    let _refresh_guard = refresh_lock.lock().await;
+
+    {
+        let guard = cache.lock().unwrap_or_else(|poison| poison.into_inner());
+        let stale = guard
+            .refreshed_at
+            .is_none_or(|at| at.elapsed() >= LOCAL_INTERFACE_CACHE_TTL);
+        if !stale {
+            return guard.ips.clone();
+        }
+    }
+
+    let refreshed = tokio::task::spawn_blocking(collect_local_interface_ips)
+        .await
+        .unwrap_or_default();
+
+    let mut guard = cache.lock().unwrap_or_else(|poison| poison.into_inner());
+    let stale = guard
+        .refreshed_at
+        .is_none_or(|at| at.elapsed() >= LOCAL_INTERFACE_CACHE_TTL);
+    if stale {
+        guard.ips = choose_interface_snapshot(&guard.ips, refreshed);
+        guard.refreshed_at = Some(StdInstant::now());
+    }
+
+    guard.ips.clone()
+}
+
+#[cfg(all(not(unix), test))]
+fn local_interface_ips() -> Vec<IpAddr> {
+    Vec::new()
+}
+
+#[cfg(not(unix))]
+async fn local_interface_ips_async() -> Vec<IpAddr> {
+    Vec::new()
+}
+
+#[cfg(test)]
+static LOCAL_INTERFACE_ENUMERATIONS: AtomicUsize = AtomicUsize::new(0);
+
+#[cfg(test)]
+fn reset_local_interface_enumerations_for_tests() {
+    LOCAL_INTERFACE_ENUMERATIONS.store(0, Ordering::Relaxed);
+
+    #[cfg(unix)]
+    if let Some(cache) = LOCAL_INTERFACE_CACHE.get() {
+        let mut guard = cache.lock().unwrap_or_else(|poison| poison.into_inner());
+        guard.ips.clear();
+        guard.refreshed_at = None;
+    }
+}
+
+#[cfg(test)]
+fn local_interface_enumerations_for_tests() -> usize {
+    LOCAL_INTERFACE_ENUMERATIONS.load(Ordering::Relaxed)
+}
+
+fn is_mask_target_local_listener_with_interfaces(
+    mask_host: &str,
+    mask_port: u16,
+    local_addr: SocketAddr,
+    resolved_override: Option<SocketAddr>,
+    interface_ips: &[IpAddr],
+) -> bool {
+    if mask_port != local_addr.port() {
+        return false;
+    }
+
+    let local_ip = canonical_ip(local_addr.ip());
+    let literal_mask_ip = parse_mask_host_ip_literal(mask_host).map(canonical_ip);
+
+    if let Some(addr) = resolved_override {
+        let resolved_ip = canonical_ip(addr.ip());
+        if resolved_ip == local_ip {
+            return true;
+        }
+
+        if local_ip.is_unspecified()
+            && (resolved_ip.is_loopback()
+                || resolved_ip.is_unspecified()
+                || interface_ips.contains(&resolved_ip))
+        {
+            return true;
+        }
+    }
+
+    if let Some(mask_ip) = literal_mask_ip {
+        if mask_ip == local_ip {
+            return true;
+        }
+
+        if local_ip.is_unspecified()
+            && (mask_ip.is_loopback()
+                || mask_ip.is_unspecified()
+                || interface_ips.contains(&mask_ip))
+        {
+            return true;
+        }
+    }
+
+    false
+}
+
+#[cfg(test)]
+fn is_mask_target_local_listener(
+    mask_host: &str,
+    mask_port: u16,
+    local_addr: SocketAddr,
+    resolved_override: Option<SocketAddr>,
+) -> bool {
+    if mask_port != local_addr.port() {
+        return false;
+    }
+
+    let interfaces = local_interface_ips();
+    is_mask_target_local_listener_with_interfaces(
+        mask_host,
+        mask_port,
+        local_addr,
+        resolved_override,
+        &interfaces,
+    )
+}
+
+async fn is_mask_target_local_listener_async(
+    mask_host: &str,
+    mask_port: u16,
+    local_addr: SocketAddr,
+    resolved_override: Option<SocketAddr>,
+) -> bool {
+    if mask_port != local_addr.port() {
+        return false;
+    }
+
+    let interfaces = local_interface_ips_async().await;
+    is_mask_target_local_listener_with_interfaces(
+        mask_host,
+        mask_port,
+        local_addr,
+        resolved_override,
+        &interfaces,
+    )
+}
+
+fn masking_beobachten_ttl(config: &ProxyConfig) -> Duration {
+    let minutes = config.general.beobachten_minutes;
+    let clamped = minutes.clamp(1, 24 * 60);
+    Duration::from_secs(clamped.saturating_mul(60))
+}
+
 fn build_mask_proxy_header(
     version: u8,
     peer: SocketAddr,
@@ -290,13 +635,14 @@ pub async fn handle_bad_client<R, W>(
 {
     let client_type = detect_client_type(initial_data);
     if config.general.beobachten {
-        let ttl = Duration::from_secs(config.general.beobachten_minutes.saturating_mul(60));
+        let ttl = masking_beobachten_ttl(config);
         beobachten.record(client_type, peer.ip(), ttl);
     }
 
     if !config.censorship.mask {
         // Masking disabled, just consume data
-        consume_client_data_with_timeout(reader).await;
+        consume_client_data_with_timeout_and_cap(reader, config.censorship.mask_relay_max_bytes)
+            .await;
         return;
     }
 
@@ -341,6 +687,7 @@ pub async fn handle_bad_client<R, W>(
                         config.censorship.mask_shape_above_cap_blur,
                         config.censorship.mask_shape_above_cap_blur_max_bytes,
                         config.censorship.mask_shape_hardening_aggressive_mode,
+                        config.censorship.mask_relay_max_bytes,
                     ),
                 )
                 .await
@@ -353,12 +700,20 @@ pub async fn handle_bad_client<R, W>(
             Ok(Err(e)) => {
                 wait_mask_connect_budget_if_needed(connect_started, config).await;
                 debug!(error = %e, "Failed to connect to mask unix socket");
-                consume_client_data_with_timeout(reader).await;
+                consume_client_data_with_timeout_and_cap(
+                    reader,
+                    config.censorship.mask_relay_max_bytes,
+                )
+                .await;
                 wait_mask_outcome_budget(outcome_started, config).await;
             }
             Err(_) => {
                 debug!("Timeout connecting to mask unix socket");
-                consume_client_data_with_timeout(reader).await;
+                consume_client_data_with_timeout_and_cap(
+                    reader,
+                    config.censorship.mask_relay_max_bytes,
+                )
+                .await;
                 wait_mask_outcome_budget(outcome_started, config).await;
             }
         }
@@ -372,6 +727,29 @@ pub async fn handle_bad_client<R, W>(
         .unwrap_or(&config.censorship.tls_domain);
     let mask_port = config.censorship.mask_port;
 
+    // Fail closed when fallback points at our own listener endpoint.
+    // Self-referential masking can create recursive proxy loops under
+    // misconfiguration and leak distinguishable load spikes to adversaries.
+    let resolved_mask_addr = resolve_socket_addr(mask_host, mask_port);
+    if is_mask_target_local_listener_async(mask_host, mask_port, local_addr, resolved_mask_addr)
+        .await
+    {
+        let outcome_started = Instant::now();
+        debug!(
+            client_type = client_type,
+            host = %mask_host,
+            port = mask_port,
+            local = %local_addr,
+            "Mask target resolves to local listener; refusing self-referential masking fallback"
+        );
+        consume_client_data_with_timeout_and_cap(reader, config.censorship.mask_relay_max_bytes)
+            .await;
+        wait_mask_outcome_budget(outcome_started, config).await;
+        return;
+    }
+
+    let outcome_started = Instant::now();
+
     debug!(
         client_type = client_type,
         host = %mask_host,
@@ -381,10 +759,9 @@ pub async fn handle_bad_client<R, W>(
     );
 
     // Apply runtime DNS override for mask target when configured.
-    let mask_addr = resolve_socket_addr(mask_host, mask_port)
+    let mask_addr = resolved_mask_addr
         .map(|addr| addr.to_string())
         .unwrap_or_else(|| format!("{}:{}", mask_host, mask_port));
-    let outcome_started = Instant::now();
     let connect_started = Instant::now();
     let connect_result = timeout(MASK_TIMEOUT, TcpStream::connect(&mask_addr)).await;
     match connect_result {
@@ -413,6 +790,7 @@ pub async fn handle_bad_client<R, W>(
                     config.censorship.mask_shape_above_cap_blur,
                     config.censorship.mask_shape_above_cap_blur_max_bytes,
                     config.censorship.mask_shape_hardening_aggressive_mode,
+                    config.censorship.mask_relay_max_bytes,
                 ),
             )
             .await
@@ -425,12 +803,20 @@ pub async fn handle_bad_client<R, W>(
         Ok(Err(e)) => {
             wait_mask_connect_budget_if_needed(connect_started, config).await;
             debug!(error = %e, "Failed to connect to mask host");
-            consume_client_data_with_timeout(reader).await;
+            consume_client_data_with_timeout_and_cap(
+                reader,
+                config.censorship.mask_relay_max_bytes,
+            )
+            .await;
             wait_mask_outcome_budget(outcome_started, config).await;
         }
         Err(_) => {
             debug!("Timeout connecting to mask host");
-            consume_client_data_with_timeout(reader).await;
+            consume_client_data_with_timeout_and_cap(
+                reader,
+                config.censorship.mask_relay_max_bytes,
+            )
+            .await;
             wait_mask_outcome_budget(outcome_started, config).await;
         }
     }
@@ -449,6 +835,7 @@ async fn relay_to_mask<R, W, MR, MW>(
     shape_above_cap_blur: bool,
     shape_above_cap_blur_max_bytes: usize,
     shape_hardening_aggressive_mode: bool,
+    mask_relay_max_bytes: usize,
 ) where
     R: AsyncRead + Unpin + Send + 'static,
     W: AsyncWrite + Unpin + Send + 'static,
@@ -464,8 +851,18 @@ async fn relay_to_mask<R, W, MR, MW>(
     }
 
     let (upstream_copy, downstream_copy) = tokio::join!(
-        async { copy_with_idle_timeout(&mut reader, &mut mask_write).await },
-        async { copy_with_idle_timeout(&mut mask_read, &mut writer).await }
+        async {
+            copy_with_idle_timeout(
+                &mut reader,
+                &mut mask_write,
+                mask_relay_max_bytes,
+                !shape_hardening_enabled,
+            )
+            .await
+        },
+        async {
+            copy_with_idle_timeout(&mut mask_read, &mut writer, mask_relay_max_bytes, true).await
+        }
     );
 
     let total_sent = initial_data.len().saturating_add(upstream_copy.total);
@@ -491,13 +888,36 @@ async fn relay_to_mask<R, W, MR, MW>(
     let _ = writer.shutdown().await;
 }
 
-/// Just consume all data from client without responding
-async fn consume_client_data<R: AsyncRead + Unpin>(mut reader: R) {
-    let mut buf = vec![0u8; MASK_BUFFER_SIZE];
-    while let Ok(n) = reader.read(&mut buf).await {
+/// Just consume all data from client without responding.
+async fn consume_client_data<R: AsyncRead + Unpin>(mut reader: R, byte_cap: usize) {
+    if byte_cap == 0 {
+        return;
+    }
+
+    // Keep drain path fail-closed under slow-loris stalls.
+    let mut buf = Box::new([0u8; MASK_BUFFER_SIZE]);
+    let mut total = 0usize;
+
+    loop {
+        let remaining_budget = byte_cap.saturating_sub(total);
+        if remaining_budget == 0 {
+            break;
+        }
+
+        let read_len = remaining_budget.min(MASK_BUFFER_SIZE);
+        let n = match timeout(MASK_RELAY_IDLE_TIMEOUT, reader.read(&mut buf[..read_len])).await {
+            Ok(Ok(n)) => n,
+            Ok(Err(_)) | Err(_) => break,
+        };
+
         if n == 0 {
             break;
         }
+
+        total = total.saturating_add(n);
+        if total >= byte_cap {
+            break;
+        }
     }
 }
 
@@ -521,6 +941,10 @@ mod masking_shape_above_cap_blur_security_tests;
 #[path = "tests/masking_timing_normalization_security_tests.rs"]
 mod masking_timing_normalization_security_tests;
 
+#[cfg(test)]
+#[path = "tests/masking_timing_budget_coupling_security_tests.rs"]
+mod masking_timing_budget_coupling_security_tests;
+
 #[cfg(test)]
 #[path = "tests/masking_ab_envelope_blur_integration_security_tests.rs"]
 mod masking_ab_envelope_blur_integration_security_tests;
@@ -548,3 +972,83 @@ mod masking_aggressive_mode_security_tests;
 #[cfg(test)]
 #[path = "tests/masking_timing_sidechannel_redteam_expected_fail_tests.rs"]
 mod masking_timing_sidechannel_redteam_expected_fail_tests;
+
+#[cfg(test)]
+#[path = "tests/masking_self_target_loop_security_tests.rs"]
+mod masking_self_target_loop_security_tests;
+
+#[cfg(test)]
+#[path = "tests/masking_classification_completeness_security_tests.rs"]
+mod masking_classification_completeness_security_tests;
+
+#[cfg(test)]
+#[path = "tests/masking_relay_guardrails_security_tests.rs"]
+mod masking_relay_guardrails_security_tests;
+
+#[cfg(test)]
+#[path = "tests/masking_connect_failure_close_matrix_security_tests.rs"]
+mod masking_connect_failure_close_matrix_security_tests;
+
+#[cfg(test)]
+#[path = "tests/masking_additional_hardening_security_tests.rs"]
+mod masking_additional_hardening_security_tests;
+
+#[cfg(test)]
+#[path = "tests/masking_consume_idle_timeout_security_tests.rs"]
+mod masking_consume_idle_timeout_security_tests;
+
+#[cfg(test)]
+#[path = "tests/masking_http2_probe_classification_security_tests.rs"]
+mod masking_http2_probe_classification_security_tests;
+
+#[cfg(test)]
+#[path = "tests/masking_http_probe_boundary_security_tests.rs"]
+mod masking_http_probe_boundary_security_tests;
+
+#[cfg(test)]
+#[path = "tests/masking_rng_hoist_perf_regression_tests.rs"]
+mod masking_rng_hoist_perf_regression_tests;
+
+#[cfg(test)]
+#[path = "tests/masking_http2_preface_integration_security_tests.rs"]
+mod masking_http2_preface_integration_security_tests;
+
+#[cfg(test)]
+#[path = "tests/masking_consume_stress_adversarial_tests.rs"]
+mod masking_consume_stress_adversarial_tests;
+
+#[cfg(test)]
+#[path = "tests/masking_interface_cache_security_tests.rs"]
+mod masking_interface_cache_security_tests;
+
+#[cfg(test)]
+#[path = "tests/masking_interface_cache_defense_in_depth_security_tests.rs"]
+mod masking_interface_cache_defense_in_depth_security_tests;
+
+#[cfg(test)]
+#[path = "tests/masking_interface_cache_concurrency_security_tests.rs"]
+mod masking_interface_cache_concurrency_security_tests;
+
+#[cfg(test)]
+#[path = "tests/masking_production_cap_regression_security_tests.rs"]
+mod masking_production_cap_regression_security_tests;
+
+#[cfg(test)]
+#[path = "tests/masking_extended_attack_surface_security_tests.rs"]
+mod masking_extended_attack_surface_security_tests;
+
+#[cfg(test)]
+#[path = "tests/masking_padding_timeout_adversarial_tests.rs"]
+mod masking_padding_timeout_adversarial_tests;
+
+#[cfg(all(test, feature = "redteam_offline_expected_fail"))]
+#[path = "tests/masking_offline_target_redteam_expected_fail_tests.rs"]
+mod masking_offline_target_redteam_expected_fail_tests;
+
+#[cfg(test)]
+#[path = "tests/masking_baseline_invariant_tests.rs"]
+mod masking_baseline_invariant_tests;
+
+#[cfg(test)]
+#[path = "tests/masking_lognormal_timing_security_tests.rs"]
+mod masking_lognormal_timing_security_tests;

src/proxy/mod.rs

@@ -4,58 +4,58 @@
 #![cfg_attr(test, allow(warnings))]
 #![cfg_attr(not(test), forbid(clippy::undocumented_unsafe_blocks))]
 #![cfg_attr(
-	not(test),
-	deny(
-		clippy::unwrap_used,
-		clippy::expect_used,
-		clippy::panic,
-		clippy::todo,
-		clippy::unimplemented,
-		clippy::correctness,
-		clippy::option_if_let_else,
-		clippy::or_fun_call,
-		clippy::branches_sharing_code,
-		clippy::single_option_map,
-		clippy::useless_let_if_seq,
-		clippy::redundant_locals,
-		clippy::cloned_ref_to_slice_refs,
-		unsafe_code,
-		clippy::await_holding_lock,
-		clippy::await_holding_refcell_ref,
-		clippy::debug_assert_with_mut_call,
-		clippy::macro_use_imports,
-		clippy::cast_ptr_alignment,
-		clippy::cast_lossless,
-		clippy::ptr_as_ptr,
-		clippy::large_stack_arrays,
-		clippy::same_functions_in_if_condition,
-		trivial_casts,
-		trivial_numeric_casts,
-		unused_extern_crates,
-		unused_import_braces,
-		rust_2018_idioms
-	)
+    not(test),
+    deny(
+        clippy::unwrap_used,
+        clippy::expect_used,
+        clippy::panic,
+        clippy::todo,
+        clippy::unimplemented,
+        clippy::correctness,
+        clippy::option_if_let_else,
+        clippy::or_fun_call,
+        clippy::branches_sharing_code,
+        clippy::single_option_map,
+        clippy::useless_let_if_seq,
+        clippy::redundant_locals,
+        clippy::cloned_ref_to_slice_refs,
+        unsafe_code,
+        clippy::await_holding_lock,
+        clippy::await_holding_refcell_ref,
+        clippy::debug_assert_with_mut_call,
+        clippy::macro_use_imports,
+        clippy::cast_ptr_alignment,
+        clippy::cast_lossless,
+        clippy::ptr_as_ptr,
+        clippy::large_stack_arrays,
+        clippy::same_functions_in_if_condition,
+        trivial_casts,
+        trivial_numeric_casts,
+        unused_extern_crates,
+        unused_import_braces,
+        rust_2018_idioms
+    )
 )]
 #![cfg_attr(
-	not(test),
-	allow(
-		clippy::use_self,
-		clippy::redundant_closure,
-		clippy::too_many_arguments,
-		clippy::doc_markdown,
-		clippy::missing_const_for_fn,
-		clippy::unnecessary_operation,
-		clippy::redundant_pub_crate,
-		clippy::derive_partial_eq_without_eq,
-		clippy::type_complexity,
-		clippy::new_ret_no_self,
-		clippy::cast_possible_truncation,
-		clippy::cast_possible_wrap,
-		clippy::significant_drop_tightening,
-		clippy::significant_drop_in_scrutinee,
-		clippy::float_cmp,
-		clippy::nursery
-	)
+    not(test),
+    allow(
+        clippy::use_self,
+        clippy::redundant_closure,
+        clippy::too_many_arguments,
+        clippy::doc_markdown,
+        clippy::missing_const_for_fn,
+        clippy::unnecessary_operation,
+        clippy::redundant_pub_crate,
+        clippy::derive_partial_eq_without_eq,
+        clippy::type_complexity,
+        clippy::new_ret_no_self,
+        clippy::cast_possible_truncation,
+        clippy::cast_possible_wrap,
+        clippy::significant_drop_tightening,
+        clippy::significant_drop_in_scrutinee,
+        clippy::float_cmp,
+        clippy::nursery
+    )
 )]
 
 pub mod adaptive_buffers;
@@ -67,6 +67,7 @@ pub mod middle_relay;
 pub mod relay;
 pub mod route_mode;
 pub mod session_eviction;
+pub mod shared_state;
 
 pub use client::ClientHandler;
 #[allow(unused_imports)]
@@ -75,3 +76,15 @@ pub use handshake::*;
 pub use masking::*;
 #[allow(unused_imports)]
 pub use relay::*;
+
+#[cfg(test)]
+#[path = "tests/test_harness_common.rs"]
+mod test_harness_common;
+
+#[cfg(test)]
+#[path = "tests/proxy_shared_state_isolation_tests.rs"]
+mod proxy_shared_state_isolation_tests;
+
+#[cfg(test)]
+#[path = "tests/proxy_shared_state_parallel_execution_tests.rs"]
+mod proxy_shared_state_parallel_execution_tests;

src/proxy/relay.rs

@@ -52,13 +52,12 @@
 //! - `SharedCounters` (atomics) let the watchdog read stats without locking
 
 use crate::error::{ProxyError, Result};
-use crate::stats::Stats;
+use crate::stats::{Stats, UserStats};
 use crate::stream::BufferPool;
-use dashmap::DashMap;
 use std::io;
 use std::pin::Pin;
+use std::sync::Arc;
 use std::sync::atomic::{AtomicBool, AtomicU64, Ordering};
-use std::sync::{Arc, Mutex, OnceLock};
 use std::task::{Context, Poll};
 use std::time::Duration;
 use tokio::io::{AsyncRead, AsyncWrite, AsyncWriteExt, ReadBuf, copy_bidirectional_with_sizes};
@@ -71,6 +70,7 @@ use tracing::{debug, trace, warn};
 ///
 /// iOS keeps Telegram connections alive in background for up to 30 minutes.
 /// Closing earlier causes unnecessary reconnects and handshake overhead.
+#[allow(dead_code)]
 const ACTIVITY_TIMEOUT: Duration = Duration::from_secs(1800);
 
 /// Watchdog check interval — also used for periodic rate logging.
@@ -209,12 +209,10 @@ struct StatsIo<S> {
     counters: Arc<SharedCounters>,
     stats: Arc<Stats>,
     user: String,
+    user_stats: Arc<UserStats>,
     quota_limit: Option<u64>,
     quota_exceeded: Arc<AtomicBool>,
-    quota_read_wake_scheduled: bool,
-    quota_write_wake_scheduled: bool,
-    quota_read_retry_active: Arc<AtomicBool>,
-    quota_write_retry_active: Arc<AtomicBool>,
+    quota_bytes_since_check: u64,
     epoch: Instant,
 }
 
@@ -230,30 +228,21 @@ impl<S> StatsIo<S> {
     ) -> Self {
         // Mark initial activity so the watchdog doesn't fire before data flows
         counters.touch(Instant::now(), epoch);
+        let user_stats = stats.get_or_create_user_stats_handle(&user);
         Self {
             inner,
             counters,
             stats,
             user,
+            user_stats,
             quota_limit,
             quota_exceeded,
-            quota_read_wake_scheduled: false,
-            quota_write_wake_scheduled: false,
-            quota_read_retry_active: Arc::new(AtomicBool::new(false)),
-            quota_write_retry_active: Arc::new(AtomicBool::new(false)),
+            quota_bytes_since_check: 0,
             epoch,
         }
     }
 }
 
-impl<S> Drop for StatsIo<S> {
-    fn drop(&mut self) {
-        self.quota_read_retry_active.store(false, Ordering::Relaxed);
-        self.quota_write_retry_active
-            .store(false, Ordering::Relaxed);
-    }
-}
-
 #[derive(Debug)]
 struct QuotaIoSentinel;
 
@@ -277,84 +266,22 @@ fn is_quota_io_error(err: &io::Error) -> bool {
             .is_some()
 }
 
-#[cfg(test)]
-const QUOTA_CONTENTION_RETRY_INTERVAL: Duration = Duration::from_millis(1);
-#[cfg(not(test))]
-const QUOTA_CONTENTION_RETRY_INTERVAL: Duration = Duration::from_millis(2);
+const QUOTA_NEAR_LIMIT_BYTES: u64 = 64 * 1024;
+const QUOTA_LARGE_CHARGE_BYTES: u64 = 16 * 1024;
+const QUOTA_ADAPTIVE_INTERVAL_MIN_BYTES: u64 = 4 * 1024;
+const QUOTA_ADAPTIVE_INTERVAL_MAX_BYTES: u64 = 64 * 1024;
 
-fn spawn_quota_retry_waker(retry_active: Arc<AtomicBool>, waker: std::task::Waker) {
-    tokio::task::spawn(async move {
-        loop {
-            if !retry_active.load(Ordering::Relaxed) {
-                break;
-            }
-            tokio::time::sleep(QUOTA_CONTENTION_RETRY_INTERVAL).await;
-            if !retry_active.load(Ordering::Relaxed) {
-                break;
-            }
-            waker.wake_by_ref();
-        }
-    });
+#[inline]
+fn quota_adaptive_interval_bytes(remaining_before: u64) -> u64 {
+    remaining_before.saturating_div(2).clamp(
+        QUOTA_ADAPTIVE_INTERVAL_MIN_BYTES,
+        QUOTA_ADAPTIVE_INTERVAL_MAX_BYTES,
+    )
 }
 
-static QUOTA_USER_LOCKS: OnceLock<DashMap<String, Arc<Mutex<()>>>> = OnceLock::new();
-static QUOTA_USER_OVERFLOW_LOCKS: OnceLock<Vec<Arc<Mutex<()>>>> = OnceLock::new();
-
-#[cfg(test)]
-const QUOTA_USER_LOCKS_MAX: usize = 64;
-#[cfg(not(test))]
-const QUOTA_USER_LOCKS_MAX: usize = 4_096;
-#[cfg(test)]
-const QUOTA_OVERFLOW_LOCK_STRIPES: usize = 16;
-#[cfg(not(test))]
-const QUOTA_OVERFLOW_LOCK_STRIPES: usize = 256;
-
-#[cfg(test)]
-fn quota_user_lock_test_guard() -> &'static Mutex<()> {
-    static TEST_LOCK: OnceLock<Mutex<()>> = OnceLock::new();
-    TEST_LOCK.get_or_init(|| Mutex::new(()))
-}
-
-#[cfg(test)]
-fn quota_user_lock_test_scope() -> std::sync::MutexGuard<'static, ()> {
-    quota_user_lock_test_guard()
-        .lock()
-        .unwrap_or_else(|poisoned| poisoned.into_inner())
-}
-
-fn quota_overflow_user_lock(user: &str) -> Arc<Mutex<()>> {
-    let stripes = QUOTA_USER_OVERFLOW_LOCKS.get_or_init(|| {
-        (0..QUOTA_OVERFLOW_LOCK_STRIPES)
-            .map(|_| Arc::new(Mutex::new(())))
-            .collect()
-    });
-
-    let hash = crc32fast::hash(user.as_bytes()) as usize;
-    Arc::clone(&stripes[hash % stripes.len()])
-}
-
-fn quota_user_lock(user: &str) -> Arc<Mutex<()>> {
-    let locks = QUOTA_USER_LOCKS.get_or_init(DashMap::new);
-    if let Some(existing) = locks.get(user) {
-        return Arc::clone(existing.value());
-    }
-
-    if locks.len() >= QUOTA_USER_LOCKS_MAX {
-        locks.retain(|_, value| Arc::strong_count(value) > 1);
-    }
-
-    if locks.len() >= QUOTA_USER_LOCKS_MAX {
-        return quota_overflow_user_lock(user);
-    }
-
-    let created = Arc::new(Mutex::new(()));
-    match locks.entry(user.to_string()) {
-        dashmap::mapref::entry::Entry::Occupied(entry) => Arc::clone(entry.get()),
-        dashmap::mapref::entry::Entry::Vacant(entry) => {
-            entry.insert(Arc::clone(&created));
-            created
-        }
-    }
+#[inline]
+fn should_immediate_quota_check(remaining_before: u64, charge_bytes: u64) -> bool {
+    remaining_before <= QUOTA_NEAR_LIMIT_BYTES || charge_bytes >= QUOTA_LARGE_CHARGE_BYTES
 }
 
 impl<S: AsyncRead + Unpin> AsyncRead for StatsIo<S> {
@@ -364,80 +291,60 @@ impl<S: AsyncRead + Unpin> AsyncRead for StatsIo<S> {
         buf: &mut ReadBuf<'_>,
     ) -> Poll<io::Result<()>> {
         let this = self.get_mut();
-        if this.quota_exceeded.load(Ordering::Relaxed) {
+        if this.quota_exceeded.load(Ordering::Acquire) {
             return Poll::Ready(Err(quota_io_error()));
         }
 
-        let quota_lock = this
-            .quota_limit
-            .is_some()
-            .then(|| quota_user_lock(&this.user));
-        let _quota_guard = if let Some(lock) = quota_lock.as_ref() {
-            match lock.try_lock() {
-                Ok(guard) => {
-                    this.quota_read_wake_scheduled = false;
-                    this.quota_read_retry_active.store(false, Ordering::Relaxed);
-                    Some(guard)
-                }
-                Err(_) => {
-                    if !this.quota_read_wake_scheduled {
-                        this.quota_read_wake_scheduled = true;
-                        this.quota_read_retry_active.store(true, Ordering::Relaxed);
-                        spawn_quota_retry_waker(
-                            Arc::clone(&this.quota_read_retry_active),
-                            cx.waker().clone(),
-                        );
-                    }
-                    return Poll::Pending;
-                }
+        let mut remaining_before = None;
+        if let Some(limit) = this.quota_limit {
+            let used_before = this.user_stats.quota_used();
+            let remaining = limit.saturating_sub(used_before);
+            if remaining == 0 {
+                this.quota_exceeded.store(true, Ordering::Release);
+                return Poll::Ready(Err(quota_io_error()));
             }
-        } else {
-            None
-        };
-
-        if let Some(limit) = this.quota_limit
-            && this.stats.get_user_total_octets(&this.user) >= limit
-        {
-            this.quota_exceeded.store(true, Ordering::Relaxed);
-            return Poll::Ready(Err(quota_io_error()));
+            remaining_before = Some(remaining);
         }
+
         let before = buf.filled().len();
 
         match Pin::new(&mut this.inner).poll_read(cx, buf) {
             Poll::Ready(Ok(())) => {
                 let n = buf.filled().len() - before;
                 if n > 0 {
-                    let mut reached_quota_boundary = false;
-                    if let Some(limit) = this.quota_limit {
-                        let used = this.stats.get_user_total_octets(&this.user);
-                        if used >= limit {
-                            this.quota_exceeded.store(true, Ordering::Relaxed);
-                            return Poll::Ready(Err(quota_io_error()));
-                        }
-
-                        let remaining = limit - used;
-                        if (n as u64) > remaining {
-                            // Fail closed: when a single read chunk would cross quota,
-                            // stop relay immediately without accounting beyond the cap.
-                            this.quota_exceeded.store(true, Ordering::Relaxed);
-                            return Poll::Ready(Err(quota_io_error()));
-                        }
-
-                        reached_quota_boundary = (n as u64) == remaining;
-                    }
+                    let n_to_charge = n as u64;
 
                     // C→S: client sent data
                     this.counters
                         .c2s_bytes
-                        .fetch_add(n as u64, Ordering::Relaxed);
+                        .fetch_add(n_to_charge, Ordering::Relaxed);
                     this.counters.c2s_ops.fetch_add(1, Ordering::Relaxed);
                     this.counters.touch(Instant::now(), this.epoch);
 
-                    this.stats.add_user_octets_from(&this.user, n as u64);
-                    this.stats.increment_user_msgs_from(&this.user);
+                    this.stats
+                        .add_user_octets_from_handle(this.user_stats.as_ref(), n_to_charge);
+                    this.stats
+                        .increment_user_msgs_from_handle(this.user_stats.as_ref());
 
-                    if reached_quota_boundary {
-                        this.quota_exceeded.store(true, Ordering::Relaxed);
+                    if let (Some(limit), Some(remaining)) = (this.quota_limit, remaining_before) {
+                        this.stats
+                            .quota_charge_post_write(this.user_stats.as_ref(), n_to_charge);
+                        if should_immediate_quota_check(remaining, n_to_charge) {
+                            this.quota_bytes_since_check = 0;
+                            if this.user_stats.quota_used() >= limit {
+                                this.quota_exceeded.store(true, Ordering::Release);
+                            }
+                        } else {
+                            this.quota_bytes_since_check =
+                                this.quota_bytes_since_check.saturating_add(n_to_charge);
+                            let interval = quota_adaptive_interval_bytes(remaining);
+                            if this.quota_bytes_since_check >= interval {
+                                this.quota_bytes_since_check = 0;
+                                if this.user_stats.quota_used() >= limit {
+                                    this.quota_exceeded.store(true, Ordering::Release);
+                                }
+                            }
+                        }
                     }
 
                     trace!(user = %this.user, bytes = n, "C->S");
@@ -456,75 +363,57 @@ impl<S: AsyncWrite + Unpin> AsyncWrite for StatsIo<S> {
         buf: &[u8],
     ) -> Poll<io::Result<usize>> {
         let this = self.get_mut();
-        if this.quota_exceeded.load(Ordering::Relaxed) {
+        if this.quota_exceeded.load(Ordering::Acquire) {
             return Poll::Ready(Err(quota_io_error()));
         }
 
-        let quota_lock = this
-            .quota_limit
-            .is_some()
-            .then(|| quota_user_lock(&this.user));
-        let _quota_guard = if let Some(lock) = quota_lock.as_ref() {
-            match lock.try_lock() {
-                Ok(guard) => {
-                    this.quota_write_wake_scheduled = false;
-                    this.quota_write_retry_active
-                        .store(false, Ordering::Relaxed);
-                    Some(guard)
-                }
-                Err(_) => {
-                    if !this.quota_write_wake_scheduled {
-                        this.quota_write_wake_scheduled = true;
-                        this.quota_write_retry_active.store(true, Ordering::Relaxed);
-                        spawn_quota_retry_waker(
-                            Arc::clone(&this.quota_write_retry_active),
-                            cx.waker().clone(),
-                        );
-                    }
-                    return Poll::Pending;
-                }
-            }
-        } else {
-            None
-        };
-
-        let write_buf = if let Some(limit) = this.quota_limit {
-            let used = this.stats.get_user_total_octets(&this.user);
-            if used >= limit {
-                this.quota_exceeded.store(true, Ordering::Relaxed);
+        let mut remaining_before = None;
+        if let Some(limit) = this.quota_limit {
+            let used_before = this.user_stats.quota_used();
+            let remaining = limit.saturating_sub(used_before);
+            if remaining == 0 {
+                this.quota_exceeded.store(true, Ordering::Release);
                 return Poll::Ready(Err(quota_io_error()));
             }
+            remaining_before = Some(remaining);
+        }
 
-            let remaining = (limit - used) as usize;
-            if buf.len() > remaining {
-                // Fail closed: do not emit partial S->C payload when remaining
-                // quota cannot accommodate the pending write request.
-                this.quota_exceeded.store(true, Ordering::Relaxed);
-                return Poll::Ready(Err(quota_io_error()));
-            }
-            buf
-        } else {
-            buf
-        };
-
-        match Pin::new(&mut this.inner).poll_write(cx, write_buf) {
+        match Pin::new(&mut this.inner).poll_write(cx, buf) {
             Poll::Ready(Ok(n)) => {
                 if n > 0 {
+                    let n_to_charge = n as u64;
+
                     // S→C: data written to client
                     this.counters
                         .s2c_bytes
-                        .fetch_add(n as u64, Ordering::Relaxed);
+                        .fetch_add(n_to_charge, Ordering::Relaxed);
                     this.counters.s2c_ops.fetch_add(1, Ordering::Relaxed);
                     this.counters.touch(Instant::now(), this.epoch);
 
-                    this.stats.add_user_octets_to(&this.user, n as u64);
-                    this.stats.increment_user_msgs_to(&this.user);
+                    this.stats
+                        .add_user_octets_to_handle(this.user_stats.as_ref(), n_to_charge);
+                    this.stats
+                        .increment_user_msgs_to_handle(this.user_stats.as_ref());
 
-                    if let Some(limit) = this.quota_limit
-                        && this.stats.get_user_total_octets(&this.user) >= limit
-                    {
-                        this.quota_exceeded.store(true, Ordering::Relaxed);
-                        return Poll::Ready(Err(quota_io_error()));
+                    if let (Some(limit), Some(remaining)) = (this.quota_limit, remaining_before) {
+                        this.stats
+                            .quota_charge_post_write(this.user_stats.as_ref(), n_to_charge);
+                        if should_immediate_quota_check(remaining, n_to_charge) {
+                            this.quota_bytes_since_check = 0;
+                            if this.user_stats.quota_used() >= limit {
+                                this.quota_exceeded.store(true, Ordering::Release);
+                            }
+                        } else {
+                            this.quota_bytes_since_check =
+                                this.quota_bytes_since_check.saturating_add(n_to_charge);
+                            let interval = quota_adaptive_interval_bytes(remaining);
+                            if this.quota_bytes_since_check >= interval {
+                                this.quota_bytes_since_check = 0;
+                                if this.user_stats.quota_used() >= limit {
+                                    this.quota_exceeded.store(true, Ordering::Release);
+                                }
+                            }
+                        }
                     }
 
                     trace!(user = %this.user, bytes = n, "S->C");
@@ -565,6 +454,7 @@ impl<S: AsyncWrite + Unpin> AsyncWrite for StatsIo<S> {
 /// - Clean shutdown: both write sides are shut down on exit
 /// - Error propagation: quota exits return `ProxyError::DataQuotaExceeded`,
 ///   other I/O failures are returned as `ProxyError::Io`
+#[allow(dead_code)]
 pub async fn relay_bidirectional<CR, CW, SR, SW>(
     client_reader: CR,
     client_writer: CW,
@@ -583,6 +473,42 @@ where
     SR: AsyncRead + Unpin + Send + 'static,
     SW: AsyncWrite + Unpin + Send + 'static,
 {
+    relay_bidirectional_with_activity_timeout(
+        client_reader,
+        client_writer,
+        server_reader,
+        server_writer,
+        c2s_buf_size,
+        s2c_buf_size,
+        user,
+        stats,
+        quota_limit,
+        _buffer_pool,
+        ACTIVITY_TIMEOUT,
+    )
+    .await
+}
+
+pub async fn relay_bidirectional_with_activity_timeout<CR, CW, SR, SW>(
+    client_reader: CR,
+    client_writer: CW,
+    server_reader: SR,
+    server_writer: SW,
+    c2s_buf_size: usize,
+    s2c_buf_size: usize,
+    user: &str,
+    stats: Arc<Stats>,
+    quota_limit: Option<u64>,
+    _buffer_pool: Arc<BufferPool>,
+    activity_timeout: Duration,
+) -> Result<()>
+where
+    CR: AsyncRead + Unpin + Send + 'static,
+    CW: AsyncWrite + Unpin + Send + 'static,
+    SR: AsyncRead + Unpin + Send + 'static,
+    SW: AsyncWrite + Unpin + Send + 'static,
+{
+    let activity_timeout = activity_timeout.max(Duration::from_secs(1));
     let epoch = Instant::now();
     let counters = Arc::new(SharedCounters::new());
     let quota_exceeded = Arc::new(AtomicBool::new(false));
@@ -618,13 +544,13 @@ where
             let now = Instant::now();
             let idle = wd_counters.idle_duration(now, epoch);
 
-            if wd_quota_exceeded.load(Ordering::Relaxed) {
+            if wd_quota_exceeded.load(Ordering::Acquire) {
                 warn!(user = %wd_user, "User data quota reached, closing relay");
                 return;
             }
 
             // ── Activity timeout ────────────────────────────────────
-            if idle >= ACTIVITY_TIMEOUT {
+            if idle >= activity_timeout {
                 let c2s = wd_counters.c2s_bytes.load(Ordering::Relaxed);
                 let s2c = wd_counters.s2c_bytes.load(Ordering::Relaxed);
                 warn!(
@@ -756,18 +682,10 @@ where
     }
 }
 
-#[cfg(test)]
-#[path = "tests/relay_security_tests.rs"]
-mod security_tests;
-
 #[cfg(test)]
 #[path = "tests/relay_adversarial_tests.rs"]
 mod adversarial_tests;
 
-#[cfg(test)]
-#[path = "tests/relay_quota_lock_pressure_adversarial_tests.rs"]
-mod relay_quota_lock_pressure_adversarial_tests;
-
 #[cfg(test)]
 #[path = "tests/relay_quota_boundary_blackhat_tests.rs"]
 mod relay_quota_boundary_blackhat_tests;
@@ -780,14 +698,18 @@ mod relay_quota_model_adversarial_tests;
 #[path = "tests/relay_quota_overflow_regression_tests.rs"]
 mod relay_quota_overflow_regression_tests;
 
+#[cfg(test)]
+#[path = "tests/relay_quota_extended_attack_surface_security_tests.rs"]
+mod relay_quota_extended_attack_surface_security_tests;
+
 #[cfg(test)]
 #[path = "tests/relay_watchdog_delta_security_tests.rs"]
 mod relay_watchdog_delta_security_tests;
 
 #[cfg(test)]
-#[path = "tests/relay_quota_waker_storm_adversarial_tests.rs"]
-mod relay_quota_waker_storm_adversarial_tests;
+#[path = "tests/relay_atomic_quota_invariant_tests.rs"]
+mod relay_atomic_quota_invariant_tests;
 
 #[cfg(test)]
-#[path = "tests/relay_quota_wake_liveness_regression_tests.rs"]
-mod relay_quota_wake_liveness_regression_tests;
+#[path = "tests/relay_baseline_invariant_tests.rs"]
+mod relay_baseline_invariant_tests;

src/proxy/shared_state.rs

@@ -0,0 +1,146 @@
+use std::collections::HashSet;
+use std::collections::hash_map::RandomState;
+use std::net::{IpAddr, SocketAddr};
+use std::sync::atomic::{AtomicBool, AtomicU64, Ordering};
+use std::sync::{Arc, Mutex};
+use std::time::Instant;
+
+use dashmap::DashMap;
+use tokio::sync::mpsc;
+
+use crate::proxy::handshake::{AuthProbeSaturationState, AuthProbeState};
+use crate::proxy::middle_relay::{DesyncDedupRotationState, RelayIdleCandidateRegistry};
+
+#[derive(Debug, Clone, Copy, PartialEq, Eq)]
+pub(crate) enum ConntrackCloseReason {
+    NormalEof,
+    Timeout,
+    Pressure,
+    Reset,
+    Other,
+}
+
+#[derive(Debug, Clone, Copy)]
+pub(crate) struct ConntrackCloseEvent {
+    pub(crate) src: SocketAddr,
+    pub(crate) dst: SocketAddr,
+    pub(crate) reason: ConntrackCloseReason,
+}
+
+#[derive(Debug, Clone, Copy, PartialEq, Eq)]
+pub(crate) enum ConntrackClosePublishResult {
+    Sent,
+    Disabled,
+    QueueFull,
+    QueueClosed,
+}
+
+pub(crate) struct HandshakeSharedState {
+    pub(crate) auth_probe: DashMap<IpAddr, AuthProbeState>,
+    pub(crate) auth_probe_saturation: Mutex<Option<AuthProbeSaturationState>>,
+    pub(crate) auth_probe_eviction_hasher: RandomState,
+    pub(crate) invalid_secret_warned: Mutex<HashSet<(String, String)>>,
+    pub(crate) unknown_sni_warn_next_allowed: Mutex<Option<Instant>>,
+}
+
+pub(crate) struct MiddleRelaySharedState {
+    pub(crate) desync_dedup: DashMap<u64, Instant>,
+    pub(crate) desync_dedup_previous: DashMap<u64, Instant>,
+    pub(crate) desync_hasher: RandomState,
+    pub(crate) desync_full_cache_last_emit_at: Mutex<Option<Instant>>,
+    pub(crate) desync_dedup_rotation_state: Mutex<DesyncDedupRotationState>,
+    pub(crate) relay_idle_registry: Mutex<RelayIdleCandidateRegistry>,
+    pub(crate) relay_idle_mark_seq: AtomicU64,
+}
+
+pub(crate) struct ProxySharedState {
+    pub(crate) handshake: HandshakeSharedState,
+    pub(crate) middle_relay: MiddleRelaySharedState,
+    pub(crate) conntrack_pressure_active: AtomicBool,
+    pub(crate) conntrack_close_tx: Mutex<Option<mpsc::Sender<ConntrackCloseEvent>>>,
+}
+
+impl ProxySharedState {
+    pub(crate) fn new() -> Arc<Self> {
+        Arc::new(Self {
+            handshake: HandshakeSharedState {
+                auth_probe: DashMap::new(),
+                auth_probe_saturation: Mutex::new(None),
+                auth_probe_eviction_hasher: RandomState::new(),
+                invalid_secret_warned: Mutex::new(HashSet::new()),
+                unknown_sni_warn_next_allowed: Mutex::new(None),
+            },
+            middle_relay: MiddleRelaySharedState {
+                desync_dedup: DashMap::new(),
+                desync_dedup_previous: DashMap::new(),
+                desync_hasher: RandomState::new(),
+                desync_full_cache_last_emit_at: Mutex::new(None),
+                desync_dedup_rotation_state: Mutex::new(DesyncDedupRotationState::default()),
+                relay_idle_registry: Mutex::new(RelayIdleCandidateRegistry::default()),
+                relay_idle_mark_seq: AtomicU64::new(0),
+            },
+            conntrack_pressure_active: AtomicBool::new(false),
+            conntrack_close_tx: Mutex::new(None),
+        })
+    }
+
+    pub(crate) fn set_conntrack_close_sender(&self, tx: mpsc::Sender<ConntrackCloseEvent>) {
+        match self.conntrack_close_tx.lock() {
+            Ok(mut guard) => {
+                *guard = Some(tx);
+            }
+            Err(poisoned) => {
+                let mut guard = poisoned.into_inner();
+                *guard = Some(tx);
+                self.conntrack_close_tx.clear_poison();
+            }
+        }
+    }
+
+    pub(crate) fn disable_conntrack_close_sender(&self) {
+        match self.conntrack_close_tx.lock() {
+            Ok(mut guard) => {
+                *guard = None;
+            }
+            Err(poisoned) => {
+                let mut guard = poisoned.into_inner();
+                *guard = None;
+                self.conntrack_close_tx.clear_poison();
+            }
+        }
+    }
+
+    pub(crate) fn publish_conntrack_close_event(
+        &self,
+        event: ConntrackCloseEvent,
+    ) -> ConntrackClosePublishResult {
+        let tx = match self.conntrack_close_tx.lock() {
+            Ok(guard) => guard.clone(),
+            Err(poisoned) => {
+                let guard = poisoned.into_inner();
+                let cloned = guard.clone();
+                self.conntrack_close_tx.clear_poison();
+                cloned
+            }
+        };
+
+        let Some(tx) = tx else {
+            return ConntrackClosePublishResult::Disabled;
+        };
+
+        match tx.try_send(event) {
+            Ok(()) => ConntrackClosePublishResult::Sent,
+            Err(mpsc::error::TrySendError::Full(_)) => ConntrackClosePublishResult::QueueFull,
+            Err(mpsc::error::TrySendError::Closed(_)) => ConntrackClosePublishResult::QueueClosed,
+        }
+    }
+
+    pub(crate) fn set_conntrack_pressure_active(&self, active: bool) {
+        self.conntrack_pressure_active
+            .store(active, Ordering::Relaxed);
+    }
+
+    pub(crate) fn conntrack_pressure_active(&self) -> bool {
+        self.conntrack_pressure_active.load(Ordering::Relaxed)
+    }
+}

src/proxy/tests/adaptive_buffers_record_race_security_tests.rs

@@ -0,0 +1,260 @@
+use super::*;
+use std::sync::Arc;
+use std::sync::atomic::{AtomicUsize, Ordering};
+use std::time::{Duration, Instant};
+
+static RACE_TEST_KEY_COUNTER: AtomicUsize = AtomicUsize::new(1_000_000);
+
+fn race_unique_key(prefix: &str) -> String {
+    let id = RACE_TEST_KEY_COUNTER.fetch_add(1, Ordering::Relaxed);
+    format!("{}_{}", prefix, id)
+}
+
+// ── TOCTOU race: concurrent record_user_tier can downgrade tier ─────────
+// Two threads call record_user_tier for the same NEW user simultaneously.
+// Thread A records Tier1, Thread B records Base. Without atomic entry API,
+// the insert() call overwrites without max(), causing Tier1 → Base downgrade.
+
+#[test]
+fn adaptive_record_concurrent_insert_no_tier_downgrade() {
+    // Run multiple rounds to increase race detection probability.
+    for round in 0..50 {
+        let key = race_unique_key(&format!("race_downgrade_{}", round));
+        let key_a = key.clone();
+        let key_b = key.clone();
+
+        let barrier = Arc::new(std::sync::Barrier::new(2));
+        let barrier_a = Arc::clone(&barrier);
+        let barrier_b = Arc::clone(&barrier);
+
+        let ha = std::thread::spawn(move || {
+            barrier_a.wait();
+            record_user_tier(&key_a, AdaptiveTier::Tier2);
+        });
+
+        let hb = std::thread::spawn(move || {
+            barrier_b.wait();
+            record_user_tier(&key_b, AdaptiveTier::Base);
+        });
+
+        ha.join().expect("thread A panicked");
+        hb.join().expect("thread B panicked");
+
+        let result = seed_tier_for_user(&key);
+        profiles().remove(&key);
+
+        // The final tier must be at least Tier2, never downgraded to Base.
+        // With correct max() semantics: max(Tier2, Base) = Tier2.
+        assert!(
+            result >= AdaptiveTier::Tier2,
+            "Round {}: concurrent insert downgraded tier from Tier2 to {:?}",
+            round,
+            result,
+        );
+    }
+}
+
+// ── TOCTOU race: three threads write three tiers, highest must survive ──
+
+#[test]
+fn adaptive_record_triple_concurrent_insert_highest_tier_survives() {
+    for round in 0..30 {
+        let key = race_unique_key(&format!("triple_race_{}", round));
+        let barrier = Arc::new(std::sync::Barrier::new(3));
+
+        let handles: Vec<_> = [AdaptiveTier::Base, AdaptiveTier::Tier1, AdaptiveTier::Tier3]
+            .into_iter()
+            .map(|tier| {
+                let k = key.clone();
+                let b = Arc::clone(&barrier);
+                std::thread::spawn(move || {
+                    b.wait();
+                    record_user_tier(&k, tier);
+                })
+            })
+            .collect();
+
+        for h in handles {
+            h.join().expect("thread panicked");
+        }
+
+        let result = seed_tier_for_user(&key);
+        profiles().remove(&key);
+
+        assert!(
+            result >= AdaptiveTier::Tier3,
+            "Round {}: triple concurrent insert didn't preserve Tier3, got {:?}",
+            round,
+            result,
+        );
+    }
+}
+
+// ── Stress: 20 threads writing different tiers to same key ──────────────
+
+#[test]
+fn adaptive_record_20_concurrent_writers_no_panic_no_downgrade() {
+    let key = race_unique_key("stress_20");
+    let barrier = Arc::new(std::sync::Barrier::new(20));
+
+    let handles: Vec<_> = (0..20u32)
+        .map(|i| {
+            let k = key.clone();
+            let b = Arc::clone(&barrier);
+            std::thread::spawn(move || {
+                b.wait();
+                let tier = match i % 4 {
+                    0 => AdaptiveTier::Base,
+                    1 => AdaptiveTier::Tier1,
+                    2 => AdaptiveTier::Tier2,
+                    _ => AdaptiveTier::Tier3,
+                };
+                for _ in 0..100 {
+                    record_user_tier(&k, tier);
+                }
+            })
+        })
+        .collect();
+
+    for h in handles {
+        h.join().expect("thread panicked");
+    }
+
+    let result = seed_tier_for_user(&key);
+    profiles().remove(&key);
+
+    // At least one thread writes Tier3, max() should preserve it
+    assert!(
+        result >= AdaptiveTier::Tier3,
+        "20 concurrent writers: expected at least Tier3, got {:?}",
+        result,
+    );
+}
+
+// ── TOCTOU: seed reads stale, concurrent record inserts fresh ───────────
+// Verifies remove_if predicate preserves fresh insertions.
+
+#[test]
+fn adaptive_seed_and_record_race_preserves_fresh_entry() {
+    for round in 0..30 {
+        let key = race_unique_key(&format!("seed_record_race_{}", round));
+
+        // Plant a stale entry
+        let stale_time = Instant::now() - Duration::from_secs(600);
+        profiles().insert(
+            key.clone(),
+            UserAdaptiveProfile {
+                tier: AdaptiveTier::Tier1,
+                seen_at: stale_time,
+            },
+        );
+
+        let key_seed = key.clone();
+        let key_record = key.clone();
+        let barrier = Arc::new(std::sync::Barrier::new(2));
+        let barrier_s = Arc::clone(&barrier);
+        let barrier_r = Arc::clone(&barrier);
+
+        let h_seed = std::thread::spawn(move || {
+            barrier_s.wait();
+            seed_tier_for_user(&key_seed)
+        });
+
+        let h_record = std::thread::spawn(move || {
+            barrier_r.wait();
+            record_user_tier(&key_record, AdaptiveTier::Tier3);
+        });
+
+        let _seed_result = h_seed.join().expect("seed thread panicked");
+        h_record.join().expect("record thread panicked");
+
+        let final_result = seed_tier_for_user(&key);
+        profiles().remove(&key);
+
+        // Fresh Tier3 entry should survive the stale-removal race.
+        // Due to non-deterministic scheduling, the outcome depends on ordering:
+        // - If record wins: Tier3 is present, seed returns Tier3
+        // - If seed wins: stale entry removed, then record inserts Tier3
+        // Either way, Tier3 should be visible after both complete.
+        assert!(
+            final_result == AdaptiveTier::Tier3 || final_result == AdaptiveTier::Base,
+            "Round {}: unexpected tier after seed+record race: {:?}",
+            round,
+            final_result,
+        );
+    }
+}
+
+// ── Eviction safety: retain() during concurrent inserts ─────────────────
+
+#[test]
+fn adaptive_eviction_during_concurrent_inserts_no_panic() {
+    let prefix = race_unique_key("evict_conc");
+    let stale_time = Instant::now() - Duration::from_secs(600);
+
+    // Pre-fill with stale entries to push past the eviction threshold
+    for i in 0..100 {
+        let k = format!("{}_{}", prefix, i);
+        profiles().insert(
+            k,
+            UserAdaptiveProfile {
+                tier: AdaptiveTier::Base,
+                seen_at: stale_time,
+            },
+        );
+    }
+
+    let barrier = Arc::new(std::sync::Barrier::new(10));
+    let handles: Vec<_> = (0..10)
+        .map(|t| {
+            let b = Arc::clone(&barrier);
+            let pfx = prefix.clone();
+            std::thread::spawn(move || {
+                b.wait();
+                for i in 0..50 {
+                    let k = format!("{}_t{}_{}", pfx, t, i);
+                    record_user_tier(&k, AdaptiveTier::Tier1);
+                }
+            })
+        })
+        .collect();
+
+    for h in handles {
+        h.join().expect("eviction thread panicked");
+    }
+
+    // Cleanup
+    profiles().retain(|k, _| !k.starts_with(&prefix));
+}
+
+// ── Adversarial: attacker races insert+seed in tight loop ───────────────
+
+#[test]
+fn adaptive_tight_loop_insert_seed_race_no_panic() {
+    let key = race_unique_key("tight_loop");
+    let key_w = key.clone();
+    let key_r = key.clone();
+
+    let done = Arc::new(std::sync::atomic::AtomicBool::new(false));
+    let done_w = Arc::clone(&done);
+    let done_r = Arc::clone(&done);
+
+    let writer = std::thread::spawn(move || {
+        while !done_w.load(Ordering::Relaxed) {
+            record_user_tier(&key_w, AdaptiveTier::Tier2);
+        }
+    });
+
+    let reader = std::thread::spawn(move || {
+        while !done_r.load(Ordering::Relaxed) {
+            let _ = seed_tier_for_user(&key_r);
+        }
+    });
+
+    std::thread::sleep(Duration::from_millis(100));
+    done.store(true, Ordering::Relaxed);
+
+    writer.join().expect("writer panicked");
+    reader.join().expect("reader panicked");
+    profiles().remove(&key);
+}

src/proxy/tests/adaptive_buffers_security_tests.rs

@@ -0,0 +1,453 @@
+use super::*;
+use std::sync::atomic::{AtomicUsize, Ordering};
+use std::time::{Duration, Instant};
+
+// Unique key generator to avoid test interference through the global DashMap.
+static TEST_KEY_COUNTER: AtomicUsize = AtomicUsize::new(0);
+
+fn unique_key(prefix: &str) -> String {
+    let id = TEST_KEY_COUNTER.fetch_add(1, Ordering::Relaxed);
+    format!("{}_{}", prefix, id)
+}
+
+// ── Positive / Lifecycle ────────────────────────────────────────────────
+
+#[test]
+fn adaptive_seed_unknown_user_returns_base() {
+    let key = unique_key("seed_unknown");
+    assert_eq!(seed_tier_for_user(&key), AdaptiveTier::Base);
+}
+
+#[test]
+fn adaptive_record_then_seed_returns_recorded_tier() {
+    let key = unique_key("record_seed");
+    record_user_tier(&key, AdaptiveTier::Tier1);
+    assert_eq!(seed_tier_for_user(&key), AdaptiveTier::Tier1);
+}
+
+#[test]
+fn adaptive_separate_users_have_independent_tiers() {
+    let key_a = unique_key("indep_a");
+    let key_b = unique_key("indep_b");
+    record_user_tier(&key_a, AdaptiveTier::Tier1);
+    record_user_tier(&key_b, AdaptiveTier::Tier2);
+    assert_eq!(seed_tier_for_user(&key_a), AdaptiveTier::Tier1);
+    assert_eq!(seed_tier_for_user(&key_b), AdaptiveTier::Tier2);
+}
+
+#[test]
+fn adaptive_record_upgrades_tier_within_ttl() {
+    let key = unique_key("upgrade");
+    record_user_tier(&key, AdaptiveTier::Base);
+    record_user_tier(&key, AdaptiveTier::Tier1);
+    assert_eq!(seed_tier_for_user(&key), AdaptiveTier::Tier1);
+}
+
+#[test]
+fn adaptive_record_does_not_downgrade_within_ttl() {
+    let key = unique_key("no_downgrade");
+    record_user_tier(&key, AdaptiveTier::Tier2);
+    record_user_tier(&key, AdaptiveTier::Base);
+    // max(Tier2, Base) = Tier2 — within TTL the higher tier is retained
+    assert_eq!(seed_tier_for_user(&key), AdaptiveTier::Tier2);
+}
+
+// ── Edge Cases ──────────────────────────────────────────────────────────
+
+#[test]
+fn adaptive_base_tier_buffers_unchanged() {
+    let (c2s, s2c) = direct_copy_buffers_for_tier(AdaptiveTier::Base, 65536, 262144);
+    assert_eq!(c2s, 65536);
+    assert_eq!(s2c, 262144);
+}
+
+#[test]
+fn adaptive_tier1_buffers_within_caps() {
+    let (c2s, s2c) = direct_copy_buffers_for_tier(AdaptiveTier::Tier1, 65536, 262144);
+    assert!(c2s > 65536, "Tier1 c2s should exceed Base");
+    assert!(
+        c2s <= 128 * 1024,
+        "Tier1 c2s should not exceed DIRECT_C2S_CAP_BYTES"
+    );
+    assert!(s2c > 262144, "Tier1 s2c should exceed Base");
+    assert!(
+        s2c <= 512 * 1024,
+        "Tier1 s2c should not exceed DIRECT_S2C_CAP_BYTES"
+    );
+}
+
+#[test]
+fn adaptive_tier3_buffers_capped() {
+    let (c2s, s2c) = direct_copy_buffers_for_tier(AdaptiveTier::Tier3, 65536, 262144);
+    assert!(c2s <= 128 * 1024, "Tier3 c2s must not exceed cap");
+    assert!(s2c <= 512 * 1024, "Tier3 s2c must not exceed cap");
+}
+
+#[test]
+fn adaptive_scale_zero_base_returns_at_least_one() {
+    // scale(0, num, den, cap) should return at least 1 (the .max(1) guard)
+    let (c2s, s2c) = direct_copy_buffers_for_tier(AdaptiveTier::Tier1, 0, 0);
+    assert!(c2s >= 1);
+    assert!(s2c >= 1);
+}
+
+// ── Stale Entry Handling ────────────────────────────────────────────────
+
+#[test]
+fn adaptive_stale_profile_returns_base_tier() {
+    let key = unique_key("stale_base");
+    // Manually insert a stale entry with seen_at in the far past.
+    // PROFILE_TTL = 300s, so 600s ago is well past expiry.
+    let stale_time = Instant::now() - Duration::from_secs(600);
+    profiles().insert(
+        key.clone(),
+        UserAdaptiveProfile {
+            tier: AdaptiveTier::Tier3,
+            seen_at: stale_time,
+        },
+    );
+    assert_eq!(
+        seed_tier_for_user(&key),
+        AdaptiveTier::Base,
+        "Stale profile should return Base"
+    );
+}
+
+// RED TEST: exposes the stale entry leak bug.
+// After seed_tier_for_user returns Base for a stale entry, the entry should be
+// removed from the cache. Currently it is NOT removed — stale entries accumulate
+// indefinitely, consuming memory.
+#[test]
+fn adaptive_stale_entry_removed_after_seed() {
+    let key = unique_key("stale_removal");
+    let stale_time = Instant::now() - Duration::from_secs(600);
+    profiles().insert(
+        key.clone(),
+        UserAdaptiveProfile {
+            tier: AdaptiveTier::Tier2,
+            seen_at: stale_time,
+        },
+    );
+    let _ = seed_tier_for_user(&key);
+    // After seeding, the stale entry should have been removed.
+    assert!(
+        !profiles().contains_key(&key),
+        "Stale entry should be removed from cache after seed_tier_for_user"
+    );
+}
+
+// ── Cardinality Attack / Unbounded Growth ───────────────────────────────
+
+// RED TEST: exposes the missing eviction cap.
+// An attacker who can trigger record_user_tier with arbitrary user keys can
+// grow the global DashMap without bound, exhausting server memory.
+// After inserting MAX_USER_PROFILES_ENTRIES + 1 stale entries, record_user_tier
+// must trigger retain()-based eviction that purges all stale entries.
+#[test]
+fn adaptive_profile_cache_bounded_under_cardinality_attack() {
+    let prefix = unique_key("cardinality");
+    let stale_time = Instant::now() - Duration::from_secs(600);
+    let n = MAX_USER_PROFILES_ENTRIES + 1;
+    for i in 0..n {
+        let key = format!("{}_{}", prefix, i);
+        profiles().insert(
+            key,
+            UserAdaptiveProfile {
+                tier: AdaptiveTier::Base,
+                seen_at: stale_time,
+            },
+        );
+    }
+    // This insert should push the cache over MAX_USER_PROFILES_ENTRIES and trigger eviction.
+    let trigger_key = unique_key("cardinality_trigger");
+    record_user_tier(&trigger_key, AdaptiveTier::Base);
+
+    // Count surviving stale entries.
+    let mut surviving_stale = 0;
+    for i in 0..n {
+        let key = format!("{}_{}", prefix, i);
+        if profiles().contains_key(&key) {
+            surviving_stale += 1;
+        }
+    }
+    // Cleanup: remove anything that survived + the trigger key.
+    for i in 0..n {
+        let key = format!("{}_{}", prefix, i);
+        profiles().remove(&key);
+    }
+    profiles().remove(&trigger_key);
+
+    // All stale entries (600s past PROFILE_TTL=300s) should have been evicted.
+    assert_eq!(
+        surviving_stale, 0,
+        "All {} stale entries should be evicted, but {} survived",
+        n, surviving_stale
+    );
+}
+
+// ── Key Length Validation ────────────────────────────────────────────────
+
+// RED TEST: exposes missing key length validation.
+// An attacker can submit arbitrarily large user keys, each consuming memory
+// for the String allocation in the DashMap key.
+#[test]
+fn adaptive_oversized_user_key_rejected_on_record() {
+    let oversized_key: String = "X".repeat(1024); // 1KB key — should be rejected
+    record_user_tier(&oversized_key, AdaptiveTier::Tier1);
+    // With key length validation, the oversized key should NOT be stored.
+    let stored = profiles().contains_key(&oversized_key);
+    // Cleanup regardless
+    profiles().remove(&oversized_key);
+    assert!(
+        !stored,
+        "Oversized user key (1024 bytes) should be rejected by record_user_tier"
+    );
+}
+
+#[test]
+fn adaptive_oversized_user_key_rejected_on_seed() {
+    let oversized_key: String = "X".repeat(1024);
+    // Insert it directly to test seed behavior
+    profiles().insert(
+        oversized_key.clone(),
+        UserAdaptiveProfile {
+            tier: AdaptiveTier::Tier3,
+            seen_at: Instant::now(),
+        },
+    );
+    let result = seed_tier_for_user(&oversized_key);
+    profiles().remove(&oversized_key);
+    assert_eq!(
+        result,
+        AdaptiveTier::Base,
+        "Oversized user key should return Base from seed_tier_for_user"
+    );
+}
+
+#[test]
+fn adaptive_empty_user_key_safe() {
+    // Empty string is a valid (if unusual) key — should not panic
+    record_user_tier("", AdaptiveTier::Tier1);
+    let tier = seed_tier_for_user("");
+    profiles().remove("");
+    assert_eq!(tier, AdaptiveTier::Tier1);
+}
+
+#[test]
+fn adaptive_max_length_key_accepted() {
+    // A key at exactly 512 bytes should be accepted
+    let key: String = "K".repeat(512);
+    record_user_tier(&key, AdaptiveTier::Tier1);
+    let tier = seed_tier_for_user(&key);
+    profiles().remove(&key);
+    assert_eq!(tier, AdaptiveTier::Tier1);
+}
+
+// ── Concurrent Access Safety ────────────────────────────────────────────
+
+#[test]
+fn adaptive_concurrent_record_and_seed_no_torn_read() {
+    let key = unique_key("concurrent_rw");
+    let key_clone = key.clone();
+
+    // Record from multiple threads simultaneously
+    let handles: Vec<_> = (0..10)
+        .map(|i| {
+            let k = key_clone.clone();
+            std::thread::spawn(move || {
+                let tier = if i % 2 == 0 {
+                    AdaptiveTier::Tier1
+                } else {
+                    AdaptiveTier::Tier2
+                };
+                record_user_tier(&k, tier);
+            })
+        })
+        .collect();
+
+    for h in handles {
+        h.join().expect("thread panicked");
+    }
+
+    let result = seed_tier_for_user(&key);
+    profiles().remove(&key);
+    // Result must be one of the recorded tiers, not a corrupted value
+    assert!(
+        result == AdaptiveTier::Tier1 || result == AdaptiveTier::Tier2,
+        "Concurrent writes produced unexpected tier: {:?}",
+        result
+    );
+}
+
+#[test]
+fn adaptive_concurrent_seed_does_not_panic() {
+    let key = unique_key("concurrent_seed");
+    record_user_tier(&key, AdaptiveTier::Tier1);
+    let key_clone = key.clone();
+
+    let handles: Vec<_> = (0..20)
+        .map(|_| {
+            let k = key_clone.clone();
+            std::thread::spawn(move || {
+                for _ in 0..100 {
+                    let _ = seed_tier_for_user(&k);
+                }
+            })
+        })
+        .collect();
+
+    for h in handles {
+        h.join().expect("concurrent seed panicked");
+    }
+    profiles().remove(&key);
+}
+
+// ── TOCTOU: Concurrent seed + record race ───────────────────────────────
+
+// RED TEST: seed_tier_for_user reads a stale entry, drops the reference,
+// then another thread inserts a fresh entry. If seed then removes unconditionally
+// (without atomic predicate), the fresh entry is lost. With remove_if, the
+// fresh entry survives.
+#[test]
+fn adaptive_remove_if_does_not_delete_fresh_concurrent_insert() {
+    let key = unique_key("toctou");
+    let stale_time = Instant::now() - Duration::from_secs(600);
+    profiles().insert(
+        key.clone(),
+        UserAdaptiveProfile {
+            tier: AdaptiveTier::Tier1,
+            seen_at: stale_time,
+        },
+    );
+
+    // Thread A: seed_tier (will see stale, should attempt removal)
+    // Thread B: record_user_tier (inserts fresh entry concurrently)
+    let key_a = key.clone();
+    let key_b = key.clone();
+
+    let handle_b = std::thread::spawn(move || {
+        // Small yield to increase chance of interleaving
+        std::thread::yield_now();
+        record_user_tier(&key_b, AdaptiveTier::Tier3);
+    });
+
+    let _ = seed_tier_for_user(&key_a);
+
+    handle_b.join().expect("thread B panicked");
+
+    // After both operations, the fresh Tier3 entry should survive.
+    // With a correct remove_if predicate, the fresh entry is NOT deleted.
+    // Without remove_if (current code), the entry may be lost.
+    let final_tier = seed_tier_for_user(&key);
+    profiles().remove(&key);
+
+    // The fresh Tier3 entry should survive the stale-removal race.
+    // Note: Due to non-deterministic scheduling, this test may pass even
+    // without the fix if thread B wins the race. Run with --test-threads=1
+    // or multiple iterations for reliable detection.
+    assert!(
+        final_tier == AdaptiveTier::Tier3 || final_tier == AdaptiveTier::Base,
+        "Unexpected tier after TOCTOU race: {:?}",
+        final_tier
+    );
+}
+
+// ── Fuzz: Random keys ──────────────────────────────────────────────────
+
+#[test]
+fn adaptive_fuzz_random_keys_no_panic() {
+    use rand::{Rng, RngExt};
+    let mut rng = rand::rng();
+    let mut keys = Vec::new();
+    for _ in 0..200 {
+        let len: usize = rng.random_range(0..=256);
+        let key: String = (0..len)
+            .map(|_| {
+                let c: u8 = rng.random_range(0x20..=0x7E);
+                c as char
+            })
+            .collect();
+        record_user_tier(&key, AdaptiveTier::Tier1);
+        let _ = seed_tier_for_user(&key);
+        keys.push(key);
+    }
+    // Cleanup
+    for key in &keys {
+        profiles().remove(key);
+    }
+}
+
+// ── average_throughput_to_tier (proposed function, tests the mapping) ────
+
+// These tests verify the function that will be added in PR-D.
+// They are written against the current code's constant definitions.
+
+#[test]
+fn adaptive_throughput_mapping_below_threshold_is_base() {
+    // 7 Mbps < 8 Mbps threshold → Base
+    // 7 Mbps = 7_000_000 bps = 875_000 bytes/s over 10s = 8_750_000 bytes
+    // max(c2s, s2c) determines direction
+    let c2s_bytes: u64 = 8_750_000;
+    let s2c_bytes: u64 = 1_000_000;
+    let duration_secs: f64 = 10.0;
+    let avg_bps = (c2s_bytes.max(s2c_bytes) as f64 * 8.0) / duration_secs;
+    // 8_750_000 * 8 / 10 = 7_000_000 bps = 7 Mbps → Base
+    assert!(
+        avg_bps < THROUGHPUT_UP_BPS,
+        "Should be below threshold: {} < {}",
+        avg_bps,
+        THROUGHPUT_UP_BPS,
+    );
+}
+
+#[test]
+fn adaptive_throughput_mapping_above_threshold_is_tier1() {
+    // 10 Mbps > 8 Mbps threshold → Tier1
+    let bytes_10mbps_10s: u64 = 12_500_000; // 10 Mbps * 10s / 8 = 12_500_000 bytes
+    let duration_secs: f64 = 10.0;
+    let avg_bps = (bytes_10mbps_10s as f64 * 8.0) / duration_secs;
+    assert!(
+        avg_bps >= THROUGHPUT_UP_BPS,
+        "Should be above threshold: {} >= {}",
+        avg_bps,
+        THROUGHPUT_UP_BPS,
+    );
+}
+
+#[test]
+fn adaptive_throughput_short_session_should_return_base() {
+    // Sessions shorter than 1 second should not promote (too little data to judge)
+    let duration_secs: f64 = 0.5;
+    // Even with high throughput, short sessions should return Base
+    assert!(
+        duration_secs < 1.0,
+        "Short session duration guard should activate"
+    );
+}
+
+// ── me_flush_policy_for_tier ────────────────────────────────────────────
+
+#[test]
+fn adaptive_me_flush_base_unchanged() {
+    let (frames, bytes, delay) =
+        me_flush_policy_for_tier(AdaptiveTier::Base, 32, 65536, Duration::from_micros(1000));
+    assert_eq!(frames, 32);
+    assert_eq!(bytes, 65536);
+    assert_eq!(delay, Duration::from_micros(1000));
+}
+
+#[test]
+fn adaptive_me_flush_tier1_delay_reduced() {
+    let (_, _, delay) =
+        me_flush_policy_for_tier(AdaptiveTier::Tier1, 32, 65536, Duration::from_micros(1000));
+    // Tier1: delay * 7/10 = 700 µs
+    assert_eq!(delay, Duration::from_micros(700));
+}
+
+#[test]
+fn adaptive_me_flush_delay_never_below_minimum() {
+    let (_, _, delay) =
+        me_flush_policy_for_tier(AdaptiveTier::Tier3, 32, 65536, Duration::from_micros(200));
+    // Tier3: 200 * 3/10 = 60, but min is ME_DELAY_MIN_US = 150
+    assert!(delay.as_micros() >= 150, "Delay must respect minimum");
+}

src/proxy/tests/client_clever_advanced_tests.rs

@@ -0,0 +1,473 @@
+use super::*;
+use crate::config::{ProxyConfig, UpstreamConfig, UpstreamType};
+use crate::protocol::constants::{MAX_TLS_PLAINTEXT_SIZE, MIN_TLS_CLIENT_HELLO_SIZE};
+use crate::stats::Stats;
+use crate::transport::UpstreamManager;
+use std::net::{IpAddr, Ipv4Addr, Ipv6Addr, SocketAddr};
+use std::pin::Pin;
+use std::sync::Arc;
+use std::task::{Context, Poll};
+use std::time::Duration;
+use tokio::io::{AsyncRead, AsyncReadExt, AsyncWriteExt, ReadBuf, duplex};
+use tokio::net::TcpListener;
+
+#[test]
+fn edge_mask_reject_delay_min_greater_than_max_does_not_panic() {
+    let mut config = ProxyConfig::default();
+    config.censorship.server_hello_delay_min_ms = 5000;
+    config.censorship.server_hello_delay_max_ms = 1000;
+
+    let rt = tokio::runtime::Runtime::new().unwrap();
+    rt.block_on(async {
+        let start = std::time::Instant::now();
+        maybe_apply_mask_reject_delay(&config).await;
+        let elapsed = start.elapsed();
+
+        assert!(elapsed >= Duration::from_millis(1000));
+        assert!(elapsed < Duration::from_millis(1500));
+    });
+}
+
+#[test]
+fn edge_handshake_timeout_with_mask_grace_saturating_add_prevents_overflow() {
+    let mut config = ProxyConfig::default();
+    config.timeouts.client_handshake = u64::MAX;
+    config.censorship.mask = true;
+
+    let timeout = handshake_timeout_with_mask_grace(&config);
+    assert_eq!(timeout.as_secs(), u64::MAX);
+}
+
+#[test]
+fn edge_tls_clienthello_len_in_bounds_exact_boundaries() {
+    assert!(tls_clienthello_len_in_bounds(MIN_TLS_CLIENT_HELLO_SIZE));
+    assert!(!tls_clienthello_len_in_bounds(
+        MIN_TLS_CLIENT_HELLO_SIZE - 1
+    ));
+    assert!(tls_clienthello_len_in_bounds(MAX_TLS_PLAINTEXT_SIZE));
+    assert!(!tls_clienthello_len_in_bounds(MAX_TLS_PLAINTEXT_SIZE + 1));
+}
+
+#[test]
+fn edge_synthetic_local_addr_boundaries() {
+    assert_eq!(synthetic_local_addr(0).port(), 0);
+    assert_eq!(synthetic_local_addr(80).port(), 80);
+    assert_eq!(synthetic_local_addr(u16::MAX).port(), u16::MAX);
+}
+
+#[test]
+fn edge_beobachten_record_handshake_failure_class_stream_error_eof() {
+    let beobachten = BeobachtenStore::new();
+    let mut config = ProxyConfig::default();
+    config.general.beobachten = true;
+    config.general.beobachten_minutes = 1;
+
+    let eof_err = ProxyError::Stream(crate::error::StreamError::UnexpectedEof);
+    let peer_ip: IpAddr = "198.51.100.100".parse().unwrap();
+
+    record_handshake_failure_class(&beobachten, &config, peer_ip, &eof_err);
+
+    let snapshot = beobachten.snapshot_text(Duration::from_secs(60));
+    assert!(snapshot.contains("[expected_64_got_0]"));
+}
+
+#[tokio::test]
+async fn adversarial_tls_handshake_timeout_during_masking_delay() {
+    let mut cfg = ProxyConfig::default();
+    cfg.general.beobachten = false;
+    cfg.timeouts.client_handshake = 1;
+    cfg.censorship.mask = true;
+    cfg.censorship.server_hello_delay_min_ms = 3000;
+    cfg.censorship.server_hello_delay_max_ms = 3000;
+
+    let config = Arc::new(cfg);
+    let stats = Arc::new(Stats::new());
+    let (server_side, mut client_side) = duplex(4096);
+
+    let handle = tokio::spawn(handle_client_stream(
+        server_side,
+        "198.51.100.1:55000".parse().unwrap(),
+        config,
+        stats.clone(),
+        Arc::new(UpstreamManager::new(
+            vec![],
+            1,
+            1,
+            1,
+            10,
+            1,
+            false,
+            stats.clone(),
+        )),
+        Arc::new(ReplayChecker::new(128, Duration::from_secs(60))),
+        Arc::new(BufferPool::new()),
+        Arc::new(SecureRandom::new()),
+        None,
+        Arc::new(RouteRuntimeController::new(RelayRouteMode::Direct)),
+        None,
+        Arc::new(UserIpTracker::new()),
+        Arc::new(BeobachtenStore::new()),
+        false,
+    ));
+
+    client_side
+        .write_all(&[0x16, 0x03, 0x01, 0xFF, 0xFF])
+        .await
+        .unwrap();
+
+    let result = tokio::time::timeout(Duration::from_secs(4), handle)
+        .await
+        .unwrap()
+        .unwrap();
+
+    assert!(matches!(result, Err(ProxyError::TgHandshakeTimeout)));
+    assert_eq!(stats.get_handshake_timeouts(), 1);
+}
+
+#[tokio::test]
+async fn blackhat_proxy_protocol_slowloris_timeout() {
+    let mut cfg = ProxyConfig::default();
+    cfg.server.proxy_protocol_header_timeout_ms = 200;
+    let config = Arc::new(cfg);
+    let stats = Arc::new(Stats::new());
+
+    let (server_side, mut client_side) = duplex(4096);
+    let handle = tokio::spawn(handle_client_stream(
+        server_side,
+        "198.51.100.2:55000".parse().unwrap(),
+        config,
+        stats.clone(),
+        Arc::new(UpstreamManager::new(
+            vec![],
+            1,
+            1,
+            1,
+            10,
+            1,
+            false,
+            stats.clone(),
+        )),
+        Arc::new(ReplayChecker::new(128, Duration::from_secs(60))),
+        Arc::new(BufferPool::new()),
+        Arc::new(SecureRandom::new()),
+        None,
+        Arc::new(RouteRuntimeController::new(RelayRouteMode::Direct)),
+        None,
+        Arc::new(UserIpTracker::new()),
+        Arc::new(BeobachtenStore::new()),
+        true,
+    ));
+
+    client_side.write_all(b"PROXY TCP4 192.").await.unwrap();
+    tokio::time::sleep(Duration::from_millis(300)).await;
+
+    let result = tokio::time::timeout(Duration::from_secs(2), handle)
+        .await
+        .unwrap()
+        .unwrap();
+
+    assert!(matches!(result, Err(ProxyError::InvalidProxyProtocol)));
+    assert_eq!(stats.get_connects_bad(), 1);
+}
+
+#[test]
+fn blackhat_ipv4_mapped_ipv6_proxy_source_bypass_attempt() {
+    let trusted = vec!["192.0.2.0/24".parse().unwrap()];
+    let peer_ip = IpAddr::V6(Ipv6Addr::new(0, 0, 0, 0, 0, 0xffff, 0xc000, 0x0201));
+    assert!(!is_trusted_proxy_source(peer_ip, &trusted));
+}
+
+#[tokio::test]
+async fn negative_proxy_protocol_enabled_but_client_sends_tls_hello() {
+    let mut cfg = ProxyConfig::default();
+    cfg.server.proxy_protocol_header_timeout_ms = 500;
+    let config = Arc::new(cfg);
+    let stats = Arc::new(Stats::new());
+
+    let (server_side, mut client_side) = duplex(4096);
+    let handle = tokio::spawn(handle_client_stream(
+        server_side,
+        "198.51.100.3:55000".parse().unwrap(),
+        config,
+        stats.clone(),
+        Arc::new(UpstreamManager::new(
+            vec![],
+            1,
+            1,
+            1,
+            10,
+            1,
+            false,
+            stats.clone(),
+        )),
+        Arc::new(ReplayChecker::new(128, Duration::from_secs(60))),
+        Arc::new(BufferPool::new()),
+        Arc::new(SecureRandom::new()),
+        None,
+        Arc::new(RouteRuntimeController::new(RelayRouteMode::Direct)),
+        None,
+        Arc::new(UserIpTracker::new()),
+        Arc::new(BeobachtenStore::new()),
+        true,
+    ));
+
+    client_side
+        .write_all(&[0x16, 0x03, 0x01, 0x02, 0x00])
+        .await
+        .unwrap();
+
+    let result = tokio::time::timeout(Duration::from_secs(2), handle)
+        .await
+        .unwrap()
+        .unwrap();
+
+    assert!(matches!(result, Err(ProxyError::InvalidProxyProtocol)));
+    assert_eq!(stats.get_connects_bad(), 1);
+}
+
+#[tokio::test]
+async fn edge_client_stream_exactly_4_bytes_eof() {
+    let config = Arc::new(ProxyConfig::default());
+    let stats = Arc::new(Stats::new());
+    let beobachten = Arc::new(BeobachtenStore::new());
+
+    let (server_side, mut client_side) = duplex(4096);
+    let handle = tokio::spawn(handle_client_stream(
+        server_side,
+        "198.51.100.4:55000".parse().unwrap(),
+        config,
+        stats.clone(),
+        Arc::new(UpstreamManager::new(
+            vec![],
+            1,
+            1,
+            1,
+            10,
+            1,
+            false,
+            stats.clone(),
+        )),
+        Arc::new(ReplayChecker::new(128, Duration::from_secs(60))),
+        Arc::new(BufferPool::new()),
+        Arc::new(SecureRandom::new()),
+        None,
+        Arc::new(RouteRuntimeController::new(RelayRouteMode::Direct)),
+        None,
+        Arc::new(UserIpTracker::new()),
+        beobachten.clone(),
+        false,
+    ));
+
+    client_side
+        .write_all(&[0x16, 0x03, 0x01, 0x00])
+        .await
+        .unwrap();
+    client_side.shutdown().await.unwrap();
+
+    let _ = tokio::time::timeout(Duration::from_secs(2), handle).await;
+
+    let snapshot = beobachten.snapshot_text(Duration::from_secs(60));
+    assert!(snapshot.contains("[expected_64_got_0]"));
+}
+
+#[tokio::test]
+async fn edge_client_stream_tls_header_valid_but_body_1_byte_short_eof() {
+    let config = Arc::new(ProxyConfig::default());
+    let stats = Arc::new(Stats::new());
+
+    let (server_side, mut client_side) = duplex(4096);
+    let handle = tokio::spawn(handle_client_stream(
+        server_side,
+        "198.51.100.5:55000".parse().unwrap(),
+        config,
+        stats.clone(),
+        Arc::new(UpstreamManager::new(
+            vec![],
+            1,
+            1,
+            1,
+            10,
+            1,
+            false,
+            stats.clone(),
+        )),
+        Arc::new(ReplayChecker::new(128, Duration::from_secs(60))),
+        Arc::new(BufferPool::new()),
+        Arc::new(SecureRandom::new()),
+        None,
+        Arc::new(RouteRuntimeController::new(RelayRouteMode::Direct)),
+        None,
+        Arc::new(UserIpTracker::new()),
+        Arc::new(BeobachtenStore::new()),
+        false,
+    ));
+
+    client_side
+        .write_all(&[0x16, 0x03, 0x01, 0x00, 100])
+        .await
+        .unwrap();
+    client_side.write_all(&vec![0x41; 99]).await.unwrap();
+    client_side.shutdown().await.unwrap();
+
+    let _ = tokio::time::timeout(Duration::from_secs(2), handle).await;
+    assert_eq!(stats.get_connects_bad(), 1);
+}
+
+#[tokio::test]
+async fn integration_non_tls_modes_disabled_immediately_masks() {
+    let mut cfg = ProxyConfig::default();
+    cfg.general.modes.classic = false;
+    cfg.general.modes.secure = false;
+    cfg.censorship.mask = true;
+    let config = Arc::new(cfg);
+    let stats = Arc::new(Stats::new());
+
+    let (server_side, mut client_side) = duplex(4096);
+    let handle = tokio::spawn(handle_client_stream(
+        server_side,
+        "198.51.100.6:55000".parse().unwrap(),
+        config,
+        stats.clone(),
+        Arc::new(UpstreamManager::new(
+            vec![],
+            1,
+            1,
+            1,
+            10,
+            1,
+            false,
+            stats.clone(),
+        )),
+        Arc::new(ReplayChecker::new(128, Duration::from_secs(60))),
+        Arc::new(BufferPool::new()),
+        Arc::new(SecureRandom::new()),
+        None,
+        Arc::new(RouteRuntimeController::new(RelayRouteMode::Direct)),
+        None,
+        Arc::new(UserIpTracker::new()),
+        Arc::new(BeobachtenStore::new()),
+        false,
+    ));
+
+    client_side.write_all(b"GET / HTTP/1.1\r\n").await.unwrap();
+    let _ = tokio::time::timeout(Duration::from_secs(2), handle).await;
+    assert_eq!(stats.get_connects_bad(), 1);
+}
+
+struct YieldingReader {
+    data: Vec<u8>,
+    pos: usize,
+    yields_left: usize,
+}
+
+impl AsyncRead for YieldingReader {
+    fn poll_read(
+        self: Pin<&mut Self>,
+        cx: &mut Context<'_>,
+        buf: &mut ReadBuf<'_>,
+    ) -> Poll<std::io::Result<()>> {
+        let this = self.get_mut();
+        if this.yields_left > 0 {
+            this.yields_left -= 1;
+            cx.waker().wake_by_ref();
+            return Poll::Pending;
+        }
+        if this.pos >= this.data.len() {
+            return Poll::Ready(Ok(()));
+        }
+        buf.put_slice(&this.data[this.pos..this.pos + 1]);
+        this.pos += 1;
+        this.yields_left = 2;
+        Poll::Ready(Ok(()))
+    }
+}
+
+#[tokio::test]
+async fn fuzz_read_with_progress_heavy_yielding() {
+    let expected_data = b"HEAVY_YIELD_TEST_DATA".to_vec();
+    let mut reader = YieldingReader {
+        data: expected_data.clone(),
+        pos: 0,
+        yields_left: 2,
+    };
+
+    let mut buf = vec![0u8; expected_data.len()];
+    let read_bytes = read_with_progress(&mut reader, &mut buf).await.unwrap();
+
+    assert_eq!(read_bytes, expected_data.len());
+    assert_eq!(buf, expected_data);
+}
+
+#[test]
+fn edge_wrap_tls_application_record_exactly_u16_max() {
+    let payload = vec![0u8; 65535];
+    let wrapped = wrap_tls_application_record(&payload);
+    assert_eq!(wrapped.len(), 65540);
+    assert_eq!(wrapped[0], TLS_RECORD_APPLICATION);
+    assert_eq!(&wrapped[3..5], &65535u16.to_be_bytes());
+}
+
+#[test]
+fn fuzz_wrap_tls_application_record_lengths() {
+    let lengths = [0, 1, 65534, 65535, 65536, 131070, 131071, 131072];
+    for len in lengths {
+        let payload = vec![0u8; len];
+        let wrapped = wrap_tls_application_record(&payload);
+        let expected_chunks = len.div_ceil(65535).max(1);
+        assert_eq!(wrapped.len(), len + 5 * expected_chunks);
+    }
+}
+
+#[tokio::test]
+async fn stress_user_connection_reservation_concurrent_same_ip_exhaustion() {
+    let user = "stress-same-ip-user";
+    let mut config = ProxyConfig::default();
+    config.access.user_max_tcp_conns.insert(user.to_string(), 5);
+
+    let config = Arc::new(config);
+    let stats = Arc::new(Stats::new());
+    let ip_tracker = Arc::new(UserIpTracker::new());
+    ip_tracker.set_user_limit(user, 10).await;
+
+    let peer = SocketAddr::new(IpAddr::V4(Ipv4Addr::new(198, 51, 100, 77)), 55000);
+
+    let mut tasks = tokio::task::JoinSet::new();
+    let mut reservations = Vec::new();
+
+    for _ in 0..10 {
+        let config = config.clone();
+        let stats = stats.clone();
+        let ip_tracker = ip_tracker.clone();
+        tasks.spawn(async move {
+            RunningClientHandler::acquire_user_connection_reservation_static(
+                user, &config, stats, peer, ip_tracker,
+            )
+            .await
+        });
+    }
+
+    let mut successes = 0;
+    let mut failures = 0;
+
+    while let Some(res) = tasks.join_next().await {
+        match res.unwrap() {
+            Ok(r) => {
+                successes += 1;
+                reservations.push(r);
+            }
+            Err(_) => failures += 1,
+        }
+    }
+
+    assert_eq!(successes, 5);
+    assert_eq!(failures, 5);
+    assert_eq!(stats.get_user_curr_connects(user), 5);
+    assert_eq!(ip_tracker.get_active_ip_count(user).await, 1);
+
+    for reservation in reservations {
+        reservation.release().await;
+    }
+
+    assert_eq!(stats.get_user_curr_connects(user), 0);
+    assert_eq!(ip_tracker.get_active_ip_count(user).await, 0);
+}

src/proxy/tests/client_deep_invariants_tests.rs

@@ -0,0 +1,224 @@
+use super::*;
+use crate::config::ProxyConfig;
+use crate::protocol::constants::MIN_TLS_CLIENT_HELLO_SIZE;
+use crate::stats::Stats;
+use crate::transport::UpstreamManager;
+use std::sync::Arc;
+use std::time::Duration;
+use tokio::io::{AsyncWriteExt, duplex};
+
+fn preload_user_quota(stats: &Stats, user: &str, bytes: u64) {
+    let user_stats = stats.get_or_create_user_stats_handle(user);
+    stats.quota_charge_post_write(user_stats.as_ref(), bytes);
+}
+
+#[test]
+fn invariant_wrap_tls_application_record_exact_multiples() {
+    let chunk_size = u16::MAX as usize;
+    let payload = vec![0xAA; chunk_size * 2];
+
+    let wrapped = wrap_tls_application_record(&payload);
+
+    assert_eq!(wrapped.len(), 2 * (5 + chunk_size));
+    assert_eq!(wrapped[0], TLS_RECORD_APPLICATION);
+    assert_eq!(&wrapped[3..5], &65535u16.to_be_bytes());
+
+    let second_header_idx = 5 + chunk_size;
+    assert_eq!(wrapped[second_header_idx], TLS_RECORD_APPLICATION);
+    assert_eq!(
+        &wrapped[second_header_idx + 3..second_header_idx + 5],
+        &65535u16.to_be_bytes()
+    );
+}
+
+#[tokio::test]
+async fn invariant_tls_clienthello_truncation_exact_boundary_triggers_masking() {
+    let config = Arc::new(ProxyConfig::default());
+    let stats = Arc::new(Stats::new());
+
+    let (server_side, mut client_side) = duplex(4096);
+    let handler = tokio::spawn(handle_client_stream(
+        server_side,
+        "198.51.100.20:55000".parse().unwrap(),
+        config,
+        stats.clone(),
+        Arc::new(UpstreamManager::new(
+            vec![],
+            1,
+            1,
+            1,
+            10,
+            1,
+            false,
+            stats.clone(),
+        )),
+        Arc::new(ReplayChecker::new(128, Duration::from_secs(60))),
+        Arc::new(BufferPool::new()),
+        Arc::new(SecureRandom::new()),
+        None,
+        Arc::new(RouteRuntimeController::new(RelayRouteMode::Direct)),
+        None,
+        Arc::new(UserIpTracker::new()),
+        Arc::new(BeobachtenStore::new()),
+        false,
+    ));
+
+    let claimed_len = MIN_TLS_CLIENT_HELLO_SIZE as u16;
+    let mut header = vec![0x16, 0x03, 0x01];
+    header.extend_from_slice(&claimed_len.to_be_bytes());
+
+    client_side.write_all(&header).await.unwrap();
+    client_side
+        .write_all(&vec![0x42; MIN_TLS_CLIENT_HELLO_SIZE - 1])
+        .await
+        .unwrap();
+    client_side.shutdown().await.unwrap();
+
+    let _ = tokio::time::timeout(Duration::from_secs(2), handler)
+        .await
+        .unwrap();
+    assert_eq!(stats.get_connects_bad(), 1);
+}
+
+#[tokio::test]
+async fn invariant_acquire_reservation_ip_limit_rollback() {
+    let user = "rollback-test-user";
+    let mut config = ProxyConfig::default();
+    config
+        .access
+        .user_max_tcp_conns
+        .insert(user.to_string(), 10);
+
+    let stats = Arc::new(Stats::new());
+    let ip_tracker = Arc::new(UserIpTracker::new());
+    ip_tracker.set_user_limit(user, 1).await;
+
+    let peer_a = "198.51.100.21:55000".parse().unwrap();
+    let _res_a = RunningClientHandler::acquire_user_connection_reservation_static(
+        user,
+        &config,
+        stats.clone(),
+        peer_a,
+        ip_tracker.clone(),
+    )
+    .await
+    .unwrap();
+
+    assert_eq!(stats.get_user_curr_connects(user), 1);
+
+    let peer_b = "203.0.113.22:55000".parse().unwrap();
+    let res_b = RunningClientHandler::acquire_user_connection_reservation_static(
+        user,
+        &config,
+        stats.clone(),
+        peer_b,
+        ip_tracker.clone(),
+    )
+    .await;
+
+    assert!(matches!(
+        res_b,
+        Err(ProxyError::ConnectionLimitExceeded { .. })
+    ));
+    assert_eq!(stats.get_user_curr_connects(user), 1);
+}
+
+#[tokio::test]
+async fn invariant_quota_exact_boundary_inclusive() {
+    let user = "quota-strict-user";
+    let mut config = ProxyConfig::default();
+    config.access.user_data_quota.insert(user.to_string(), 1000);
+
+    let stats = Arc::new(Stats::new());
+    let ip_tracker = Arc::new(UserIpTracker::new());
+    let peer = "198.51.100.23:55000".parse().unwrap();
+
+    preload_user_quota(stats.as_ref(), user, 999);
+    let res1 = RunningClientHandler::acquire_user_connection_reservation_static(
+        user,
+        &config,
+        stats.clone(),
+        peer,
+        ip_tracker.clone(),
+    )
+    .await;
+    assert!(res1.is_ok());
+    res1.unwrap().release().await;
+
+    preload_user_quota(stats.as_ref(), user, 1);
+    let res2 = RunningClientHandler::acquire_user_connection_reservation_static(
+        user,
+        &config,
+        stats.clone(),
+        peer,
+        ip_tracker.clone(),
+    )
+    .await;
+    assert!(matches!(res2, Err(ProxyError::DataQuotaExceeded { .. })));
+}
+
+#[tokio::test]
+async fn invariant_direct_mode_partial_header_eof_is_error_not_bad_connect() {
+    let mut cfg = ProxyConfig::default();
+    cfg.general.beobachten = true;
+    cfg.general.beobachten_minutes = 1;
+
+    let config = Arc::new(cfg);
+    let stats = Arc::new(Stats::new());
+    let beobachten = Arc::new(BeobachtenStore::new());
+
+    let (server_side, mut client_side) = duplex(4096);
+    let handler = tokio::spawn(handle_client_stream(
+        server_side,
+        "198.51.100.25:55000".parse().unwrap(),
+        config,
+        stats.clone(),
+        Arc::new(UpstreamManager::new(
+            vec![],
+            1,
+            1,
+            1,
+            10,
+            1,
+            false,
+            stats.clone(),
+        )),
+        Arc::new(ReplayChecker::new(128, Duration::from_secs(60))),
+        Arc::new(BufferPool::new()),
+        Arc::new(SecureRandom::new()),
+        None,
+        Arc::new(RouteRuntimeController::new(RelayRouteMode::Direct)),
+        None,
+        Arc::new(UserIpTracker::new()),
+        beobachten.clone(),
+        false,
+    ));
+
+    client_side.write_all(&[0xEF, 0xEF, 0xEF]).await.unwrap();
+    client_side.shutdown().await.unwrap();
+
+    let result = tokio::time::timeout(Duration::from_secs(2), handler)
+        .await
+        .unwrap()
+        .unwrap();
+
+    assert!(result.is_err());
+    assert_eq!(stats.get_connects_bad(), 0);
+    let snapshot = beobachten.snapshot_text(Duration::from_secs(60));
+    assert!(snapshot.contains("[expected_64_got_0]"));
+}
+
+#[tokio::test]
+async fn invariant_route_mode_snapshot_picks_up_latest_mode() {
+    let route_runtime = Arc::new(RouteRuntimeController::new(RelayRouteMode::Direct));
+    assert!(matches!(
+        route_runtime.snapshot().mode,
+        RelayRouteMode::Direct
+    ));
+
+    route_runtime.set_mode(RelayRouteMode::Middle);
+    assert!(matches!(
+        route_runtime.snapshot().mode,
+        RelayRouteMode::Middle
+    ));
+}

src/proxy/tests/client_masking_blackhat_campaign_tests.rs

@@ -40,6 +40,7 @@ fn new_upstream_manager(stats: Arc<Stats>) -> Arc<UpstreamManager> {
         1,
         1,
         1,
+        10,
         1,
         false,
         stats,

src/proxy/tests/client_masking_budget_security_tests.rs

@@ -36,6 +36,7 @@ fn build_harness(config: ProxyConfig) -> PipelineHarness {
         1,
         1,
         1,
+        10,
         1,
         false,
         stats.clone(),

src/proxy/tests/client_masking_diagnostics_security_tests.rs

@@ -20,6 +20,7 @@ fn new_upstream_manager(stats: Arc<Stats>) -> Arc<UpstreamManager> {
         1,
         1,
         1,
+        10,
         1,
         false,
         stats,

src/proxy/tests/client_masking_fragmented_classifier_security_tests.rs

@@ -0,0 +1,101 @@
+use super::*;
+use crate::config::{UpstreamConfig, UpstreamType};
+use std::sync::Arc;
+use tokio::io::{AsyncReadExt, AsyncWriteExt, duplex};
+use tokio::net::TcpListener;
+use tokio::time::Duration;
+
+fn new_upstream_manager(stats: Arc<Stats>) -> Arc<UpstreamManager> {
+    Arc::new(UpstreamManager::new(
+        vec![UpstreamConfig {
+            upstream_type: UpstreamType::Direct {
+                interface: None,
+                bind_addresses: None,
+            },
+            weight: 1,
+            enabled: true,
+            scopes: String::new(),
+            selected_scope: String::new(),
+        }],
+        1,
+        1,
+        1,
+        10,
+        1,
+        false,
+        stats,
+    ))
+}
+
+#[tokio::test]
+async fn fragmented_connect_probe_is_classified_as_http_via_prefetch_window() {
+    let listener = TcpListener::bind("127.0.0.1:0").await.unwrap();
+    let backend_addr = listener.local_addr().unwrap();
+
+    let accept_task = tokio::spawn(async move {
+        let (mut stream, _) = listener.accept().await.unwrap();
+        let mut got = Vec::new();
+        stream.read_to_end(&mut got).await.unwrap();
+        got
+    });
+
+    let mut cfg = ProxyConfig::default();
+    cfg.general.beobachten = true;
+    cfg.general.beobachten_minutes = 1;
+    cfg.censorship.mask = true;
+    cfg.censorship.mask_unix_sock = None;
+    cfg.censorship.mask_host = Some("127.0.0.1".to_string());
+    cfg.censorship.mask_port = backend_addr.port();
+    cfg.general.modes.classic = false;
+    cfg.general.modes.secure = false;
+
+    let config = Arc::new(cfg);
+    let stats = Arc::new(Stats::new());
+    let beobachten = Arc::new(BeobachtenStore::new());
+
+    let (server_side, mut client_side) = duplex(4096);
+    let peer: SocketAddr = "198.51.100.251:57501".parse().unwrap();
+
+    let handler = tokio::spawn(handle_client_stream(
+        server_side,
+        peer,
+        config,
+        stats.clone(),
+        new_upstream_manager(stats),
+        Arc::new(ReplayChecker::new(128, Duration::from_secs(60))),
+        Arc::new(BufferPool::new()),
+        Arc::new(SecureRandom::new()),
+        None,
+        Arc::new(RouteRuntimeController::new(RelayRouteMode::Direct)),
+        None,
+        Arc::new(UserIpTracker::new()),
+        beobachten.clone(),
+        false,
+    ));
+
+    client_side.write_all(b"CONNE").await.unwrap();
+    client_side
+        .write_all(b"CT example.org:443 HTTP/1.1\r\nHost: example.org\r\n\r\n")
+        .await
+        .unwrap();
+    client_side.shutdown().await.unwrap();
+
+    let forwarded = tokio::time::timeout(Duration::from_secs(3), accept_task)
+        .await
+        .unwrap()
+        .unwrap();
+    assert!(
+        forwarded.starts_with(b"CONNECT example.org:443 HTTP/1.1"),
+        "mask backend must receive the full fragmented CONNECT probe"
+    );
+
+    let result = tokio::time::timeout(Duration::from_secs(3), handler)
+        .await
+        .unwrap()
+        .unwrap();
+    assert!(result.is_ok());
+
+    let snapshot = beobachten.snapshot_text(Duration::from_secs(60));
+    assert!(snapshot.contains("[HTTP]"));
+    assert!(snapshot.contains("198.51.100.251-1"));
+}

src/proxy/tests/client_masking_hard_adversarial_tests.rs

@@ -34,6 +34,7 @@ fn new_upstream_manager(stats: Arc<Stats>) -> Arc<UpstreamManager> {
         1,
         1,
         1,
+        10,
         1,
         false,
         stats,

src/proxy/tests/client_masking_http2_fragmented_preface_security_tests.rs

@@ -0,0 +1,123 @@
+use super::*;
+use crate::config::{UpstreamConfig, UpstreamType};
+use std::sync::Arc;
+use tokio::io::{AsyncReadExt, AsyncWriteExt, duplex};
+use tokio::net::TcpListener;
+use tokio::time::{Duration, sleep};
+
+fn new_upstream_manager(stats: Arc<Stats>) -> Arc<UpstreamManager> {
+    Arc::new(UpstreamManager::new(
+        vec![UpstreamConfig {
+            upstream_type: UpstreamType::Direct {
+                interface: None,
+                bind_addresses: None,
+            },
+            weight: 1,
+            enabled: true,
+            scopes: String::new(),
+            selected_scope: String::new(),
+        }],
+        1,
+        1,
+        1,
+        10,
+        1,
+        false,
+        stats,
+    ))
+}
+
+async fn run_http2_fragment_case(split_at: usize, delay_ms: u64, peer: SocketAddr) {
+    let listener = TcpListener::bind("127.0.0.1:0").await.unwrap();
+    let backend_addr = listener.local_addr().unwrap();
+    let preface = b"PRI * HTTP/2.0\r\n\r\nSM\r\n\r\n".to_vec();
+
+    let accept_task = tokio::spawn(async move {
+        let (mut stream, _) = listener.accept().await.unwrap();
+        let mut got = Vec::new();
+        stream.read_to_end(&mut got).await.unwrap();
+        got
+    });
+
+    let mut cfg = ProxyConfig::default();
+    cfg.general.beobachten = true;
+    cfg.general.beobachten_minutes = 1;
+    cfg.censorship.mask = true;
+    cfg.censorship.mask_unix_sock = None;
+    cfg.censorship.mask_host = Some("127.0.0.1".to_string());
+    cfg.censorship.mask_port = backend_addr.port();
+    cfg.general.modes.classic = false;
+    cfg.general.modes.secure = false;
+
+    let config = Arc::new(cfg);
+    let stats = Arc::new(Stats::new());
+    let beobachten = Arc::new(BeobachtenStore::new());
+
+    let (server_side, mut client_side) = duplex(4096);
+    let handler = tokio::spawn(handle_client_stream(
+        server_side,
+        peer,
+        config,
+        stats.clone(),
+        new_upstream_manager(stats),
+        Arc::new(ReplayChecker::new(128, Duration::from_secs(60))),
+        Arc::new(BufferPool::new()),
+        Arc::new(SecureRandom::new()),
+        None,
+        Arc::new(RouteRuntimeController::new(RelayRouteMode::Direct)),
+        None,
+        Arc::new(UserIpTracker::new()),
+        beobachten.clone(),
+        false,
+    ));
+
+    let first = split_at.min(preface.len());
+    client_side.write_all(&preface[..first]).await.unwrap();
+    if first < preface.len() {
+        sleep(Duration::from_millis(delay_ms)).await;
+        client_side.write_all(&preface[first..]).await.unwrap();
+    }
+    client_side.shutdown().await.unwrap();
+
+    let forwarded = tokio::time::timeout(Duration::from_secs(3), accept_task)
+        .await
+        .unwrap()
+        .unwrap();
+    assert!(
+        forwarded.starts_with(&preface),
+        "mask backend must receive an intact HTTP/2 preface prefix"
+    );
+
+    let result = tokio::time::timeout(Duration::from_secs(3), handler)
+        .await
+        .unwrap()
+        .unwrap();
+    assert!(result.is_ok());
+
+    let snapshot = beobachten.snapshot_text(Duration::from_secs(60));
+    assert!(snapshot.contains("[HTTP]"));
+    assert!(snapshot.contains(&format!("{}-1", peer.ip())));
+}
+
+#[tokio::test]
+async fn http2_preface_fragmentation_matrix_is_classified_and_forwarded() {
+    let cases = [(2usize, 0u64), (3, 0), (4, 0), (2, 7), (3, 7), (8, 1)];
+
+    for (i, (split_at, delay_ms)) in cases.into_iter().enumerate() {
+        let peer: SocketAddr = format!("198.51.100.{}:58{}", 140 + i, 100 + i)
+            .parse()
+            .unwrap();
+        run_http2_fragment_case(split_at, delay_ms, peer).await;
+    }
+}
+
+#[tokio::test]
+async fn http2_preface_splitpoint_light_fuzz_classifies_http() {
+    for split_at in 2usize..=12 {
+        let delay_ms = if split_at % 3 == 0 { 7 } else { 1 };
+        let peer: SocketAddr = format!("198.51.101.{}:59{}", split_at, 10 + split_at)
+            .parse()
+            .unwrap();
+        run_http2_fragment_case(split_at, delay_ms, peer).await;
+    }
+}

src/proxy/tests/client_masking_prefetch_config_pipeline_integration_security_tests.rs

@@ -0,0 +1,151 @@
+use super::*;
+use crate::config::{UpstreamConfig, UpstreamType};
+use std::sync::Arc;
+use tokio::io::{AsyncReadExt, AsyncWriteExt, duplex};
+use tokio::net::TcpListener;
+use tokio::time::{Duration, sleep};
+
+fn new_upstream_manager(stats: Arc<Stats>) -> Arc<UpstreamManager> {
+    Arc::new(UpstreamManager::new(
+        vec![UpstreamConfig {
+            upstream_type: UpstreamType::Direct {
+                interface: None,
+                bind_addresses: None,
+            },
+            weight: 1,
+            enabled: true,
+            scopes: String::new(),
+            selected_scope: String::new(),
+        }],
+        1,
+        1,
+        1,
+        10,
+        1,
+        false,
+        stats,
+    ))
+}
+
+async fn run_pipeline_prefetch_case(
+    prefetch_timeout_ms: u64,
+    delayed_tail_ms: u64,
+    peer: SocketAddr,
+) -> (Vec<u8>, String) {
+    let listener = TcpListener::bind("127.0.0.1:0").await.unwrap();
+    let backend_addr = listener.local_addr().unwrap();
+
+    let accept_task = tokio::spawn(async move {
+        let (mut stream, _) = listener.accept().await.unwrap();
+        let mut got = Vec::new();
+        stream.read_to_end(&mut got).await.unwrap();
+        got
+    });
+
+    let mut cfg = ProxyConfig::default();
+    cfg.general.beobachten = true;
+    cfg.general.beobachten_minutes = 1;
+    cfg.censorship.mask = true;
+    cfg.censorship.mask_unix_sock = None;
+    cfg.censorship.mask_host = Some("127.0.0.1".to_string());
+    cfg.censorship.mask_port = backend_addr.port();
+    cfg.censorship.mask_classifier_prefetch_timeout_ms = prefetch_timeout_ms;
+    cfg.general.modes.classic = false;
+    cfg.general.modes.secure = false;
+
+    let config = Arc::new(cfg);
+    let stats = Arc::new(Stats::new());
+    let beobachten = Arc::new(BeobachtenStore::new());
+
+    let (server_side, mut client_side) = duplex(4096);
+    let handler = tokio::spawn(handle_client_stream(
+        server_side,
+        peer,
+        config,
+        stats.clone(),
+        new_upstream_manager(stats),
+        Arc::new(ReplayChecker::new(128, Duration::from_secs(60))),
+        Arc::new(BufferPool::new()),
+        Arc::new(SecureRandom::new()),
+        None,
+        Arc::new(RouteRuntimeController::new(RelayRouteMode::Direct)),
+        None,
+        Arc::new(UserIpTracker::new()),
+        beobachten.clone(),
+        false,
+    ));
+
+    client_side.write_all(b"C").await.unwrap();
+    sleep(Duration::from_millis(delayed_tail_ms)).await;
+
+    client_side
+        .write_all(b"ONNECT example.org:443 HTTP/1.1\r\nHost: example.org\r\n\r\n")
+        .await
+        .unwrap();
+    client_side.shutdown().await.unwrap();
+
+    let forwarded = tokio::time::timeout(Duration::from_secs(3), accept_task)
+        .await
+        .unwrap()
+        .unwrap();
+
+    let result = tokio::time::timeout(Duration::from_secs(3), handler)
+        .await
+        .unwrap()
+        .unwrap();
+    assert!(result.is_ok());
+
+    let snapshot = beobachten.snapshot_text(Duration::from_secs(60));
+    (forwarded, snapshot)
+}
+
+#[tokio::test]
+async fn tdd_pipeline_prefetch_5ms_misses_15ms_tail_and_classifies_as_port_scanner() {
+    let peer: SocketAddr = "198.51.100.171:58071".parse().unwrap();
+    let (forwarded, snapshot) = run_pipeline_prefetch_case(5, 15, peer).await;
+
+    assert!(
+        forwarded.starts_with(b"CONNECT"),
+        "mask backend must still receive full payload bytes in-order"
+    );
+    assert!(
+        snapshot.contains("[HTTP]") || snapshot.contains("[port-scanner]"),
+        "unexpected classifier snapshot for 5ms delayed-tail case: {snapshot}"
+    );
+}
+
+#[tokio::test]
+async fn tdd_pipeline_prefetch_20ms_recovers_15ms_tail_and_classifies_as_http() {
+    let peer: SocketAddr = "198.51.100.172:58072".parse().unwrap();
+    let (forwarded, snapshot) = run_pipeline_prefetch_case(20, 15, peer).await;
+
+    assert!(
+        forwarded.starts_with(b"CONNECT"),
+        "mask backend must receive full CONNECT payload"
+    );
+    assert!(
+        snapshot.contains("[HTTP]"),
+        "20ms budget should recover delayed fragmented prefix and classify as HTTP"
+    );
+}
+
+#[tokio::test]
+async fn matrix_pipeline_prefetch_budget_behavior_5_20_50ms() {
+    let peer5: SocketAddr = "198.51.100.173:58073".parse().unwrap();
+    let peer20: SocketAddr = "198.51.100.174:58074".parse().unwrap();
+    let peer50: SocketAddr = "198.51.100.175:58075".parse().unwrap();
+
+    let (_, snap5) = run_pipeline_prefetch_case(5, 35, peer5).await;
+    let (_, snap20) = run_pipeline_prefetch_case(20, 35, peer20).await;
+    let (_, snap50) = run_pipeline_prefetch_case(50, 35, peer50).await;
+
+    assert!(
+        snap5.contains("[HTTP]") || snap5.contains("[port-scanner]"),
+        "unexpected 5ms snapshot: {snap5}"
+    );
+    assert!(
+        snap20.contains("[HTTP]") || snap20.contains("[port-scanner]"),
+        "unexpected 20ms snapshot: {snap20}"
+    );
+    assert!(snap50.contains("[HTTP]"));
+}

src/proxy/tests/client_masking_prefetch_config_runtime_security_tests.rs

@@ -0,0 +1,88 @@
+use super::*;
+use tokio::io::{AsyncWriteExt, duplex};
+use tokio::time::{Duration, sleep};
+
+#[test]
+fn prefetch_timeout_budget_reads_from_config() {
+    let mut cfg = ProxyConfig::default();
+    assert_eq!(
+        mask_classifier_prefetch_timeout(&cfg),
+        Duration::from_millis(5),
+        "default prefetch timeout budget must remain 5ms"
+    );
+
+    cfg.censorship.mask_classifier_prefetch_timeout_ms = 20;
+    assert_eq!(
+        mask_classifier_prefetch_timeout(&cfg),
+        Duration::from_millis(20),
+        "runtime prefetch timeout budget must follow configured value"
+    );
+}
+
+#[tokio::test]
+async fn configured_prefetch_budget_20ms_recovers_tail_delayed_15ms() {
+    let (mut reader, mut writer) = duplex(1024);
+
+    let writer_task = tokio::spawn(async move {
+        sleep(Duration::from_millis(15)).await;
+        writer
+            .write_all(b"ONNECT example.org:443 HTTP/1.1\r\n")
+            .await
+            .expect("tail bytes must be writable");
+        writer
+            .shutdown()
+            .await
+            .expect("writer shutdown must succeed");
+    });
+
+    let mut initial_data = b"C".to_vec();
+    extend_masking_initial_window_with_timeout(
+        &mut reader,
+        &mut initial_data,
+        Duration::from_millis(20),
+    )
+    .await;
+
+    writer_task
+        .await
+        .expect("writer task must not panic in runtime timeout test");
+
+    assert!(
+        initial_data.starts_with(b"CONNECT"),
+        "20ms configured prefetch budget should recover 15ms delayed CONNECT tail"
+    );
+}
+
+#[tokio::test]
+async fn configured_prefetch_budget_5ms_misses_tail_delayed_15ms() {
+    let (mut reader, mut writer) = duplex(1024);
+
+    let writer_task = tokio::spawn(async move {
+        sleep(Duration::from_millis(15)).await;
+        writer
+            .write_all(b"ONNECT example.org:443 HTTP/1.1\r\n")
+            .await
+            .expect("tail bytes must be writable");
+        writer
+            .shutdown()
+            .await
+            .expect("writer shutdown must succeed");
+    });
+
+    let mut initial_data = b"C".to_vec();
+    extend_masking_initial_window_with_timeout(
+        &mut reader,
+        &mut initial_data,
+        Duration::from_millis(5),
+    )
+    .await;
+
+    writer_task
+        .await
+        .expect("writer task must not panic in runtime timeout test");
+
+    assert!(
+        !initial_data.starts_with(b"CONNECT"),
+        "5ms configured prefetch budget should miss 15ms delayed CONNECT tail"
+    );
+}

src/proxy/tests/client_masking_prefetch_invariant_security_tests.rs

@@ -0,0 +1,265 @@
+use super::*;
+use crate::config::{UpstreamConfig, UpstreamType};
+use crate::crypto::sha256_hmac;
+use crate::protocol::constants::{HANDSHAKE_LEN, TLS_VERSION};
+use crate::protocol::tls;
+use tokio::io::{AsyncReadExt, AsyncWriteExt, duplex};
+use tokio::net::TcpListener;
+
+struct PipelineHarness {
+    config: Arc<ProxyConfig>,
+    stats: Arc<Stats>,
+    upstream_manager: Arc<UpstreamManager>,
+    replay_checker: Arc<ReplayChecker>,
+    buffer_pool: Arc<BufferPool>,
+    rng: Arc<SecureRandom>,
+    route_runtime: Arc<RouteRuntimeController>,
+    ip_tracker: Arc<UserIpTracker>,
+    beobachten: Arc<BeobachtenStore>,
+}
+
+fn build_harness(secret_hex: &str, mask_port: u16) -> PipelineHarness {
+    let mut cfg = ProxyConfig::default();
+    cfg.general.beobachten = false;
+    cfg.censorship.mask = true;
+    cfg.censorship.mask_unix_sock = None;
+    cfg.censorship.mask_host = Some("127.0.0.1".to_string());
+    cfg.censorship.mask_port = mask_port;
+    cfg.censorship.mask_proxy_protocol = 0;
+    cfg.access.ignore_time_skew = true;
+    cfg.access
+        .users
+        .insert("user".to_string(), secret_hex.to_string());
+
+    let config = Arc::new(cfg);
+    let stats = Arc::new(Stats::new());
+    let upstream_manager = Arc::new(UpstreamManager::new(
+        vec![UpstreamConfig {
+            upstream_type: UpstreamType::Direct {
+                interface: None,
+                bind_addresses: None,
+            },
+            weight: 1,
+            enabled: true,
+            scopes: String::new(),
+            selected_scope: String::new(),
+        }],
+        1,
+        1,
+        1,
+        10,
+        1,
+        false,
+        stats.clone(),
+    ));
+
+    PipelineHarness {
+        config,
+        stats,
+        upstream_manager,
+        replay_checker: Arc::new(ReplayChecker::new(256, Duration::from_secs(60))),
+        buffer_pool: Arc::new(BufferPool::new()),
+        rng: Arc::new(SecureRandom::new()),
+        route_runtime: Arc::new(RouteRuntimeController::new(RelayRouteMode::Direct)),
+        ip_tracker: Arc::new(UserIpTracker::new()),
+        beobachten: Arc::new(BeobachtenStore::new()),
+    }
+}
+
+fn make_valid_tls_client_hello(secret: &[u8], timestamp: u32, tls_len: usize, fill: u8) -> Vec<u8> {
+    let total_len = 5 + tls_len;
+    let mut handshake = vec![fill; total_len];
+
+    handshake[0] = 0x16;
+    handshake[1] = 0x03;
+    handshake[2] = 0x01;
+    handshake[3..5].copy_from_slice(&(tls_len as u16).to_be_bytes());
+
+    let session_id_len: usize = 32;
+    handshake[tls::TLS_DIGEST_POS + tls::TLS_DIGEST_LEN] = session_id_len as u8;
+
+    handshake[tls::TLS_DIGEST_POS..tls::TLS_DIGEST_POS + tls::TLS_DIGEST_LEN].fill(0);
+    let computed = sha256_hmac(secret, &handshake);
+    let mut digest = computed;
+    let ts = timestamp.to_le_bytes();
+    for i in 0..4 {
+        digest[28 + i] ^= ts[i];
+    }
+    handshake[tls::TLS_DIGEST_POS..tls::TLS_DIGEST_POS + tls::TLS_DIGEST_LEN]
+        .copy_from_slice(&digest);
+
+    handshake
+}
+
+fn wrap_tls_application_data(payload: &[u8]) -> Vec<u8> {
+    let mut record = Vec::with_capacity(5 + payload.len());
+    record.push(0x17);
+    record.extend_from_slice(&TLS_VERSION);
+    record.extend_from_slice(&(payload.len() as u16).to_be_bytes());
+    record.extend_from_slice(payload);
+    record
+}
+
+async fn read_and_discard_tls_record_body<T>(stream: &mut T, header: [u8; 5])
+where
+    T: tokio::io::AsyncRead + Unpin,
+{
+    let len = u16::from_be_bytes([header[3], header[4]]) as usize;
+    let mut body = vec![0u8; len];
+    stream.read_exact(&mut body).await.unwrap();
+}
+
+#[test]
+fn empty_initial_data_prefetch_gate_is_fail_closed() {
+    assert!(
+        !should_prefetch_mask_classifier_window(&[]),
+        "empty initial_data must not trigger classifier prefetch"
+    );
+}
+
+#[tokio::test]
+async fn blackhat_empty_initial_data_prefetch_must_not_consume_fallback_payload() {
+    let payload = b"\x17\x03\x03\x00\x10coalesced-tail-bytes".to_vec();
+    let (mut reader, mut writer) = duplex(1024);
+
+    writer.write_all(&payload).await.unwrap();
+    writer.shutdown().await.unwrap();
+
+    let mut initial_data = Vec::new();
+    extend_masking_initial_window(&mut reader, &mut initial_data).await;
+
+    assert!(
+        initial_data.is_empty(),
+        "empty initial_data must remain empty after prefetch stage"
+    );
+
+    let mut remaining = Vec::new();
+    reader.read_to_end(&mut remaining).await.unwrap();
+    assert_eq!(
+        remaining, payload,
+        "prefetch stage must not consume fallback payload when initial_data is empty"
+    );
+}
+
+#[tokio::test]
+async fn positive_fragmented_http_prefix_still_prefetches_within_window() {
+    let (mut reader, mut writer) = duplex(1024);
+    writer
+        .write_all(b"NECT example.org:443 HTTP/1.1\r\n")
+        .await
+        .unwrap();
+    writer.shutdown().await.unwrap();
+
+    let mut initial_data = b"CON".to_vec();
+    extend_masking_initial_window(&mut reader, &mut initial_data).await;
+
+    assert!(
+        initial_data.starts_with(b"CONNECT"),
+        "fragmented HTTP method prefix should still be recoverable by prefetch"
+    );
+    assert!(
+        initial_data.len() <= 16,
+        "prefetch window must remain bounded"
+    );
+}
+
+#[tokio::test]
+async fn light_fuzz_empty_initial_data_never_prefetches_any_bytes() {
+    let mut seed = 0xD15C_A11E_2026_0322u64;
+
+    for _ in 0..128 {
+        seed ^= seed << 7;
+        seed ^= seed >> 9;
+        seed ^= seed << 8;
+
+        let len = ((seed & 0x3f) as usize).saturating_add(1);
+        let mut payload = vec![0u8; len];
+        for (idx, byte) in payload.iter_mut().enumerate() {
+            *byte = (seed as u8).wrapping_add(idx as u8).wrapping_mul(17);
+        }
+
+        let (mut reader, mut writer) = duplex(1024);
+        writer.write_all(&payload).await.unwrap();
+        writer.shutdown().await.unwrap();
+
+        let mut initial_data = Vec::new();
+        extend_masking_initial_window(&mut reader, &mut initial_data).await;
+        assert!(initial_data.is_empty());
+
+        let mut remaining = Vec::new();
+        reader.read_to_end(&mut remaining).await.unwrap();
+        assert_eq!(remaining, payload);
+    }
+}
+
+#[tokio::test]
+async fn blackhat_integration_empty_initial_data_path_is_byte_exact_and_eof_clean() {
+    let listener = TcpListener::bind("127.0.0.1:0").await.unwrap();
+    let backend_addr = listener.local_addr().unwrap();
+
+    let secret = [0xD3u8; 16];
+    let client_hello = make_valid_tls_client_hello(&secret, 411, 600, 0x2B);
+    let mut invalid_payload = vec![0u8; HANDSHAKE_LEN];
+    invalid_payload[0] = 0xFF;
+    let invalid_mtproto_record = wrap_tls_application_data(&invalid_payload);
+    let trailing_record = wrap_tls_application_data(b"empty-prefetch-invariant");
+    let expected = trailing_record.clone();
+
+    let accept_task = tokio::spawn(async move {
+        let (mut stream, _) = listener.accept().await.unwrap();
+
+        let mut got = vec![0u8; expected.len()];
+        stream.read_exact(&mut got).await.unwrap();
+        assert_eq!(got, expected);
+
+        let mut one = [0u8; 1];
+        let n = stream.read(&mut one).await.unwrap();
+        assert_eq!(
+            n, 0,
+            "fallback stream must not append synthetic bytes on empty initial_data path"
+        );
+    });
+
+    let harness = build_harness("d3d3d3d3d3d3d3d3d3d3d3d3d3d3d3d3", backend_addr.port());
+    let (server_side, mut client_side) = duplex(131072);
+
+    let handler = tokio::spawn(handle_client_stream(
+        server_side,
+        "198.51.100.245:56145".parse().unwrap(),
+        harness.config,
+        harness.stats,
+        harness.upstream_manager,
+        harness.replay_checker,
+        harness.buffer_pool,
+        harness.rng,
+        None,
+        harness.route_runtime,
+        None,
+        harness.ip_tracker,
+        harness.beobachten,
+        false,
+    ));
+
+    client_side.write_all(&client_hello).await.unwrap();
+    let mut head = [0u8; 5];
+    client_side.read_exact(&mut head).await.unwrap();
+    assert_eq!(head[0], 0x16);
+    read_and_discard_tls_record_body(&mut client_side, head).await;
+
+    client_side
+        .write_all(&invalid_mtproto_record)
+        .await
+        .unwrap();
+    client_side.write_all(&trailing_record).await.unwrap();
+    client_side.shutdown().await.unwrap();
+
+    tokio::time::timeout(Duration::from_secs(3), accept_task)
+        .await
+        .unwrap()
+        .unwrap();
+
+    let _ = tokio::time::timeout(Duration::from_secs(3), handler)
+        .await
+        .unwrap()
+        .unwrap();
+}

src/proxy/tests/client_masking_prefetch_strict_boundary_security_tests.rs

@@ -0,0 +1,72 @@
+use super::*;
+use tokio::io::{AsyncWriteExt, duplex};
+use tokio::time::{Duration, advance, sleep};
+
+async fn run_strict_prefetch_case(prefetch_ms: u64, tail_delay_ms: u64) -> Vec<u8> {
+    let (mut reader, mut writer) = duplex(1024);
+
+    let writer_task = tokio::spawn(async move {
+        sleep(Duration::from_millis(tail_delay_ms)).await;
+        let _ = writer
+            .write_all(b"ONNECT example.org:443 HTTP/1.1\r\n")
+            .await;
+        let _ = writer.shutdown().await;
+    });
+
+    let mut initial_data = b"C".to_vec();
+    let mut prefetch_task = tokio::spawn(async move {
+        extend_masking_initial_window_with_timeout(
+            &mut reader,
+            &mut initial_data,
+            Duration::from_millis(prefetch_ms),
+        )
+        .await;
+        initial_data
+    });
+
+    tokio::task::yield_now().await;
+
+    if tail_delay_ms > 0 {
+        advance(Duration::from_millis(tail_delay_ms)).await;
+        tokio::task::yield_now().await;
+    }
+
+    if prefetch_ms > tail_delay_ms {
+        advance(Duration::from_millis(prefetch_ms - tail_delay_ms)).await;
+        tokio::task::yield_now().await;
+    }
+
+    let result = prefetch_task.await.expect("prefetch task must not panic");
+    writer_task.await.expect("writer task must not panic");
+    result
+}
+
+#[tokio::test(start_paused = true)]
+async fn strict_prefetch_5ms_misses_15ms_tail() {
+    let got = run_strict_prefetch_case(5, 15).await;
+    assert_eq!(got, b"C".to_vec());
+}
+
+#[tokio::test(start_paused = true)]
+async fn strict_prefetch_20ms_recovers_15ms_tail() {
+    let got = run_strict_prefetch_case(20, 15).await;
+    assert!(got.starts_with(b"CONNECT"));
+}
+
+#[tokio::test(start_paused = true)]
+async fn strict_prefetch_50ms_recovers_35ms_tail() {
+    let got = run_strict_prefetch_case(50, 35).await;
+    assert!(got.starts_with(b"CONNECT"));
+}
+
+#[tokio::test(start_paused = true)]
+async fn strict_prefetch_equal_budget_and_delay_recovers_tail() {
+    let got = run_strict_prefetch_case(20, 20).await;
+    assert!(got.starts_with(b"CONNECT"));
+}
+
+#[tokio::test(start_paused = true)]
+async fn strict_prefetch_one_ms_after_budget_misses_tail() {
+    let got = run_strict_prefetch_case(20, 21).await;
+    assert_eq!(got, b"C".to_vec());
+}

src/proxy/tests/client_masking_prefetch_timing_matrix_security_tests.rs

@@ -0,0 +1,98 @@
+use super::*;
+use tokio::io::{AsyncRead, AsyncReadExt, AsyncWriteExt, duplex};
+use tokio::time::{Duration, sleep, timeout};
+
+async fn extend_masking_initial_window_with_budget<R>(
+    reader: &mut R,
+    initial_data: &mut Vec<u8>,
+    prefetch_timeout: Duration,
+) where
+    R: AsyncRead + Unpin,
+{
+    if !should_prefetch_mask_classifier_window(initial_data) {
+        return;
+    }
+
+    let need = 16usize.saturating_sub(initial_data.len());
+    if need == 0 {
+        return;
+    }
+
+    let mut extra = [0u8; 16];
+    if let Ok(Ok(n)) = timeout(prefetch_timeout, reader.read(&mut extra[..need])).await
+        && n > 0
+    {
+        initial_data.extend_from_slice(&extra[..n]);
+    }
+}
+
+async fn run_prefetch_budget_case(prefetch_budget_ms: u64, delayed_tail_ms: u64) -> bool {
+    let (mut reader, mut writer) = duplex(1024);
+
+    let writer_task = tokio::spawn(async move {
+        sleep(Duration::from_millis(delayed_tail_ms)).await;
+        writer
+            .write_all(b"ONNECT example.org:443 HTTP/1.1\r\n")
+            .await
+            .expect("tail bytes must be writable");
+        writer
+            .shutdown()
+            .await
+            .expect("writer shutdown must succeed");
+    });
+
+    let mut initial_data = b"C".to_vec();
+    extend_masking_initial_window_with_budget(
+        &mut reader,
+        &mut initial_data,
+        Duration::from_millis(prefetch_budget_ms),
+    )
+    .await;
+
+    writer_task
+        .await
+        .expect("writer task must not panic during matrix case");
+
+    initial_data.starts_with(b"CONNECT")
+}
+
+#[tokio::test]
+async fn adversarial_prefetch_budget_matrix_5_20_50ms_for_fragmented_connect_tail() {
+    let cases = [
+        // (tail-delay-ms, expected CONNECT recovery for budgets [5, 20, 50])
+        (2u64, [true, true, true]),
+        (15u64, [false, true, true]),
+        (35u64, [false, false, true]),
+    ];
+
+    for (tail_delay_ms, expected) in cases {
+        let got_5 = run_prefetch_budget_case(5, tail_delay_ms).await;
+        let got_20 = run_prefetch_budget_case(20, tail_delay_ms).await;
+        let got_50 = run_prefetch_budget_case(50, tail_delay_ms).await;
+
+        assert_eq!(
+            got_5, expected[0],
+            "5ms prefetch budget mismatch for tail delay {}ms",
+            tail_delay_ms
+        );
+        assert_eq!(
+            got_20, expected[1],
+            "20ms prefetch budget mismatch for tail delay {}ms",
+            tail_delay_ms
+        );
+        assert_eq!(
+            got_50, expected[2],
+            "50ms prefetch budget mismatch for tail delay {}ms",
+            tail_delay_ms
+        );
+    }
+}
+
+#[tokio::test]
+async fn control_current_runtime_prefetch_budget_is_5ms() {
+    assert_eq!(
+        MASK_CLASSIFIER_PREFETCH_TIMEOUT,
+        Duration::from_millis(5),
+        "matrix assumptions require current runtime prefetch budget to stay at 5ms"
+    );
+}

src/proxy/tests/client_masking_probe_evasion_blackhat_tests.rs

@@ -25,6 +25,7 @@ fn make_test_upstream_manager(stats: Arc<Stats>) -> Arc<UpstreamManager> {
         1,
         1,
         1,
+        10,
         1,
         false,
         stats,

src/proxy/tests/client_masking_redteam_expected_fail_tests.rs

@@ -48,6 +48,7 @@ fn build_harness(secret_hex: &str, mask_port: u16) -> RedTeamHarness {
         1,
         1,
         1,
+        10,
         1,
         false,
         stats.clone(),
@@ -237,6 +238,7 @@ async fn redteam_03_masking_duration_must_be_less_than_1ms_when_backend_down() {
             1,
             1,
             1,
+            10,
             1,
             false,
             Arc::new(Stats::new()),
@@ -477,6 +479,7 @@ async fn measure_invalid_probe_duration_ms(delay_ms: u64, tls_len: u16, body_sen
             1,
             1,
             1,
+            10,
             1,
             false,
             Arc::new(Stats::new()),
@@ -550,6 +553,7 @@ async fn capture_forwarded_probe_len(tls_len: u16, body_sent: usize) -> usize {
             1,
             1,
             1,
+            10,
             1,
             false,
             Arc::new(Stats::new()),

src/proxy/tests/client_masking_replay_timing_security_tests.rs

@@ -0,0 +1,168 @@
+use super::*;
+use crate::config::{UpstreamConfig, UpstreamType};
+use crate::crypto::sha256_hmac;
+use crate::protocol::constants::{HANDSHAKE_LEN, TLS_VERSION};
+use crate::protocol::tls;
+use std::sync::Arc;
+use tokio::io::{AsyncReadExt, AsyncWriteExt, duplex};
+use tokio::time::{Duration, Instant};
+
+fn new_upstream_manager(stats: Arc<Stats>) -> Arc<UpstreamManager> {
+    Arc::new(UpstreamManager::new(
+        vec![UpstreamConfig {
+            upstream_type: UpstreamType::Direct {
+                interface: None,
+                bind_addresses: None,
+            },
+            weight: 1,
+            enabled: true,
+            scopes: String::new(),
+            selected_scope: String::new(),
+        }],
+        1,
+        1,
+        1,
+        10,
+        1,
+        false,
+        stats,
+    ))
+}
+
+fn make_valid_tls_client_hello(secret: &[u8], timestamp: u32, tls_len: usize, fill: u8) -> Vec<u8> {
+    let total_len = 5 + tls_len;
+    let mut handshake = vec![fill; total_len];
+
+    handshake[0] = 0x16;
+    handshake[1] = 0x03;
+    handshake[2] = 0x01;
+    handshake[3..5].copy_from_slice(&(tls_len as u16).to_be_bytes());
+
+    let session_id_len: usize = 32;
+    handshake[tls::TLS_DIGEST_POS + tls::TLS_DIGEST_LEN] = session_id_len as u8;
+
+    handshake[tls::TLS_DIGEST_POS..tls::TLS_DIGEST_POS + tls::TLS_DIGEST_LEN].fill(0);
+    let computed = sha256_hmac(secret, &handshake);
+    let mut digest = computed;
+    let ts = timestamp.to_le_bytes();
+    for i in 0..4 {
+        digest[28 + i] ^= ts[i];
+    }
+
+    handshake[tls::TLS_DIGEST_POS..tls::TLS_DIGEST_POS + tls::TLS_DIGEST_LEN]
+        .copy_from_slice(&digest);
+    handshake
+}
+
+async fn run_replay_candidate_session(
+    replay_checker: Arc<ReplayChecker>,
+    hello: &[u8],
+    peer: SocketAddr,
+    drive_mtproto_fail: bool,
+) -> Duration {
+    let mut cfg = ProxyConfig::default();
+    cfg.general.beobachten = false;
+    cfg.censorship.mask = true;
+    cfg.censorship.mask_unix_sock = None;
+    cfg.censorship.mask_host = Some("127.0.0.1".to_string());
+    cfg.censorship.mask_port = 1;
+    cfg.censorship.mask_timing_normalization_enabled = false;
+    cfg.access.ignore_time_skew = true;
+    cfg.access.users.insert(
+        "user".to_string(),
+        "abababababababababababababababab".to_string(),
+    );
+
+    let config = Arc::new(cfg);
+    let stats = Arc::new(Stats::new());
+    let beobachten = Arc::new(BeobachtenStore::new());
+
+    let (server_side, mut client_side) = duplex(65536);
+    let started = Instant::now();
+
+    let task = tokio::spawn(handle_client_stream(
+        server_side,
+        peer,
+        config,
+        stats.clone(),
+        new_upstream_manager(stats),
+        replay_checker,
+        Arc::new(BufferPool::new()),
+        Arc::new(SecureRandom::new()),
+        None,
+        Arc::new(RouteRuntimeController::new(RelayRouteMode::Direct)),
+        None,
+        Arc::new(UserIpTracker::new()),
+        beobachten,
+        false,
+    ));
+
+    client_side.write_all(hello).await.unwrap();
+
+    if drive_mtproto_fail {
+        let mut server_hello_head = [0u8; 5];
+        client_side
+            .read_exact(&mut server_hello_head)
+            .await
+            .unwrap();
+        assert_eq!(server_hello_head[0], 0x16);
+        let body_len = u16::from_be_bytes([server_hello_head[3], server_hello_head[4]]) as usize;
+        let mut body = vec![0u8; body_len];
+        client_side.read_exact(&mut body).await.unwrap();
+
+        let mut invalid_mtproto_record = Vec::with_capacity(5 + HANDSHAKE_LEN);
+        invalid_mtproto_record.push(0x17);
+        invalid_mtproto_record.extend_from_slice(&TLS_VERSION);
+        invalid_mtproto_record.extend_from_slice(&(HANDSHAKE_LEN as u16).to_be_bytes());
+        invalid_mtproto_record.extend_from_slice(&vec![0u8; HANDSHAKE_LEN]);
+        client_side
+            .write_all(&invalid_mtproto_record)
+            .await
+            .unwrap();
+        client_side
+            .write_all(b"GET /replay-fallback HTTP/1.1\r\nHost: x\r\n\r\n")
+            .await
+            .unwrap();
+    }
+
+    client_side.shutdown().await.unwrap();
+
+    let _ = tokio::time::timeout(Duration::from_secs(4), task)
+        .await
+        .unwrap()
+        .unwrap();
+
+    started.elapsed()
+}
+
+#[tokio::test]
+async fn replay_reject_still_honors_masking_timing_budget() {
+    let replay_checker = Arc::new(ReplayChecker::new(256, Duration::from_secs(60)));
+    let hello = make_valid_tls_client_hello(&[0xAB; 16], 7, 600, 0x51);
+
+    let seed_elapsed = run_replay_candidate_session(
+        Arc::clone(&replay_checker),
+        &hello,
+        "198.51.100.201:58001".parse().unwrap(),
+        true,
+    )
+    .await;
+
+    assert!(
+        seed_elapsed >= Duration::from_millis(40) && seed_elapsed < Duration::from_millis(250),
+        "seed replay-candidate run must honor masking timing budget without unbounded delay"
+    );
+
+    let replay_elapsed = run_replay_candidate_session(
+        Arc::clone(&replay_checker),
+        &hello,
+        "198.51.100.202:58002".parse().unwrap(),
+        false,
+    )
+    .await;
+
+    assert!(
+        replay_elapsed >= Duration::from_millis(40) && replay_elapsed < Duration::from_millis(250),
+        "replay rejection path must still satisfy masking timing budget without unbounded DB/CPU delay"
+    );
+}

src/proxy/tests/client_masking_shape_classifier_fuzz_redteam_expected_fail_tests.rs

@@ -20,6 +20,7 @@ fn new_upstream_manager(stats: Arc<Stats>) -> Arc<UpstreamManager> {
         1,
         1,
         1,
+        10,
         1,
         false,
         stats,

src/proxy/tests/client_masking_shape_hardening_adversarial_tests.rs

@@ -20,6 +20,7 @@ fn new_upstream_manager(stats: Arc<Stats>) -> Arc<UpstreamManager> {
         1,
         1,
         1,
+        10,
         1,
         false,
         stats,

src/proxy/tests/client_masking_shape_hardening_redteam_expected_fail_tests.rs

@@ -20,6 +20,7 @@ fn new_upstream_manager(stats: Arc<Stats>) -> Arc<UpstreamManager> {
         1,
         1,
         1,
+        10,
         1,
         false,
         stats,

src/proxy/tests/client_masking_shape_hardening_security_tests.rs

@@ -20,6 +20,7 @@ fn new_upstream_manager(stats: Arc<Stats>) -> Arc<UpstreamManager> {
         1,
         1,
         1,
+        10,
         1,
         false,
         stats,

src/proxy/tests/client_masking_stress_adversarial_tests.rs

@@ -34,6 +34,7 @@ fn new_upstream_manager(stats: Arc<Stats>) -> Arc<UpstreamManager> {
         1,
         1,
         1,
+        10,
         1,
         false,
         stats,

src/proxy/tests/client_more_advanced_tests.rs

@@ -0,0 +1,291 @@
+use super::*;
+use crate::config::ProxyConfig;
+use crate::stats::Stats;
+use crate::transport::UpstreamManager;
+use std::sync::Arc;
+use std::time::Duration;
+use tokio::io::{AsyncReadExt, AsyncWriteExt, duplex};
+
+fn preload_user_quota(stats: &Stats, user: &str, bytes: u64) {
+    let user_stats = stats.get_or_create_user_stats_handle(user);
+    stats.quota_charge_post_write(user_stats.as_ref(), bytes);
+}
+
+#[tokio::test]
+async fn edge_mask_delay_bypassed_if_max_is_zero() {
+    let mut config = ProxyConfig::default();
+    config.censorship.server_hello_delay_min_ms = 10_000;
+    config.censorship.server_hello_delay_max_ms = 0;
+
+    let start = std::time::Instant::now();
+    maybe_apply_mask_reject_delay(&config).await;
+    assert!(start.elapsed() < Duration::from_millis(50));
+}
+
+#[test]
+fn edge_beobachten_ttl_clamps_exactly_to_24_hours() {
+    let mut config = ProxyConfig::default();
+    config.general.beobachten = true;
+    config.general.beobachten_minutes = 100_000;
+
+    let ttl = beobachten_ttl(&config);
+    assert_eq!(ttl.as_secs(), 24 * 60 * 60);
+}
+
+#[test]
+fn edge_wrap_tls_application_record_empty_payload() {
+    let wrapped = wrap_tls_application_record(&[]);
+    assert_eq!(wrapped.len(), 5);
+    assert_eq!(wrapped[0], TLS_RECORD_APPLICATION);
+    assert_eq!(&wrapped[3..5], &[0, 0]);
+}
+
+#[tokio::test]
+async fn boundary_user_data_quota_exact_match_rejects() {
+    let user = "quota-boundary-user";
+    let mut config = ProxyConfig::default();
+    config.access.user_data_quota.insert(user.to_string(), 1024);
+
+    let stats = Arc::new(Stats::new());
+    preload_user_quota(stats.as_ref(), user, 1024);
+
+    let ip_tracker = Arc::new(UserIpTracker::new());
+    let peer = "198.51.100.10:55000".parse().unwrap();
+
+    let result = RunningClientHandler::acquire_user_connection_reservation_static(
+        user, &config, stats, peer, ip_tracker,
+    )
+    .await;
+
+    assert!(matches!(result, Err(ProxyError::DataQuotaExceeded { .. })));
+}
+
+#[tokio::test]
+async fn boundary_user_expiration_in_past_rejects() {
+    let user = "expired-boundary-user";
+    let mut config = ProxyConfig::default();
+    let expired_time = chrono::Utc::now() - chrono::Duration::milliseconds(1);
+    config
+        .access
+        .user_expirations
+        .insert(user.to_string(), expired_time);
+
+    let stats = Arc::new(Stats::new());
+    let ip_tracker = Arc::new(UserIpTracker::new());
+    let peer = "198.51.100.11:55000".parse().unwrap();
+
+    let result = RunningClientHandler::acquire_user_connection_reservation_static(
+        user, &config, stats, peer, ip_tracker,
+    )
+    .await;
+
+    assert!(matches!(result, Err(ProxyError::UserExpired { .. })));
+}
+
+#[tokio::test]
+async fn blackhat_proxy_protocol_massive_garbage_rejected_quickly() {
+    let mut cfg = ProxyConfig::default();
+    cfg.server.proxy_protocol_header_timeout_ms = 300;
+    let config = Arc::new(cfg);
+    let stats = Arc::new(Stats::new());
+
+    let (server_side, mut client_side) = duplex(4096);
+    let handler = tokio::spawn(handle_client_stream(
+        server_side,
+        "198.51.100.12:55000".parse().unwrap(),
+        config,
+        stats.clone(),
+        Arc::new(UpstreamManager::new(
+            vec![],
+            1,
+            1,
+            1,
+            10,
+            1,
+            false,
+            stats.clone(),
+        )),
+        Arc::new(ReplayChecker::new(128, Duration::from_secs(60))),
+        Arc::new(BufferPool::new()),
+        Arc::new(SecureRandom::new()),
+        None,
+        Arc::new(RouteRuntimeController::new(RelayRouteMode::Direct)),
+        None,
+        Arc::new(UserIpTracker::new()),
+        Arc::new(BeobachtenStore::new()),
+        true,
+    ));
+
+    client_side.write_all(&vec![b'A'; 2000]).await.unwrap();
+
+    let result = tokio::time::timeout(Duration::from_secs(2), handler)
+        .await
+        .unwrap()
+        .unwrap();
+    assert!(matches!(result, Err(ProxyError::InvalidProxyProtocol)));
+    assert_eq!(stats.get_connects_bad(), 1);
+}
+
+#[tokio::test]
+async fn edge_tls_body_immediate_eof_triggers_masking_and_bad_connect() {
+    let mut cfg = ProxyConfig::default();
+    cfg.general.beobachten = true;
+    cfg.general.beobachten_minutes = 1;
+
+    let config = Arc::new(cfg);
+    let stats = Arc::new(Stats::new());
+    let beobachten = Arc::new(BeobachtenStore::new());
+
+    let (server_side, mut client_side) = duplex(4096);
+    let handler = tokio::spawn(handle_client_stream(
+        server_side,
+        "198.51.100.13:55000".parse().unwrap(),
+        config,
+        stats.clone(),
+        Arc::new(UpstreamManager::new(
+            vec![],
+            1,
+            1,
+            1,
+            10,
+            1,
+            false,
+            stats.clone(),
+        )),
+        Arc::new(ReplayChecker::new(128, Duration::from_secs(60))),
+        Arc::new(BufferPool::new()),
+        Arc::new(SecureRandom::new()),
+        None,
+        Arc::new(RouteRuntimeController::new(RelayRouteMode::Direct)),
+        None,
+        Arc::new(UserIpTracker::new()),
+        beobachten.clone(),
+        false,
+    ));
+
+    client_side
+        .write_all(&[0x16, 0x03, 0x01, 0x00, 100])
+        .await
+        .unwrap();
+    client_side.shutdown().await.unwrap();
+
+    let _ = tokio::time::timeout(Duration::from_secs(2), handler)
+        .await
+        .unwrap();
+
+    assert_eq!(stats.get_connects_bad(), 1);
+}
+
+#[tokio::test]
+async fn security_classic_mode_disabled_masks_valid_length_payload() {
+    let mut cfg = ProxyConfig::default();
+    cfg.general.modes.classic = false;
+    cfg.general.modes.secure = false;
+    cfg.censorship.mask = true;
+
+    let config = Arc::new(cfg);
+    let stats = Arc::new(Stats::new());
+
+    let (server_side, mut client_side) = duplex(4096);
+    let handler = tokio::spawn(handle_client_stream(
+        server_side,
+        "198.51.100.15:55000".parse().unwrap(),
+        config,
+        stats.clone(),
+        Arc::new(UpstreamManager::new(
+            vec![],
+            1,
+            1,
+            1,
+            10,
+            1,
+            false,
+            stats.clone(),
+        )),
+        Arc::new(ReplayChecker::new(128, Duration::from_secs(60))),
+        Arc::new(BufferPool::new()),
+        Arc::new(SecureRandom::new()),
+        None,
+        Arc::new(RouteRuntimeController::new(RelayRouteMode::Direct)),
+        None,
+        Arc::new(UserIpTracker::new()),
+        Arc::new(BeobachtenStore::new()),
+        false,
+    ));
+
+    client_side.write_all(&vec![0xEF; 64]).await.unwrap();
+    client_side.shutdown().await.unwrap();
+
+    let _ = tokio::time::timeout(Duration::from_secs(2), handler)
+        .await
+        .unwrap();
+    assert_eq!(stats.get_connects_bad(), 1);
+}
+
+#[tokio::test]
+async fn concurrency_ip_tracker_strict_limit_one_rapid_churn() {
+    let user = "rapid-churn-user";
+    let mut config = ProxyConfig::default();
+    config
+        .access
+        .user_max_tcp_conns
+        .insert(user.to_string(), 10);
+
+    let stats = Arc::new(Stats::new());
+    let ip_tracker = Arc::new(UserIpTracker::new());
+    ip_tracker.set_user_limit(user, 1).await;
+
+    let peer = "198.51.100.16:55000".parse().unwrap();
+
+    for _ in 0..500 {
+        let reservation = RunningClientHandler::acquire_user_connection_reservation_static(
+            user,
+            &config,
+            stats.clone(),
+            peer,
+            ip_tracker.clone(),
+        )
+        .await
+        .unwrap();
+        reservation.release().await;
+    }
+
+    assert_eq!(stats.get_user_curr_connects(user), 0);
+    assert_eq!(ip_tracker.get_active_ip_count(user).await, 0);
+}
+
+#[tokio::test]
+async fn quirk_read_with_progress_zero_length_buffer_returns_zero_immediately() {
+    let (mut server_side, _client_side) = duplex(4096);
+    let mut empty_buf = &mut [][..];
+
+    let result = tokio::time::timeout(
+        Duration::from_millis(50),
+        read_with_progress(&mut server_side, &mut empty_buf),
+    )
+    .await;
+
+    assert!(result.is_ok());
+    assert_eq!(result.unwrap().unwrap(), 0);
+}
+
+#[tokio::test]
+async fn stress_read_with_progress_cancellation_safety() {
+    let (mut server_side, mut client_side) = duplex(4096);
+
+    client_side.write_all(b"12345").await.unwrap();
+
+    let mut buf = [0u8; 10];
+    let result = tokio::time::timeout(
+        Duration::from_millis(50),
+        read_with_progress(&mut server_side, &mut buf),
+    )
+    .await;
+
+    assert!(result.is_err());
+
+    client_side.write_all(b"67890").await.unwrap();
+    let mut buf2 = [0u8; 5];
+    server_side.read_exact(&mut buf2).await.unwrap();
+    assert_eq!(&buf2, b"67890");
+}

src/proxy/tests/client_security_tests.rs

@@ -1,12 +1,17 @@
 use super::*;
 use crate::config::{UpstreamConfig, UpstreamType};
-use crate::crypto::AesCtr;
-use crate::crypto::sha256_hmac;
-use crate::protocol::constants::ProtoTag;
+use crate::crypto::{AesCtr, sha256, sha256_hmac};
+use crate::protocol::constants::{
+    DC_IDX_POS, HANDSHAKE_LEN, IV_LEN, PREKEY_LEN, PROTO_TAG_POS, ProtoTag, SKIP_LEN,
+    TLS_RECORD_CHANGE_CIPHER,
+};
 use crate::protocol::tls;
 use crate::proxy::handshake::HandshakeSuccess;
 use crate::stream::{CryptoReader, CryptoWriter};
 use crate::transport::proxy_protocol::ProxyProtocolV1Builder;
+use rand::Rng;
+use rand::SeedableRng;
+use rand::rngs::StdRng;
 use std::net::Ipv4Addr;
 use tokio::io::{AsyncReadExt, AsyncWriteExt, duplex};
 use tokio::net::{TcpListener, TcpStream};
@@ -25,6 +30,220 @@ fn synthetic_local_addr_uses_configured_port_for_max() {
     assert_eq!(addr.port(), u16::MAX);
 }
 
+#[test]
+fn handshake_timeout_with_mask_grace_includes_mask_margin() {
+    let mut config = ProxyConfig::default();
+    config.timeouts.client_handshake = 2;
+
+    config.censorship.mask = false;
+    assert_eq!(
+        handshake_timeout_with_mask_grace(&config),
+        Duration::from_secs(2)
+    );
+
+    config.censorship.mask = true;
+    assert_eq!(
+        handshake_timeout_with_mask_grace(&config),
+        Duration::from_millis(2750),
+        "mask mode extends handshake timeout by 750 ms"
+    );
+}
+
+#[tokio::test]
+async fn read_with_progress_reads_partial_buffers_before_eof() {
+    let data = vec![0xAA, 0xBB, 0xCC];
+    let mut reader = std::io::Cursor::new(data);
+    let mut buf = [0u8; 5];
+
+    let read = read_with_progress(&mut reader, &mut buf).await.unwrap();
+    assert_eq!(read, 3);
+    assert_eq!(&buf[..3], &[0xAA, 0xBB, 0xCC]);
+}
+
+#[test]
+fn is_trusted_proxy_source_respects_cidr_list_and_empty_rejects_all() {
+    let peer: IpAddr = "10.10.10.10".parse().unwrap();
+    assert!(!is_trusted_proxy_source(peer, &[]));
+
+    let trusted = vec!["10.0.0.0/8".parse().unwrap()];
+    assert!(is_trusted_proxy_source(peer, &trusted));
+
+    let not_trusted = vec!["192.0.2.0/24".parse().unwrap()];
+    assert!(!is_trusted_proxy_source(peer, &not_trusted));
+}
+
+#[test]
+fn is_trusted_proxy_source_accepts_cidr_zero_zero_as_global_cidr() {
+    let peer: IpAddr = "203.0.113.42".parse().unwrap();
+    let trust_all = vec!["0.0.0.0/0".parse().unwrap()];
+    assert!(is_trusted_proxy_source(peer, &trust_all));
+
+    let peer_v6: IpAddr = "2001:db8::1".parse().unwrap();
+    let trust_all_v6 = vec!["::/0".parse().unwrap()];
+    assert!(is_trusted_proxy_source(peer_v6, &trust_all_v6));
+}
+
+struct ErrorReader;
+
+impl tokio::io::AsyncRead for ErrorReader {
+    fn poll_read(
+        self: std::pin::Pin<&mut Self>,
+        _cx: &mut std::task::Context<'_>,
+        _buf: &mut tokio::io::ReadBuf<'_>,
+    ) -> std::task::Poll<std::io::Result<()>> {
+        std::task::Poll::Ready(Err(std::io::Error::new(
+            std::io::ErrorKind::UnexpectedEof,
+            "fake error",
+        )))
+    }
+}
+
+#[tokio::test]
+async fn read_with_progress_returns_error_from_failed_reader() {
+    let mut reader = ErrorReader;
+    let mut buf = [0u8; 8];
+    let err = read_with_progress(&mut reader, &mut buf).await.unwrap_err();
+    assert_eq!(err.kind(), std::io::ErrorKind::UnexpectedEof);
+}
+
+#[test]
+fn handshake_timeout_with_mask_grace_handles_maximum_values_without_overflow() {
+    let mut config = ProxyConfig::default();
+    config.timeouts.client_handshake = u64::MAX;
+    config.censorship.mask = true;
+
+    let timeout = handshake_timeout_with_mask_grace(&config);
+    assert!(timeout >= Duration::from_secs(u64::MAX));
+}
+
+#[tokio::test]
+async fn read_with_progress_zero_length_buffer_returns_zero() {
+    let data = vec![1, 2, 3];
+    let mut reader = std::io::Cursor::new(data);
+    let mut buf = [];
+
+    let read = read_with_progress(&mut reader, &mut buf).await.unwrap();
+    assert_eq!(read, 0);
+}
+
+#[test]
+fn handshake_timeout_without_mask_is_exact_base() {
+    let mut config = ProxyConfig::default();
+    config.timeouts.client_handshake = 7;
+    config.censorship.mask = false;
+
+    assert_eq!(
+        handshake_timeout_with_mask_grace(&config),
+        Duration::from_secs(7)
+    );
+}
+
+#[test]
+fn handshake_timeout_mask_enabled_adds_750ms() {
+    let mut config = ProxyConfig::default();
+    config.timeouts.client_handshake = 3;
+    config.censorship.mask = true;
+
+    assert_eq!(
+        handshake_timeout_with_mask_grace(&config),
+        Duration::from_millis(3750)
+    );
+}
+
+#[tokio::test]
+async fn read_with_progress_full_then_empty_transition() {
+    let data = vec![0x10, 0x20];
+    let mut cursor = std::io::Cursor::new(data);
+    let mut buf = [0u8; 2];
+
+    assert_eq!(read_with_progress(&mut cursor, &mut buf).await.unwrap(), 2);
+    assert_eq!(read_with_progress(&mut cursor, &mut buf).await.unwrap(), 0);
+}
+
+#[tokio::test]
+async fn read_with_progress_fragmented_io_works_over_multiple_calls() {
+    let mut cursor = std::io::Cursor::new(vec![1, 2, 3, 4, 5]);
+    let mut result = Vec::new();
+
+    for chunk_size in 1..=5 {
+        let mut b = vec![0u8; chunk_size];
+        let n = read_with_progress(&mut cursor, &mut b).await.unwrap();
+        result.extend_from_slice(&b[..n]);
+        if n == 0 {
+            break;
+        }
+    }
+
+    assert_eq!(result, vec![1, 2, 3, 4, 5]);
+}
+
+#[tokio::test]
+async fn read_with_progress_stress_randomized_chunk_sizes() {
+    for i in 0..128 {
+        let mut rng = StdRng::seed_from_u64(i as u64 + 1);
+        let mut input: Vec<u8> = (0..(i % 41)).map(|_| rng.next_u32() as u8).collect();
+        let mut cursor = std::io::Cursor::new(input.clone());
+        let mut collected = Vec::new();
+
+        while cursor.position() < cursor.get_ref().len() as u64 {
+            let chunk = 1 + (rng.next_u32() as usize % 8);
+            let mut b = vec![0u8; chunk];
+            let read = read_with_progress(&mut cursor, &mut b).await.unwrap();
+            collected.extend_from_slice(&b[..read]);
+            if read == 0 {
+                break;
+            }
+        }
+
+        assert_eq!(collected, input);
+    }
+}
+
+#[test]
+fn is_trusted_proxy_source_boundary_narrow_ipv4() {
+    let matching = "172.16.0.1".parse().unwrap();
+    let not_matching = "172.15.255.255".parse().unwrap();
+    let cidr = vec!["172.16.0.0/12".parse().unwrap()];
+    assert!(is_trusted_proxy_source(matching, &cidr));
+    assert!(!is_trusted_proxy_source(not_matching, &cidr));
+}
+
+#[test]
+fn is_trusted_proxy_source_rejects_out_of_family_ipv6_v4_cidr() {
+    let peer = "2001:db8::1".parse().unwrap();
+    let cidr = vec!["10.0.0.0/8".parse().unwrap()];
+    assert!(!is_trusted_proxy_source(peer, &cidr));
+}
+
+#[test]
+fn wrap_tls_application_record_reserved_chunks_look_reasonable() {
+    let payload = vec![0xAA; 1 + (u16::MAX as usize) + 2];
+    let wrapped = wrap_tls_application_record(&payload);
+    assert!(wrapped.len() > payload.len());
+    assert!(wrapped.contains(&0x17));
+}
+
+#[test]
+fn wrap_tls_application_record_roundtrip_size_check() {
+    let payload_len = 3000;
+    let payload = vec![0x55; payload_len];
+    let wrapped = wrap_tls_application_record(&payload);
+
+    let mut idx = 0;
+    let mut consumed = 0;
+    while idx + 5 <= wrapped.len() {
+        assert_eq!(wrapped[idx], 0x17);
+        let len = u16::from_be_bytes([wrapped[idx + 3], wrapped[idx + 4]]) as usize;
+        consumed += len;
+        idx += 5 + len;
+        if idx >= wrapped.len() {
+            break;
+        }
+    }
+
+    assert_eq!(consumed, payload_len);
+}
+
 fn make_crypto_reader<R>(reader: R) -> CryptoReader<R>
 where
     R: tokio::io::AsyncRead + Unpin,
@@ -43,6 +262,11 @@ where
     CryptoWriter::new(writer, AesCtr::new(&key, iv), 8 * 1024)
 }
 
+fn preload_user_quota(stats: &Stats, user: &str, bytes: u64) {
+    let user_stats = stats.get_or_create_user_stats_handle(user);
+    stats.quota_charge_post_write(user_stats.as_ref(), bytes);
+}
+
 #[tokio::test]
 async fn user_connection_reservation_drop_enqueues_cleanup_synchronously() {
     let ip_tracker = Arc::new(crate::ip_tracker::UserIpTracker::new());
@@ -117,6 +341,7 @@ async fn relay_task_abort_releases_user_gate_and_ip_reservation() {
         1,
         1,
         1,
+        10,
         1,
         false,
         stats.clone(),
@@ -230,6 +455,7 @@ async fn relay_cutover_releases_user_gate_and_ip_reservation() {
         1,
         1,
         1,
+        10,
         1,
         false,
         stats.clone(),
@@ -353,6 +579,7 @@ async fn integration_route_cutover_and_quota_overlap_fails_closed_and_releases_s
         1,
         1,
         1,
+        10,
         1,
         false,
         stats.clone(),
@@ -522,6 +749,7 @@ async fn proxy_protocol_header_is_rejected_when_trust_list_is_empty() {
         1,
         1,
         1,
+        10,
         1,
         false,
         stats.clone(),
@@ -598,6 +826,7 @@ async fn proxy_protocol_header_from_untrusted_peer_range_is_rejected_under_load(
             1,
             1,
             1,
+            10,
             1,
             false,
             stats.clone(),
@@ -757,6 +986,7 @@ async fn short_tls_probe_is_masked_through_client_pipeline() {
         1,
         1,
         1,
+        10,
         1,
         false,
         stats.clone(),
@@ -844,6 +1074,7 @@ async fn tls12_record_probe_is_masked_through_client_pipeline() {
         1,
         1,
         1,
+        10,
         1,
         false,
         stats.clone(),
@@ -929,6 +1160,7 @@ async fn handle_client_stream_increments_connects_all_exactly_once() {
         1,
         1,
         1,
+        10,
         1,
         false,
         stats.clone(),
@@ -1021,6 +1253,7 @@ async fn running_client_handler_increments_connects_all_exactly_once() {
         1,
         1,
         1,
+        10,
         1,
         false,
         stats.clone(),
@@ -1088,6 +1321,163 @@ async fn running_client_handler_increments_connects_all_exactly_once() {
     );
 }
 
+#[tokio::test(start_paused = true)]
+async fn idle_pooled_connection_closes_cleanly_in_generic_stream_path() {
+    let mut cfg = ProxyConfig::default();
+    cfg.general.beobachten = false;
+    cfg.timeouts.client_first_byte_idle_secs = 1;
+
+    let config = Arc::new(cfg);
+    let stats = Arc::new(Stats::new());
+    let upstream_manager = Arc::new(UpstreamManager::new(
+        vec![UpstreamConfig {
+            upstream_type: UpstreamType::Direct {
+                interface: None,
+                bind_addresses: None,
+            },
+            weight: 1,
+            enabled: true,
+            scopes: String::new(),
+            selected_scope: String::new(),
+        }],
+        1,
+        1,
+        1,
+        10,
+        1,
+        false,
+        stats.clone(),
+    ));
+    let replay_checker = Arc::new(ReplayChecker::new(128, Duration::from_secs(60)));
+    let buffer_pool = Arc::new(BufferPool::new());
+    let rng = Arc::new(SecureRandom::new());
+    let route_runtime = Arc::new(RouteRuntimeController::new(RelayRouteMode::Direct));
+    let ip_tracker = Arc::new(UserIpTracker::new());
+    let beobachten = Arc::new(BeobachtenStore::new());
+
+    let (server_side, _client_side) = duplex(4096);
+    let peer: SocketAddr = "198.51.100.169:55200".parse().unwrap();
+
+    let handler = tokio::spawn(handle_client_stream(
+        server_side,
+        peer,
+        config,
+        stats.clone(),
+        upstream_manager,
+        replay_checker,
+        buffer_pool,
+        rng,
+        None,
+        route_runtime,
+        None,
+        ip_tracker,
+        beobachten,
+        false,
+    ));
+
+    // Let the spawned handler arm the idle-phase timeout before advancing paused time.
+    tokio::task::yield_now().await;
+    tokio::time::advance(Duration::from_secs(2)).await;
+    tokio::task::yield_now().await;
+
+    let result = tokio::time::timeout(Duration::from_secs(1), handler)
+        .await
+        .unwrap()
+        .unwrap();
+    assert!(result.is_ok());
+    assert_eq!(stats.get_handshake_timeouts(), 0);
+    assert_eq!(stats.get_connects_bad(), 0);
+}
+
+#[tokio::test(start_paused = true)]
+async fn idle_pooled_connection_closes_cleanly_in_client_handler_path() {
+    let front_listener = TcpListener::bind("127.0.0.1:0").await.unwrap();
+    let front_addr = front_listener.local_addr().unwrap();
+
+    let mut cfg = ProxyConfig::default();
+    cfg.general.beobachten = false;
+    cfg.timeouts.client_first_byte_idle_secs = 1;
+
+    let config = Arc::new(cfg);
+    let stats = Arc::new(Stats::new());
+    let upstream_manager = Arc::new(UpstreamManager::new(
+        vec![UpstreamConfig {
+            upstream_type: UpstreamType::Direct {
+                interface: None,
+                bind_addresses: None,
+            },
+            weight: 1,
+            enabled: true,
+            scopes: String::new(),
+            selected_scope: String::new(),
+        }],
+        1,
+        1,
+        1,
+        10,
+        1,
+        false,
+        stats.clone(),
+    ));
+    let replay_checker = Arc::new(ReplayChecker::new(128, Duration::from_secs(60)));
+    let buffer_pool = Arc::new(BufferPool::new());
+    let rng = Arc::new(SecureRandom::new());
+    let route_runtime = Arc::new(RouteRuntimeController::new(RelayRouteMode::Direct));
+    let ip_tracker = Arc::new(UserIpTracker::new());
+    let beobachten = Arc::new(BeobachtenStore::new());
+
+    let server_task = {
+        let config = config.clone();
+        let stats = stats.clone();
+        let upstream_manager = upstream_manager.clone();
+        let replay_checker = replay_checker.clone();
+        let buffer_pool = buffer_pool.clone();
+        let rng = rng.clone();
+        let route_runtime = route_runtime.clone();
+        let ip_tracker = ip_tracker.clone();
+        let beobachten = beobachten.clone();
+
+        tokio::spawn(async move {
+            let (stream, peer) = front_listener.accept().await.unwrap();
+            let real_peer_report = Arc::new(std::sync::Mutex::new(None));
+            ClientHandler::new(
+                stream,
+                peer,
+                config,
+                stats,
+                upstream_manager,
+                replay_checker,
+                buffer_pool,
+                rng,
+                None,
+                route_runtime,
+                None,
+                ip_tracker,
+                beobachten,
+                false,
+                real_peer_report,
+            )
+            .run()
+            .await
+        })
+    };
+
+    let _client = TcpStream::connect(front_addr).await.unwrap();
+
+    // Let the accepted connection reach the idle wait before advancing paused time.
+    tokio::task::yield_now().await;
+    tokio::time::advance(Duration::from_secs(2)).await;
+    tokio::task::yield_now().await;
+
+    let result = tokio::time::timeout(Duration::from_secs(1), server_task)
+        .await
+        .unwrap()
+        .unwrap();
+    assert!(result.is_ok());
+    assert_eq!(stats.get_handshake_timeouts(), 0);
+    assert_eq!(stats.get_connects_bad(), 0);
+}
+
 #[tokio::test]
 async fn partial_tls_header_stall_triggers_handshake_timeout() {
     let mut cfg = ProxyConfig::default();
@@ -1110,6 +1500,7 @@ async fn partial_tls_header_stall_triggers_handshake_timeout() {
         1,
         1,
         1,
+        10,
         1,
         false,
         stats.clone(),
@@ -1255,6 +1646,148 @@ fn wrap_tls_application_data(payload: &[u8]) -> Vec<u8> {
     record
 }
 
+fn wrap_tls_ccs_record() -> Vec<u8> {
+    let mut record = Vec::with_capacity(6);
+    record.push(TLS_RECORD_CHANGE_CIPHER);
+    record.extend_from_slice(&[0x03, 0x03]);
+    record.extend_from_slice(&1u16.to_be_bytes());
+    record.push(0x01);
+    record
+}
+
+fn make_valid_mtproto_handshake(
+    secret_hex: &str,
+    proto_tag: ProtoTag,
+    dc_idx: i16,
+) -> [u8; HANDSHAKE_LEN] {
+    let secret = hex::decode(secret_hex).expect("secret hex must decode for mtproto test helper");
+
+    let mut handshake = [0x5Au8; HANDSHAKE_LEN];
+    for (idx, b) in handshake[SKIP_LEN..SKIP_LEN + PREKEY_LEN + IV_LEN]
+        .iter_mut()
+        .enumerate()
+    {
+        *b = (idx as u8).wrapping_add(1);
+    }
+
+    let dec_prekey = &handshake[SKIP_LEN..SKIP_LEN + PREKEY_LEN];
+    let dec_iv_bytes = &handshake[SKIP_LEN + PREKEY_LEN..SKIP_LEN + PREKEY_LEN + IV_LEN];
+
+    let mut dec_key_input = Vec::with_capacity(PREKEY_LEN + secret.len());
+    dec_key_input.extend_from_slice(dec_prekey);
+    dec_key_input.extend_from_slice(&secret);
+    let dec_key = sha256(&dec_key_input);
+
+    let mut dec_iv_arr = [0u8; IV_LEN];
+    dec_iv_arr.copy_from_slice(dec_iv_bytes);
+    let dec_iv = u128::from_be_bytes(dec_iv_arr);
+
+    let mut stream = AesCtr::new(&dec_key, dec_iv);
+    let keystream = stream.encrypt(&[0u8; HANDSHAKE_LEN]);
+
+    let mut target_plain = [0u8; HANDSHAKE_LEN];
+    target_plain[PROTO_TAG_POS..PROTO_TAG_POS + 4].copy_from_slice(&proto_tag.to_bytes());
+    target_plain[DC_IDX_POS..DC_IDX_POS + 2].copy_from_slice(&dc_idx.to_le_bytes());
+
+    for idx in PROTO_TAG_POS..HANDSHAKE_LEN {
+        handshake[idx] = target_plain[idx] ^ keystream[idx];
+    }
+
+    handshake
+}
+
+#[tokio::test]
+async fn fragmented_tls_mtproto_with_interleaved_ccs_is_accepted() {
+    let secret_hex = "55555555555555555555555555555555";
+    let secret = [0x55u8; 16];
+    let client_hello = make_valid_tls_client_hello(&secret, 0);
+    let mtproto_handshake = make_valid_mtproto_handshake(secret_hex, ProtoTag::Secure, 2);
+
+    let mut cfg = ProxyConfig::default();
+    cfg.general.beobachten = false;
+    cfg.access.ignore_time_skew = true;
+    cfg.access
+        .users
+        .insert("user".to_string(), secret_hex.to_string());
+
+    let config = Arc::new(cfg);
+    let replay_checker = Arc::new(ReplayChecker::new(128, Duration::from_secs(60)));
+    let rng = SecureRandom::new();
+
+    let (server_side, mut client_side) = duplex(131072);
+    let peer: SocketAddr = "198.51.100.85:55007".parse().unwrap();
+    let (read_half, write_half) = tokio::io::split(server_side);
+
+    let (mut tls_reader, tls_writer, tls_user) = match handle_tls_handshake(
+        &client_hello,
+        read_half,
+        write_half,
+        peer,
+        &config,
+        &replay_checker,
+        &rng,
+        None,
+    )
+    .await
+    {
+        HandshakeResult::Success(result) => result,
+        _ => panic!("expected successful TLS handshake"),
+    };
+
+    let mut tls_response_head = [0u8; 5];
+    client_side
+        .read_exact(&mut tls_response_head)
+        .await
+        .unwrap();
+    assert_eq!(tls_response_head[0], 0x16);
+    let tls_response_len =
+        u16::from_be_bytes([tls_response_head[3], tls_response_head[4]]) as usize;
+    let mut tls_response_body = vec![0u8; tls_response_len];
+    client_side
+        .read_exact(&mut tls_response_body)
+        .await
+        .unwrap();
+
+    client_side
+        .write_all(&wrap_tls_application_data(&mtproto_handshake[..13]))
+        .await
+        .unwrap();
+    client_side.write_all(&wrap_tls_ccs_record()).await.unwrap();
+    client_side
+        .write_all(&wrap_tls_application_data(&mtproto_handshake[13..37]))
+        .await
+        .unwrap();
+    client_side.write_all(&wrap_tls_ccs_record()).await.unwrap();
+    client_side
+        .write_all(&wrap_tls_application_data(&mtproto_handshake[37..]))
+        .await
+        .unwrap();
+
+    let mtproto_data = tls_reader.read_exact(HANDSHAKE_LEN).await.unwrap();
+    assert_eq!(&mtproto_data[..], &mtproto_handshake);
+
+    let mtproto_handshake: [u8; HANDSHAKE_LEN] = mtproto_data[..].try_into().unwrap();
+    let (_, _, success) = match handle_mtproto_handshake(
+        &mtproto_handshake,
+        tls_reader,
+        tls_writer,
+        peer,
+        &config,
+        &replay_checker,
+        true,
+        Some(tls_user.as_str()),
+    )
+    .await
+    {
+        HandshakeResult::Success(result) => result,
+        _ => panic!("expected successful MTProto handshake"),
+    };
+
+    assert_eq!(success.user, "user");
+    assert_eq!(success.proto_tag, ProtoTag::Secure);
+    assert_eq!(success.dc_idx, 2);
+}
+
 #[tokio::test]
 async fn valid_tls_path_does_not_fall_back_to_mask_backend() {
     let listener = TcpListener::bind("127.0.0.1:0").await.unwrap();
@@ -1292,6 +1825,7 @@ async fn valid_tls_path_does_not_fall_back_to_mask_backend() {
         1,
         1,
         1,
+        10,
         1,
         false,
         stats.clone(),
@@ -1400,6 +1934,7 @@ async fn valid_tls_with_invalid_mtproto_falls_back_to_mask_backend() {
         1,
         1,
         1,
+        10,
         1,
         false,
         stats.clone(),
@@ -1506,6 +2041,7 @@ async fn client_handler_tls_bad_mtproto_is_forwarded_to_mask_backend() {
         1,
         1,
         1,
+        10,
         1,
         false,
         stats.clone(),
@@ -1627,6 +2163,7 @@ async fn alpn_mismatch_tls_probe_is_masked_through_client_pipeline() {
         1,
         1,
         1,
+        10,
         1,
         false,
         stats.clone(),
@@ -1719,6 +2256,7 @@ async fn invalid_hmac_tls_probe_is_masked_through_client_pipeline() {
         1,
         1,
         1,
+        10,
         1,
         false,
         stats.clone(),
@@ -1817,6 +2355,7 @@ async fn burst_invalid_tls_probes_are_masked_verbatim() {
             1,
             1,
             1,
+            10,
             1,
             false,
             stats.clone(),
@@ -1995,14 +2534,16 @@ async fn tcp_limit_rejection_does_not_reserve_ip_or_trigger_rollback() {
 }
 
 #[tokio::test]
-async fn zero_tcp_limit_rejects_without_ip_or_counter_side_effects() {
+async fn zero_tcp_limit_uses_global_fallback_and_rejects_without_side_effects() {
     let mut config = ProxyConfig::default();
     config
         .access
         .user_max_tcp_conns
         .insert("user".to_string(), 0);
+    config.access.user_max_tcp_conns_global_each = 1;
 
     let stats = Stats::new();
+    stats.increment_user_curr_connects("user");
     let ip_tracker = UserIpTracker::new();
     let peer_addr: SocketAddr = "198.51.100.211:50001".parse().unwrap();
 
@@ -2019,10 +2560,75 @@ async fn zero_tcp_limit_rejects_without_ip_or_counter_side_effects() {
         result,
         Err(ProxyError::ConnectionLimitExceeded { user }) if user == "user"
     ));
+    assert_eq!(
+        stats.get_user_curr_connects("user"),
+        1,
+        "TCP-limit rejection must keep pre-existing in-flight connection count unchanged"
+    );
+    assert_eq!(ip_tracker.get_active_ip_count("user").await, 0);
+}
+
+#[tokio::test]
+async fn zero_tcp_limit_with_disabled_global_fallback_is_unlimited() {
+    let mut config = ProxyConfig::default();
+    config
+        .access
+        .user_max_tcp_conns
+        .insert("user".to_string(), 0);
+    config.access.user_max_tcp_conns_global_each = 0;
+
+    let stats = Stats::new();
+    let ip_tracker = UserIpTracker::new();
+    let peer_addr: SocketAddr = "198.51.100.212:50002".parse().unwrap();
+
+    let result = RunningClientHandler::check_user_limits_static(
+        "user",
+        &config,
+        &stats,
+        peer_addr,
+        &ip_tracker,
+    )
+    .await;
+
+    assert!(
+        result.is_ok(),
+        "per-user zero with global fallback disabled must not enforce a TCP limit"
+    );
     assert_eq!(stats.get_user_curr_connects("user"), 0);
     assert_eq!(ip_tracker.get_active_ip_count("user").await, 0);
 }
 
+#[tokio::test]
+async fn global_tcp_fallback_applies_when_per_user_limit_is_missing() {
+    let mut config = ProxyConfig::default();
+    config.access.user_max_tcp_conns_global_each = 1;
+
+    let stats = Stats::new();
+    stats.increment_user_curr_connects("user");
+    let ip_tracker = UserIpTracker::new();
+    let peer_addr: SocketAddr = "198.51.100.213:50003".parse().unwrap();
+
+    let result = RunningClientHandler::check_user_limits_static(
+        "user",
+        &config,
+        &stats,
+        peer_addr,
+        &ip_tracker,
+    )
+    .await;
+
+    assert!(matches!(
+        result,
+        Err(ProxyError::ConnectionLimitExceeded { user }) if user == "user"
+    ));
+    assert_eq!(
+        stats.get_user_curr_connects("user"),
+        1,
+        "Global fallback TCP-limit rejection must keep pre-existing counter unchanged"
+    );
+    assert_eq!(ip_tracker.get_active_ip_count("user").await, 0);
+}
+
 #[tokio::test]
 async fn check_user_limits_static_success_does_not_leak_counter_or_ip_reservation() {
     let user = "check-helper-user";
@@ -2654,6 +3260,7 @@ async fn relay_connect_error_releases_user_and_ip_before_return() {
         1,
         1,
         1,
+        10,
         1,
         false,
         stats.clone(),
@@ -2841,7 +3448,7 @@ async fn quota_rejection_does_not_reserve_ip_or_trigger_rollback() {
         .insert("user".to_string(), 1024);
 
     let stats = Stats::new();
-    stats.add_user_octets_from("user", 1024);
+    preload_user_quota(&stats, "user", 1024);
 
     let ip_tracker = UserIpTracker::new();
     let peer_addr: SocketAddr = "203.0.113.211:50001".parse().unwrap();
@@ -3214,6 +3821,7 @@ async fn untrusted_proxy_header_source_is_rejected() {
         1,
         1,
         1,
+        10,
         1,
         false,
         stats.clone(),
@@ -3283,6 +3891,7 @@ async fn empty_proxy_trusted_cidrs_rejects_proxy_header_by_default() {
         1,
         1,
         1,
+        10,
         1,
         false,
         stats.clone(),
@@ -3379,6 +3988,7 @@ async fn oversized_tls_record_is_masked_in_generic_stream_pipeline() {
         1,
         1,
         1,
+        10,
         1,
         false,
         stats.clone(),
@@ -3481,6 +4091,7 @@ async fn oversized_tls_record_is_masked_in_client_handler_pipeline() {
         1,
         1,
         1,
+        10,
         1,
         false,
         stats.clone(),
@@ -3597,6 +4208,7 @@ async fn tls_record_len_min_minus_1_is_rejected_in_generic_stream_pipeline() {
         1,
         1,
         1,
+        10,
         1,
         false,
         stats.clone(),
@@ -3699,6 +4311,7 @@ async fn tls_record_len_min_minus_1_is_rejected_in_client_handler_pipeline() {
         1,
         1,
         1,
+        10,
         1,
         false,
         stats.clone(),
@@ -3804,6 +4417,7 @@ async fn tls_record_len_16384_is_accepted_in_generic_stream_pipeline() {
         1,
         1,
         1,
+        10,
         1,
         false,
         stats.clone(),
@@ -3904,6 +4518,7 @@ async fn tls_record_len_16384_is_accepted_in_client_handler_pipeline() {
         1,
         1,
         1,
+        10,
         1,
         false,
         stats.clone(),

src/proxy/tests/client_timing_profile_adversarial_tests.rs

@@ -33,6 +33,7 @@ fn make_test_upstream_manager(stats: Arc<Stats>) -> Arc<UpstreamManager> {
         1,
         1,
         1,
+        10,
         1,
         false,
         stats,

src/proxy/tests/client_tls_clienthello_size_security_tests.rs

@@ -35,6 +35,7 @@ fn make_test_upstream_manager(stats: Arc<Stats>) -> Arc<UpstreamManager> {
         1,
         1,
         1,
+        10,
         1,
         false,
         stats,

src/proxy/tests/client_tls_clienthello_truncation_adversarial_tests.rs

@@ -36,6 +36,7 @@ fn make_test_upstream_manager(stats: Arc<Stats>) -> Arc<UpstreamManager> {
         1,
         1,
         1,
+        10,
         1,
         false,
         stats,

src/proxy/tests/client_tls_mtproto_fallback_security_tests.rs

@@ -50,6 +50,7 @@ fn build_harness(secret_hex: &str, mask_port: u16) -> PipelineHarness {
         1,
         1,
         1,
+        10,
         1,
         false,
         stats.clone(),

src/proxy/tests/client_tls_record_wrap_hardening_security_tests.rs

@@ -25,13 +25,26 @@ fn wrap_tls_application_record_oversized_payload_is_chunked_without_truncation()
         let len = u16::from_be_bytes([record[offset + 3], record[offset + 4]]) as usize;
         let body_start = offset + 5;
         let body_end = body_start + len;
-        assert!(body_end <= record.len(), "declared TLS record length must be in-bounds");
+        assert!(
+            body_end <= record.len(),
+            "declared TLS record length must be in-bounds"
+        );
         recovered.extend_from_slice(&record[body_start..body_end]);
         offset = body_end;
         frames += 1;
     }
 
-    assert_eq!(offset, record.len(), "record parser must consume exact output size");
-    assert_eq!(frames, 2, "oversized payload should split into exactly two records");
-    assert_eq!(recovered, payload, "chunked records must preserve full payload");
+    assert_eq!(
+        offset,
+        record.len(),
+        "record parser must consume exact output size"
+    );
+    assert_eq!(
+        frames, 2,
+        "oversized payload should split into exactly two records"
+    );
+    assert_eq!(
+        recovered, payload,
+        "chunked records must preserve full payload"
+    );
 }

src/proxy/tests/direct_relay_security_tests.rs

@@ -773,8 +773,7 @@ fn anchored_open_nix_path_writes_expected_lines() {
         "target/telemt-unknown-dc-anchored-open-ok-{}/unknown-dc.log",
         std::process::id()
     );
-    let sanitized =
-        sanitize_unknown_dc_log_path(&rel_candidate).expect("candidate must sanitize");
+    let sanitized = sanitize_unknown_dc_log_path(&rel_candidate).expect("candidate must sanitize");
     let _ = fs::remove_file(&sanitized.resolved_path);
 
     let mut first = open_unknown_dc_log_append_anchored(&sanitized)
@@ -787,7 +786,10 @@ fn anchored_open_nix_path_writes_expected_lines() {
 
     let content =
         fs::read_to_string(&sanitized.resolved_path).expect("anchored log file must be readable");
-    let lines: Vec<&str> = content.lines().filter(|line| !line.trim().is_empty()).collect();
+    let lines: Vec<&str> = content
+        .lines()
+        .filter(|line| !line.trim().is_empty())
+        .collect();
     assert_eq!(lines.len(), 2, "expected one line per anchored append call");
     assert!(
         lines.contains(&"dc_idx=31200") && lines.contains(&"dc_idx=31201"),
@@ -811,8 +813,7 @@ fn anchored_open_parallel_appends_preserve_line_integrity() {
         "target/telemt-unknown-dc-anchored-open-parallel-{}/unknown-dc.log",
         std::process::id()
     );
-    let sanitized =
-        sanitize_unknown_dc_log_path(&rel_candidate).expect("candidate must sanitize");
+    let sanitized = sanitize_unknown_dc_log_path(&rel_candidate).expect("candidate must sanitize");
     let _ = fs::remove_file(&sanitized.resolved_path);
 
     let mut workers = Vec::new();
@@ -831,8 +832,15 @@ fn anchored_open_parallel_appends_preserve_line_integrity() {
 
     let content =
         fs::read_to_string(&sanitized.resolved_path).expect("parallel log file must be readable");
-    let lines: Vec<&str> = content.lines().filter(|line| !line.trim().is_empty()).collect();
-    assert_eq!(lines.len(), 64, "expected one complete line per worker append");
+    let lines: Vec<&str> = content
+        .lines()
+        .filter(|line| !line.trim().is_empty())
+        .collect();
+    assert_eq!(
+        lines.len(),
+        64,
+        "expected one complete line per worker append"
+    );
     for line in lines {
         assert!(
             line.starts_with("dc_idx="),
@@ -867,8 +875,7 @@ fn anchored_open_creates_private_0600_file_permissions() {
         "target/telemt-unknown-dc-anchored-perms-{}/unknown-dc.log",
         std::process::id()
     );
-    let sanitized =
-        sanitize_unknown_dc_log_path(&rel_candidate).expect("candidate must sanitize");
+    let sanitized = sanitize_unknown_dc_log_path(&rel_candidate).expect("candidate must sanitize");
     let _ = fs::remove_file(&sanitized.resolved_path);
 
     let mut file = open_unknown_dc_log_append_anchored(&sanitized)
@@ -905,8 +912,7 @@ fn anchored_open_rejects_existing_symlink_target() {
         "target/telemt-unknown-dc-anchored-symlink-target-{}/unknown-dc.log",
         std::process::id()
     );
-    let sanitized =
-        sanitize_unknown_dc_log_path(&rel_candidate).expect("candidate must sanitize");
+    let sanitized = sanitize_unknown_dc_log_path(&rel_candidate).expect("candidate must sanitize");
 
     let outside = std::env::temp_dir().join(format!(
         "telemt-unknown-dc-anchored-symlink-outside-{}.log",
@@ -943,8 +949,7 @@ fn anchored_open_high_contention_multi_write_preserves_complete_lines() {
         "target/telemt-unknown-dc-anchored-contention-{}/unknown-dc.log",
         std::process::id()
     );
-    let sanitized =
-        sanitize_unknown_dc_log_path(&rel_candidate).expect("candidate must sanitize");
+    let sanitized = sanitize_unknown_dc_log_path(&rel_candidate).expect("candidate must sanitize");
     let _ = fs::remove_file(&sanitized.resolved_path);
 
     let workers = 24usize;
@@ -970,7 +975,10 @@ fn anchored_open_high_contention_multi_write_preserves_complete_lines() {
 
     let content = fs::read_to_string(&sanitized.resolved_path)
         .expect("contention output file must be readable");
-    let lines: Vec<&str> = content.lines().filter(|line| !line.trim().is_empty()).collect();
+    let lines: Vec<&str> = content
+        .lines()
+        .filter(|line| !line.trim().is_empty())
+        .collect();
     assert_eq!(
         lines.len(),
         workers * rounds,
@@ -1014,8 +1022,7 @@ fn append_unknown_dc_line_returns_error_for_read_only_descriptor() {
         "target/telemt-unknown-dc-append-ro-{}/unknown-dc.log",
         std::process::id()
     );
-    let sanitized =
-        sanitize_unknown_dc_log_path(&rel_candidate).expect("candidate must sanitize");
+    let sanitized = sanitize_unknown_dc_log_path(&rel_candidate).expect("candidate must sanitize");
     fs::write(&sanitized.resolved_path, "seed\n").expect("seed file must be writable");
 
     let mut readonly = std::fs::OpenOptions::new()
@@ -1295,6 +1302,7 @@ async fn direct_relay_abort_midflight_releases_route_gauge() {
         1,
         1,
         1,
+        10,
         1,
         false,
         stats.clone(),
@@ -1401,6 +1409,7 @@ async fn direct_relay_cutover_midflight_releases_route_gauge() {
         1,
         1,
         1,
+        10,
         1,
         false,
         stats.clone(),
@@ -1522,6 +1531,7 @@ async fn direct_relay_cutover_storm_multi_session_keeps_generic_errors_and_relea
         1,
         1,
         1,
+        10,
         1,
         false,
         stats.clone(),
@@ -1754,6 +1764,7 @@ async fn negative_direct_relay_dc_connection_refused_fails_fast() {
         1,
         100,
         5000,
+        10,
         3,
         false,
         stats.clone(),
@@ -1844,6 +1855,7 @@ async fn adversarial_direct_relay_cutover_integrity() {
         1,
         100,
         5000,
+        10,
         3,
         false,
         stats.clone(),

src/proxy/tests/handshake_advanced_clever_tests.rs

@@ -0,0 +1,713 @@
+use super::*;
+use crate::crypto::{AesCtr, sha256, sha256_hmac};
+use crate::protocol::constants::{ProtoTag, RESERVED_NONCE_BEGINNINGS, RESERVED_NONCE_FIRST_BYTES};
+use std::net::{IpAddr, Ipv4Addr, Ipv6Addr, SocketAddr};
+use std::sync::Arc;
+use std::time::{Duration, Instant};
+
+// --- Helpers ---
+
+fn test_config_with_secret_hex(secret_hex: &str) -> ProxyConfig {
+    let mut cfg = ProxyConfig::default();
+    cfg.access.users.clear();
+    cfg.access
+        .users
+        .insert("user".to_string(), secret_hex.to_string());
+    cfg.access.ignore_time_skew = true;
+    cfg.general.modes.secure = true;
+    cfg.general.modes.classic = true;
+    cfg.general.modes.tls = true;
+    cfg
+}
+
+fn make_valid_tls_handshake(secret: &[u8], timestamp: u32) -> Vec<u8> {
+    let session_id_len: usize = 32;
+    let len = tls::TLS_DIGEST_POS + tls::TLS_DIGEST_LEN + 1 + session_id_len;
+    let mut handshake = vec![0x42u8; len];
+
+    handshake[tls::TLS_DIGEST_POS + tls::TLS_DIGEST_LEN] = session_id_len as u8;
+    handshake[tls::TLS_DIGEST_POS..tls::TLS_DIGEST_POS + tls::TLS_DIGEST_LEN].fill(0);
+
+    let computed = sha256_hmac(secret, &handshake);
+    let mut digest = computed;
+    let ts = timestamp.to_le_bytes();
+    for i in 0..4 {
+        digest[28 + i] ^= ts[i];
+    }
+
+    handshake[tls::TLS_DIGEST_POS..tls::TLS_DIGEST_POS + tls::TLS_DIGEST_LEN]
+        .copy_from_slice(&digest);
+    handshake
+}
+
+fn make_valid_mtproto_handshake(
+    secret_hex: &str,
+    proto_tag: ProtoTag,
+    dc_idx: i16,
+) -> [u8; HANDSHAKE_LEN] {
+    let secret = hex::decode(secret_hex).expect("secret hex must decode");
+    let mut handshake = [0x5Au8; HANDSHAKE_LEN];
+    for (idx, b) in handshake[SKIP_LEN..SKIP_LEN + PREKEY_LEN + IV_LEN]
+        .iter_mut()
+        .enumerate()
+    {
+        *b = (idx as u8).wrapping_add(1);
+    }
+
+    let dec_prekey = &handshake[SKIP_LEN..SKIP_LEN + PREKEY_LEN];
+    let dec_iv_bytes = &handshake[SKIP_LEN + PREKEY_LEN..SKIP_LEN + PREKEY_LEN + IV_LEN];
+
+    let mut dec_key_input = Vec::with_capacity(PREKEY_LEN + secret.len());
+    dec_key_input.extend_from_slice(dec_prekey);
+    dec_key_input.extend_from_slice(&secret);
+    let dec_key = sha256(&dec_key_input);
+
+    let mut dec_iv_arr = [0u8; IV_LEN];
+    dec_iv_arr.copy_from_slice(dec_iv_bytes);
+    let dec_iv = u128::from_be_bytes(dec_iv_arr);
+
+    let mut stream = AesCtr::new(&dec_key, dec_iv);
+    let keystream = stream.encrypt(&[0u8; HANDSHAKE_LEN]);
+
+    let mut target_plain = [0u8; HANDSHAKE_LEN];
+    target_plain[PROTO_TAG_POS..PROTO_TAG_POS + 4].copy_from_slice(&proto_tag.to_bytes());
+    target_plain[DC_IDX_POS..DC_IDX_POS + 2].copy_from_slice(&dc_idx.to_le_bytes());
+
+    for idx in PROTO_TAG_POS..HANDSHAKE_LEN {
+        handshake[idx] = target_plain[idx] ^ keystream[idx];
+    }
+
+    handshake
+}
+
+fn make_valid_tls_client_hello_with_alpn(
+    secret: &[u8],
+    timestamp: u32,
+    alpn_protocols: &[&[u8]],
+) -> Vec<u8> {
+    let mut body = Vec::new();
+    body.extend_from_slice(&TLS_VERSION);
+    body.extend_from_slice(&[0u8; 32]);
+    body.push(32);
+    body.extend_from_slice(&[0x42u8; 32]);
+    body.extend_from_slice(&2u16.to_be_bytes());
+    body.extend_from_slice(&[0x13, 0x01]);
+    body.push(1);
+    body.push(0);
+
+    let mut ext_blob = Vec::new();
+    if !alpn_protocols.is_empty() {
+        let mut alpn_list = Vec::new();
+        for proto in alpn_protocols {
+            alpn_list.push(proto.len() as u8);
+            alpn_list.extend_from_slice(proto);
+        }
+        let mut alpn_data = Vec::new();
+        alpn_data.extend_from_slice(&(alpn_list.len() as u16).to_be_bytes());
+        alpn_data.extend_from_slice(&alpn_list);
+
+        ext_blob.extend_from_slice(&0x0010u16.to_be_bytes());
+        ext_blob.extend_from_slice(&(alpn_data.len() as u16).to_be_bytes());
+        ext_blob.extend_from_slice(&alpn_data);
+    }
+    body.extend_from_slice(&(ext_blob.len() as u16).to_be_bytes());
+    body.extend_from_slice(&ext_blob);
+
+    let mut handshake = Vec::new();
+    handshake.push(0x01);
+    let body_len = (body.len() as u32).to_be_bytes();
+    handshake.extend_from_slice(&body_len[1..4]);
+    handshake.extend_from_slice(&body);
+
+    let mut record = Vec::new();
+    record.push(TLS_RECORD_HANDSHAKE);
+    record.extend_from_slice(&[0x03, 0x01]);
+    record.extend_from_slice(&(handshake.len() as u16).to_be_bytes());
+    record.extend_from_slice(&handshake);
+
+    record[tls::TLS_DIGEST_POS..tls::TLS_DIGEST_POS + tls::TLS_DIGEST_LEN].fill(0);
+    let computed = sha256_hmac(secret, &record);
+    let mut digest = computed;
+    let ts = timestamp.to_le_bytes();
+    for i in 0..4 {
+        digest[28 + i] ^= ts[i];
+    }
+    record[tls::TLS_DIGEST_POS..tls::TLS_DIGEST_POS + tls::TLS_DIGEST_LEN].copy_from_slice(&digest);
+
+    record
+}
+
+// --- Category 1: Edge Cases & Protocol Boundaries ---
+
+#[tokio::test]
+async fn tls_minimum_viable_length_boundary() {
+    let shared = ProxySharedState::new();
+    clear_auth_probe_state_for_testing_in_shared(shared.as_ref());
+
+    let secret = [0x11u8; 16];
+    let config = test_config_with_secret_hex("11111111111111111111111111111111");
+    let replay_checker = ReplayChecker::new(128, Duration::from_secs(60));
+    let rng = SecureRandom::new();
+    let peer: SocketAddr = "192.0.2.1:12345".parse().unwrap();
+
+    let min_len = tls::TLS_DIGEST_POS + tls::TLS_DIGEST_LEN + 1;
+    let mut exact_min_handshake = vec![0x42u8; min_len];
+    exact_min_handshake[min_len - 1] = 0;
+    exact_min_handshake[tls::TLS_DIGEST_POS..tls::TLS_DIGEST_POS + tls::TLS_DIGEST_LEN].fill(0);
+    let digest = sha256_hmac(&secret, &exact_min_handshake);
+    exact_min_handshake[tls::TLS_DIGEST_POS..tls::TLS_DIGEST_POS + tls::TLS_DIGEST_LEN]
+        .copy_from_slice(&digest);
+
+    let res = handle_tls_handshake(
+        &exact_min_handshake,
+        tokio::io::empty(),
+        tokio::io::sink(),
+        peer,
+        &config,
+        &replay_checker,
+        &rng,
+        None,
+    )
+    .await;
+    assert!(
+        matches!(res, HandshakeResult::Success(_)),
+        "Exact minimum length TLS handshake must succeed"
+    );
+
+    let short_handshake = vec![0x42u8; min_len - 1];
+    let res_short = handle_tls_handshake(
+        &short_handshake,
+        tokio::io::empty(),
+        tokio::io::sink(),
+        peer,
+        &config,
+        &replay_checker,
+        &rng,
+        None,
+    )
+    .await;
+    assert!(
+        matches!(res_short, HandshakeResult::BadClient { .. }),
+        "Handshake 1 byte shorter than minimum must fail closed"
+    );
+}
+
+#[tokio::test]
+async fn mtproto_extreme_dc_index_serialization() {
+    let shared = ProxySharedState::new();
+    clear_auth_probe_state_for_testing_in_shared(shared.as_ref());
+
+    let secret_hex = "22222222222222222222222222222222";
+    let config = test_config_with_secret_hex(secret_hex);
+    for (idx, extreme_dc) in [i16::MIN, i16::MAX, -1, 0].into_iter().enumerate() {
+        // Keep replay state independent per case so we validate dc_idx encoding,
+        // not duplicate-handshake rejection behavior.
+        let replay_checker = ReplayChecker::new(128, Duration::from_secs(60));
+        let peer = SocketAddr::new(IpAddr::V4(Ipv4Addr::new(192, 0, 2, 2)), 12345 + idx as u16);
+        let handshake = make_valid_mtproto_handshake(secret_hex, ProtoTag::Secure, extreme_dc);
+        let res = handle_mtproto_handshake(
+            &handshake,
+            tokio::io::empty(),
+            tokio::io::sink(),
+            peer,
+            &config,
+            &replay_checker,
+            false,
+            None,
+        )
+        .await;
+
+        match res {
+            HandshakeResult::Success((_, _, success)) => {
+                assert_eq!(
+                    success.dc_idx, extreme_dc,
+                    "Extreme DC index {} must serialize/deserialize perfectly",
+                    extreme_dc
+                );
+            }
+            _ => panic!(
+                "MTProto handshake with extreme DC index {} failed",
+                extreme_dc
+            ),
+        }
+    }
+}
+
+#[tokio::test]
+async fn alpn_strict_case_and_padding_rejection() {
+    let shared = ProxySharedState::new();
+    clear_auth_probe_state_for_testing_in_shared(shared.as_ref());
+
+    let secret = [0x33u8; 16];
+    let mut config = test_config_with_secret_hex("33333333333333333333333333333333");
+    config.censorship.alpn_enforce = true;
+    let replay_checker = ReplayChecker::new(128, Duration::from_secs(60));
+    let rng = SecureRandom::new();
+    let peer: SocketAddr = "192.0.2.3:12345".parse().unwrap();
+
+    let bad_alpns: &[&[u8]] = &[b"H2", b"h2\0", b" http/1.1", b"http/1.1\n"];
+
+    for bad_alpn in bad_alpns {
+        let handshake = make_valid_tls_client_hello_with_alpn(&secret, 0, &[*bad_alpn]);
+        let res = handle_tls_handshake(
+            &handshake,
+            tokio::io::empty(),
+            tokio::io::sink(),
+            peer,
+            &config,
+            &replay_checker,
+            &rng,
+            None,
+        )
+        .await;
+        assert!(
+            matches!(res, HandshakeResult::BadClient { .. }),
+            "ALPN strict enforcement must reject {:?}",
+            bad_alpn
+        );
+    }
+}
+
+#[test]
+fn ipv4_mapped_ipv6_bucketing_anomaly() {
+    let ipv4_mapped_1 = IpAddr::V6(Ipv6Addr::new(0, 0, 0, 0, 0, 0xffff, 0xc000, 0x0201));
+    let ipv4_mapped_2 = IpAddr::V6(Ipv6Addr::new(0, 0, 0, 0, 0, 0xffff, 0xc633, 0x6402));
+
+    let norm_1 = normalize_auth_probe_ip(ipv4_mapped_1);
+    let norm_2 = normalize_auth_probe_ip(ipv4_mapped_2);
+
+    assert_eq!(
+        norm_1, norm_2,
+        "IPv4-mapped IPv6 addresses must collapse into the same /64 bucket (::0)"
+    );
+    assert_eq!(
+        norm_1,
+        IpAddr::V6(Ipv6Addr::new(0, 0, 0, 0, 0, 0, 0, 0)),
+        "The bucket must be exactly ::0"
+    );
+}
+
+// --- Category 2: Adversarial & Black Hat ---
+
+#[tokio::test]
+async fn mtproto_invalid_ciphertext_does_not_poison_replay_cache() {
+    let shared = ProxySharedState::new();
+    clear_auth_probe_state_for_testing_in_shared(shared.as_ref());
+
+    let secret_hex = "55555555555555555555555555555555";
+    let config = test_config_with_secret_hex(secret_hex);
+    let replay_checker = ReplayChecker::new(128, Duration::from_secs(60));
+    let peer: SocketAddr = "192.0.2.5:12345".parse().unwrap();
+
+    let valid_handshake = make_valid_mtproto_handshake(secret_hex, ProtoTag::Secure, 1);
+    let mut invalid_handshake = valid_handshake;
+    invalid_handshake[SKIP_LEN + PREKEY_LEN + IV_LEN + 1] ^= 0xFF;
+
+    let res_invalid = handle_mtproto_handshake(
+        &invalid_handshake,
+        tokio::io::empty(),
+        tokio::io::sink(),
+        peer,
+        &config,
+        &replay_checker,
+        false,
+        None,
+    )
+    .await;
+    assert!(matches!(res_invalid, HandshakeResult::BadClient { .. }));
+
+    let res_valid = handle_mtproto_handshake(
+        &valid_handshake,
+        tokio::io::empty(),
+        tokio::io::sink(),
+        peer,
+        &config,
+        &replay_checker,
+        false,
+        None,
+    )
+    .await;
+    assert!(
+        matches!(res_valid, HandshakeResult::Success(_)),
+        "Invalid MTProto ciphertext must not poison the replay cache"
+    );
+}
+
+#[tokio::test]
+async fn tls_invalid_session_does_not_poison_replay_cache() {
+    let shared = ProxySharedState::new();
+    clear_auth_probe_state_for_testing_in_shared(shared.as_ref());
+
+    let secret = [0x66u8; 16];
+    let config = test_config_with_secret_hex("66666666666666666666666666666666");
+    let replay_checker = ReplayChecker::new(128, Duration::from_secs(60));
+    let rng = SecureRandom::new();
+    let peer: SocketAddr = "192.0.2.6:12345".parse().unwrap();
+
+    let valid_handshake = make_valid_tls_handshake(&secret, 0);
+    let mut invalid_handshake = valid_handshake.clone();
+    let session_idx = tls::TLS_DIGEST_POS + tls::TLS_DIGEST_LEN + 1;
+    invalid_handshake[session_idx] ^= 0xFF;
+
+    let res_invalid = handle_tls_handshake(
+        &invalid_handshake,
+        tokio::io::empty(),
+        tokio::io::sink(),
+        peer,
+        &config,
+        &replay_checker,
+        &rng,
+        None,
+    )
+    .await;
+    assert!(matches!(res_invalid, HandshakeResult::BadClient { .. }));
+
+    let res_valid = handle_tls_handshake(
+        &valid_handshake,
+        tokio::io::empty(),
+        tokio::io::sink(),
+        peer,
+        &config,
+        &replay_checker,
+        &rng,
+        None,
+    )
+    .await;
+    assert!(
+        matches!(res_valid, HandshakeResult::Success(_)),
+        "Invalid TLS payload must not poison the replay cache"
+    );
+}
+
+#[tokio::test]
+async fn server_hello_delay_timing_neutrality_on_hmac_failure() {
+    let shared = ProxySharedState::new();
+    clear_auth_probe_state_for_testing_in_shared(shared.as_ref());
+
+    let secret = [0x77u8; 16];
+    let mut config = test_config_with_secret_hex("77777777777777777777777777777777");
+    config.censorship.server_hello_delay_min_ms = 50;
+    config.censorship.server_hello_delay_max_ms = 50;
+
+    let replay_checker = ReplayChecker::new(128, Duration::from_secs(60));
+    let rng = SecureRandom::new();
+    let peer: SocketAddr = "192.0.2.7:12345".parse().unwrap();
+
+    let mut invalid_handshake = make_valid_tls_handshake(&secret, 0);
+    invalid_handshake[tls::TLS_DIGEST_POS] ^= 0xFF;
+
+    let start = Instant::now();
+    let res = handle_tls_handshake(
+        &invalid_handshake,
+        tokio::io::empty(),
+        tokio::io::sink(),
+        peer,
+        &config,
+        &replay_checker,
+        &rng,
+        None,
+    )
+    .await;
+    let elapsed = start.elapsed();
+
+    assert!(matches!(res, HandshakeResult::BadClient { .. }));
+    assert!(
+        elapsed >= Duration::from_millis(45),
+        "Invalid HMAC must still incur the configured ServerHello delay to prevent timing side-channels"
+    );
+}
+
+#[tokio::test]
+async fn server_hello_delay_inversion_resilience() {
+    let shared = ProxySharedState::new();
+    clear_auth_probe_state_for_testing_in_shared(shared.as_ref());
+
+    let secret = [0x88u8; 16];
+    let mut config = test_config_with_secret_hex("88888888888888888888888888888888");
+    config.censorship.server_hello_delay_min_ms = 100;
+    config.censorship.server_hello_delay_max_ms = 10;
+
+    let replay_checker = ReplayChecker::new(128, Duration::from_secs(60));
+    let rng = SecureRandom::new();
+    let peer: SocketAddr = "192.0.2.8:12345".parse().unwrap();
+
+    let valid_handshake = make_valid_tls_handshake(&secret, 0);
+
+    let start = Instant::now();
+    let res = handle_tls_handshake(
+        &valid_handshake,
+        tokio::io::empty(),
+        tokio::io::sink(),
+        peer,
+        &config,
+        &replay_checker,
+        &rng,
+        None,
+    )
+    .await;
+    let elapsed = start.elapsed();
+
+    assert!(matches!(res, HandshakeResult::Success(_)));
+    assert!(
+        elapsed >= Duration::from_millis(90),
+        "Delay logic must gracefully handle min > max inversions via max.max(min)"
+    );
+}
+
+#[tokio::test]
+async fn mixed_valid_and_invalid_user_secrets_configuration() {
+    let shared = ProxySharedState::new();
+    clear_auth_probe_state_for_testing_in_shared(shared.as_ref());
+    clear_warned_secrets_for_testing_in_shared(shared.as_ref());
+
+    let mut config = ProxyConfig::default();
+    config.access.ignore_time_skew = true;
+
+    for i in 0..9 {
+        let bad_secret = if i % 2 == 0 { "badhex!" } else { "1122" };
+        config
+            .access
+            .users
+            .insert(format!("bad_user_{}", i), bad_secret.to_string());
+    }
+    let valid_secret_hex = "99999999999999999999999999999999";
+    config
+        .access
+        .users
+        .insert("good_user".to_string(), valid_secret_hex.to_string());
+    config.general.modes.secure = true;
+    config.general.modes.classic = true;
+    config.general.modes.tls = true;
+
+    let secret = [0x99u8; 16];
+    let replay_checker = ReplayChecker::new(128, Duration::from_secs(60));
+    let rng = SecureRandom::new();
+    let peer: SocketAddr = "192.0.2.9:12345".parse().unwrap();
+
+    let valid_handshake = make_valid_tls_handshake(&secret, 0);
+
+    let res = handle_tls_handshake(
+        &valid_handshake,
+        tokio::io::empty(),
+        tokio::io::sink(),
+        peer,
+        &config,
+        &replay_checker,
+        &rng,
+        None,
+    )
+    .await;
+
+    assert!(
+        matches!(res, HandshakeResult::Success(_)),
+        "Proxy must gracefully skip invalid secrets and authenticate the valid one"
+    );
+}
+
+#[tokio::test]
+async fn tls_emulation_fallback_when_cache_missing() {
+    let shared = ProxySharedState::new();
+    clear_auth_probe_state_for_testing_in_shared(shared.as_ref());
+
+    let secret = [0xAAu8; 16];
+    let mut config = test_config_with_secret_hex("aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa");
+    config.censorship.tls_emulation = true;
+    config.general.modes.tls = true;
+
+    let replay_checker = ReplayChecker::new(128, Duration::from_secs(60));
+    let rng = SecureRandom::new();
+    let peer: SocketAddr = "192.0.2.10:12345".parse().unwrap();
+
+    let valid_handshake = make_valid_tls_handshake(&secret, 0);
+
+    let res = handle_tls_handshake(
+        &valid_handshake,
+        tokio::io::empty(),
+        tokio::io::sink(),
+        peer,
+        &config,
+        &replay_checker,
+        &rng,
+        None,
+    )
+    .await;
+
+    assert!(
+        matches!(res, HandshakeResult::Success(_)),
+        "TLS emulation must gracefully fall back to standard ServerHello if cache is missing"
+    );
+}
+
+#[tokio::test]
+async fn classic_mode_over_tls_transport_protocol_confusion() {
+    let shared = ProxySharedState::new();
+    clear_auth_probe_state_for_testing_in_shared(shared.as_ref());
+
+    let secret_hex = "bbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbb";
+    let mut config = test_config_with_secret_hex(secret_hex);
+    config.general.modes.classic = true;
+    config.general.modes.tls = true;
+
+    let replay_checker = ReplayChecker::new(128, Duration::from_secs(60));
+    let peer: SocketAddr = "192.0.2.11:12345".parse().unwrap();
+
+    let handshake = make_valid_mtproto_handshake(secret_hex, ProtoTag::Intermediate, 1);
+
+    let res = handle_mtproto_handshake(
+        &handshake,
+        tokio::io::empty(),
+        tokio::io::sink(),
+        peer,
+        &config,
+        &replay_checker,
+        true,
+        None,
+    )
+    .await;
+
+    assert!(
+        matches!(res, HandshakeResult::Success(_)),
+        "Intermediate tag over TLS must succeed if classic mode is enabled, locking in cross-transport behavior"
+    );
+}
+
+#[test]
+fn generate_tg_nonce_never_emits_reserved_bytes() {
+    let client_enc_key = [0xCCu8; 32];
+    let client_enc_iv = 123456789u128;
+    let rng = SecureRandom::new();
+
+    for _ in 0..10_000 {
+        let (nonce, _, _, _, _) = generate_tg_nonce(
+            ProtoTag::Secure,
+            1,
+            &client_enc_key,
+            client_enc_iv,
+            &rng,
+            false,
+        );
+
+        assert!(
+            !RESERVED_NONCE_FIRST_BYTES.contains(&nonce[0]),
+            "Nonce must never start with reserved bytes"
+        );
+        let first_four: [u8; 4] = [nonce[0], nonce[1], nonce[2], nonce[3]];
+        assert!(
+            !RESERVED_NONCE_BEGINNINGS.contains(&first_four),
+            "Nonce must never match reserved 4-byte beginnings"
+        );
+    }
+}
+
+#[tokio::test(flavor = "multi_thread", worker_threads = 4)]
+async fn dashmap_concurrent_saturation_stress() {
+    let shared = ProxySharedState::new();
+    clear_auth_probe_state_for_testing_in_shared(shared.as_ref());
+
+    let ip_a: IpAddr = "192.0.2.13".parse().unwrap();
+    let ip_b: IpAddr = "198.51.100.13".parse().unwrap();
+    let mut tasks = Vec::new();
+
+    for i in 0..100 {
+        let target_ip = if i % 2 == 0 { ip_a } else { ip_b };
+        let shared = shared.clone();
+        tasks.push(tokio::spawn(async move {
+            for _ in 0..50 {
+                auth_probe_record_failure_in(shared.as_ref(), target_ip, Instant::now());
+            }
+        }));
+    }
+
+    for task in tasks {
+        task.await
+            .expect("Task panicked during concurrent DashMap stress");
+    }
+
+    assert!(
+        auth_probe_is_throttled_for_testing_in_shared(shared.as_ref(), ip_a),
+        "IP A must be throttled after concurrent stress"
+    );
+    assert!(
+        auth_probe_is_throttled_for_testing_in_shared(shared.as_ref(), ip_b),
+        "IP B must be throttled after concurrent stress"
+    );
+}
+
+#[test]
+fn prototag_invalid_bytes_fail_closed() {
+    let invalid_tags: [[u8; 4]; 5] = [
+        [0, 0, 0, 0],
+        [0xFF, 0xFF, 0xFF, 0xFF],
+        [0xDE, 0xAD, 0xBE, 0xEF],
+        [0xDD, 0xDD, 0xDD, 0xDE],
+        [0x11, 0x22, 0x33, 0x44],
+    ];
+
+    for tag in invalid_tags {
+        assert_eq!(
+            ProtoTag::from_bytes(tag),
+            None,
+            "Invalid ProtoTag bytes {:?} must fail closed",
+            tag
+        );
+    }
+}
+
+#[test]
+fn auth_probe_eviction_hash_collision_stress() {
+    let shared = ProxySharedState::new();
+    clear_auth_probe_state_for_testing_in_shared(shared.as_ref());
+
+    let state = auth_probe_state_for_testing_in_shared(shared.as_ref());
+    let now = Instant::now();
+
+    for i in 0..10_000u32 {
+        let ip = IpAddr::V4(Ipv4Addr::new(10, 0, (i >> 8) as u8, (i & 0xFF) as u8));
+        auth_probe_record_failure_with_state_in(shared.as_ref(), state, ip, now);
+    }
+
+    assert!(
+        state.len() <= AUTH_PROBE_TRACK_MAX_ENTRIES,
+        "Eviction logic must successfully bound the map size under heavy insertion stress"
+    );
+}
+
+#[test]
+fn encrypt_tg_nonce_with_ciphers_advances_counter_correctly() {
+    let client_enc_key = [0xDDu8; 32];
+    let client_enc_iv = 987654321u128;
+    let rng = SecureRandom::new();
+
+    let (nonce, _, _, _, _) = generate_tg_nonce(
+        ProtoTag::Secure,
+        2,
+        &client_enc_key,
+        client_enc_iv,
+        &rng,
+        false,
+    );
+
+    let (_, mut returned_encryptor, _) = encrypt_tg_nonce_with_ciphers(&nonce);
+    let zeros = [0u8; 64];
+    let returned_keystream = returned_encryptor.encrypt(&zeros);
+
+    let enc_key_iv = &nonce[SKIP_LEN..SKIP_LEN + KEY_LEN + IV_LEN];
+    let mut expected_enc_key = [0u8; 32];
+    expected_enc_key.copy_from_slice(&enc_key_iv[..KEY_LEN]);
+    let mut expected_enc_iv_arr = [0u8; IV_LEN];
+    expected_enc_iv_arr.copy_from_slice(&enc_key_iv[KEY_LEN..]);
+    let expected_enc_iv = u128::from_be_bytes(expected_enc_iv_arr);
+
+    let mut manual_encryptor = AesCtr::new(&expected_enc_key, expected_enc_iv);
+
+    let mut manual_input = Vec::new();
+    manual_input.extend_from_slice(&nonce);
+    manual_input.extend_from_slice(&zeros);
+    let manual_output = manual_encryptor.encrypt(&manual_input);
+
+    assert_eq!(
+        returned_keystream,
+        &manual_output[64..128],
+        "encrypt_tg_nonce_with_ciphers must correctly advance the AES-CTR counter by exactly the nonce length"
+    );
+}

src/proxy/tests/handshake_adversarial_tests.rs

@@ -44,12 +44,6 @@ fn make_valid_mtproto_handshake(
     handshake
 }
 
-fn auth_probe_test_guard() -> std::sync::MutexGuard<'static, ()> {
-    auth_probe_test_lock()
-        .lock()
-        .unwrap_or_else(|poisoned| poisoned.into_inner())
-}
-
 fn test_config_with_secret_hex(secret_hex: &str) -> ProxyConfig {
     let mut cfg = ProxyConfig::default();
     cfg.access.users.clear();
@@ -67,8 +61,8 @@ fn test_config_with_secret_hex(secret_hex: &str) -> ProxyConfig {
 
 #[tokio::test]
 async fn mtproto_handshake_bit_flip_anywhere_rejected() {
-    let _guard = auth_probe_test_guard();
-    clear_auth_probe_state_for_testing();
+    let shared = ProxySharedState::new();
+    clear_auth_probe_state_for_testing_in_shared(shared.as_ref());
 
     let secret_hex = "11223344556677889900aabbccddeeff";
     let base = make_valid_mtproto_handshake(secret_hex, ProtoTag::Secure, 2);
@@ -181,26 +175,26 @@ async fn mtproto_handshake_timing_neutrality_mocked() {
 
 #[tokio::test]
 async fn auth_probe_throttle_saturation_stress() {
-    let _guard = auth_probe_test_guard();
-    clear_auth_probe_state_for_testing();
+    let shared = ProxySharedState::new();
+    clear_auth_probe_state_for_testing_in_shared(shared.as_ref());
 
     let now = Instant::now();
 
     // Record enough failures for one IP to trigger backoff
     let target_ip = IpAddr::V4(Ipv4Addr::new(1, 1, 1, 1));
     for _ in 0..AUTH_PROBE_BACKOFF_START_FAILS {
-        auth_probe_record_failure(target_ip, now);
+        auth_probe_record_failure_in(shared.as_ref(), target_ip, now);
     }
 
-    assert!(auth_probe_is_throttled(target_ip, now));
+    assert!(auth_probe_is_throttled_in(shared.as_ref(), target_ip, now));
 
     // Stress test with many unique IPs
     for i in 0..500u32 {
         let ip = IpAddr::V4(Ipv4Addr::new(203, 0, 113, (i % 256) as u8));
-        auth_probe_record_failure(ip, now);
+        auth_probe_record_failure_in(shared.as_ref(), ip, now);
     }
 
-    let tracked = AUTH_PROBE_STATE.get().map(|state| state.len()).unwrap_or(0);
+    let tracked = auth_probe_state_for_testing_in_shared(shared.as_ref()).len();
     assert!(
         tracked <= AUTH_PROBE_TRACK_MAX_ENTRIES,
         "auth probe state grew past hard cap: {tracked} > {AUTH_PROBE_TRACK_MAX_ENTRIES}"
@@ -209,8 +203,8 @@ async fn auth_probe_throttle_saturation_stress() {
 
 #[tokio::test]
 async fn mtproto_handshake_abridged_prefix_rejected() {
-    let _guard = auth_probe_test_guard();
-    clear_auth_probe_state_for_testing();
+    let shared = ProxySharedState::new();
+    clear_auth_probe_state_for_testing_in_shared(shared.as_ref());
 
     let mut handshake = [0x5Au8; HANDSHAKE_LEN];
     handshake[0] = 0xef; // Abridged prefix
@@ -235,8 +229,8 @@ async fn mtproto_handshake_abridged_prefix_rejected() {
 
 #[tokio::test]
 async fn mtproto_handshake_preferred_user_mismatch_continues() {
-    let _guard = auth_probe_test_guard();
-    clear_auth_probe_state_for_testing();
+    let shared = ProxySharedState::new();
+    clear_auth_probe_state_for_testing_in_shared(shared.as_ref());
 
     let secret1_hex = "11111111111111111111111111111111";
     let secret2_hex = "22222222222222222222222222222222";
@@ -278,8 +272,8 @@ async fn mtproto_handshake_preferred_user_mismatch_continues() {
 
 #[tokio::test]
 async fn mtproto_handshake_concurrent_flood_stability() {
-    let _guard = auth_probe_test_guard();
-    clear_auth_probe_state_for_testing();
+    let shared = ProxySharedState::new();
+    clear_auth_probe_state_for_testing_in_shared(shared.as_ref());
 
     let secret_hex = "00112233445566778899aabbccddeeff";
     let base = make_valid_mtproto_handshake(secret_hex, ProtoTag::Secure, 1);
@@ -320,8 +314,8 @@ async fn mtproto_handshake_concurrent_flood_stability() {
 
 #[tokio::test]
 async fn mtproto_replay_is_rejected_across_distinct_peers() {
-    let _guard = auth_probe_test_guard();
-    clear_auth_probe_state_for_testing();
+    let shared = ProxySharedState::new();
+    clear_auth_probe_state_for_testing_in_shared(shared.as_ref());
 
     let secret_hex = "0123456789abcdeffedcba9876543210";
     let handshake = make_valid_mtproto_handshake(secret_hex, ProtoTag::Secure, 2);
@@ -360,8 +354,8 @@ async fn mtproto_replay_is_rejected_across_distinct_peers() {
 
 #[tokio::test]
 async fn mtproto_blackhat_mutation_corpus_never_panics_and_stays_fail_closed() {
-    let _guard = auth_probe_test_guard();
-    clear_auth_probe_state_for_testing();
+    let shared = ProxySharedState::new();
+    clear_auth_probe_state_for_testing_in_shared(shared.as_ref());
 
     let secret_hex = "89abcdef012345670123456789abcdef";
     let base = make_valid_mtproto_handshake(secret_hex, ProtoTag::Secure, 2);
@@ -405,27 +399,27 @@ async fn mtproto_blackhat_mutation_corpus_never_panics_and_stays_fail_closed() {
 
 #[tokio::test]
 async fn auth_probe_success_clears_throttled_peer_state() {
-    let _guard = auth_probe_test_guard();
-    clear_auth_probe_state_for_testing();
+    let shared = ProxySharedState::new();
+    clear_auth_probe_state_for_testing_in_shared(shared.as_ref());
 
     let target_ip = IpAddr::V4(Ipv4Addr::new(203, 0, 113, 90));
     let now = Instant::now();
     for _ in 0..AUTH_PROBE_BACKOFF_START_FAILS {
-        auth_probe_record_failure(target_ip, now);
+        auth_probe_record_failure_in(shared.as_ref(), target_ip, now);
     }
-    assert!(auth_probe_is_throttled(target_ip, now));
+    assert!(auth_probe_is_throttled_in(shared.as_ref(), target_ip, now));
 
-    auth_probe_record_success(target_ip);
+    auth_probe_record_success_in(shared.as_ref(), target_ip);
     assert!(
-        !auth_probe_is_throttled(target_ip, now + Duration::from_millis(1)),
+        !auth_probe_is_throttled_in(shared.as_ref(), target_ip, now + Duration::from_millis(1)),
         "successful auth must clear per-peer throttle state"
     );
 }
 
 #[tokio::test]
 async fn mtproto_invalid_storm_over_cap_keeps_probe_map_hard_bounded() {
-    let _guard = auth_probe_test_guard();
-    clear_auth_probe_state_for_testing();
+    let shared = ProxySharedState::new();
+    clear_auth_probe_state_for_testing_in_shared(shared.as_ref());
 
     let secret_hex = "00112233445566778899aabbccddeeff";
     let mut invalid = make_valid_mtproto_handshake(secret_hex, ProtoTag::Secure, 2);
@@ -458,7 +452,7 @@ async fn mtproto_invalid_storm_over_cap_keeps_probe_map_hard_bounded() {
         assert!(matches!(res, HandshakeResult::BadClient { .. }));
     }
 
-    let tracked = AUTH_PROBE_STATE.get().map(|state| state.len()).unwrap_or(0);
+    let tracked = auth_probe_state_for_testing_in_shared(shared.as_ref()).len();
     assert!(
         tracked <= AUTH_PROBE_TRACK_MAX_ENTRIES,
         "probe map must remain bounded under invalid storm: {tracked}"
@@ -467,8 +461,8 @@ async fn mtproto_invalid_storm_over_cap_keeps_probe_map_hard_bounded() {
 
 #[tokio::test]
 async fn mtproto_property_style_multi_bit_mutations_fail_closed_or_auth_only() {
-    let _guard = auth_probe_test_guard();
-    clear_auth_probe_state_for_testing();
+    let shared = ProxySharedState::new();
+    clear_auth_probe_state_for_testing_in_shared(shared.as_ref());
 
     let secret_hex = "f0e1d2c3b4a5968778695a4b3c2d1e0f";
     let base = make_valid_mtproto_handshake(secret_hex, ProtoTag::Secure, 2);
@@ -520,8 +514,8 @@ async fn mtproto_property_style_multi_bit_mutations_fail_closed_or_auth_only() {
 #[tokio::test]
 #[ignore = "heavy soak; run manually"]
 async fn mtproto_blackhat_20k_mutation_soak_never_panics() {
-    let _guard = auth_probe_test_guard();
-    clear_auth_probe_state_for_testing();
+    let shared = ProxySharedState::new();
+    clear_auth_probe_state_for_testing_in_shared(shared.as_ref());
 
     let secret_hex = "aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa";
     let base = make_valid_mtproto_handshake(secret_hex, ProtoTag::Secure, 2);

src/proxy/tests/handshake_auth_probe_eviction_bias_security_tests.rs

@@ -0,0 +1,93 @@
+use super::*;
+use std::collections::HashSet;
+use std::net::{IpAddr, Ipv4Addr};
+use std::time::{Duration, Instant};
+
+#[test]
+fn adversarial_large_state_offsets_escape_first_scan_window() {
+    let shared = ProxySharedState::new();
+    let base = Instant::now();
+    let state_len = 65_536usize;
+    let scan_limit = 1_024usize;
+
+    let mut saw_offset_outside_first_window = false;
+    for i in 0..8_192u64 {
+        let ip = IpAddr::V4(Ipv4Addr::new(
+            ((i >> 16) & 0xff) as u8,
+            ((i >> 8) & 0xff) as u8,
+            (i & 0xff) as u8,
+            ((i.wrapping_mul(131)) & 0xff) as u8,
+        ));
+        let now = base + Duration::from_nanos(i);
+        let start =
+            auth_probe_scan_start_offset_in(shared.as_ref(), ip, now, state_len, scan_limit);
+        if start >= scan_limit {
+            saw_offset_outside_first_window = true;
+            break;
+        }
+    }
+
+    assert!(
+        saw_offset_outside_first_window,
+        "scan start offset must cover the full auth-probe state, not only the first scan window"
+    );
+}
+
+#[test]
+fn stress_large_state_offsets_cover_many_scan_windows() {
+    let shared = ProxySharedState::new();
+    let base = Instant::now();
+    let state_len = 65_536usize;
+    let scan_limit = 1_024usize;
+
+    let mut covered_windows = HashSet::new();
+    for i in 0..16_384u64 {
+        let ip = IpAddr::V4(Ipv4Addr::new(
+            ((i >> 16) & 0xff) as u8,
+            ((i >> 8) & 0xff) as u8,
+            (i & 0xff) as u8,
+            ((i.wrapping_mul(17)) & 0xff) as u8,
+        ));
+        let now = base + Duration::from_micros(i);
+        let start =
+            auth_probe_scan_start_offset_in(shared.as_ref(), ip, now, state_len, scan_limit);
+        covered_windows.insert(start / scan_limit);
+    }
+
+    assert!(
+        covered_windows.len() >= 16,
+        "eviction scan must not collapse to a tiny hot zone; covered windows={} out of {}",
+        covered_windows.len(),
+        state_len / scan_limit
+    );
+}
+
+#[test]
+fn light_fuzz_offset_always_stays_inside_state_len() {
+    let shared = ProxySharedState::new();
+    let mut seed = 0xC0FF_EE12_3456_789Au64;
+    let base = Instant::now();
+
+    for _ in 0..8_192usize {
+        seed ^= seed << 7;
+        seed ^= seed >> 9;
+        seed ^= seed << 8;
+
+        let ip = IpAddr::V4(Ipv4Addr::new(
+            (seed >> 24) as u8,
+            (seed >> 16) as u8,
+            (seed >> 8) as u8,
+            seed as u8,
+        ));
+        let state_len = ((seed >> 16) as usize % 200_000).saturating_add(1);
+        let scan_limit = ((seed >> 40) as usize % 2_048).saturating_add(1);
+        let now = base + Duration::from_nanos(seed & 0x0fff);
+        let start =
+            auth_probe_scan_start_offset_in(shared.as_ref(), ip, now, state_len, scan_limit);
+
+        assert!(
+            start < state_len,
+            "scan offset must stay inside state length"
+        );
+    }
+}

src/proxy/tests/handshake_auth_probe_hardening_adversarial_tests.rs

@@ -2,68 +2,62 @@ use super::*;
 use std::net::{IpAddr, Ipv4Addr};
 use std::time::{Duration, Instant};
 
-fn auth_probe_test_guard() -> std::sync::MutexGuard<'static, ()> {
-    auth_probe_test_lock()
-        .lock()
-        .unwrap_or_else(|poisoned| poisoned.into_inner())
-}
-
 #[test]
 fn positive_preauth_throttle_activates_after_failure_threshold() {
-    let _guard = auth_probe_test_guard();
-    clear_auth_probe_state_for_testing();
+    let shared = ProxySharedState::new();
+    clear_auth_probe_state_for_testing_in_shared(shared.as_ref());
 
     let ip = IpAddr::V4(Ipv4Addr::new(198, 51, 100, 20));
     let now = Instant::now();
 
     for _ in 0..AUTH_PROBE_BACKOFF_START_FAILS {
-        auth_probe_record_failure(ip, now);
+        auth_probe_record_failure_in(shared.as_ref(), ip, now);
     }
 
     assert!(
-        auth_probe_is_throttled(ip, now),
+        auth_probe_is_throttled_in(shared.as_ref(), ip, now),
         "peer must be throttled once fail streak reaches threshold"
     );
 }
 
 #[test]
 fn negative_unrelated_peer_remains_unthrottled() {
-    let _guard = auth_probe_test_guard();
-    clear_auth_probe_state_for_testing();
+    let shared = ProxySharedState::new();
+    clear_auth_probe_state_for_testing_in_shared(shared.as_ref());
 
     let attacker = IpAddr::V4(Ipv4Addr::new(203, 0, 113, 12));
     let benign = IpAddr::V4(Ipv4Addr::new(203, 0, 113, 13));
     let now = Instant::now();
 
     for _ in 0..AUTH_PROBE_BACKOFF_START_FAILS {
-        auth_probe_record_failure(attacker, now);
+        auth_probe_record_failure_in(shared.as_ref(), attacker, now);
     }
 
-    assert!(auth_probe_is_throttled(attacker, now));
+    assert!(auth_probe_is_throttled_in(shared.as_ref(), attacker, now));
     assert!(
-        !auth_probe_is_throttled(benign, now),
+        !auth_probe_is_throttled_in(shared.as_ref(), benign, now),
         "throttle state must stay scoped to normalized peer key"
     );
 }
 
 #[test]
 fn edge_expired_entry_is_pruned_and_no_longer_throttled() {
-    let _guard = auth_probe_test_guard();
-    clear_auth_probe_state_for_testing();
+    let shared = ProxySharedState::new();
+    clear_auth_probe_state_for_testing_in_shared(shared.as_ref());
 
     let ip = IpAddr::V4(Ipv4Addr::new(192, 0, 2, 41));
     let base = Instant::now();
     for _ in 0..AUTH_PROBE_BACKOFF_START_FAILS {
-        auth_probe_record_failure(ip, base);
+        auth_probe_record_failure_in(shared.as_ref(), ip, base);
     }
 
     let expired_at = base + Duration::from_secs(AUTH_PROBE_TRACK_RETENTION_SECS + 1);
     assert!(
-        !auth_probe_is_throttled(ip, expired_at),
+        !auth_probe_is_throttled_in(shared.as_ref(), ip, expired_at),
         "expired entries must not keep throttling peers"
     );
 
-    let state = auth_probe_state_map();
+    let state = auth_probe_state_for_testing_in_shared(shared.as_ref());
     assert!(
         state.get(&normalize_auth_probe_ip(ip)).is_none(),
         "expired lookup should prune stale state"
@@ -72,36 +66,40 @@ fn edge_expired_entry_is_pruned_and_no_longer_throttled() {
 
 #[test]
 fn adversarial_saturation_grace_requires_extra_failures_before_preauth_throttle() {
-    let _guard = auth_probe_test_guard();
-    clear_auth_probe_state_for_testing();
+    let shared = ProxySharedState::new();
+    clear_auth_probe_state_for_testing_in_shared(shared.as_ref());
 
     let ip = IpAddr::V4(Ipv4Addr::new(198, 18, 0, 7));
     let now = Instant::now();
 
     for _ in 0..AUTH_PROBE_BACKOFF_START_FAILS {
-        auth_probe_record_failure(ip, now);
+        auth_probe_record_failure_in(shared.as_ref(), ip, now);
     }
-    auth_probe_note_saturation(now);
+    auth_probe_note_saturation_in(shared.as_ref(), now);
 
     assert!(
-        !auth_probe_should_apply_preauth_throttle(ip, now),
+        !auth_probe_should_apply_preauth_throttle_in(shared.as_ref(), ip, now),
         "during global saturation, peer must receive configured grace window"
     );
 
     for _ in 0..AUTH_PROBE_SATURATION_GRACE_FAILS {
-        auth_probe_record_failure(ip, now + Duration::from_millis(1));
+        auth_probe_record_failure_in(shared.as_ref(), ip, now + Duration::from_millis(1));
     }
 
     assert!(
-        auth_probe_should_apply_preauth_throttle(ip, now + Duration::from_millis(1)),
+        auth_probe_should_apply_preauth_throttle_in(
+            shared.as_ref(),
+            ip,
+            now + Duration::from_millis(1)
+        ),
         "after grace failures are exhausted, preauth throttle must activate"
     );
 }
 
 #[test]
 fn integration_over_cap_insertion_keeps_probe_map_bounded() {
-    let _guard = auth_probe_test_guard();
-    clear_auth_probe_state_for_testing();
+    let shared = ProxySharedState::new();
+    clear_auth_probe_state_for_testing_in_shared(shared.as_ref());
 
     let now = Instant::now();
     for idx in 0..(AUTH_PROBE_TRACK_MAX_ENTRIES + 1024) {
@@ -111,10 +109,10 @@ fn integration_over_cap_insertion_keeps_probe_map_bounded() {
             ((idx / 256) % 256) as u8,
             (idx % 256) as u8,
         ));
-        auth_probe_record_failure(ip, now);
+        auth_probe_record_failure_in(shared.as_ref(), ip, now);
     }
 
-    let tracked = auth_probe_state_map().len();
+    let tracked = auth_probe_state_for_testing_in_shared(shared.as_ref()).len();
     assert!(
         tracked <= AUTH_PROBE_TRACK_MAX_ENTRIES,
         "probe map must remain hard bounded under insertion storm"
@@ -123,8 +121,8 @@ fn integration_over_cap_insertion_keeps_probe_map_bounded() {
 
 #[test]
 fn light_fuzz_randomized_failures_preserve_cap_and_nonzero_streaks() {
-    let _guard = auth_probe_test_guard();
-    clear_auth_probe_state_for_testing();
+    let shared = ProxySharedState::new();
+    clear_auth_probe_state_for_testing_in_shared(shared.as_ref());
 
     let mut seed = 0x4D53_5854_6F66_6175u64;
     let now = Instant::now();
@@ -140,10 +138,14 @@ fn light_fuzz_randomized_failures_preserve_cap_and_nonzero_streaks() {
             (seed >> 8) as u8,
             seed as u8,
         ));
-        auth_probe_record_failure(ip, now + Duration::from_millis((seed & 0x3f) as u64));
+        auth_probe_record_failure_in(
+            shared.as_ref(),
+            ip,
+            now + Duration::from_millis((seed & 0x3f) as u64),
+        );
     }
 
-    let state = auth_probe_state_map();
+    let state = auth_probe_state_for_testing_in_shared(shared.as_ref());
     assert!(state.len() <= AUTH_PROBE_TRACK_MAX_ENTRIES);
     for entry in state.iter() {
         assert!(entry.value().fail_streak > 0);
@@ -152,13 +154,14 @@ fn light_fuzz_randomized_failures_preserve_cap_and_nonzero_streaks() {
 
 #[tokio::test(flavor = "multi_thread", worker_threads = 4)]
 async fn stress_parallel_failure_flood_keeps_state_hard_capped() {
-    let _guard = auth_probe_test_guard();
-    clear_auth_probe_state_for_testing();
+    let shared = ProxySharedState::new();
+    clear_auth_probe_state_for_testing_in_shared(shared.as_ref());
 
     let start = Instant::now();
     let mut tasks = Vec::new();
 
     for worker in 0..8u8 {
+        let shared = shared.clone();
         tasks.push(tokio::spawn(async move {
             for i in 0..4096u32 {
                 let ip = IpAddr::V4(Ipv4Addr::new(
@@ -167,7 +170,11 @@ async fn stress_parallel_failure_flood_keeps_state_hard_capped() {
                     ((i >> 8) & 0xff) as u8,
                     (i & 0xff) as u8,
                 ));
-                auth_probe_record_failure(ip, start + Duration::from_millis((i % 4) as u64));
+                auth_probe_record_failure_in(
+                    shared.as_ref(),
+                    ip,
+                    start + Duration::from_millis((i % 4) as u64),
+                );
             }
         }));
     }
@@ -176,12 +183,12 @@ async fn stress_parallel_failure_flood_keeps_state_hard_capped() {
         task.await.expect("stress worker must not panic");
     }
 
-    let tracked = auth_probe_state_map().len();
+    let tracked = auth_probe_state_for_testing_in_shared(shared.as_ref()).len();
     assert!(
         tracked <= AUTH_PROBE_TRACK_MAX_ENTRIES,
         "parallel failure flood must not exceed cap"
     );
 
     let probe = IpAddr::V4(Ipv4Addr::new(172, 3, 4, 5));
-    let _ = auth_probe_is_throttled(probe, start + Duration::from_millis(2));
+    let _ = auth_probe_is_throttled_in(shared.as_ref(), probe, start + Duration::from_millis(2));
 }

src/proxy/tests/handshake_auth_probe_scan_budget_security_tests.rs

@@ -0,0 +1,95 @@
+use super::*;
+use std::net::{IpAddr, Ipv4Addr};
+use std::time::{Duration, Instant};
+
+#[test]
+fn edge_zero_state_len_yields_zero_start_offset() {
+    let shared = ProxySharedState::new();
+    let ip = IpAddr::V4(Ipv4Addr::new(198, 51, 100, 44));
+    let now = Instant::now();
+
+    assert_eq!(
+        auth_probe_scan_start_offset_in(shared.as_ref(), ip, now, 0, 16),
+        0,
+        "empty map must not produce non-zero scan offset"
+    );
+}
+
+#[test]
+fn adversarial_large_state_must_allow_start_offset_outside_scan_budget_window() {
+    let shared = ProxySharedState::new();
+    let base = Instant::now();
+    let scan_limit = 16usize;
+    let state_len = 65_536usize;
+
+    let mut saw_offset_outside_window = false;
+    for i in 0..2048u32 {
+        let ip = IpAddr::V4(Ipv4Addr::new(
+            203,
+            ((i >> 16) & 0xff) as u8,
+            ((i >> 8) & 0xff) as u8,
+            (i & 0xff) as u8,
+        ));
+        let now = base + Duration::from_micros(i as u64);
+        let start =
+            auth_probe_scan_start_offset_in(shared.as_ref(), ip, now, state_len, scan_limit);
+        assert!(
+            start < state_len,
+            "start offset must stay within state length; start={start}, len={state_len}"
+        );
+        if start >= scan_limit {
+            saw_offset_outside_window = true;
+            break;
+        }
+    }
+
+    assert!(
+        saw_offset_outside_window,
+        "large-state eviction must sample beyond the first scan window"
+    );
+}
+
+#[test]
+fn positive_state_smaller_than_scan_limit_caps_to_state_len() {
+    let shared = ProxySharedState::new();
+    let ip = IpAddr::V4(Ipv4Addr::new(192, 0, 2, 17));
+    let now = Instant::now();
+
+    for state_len in 1..32usize {
+        let start = auth_probe_scan_start_offset_in(shared.as_ref(), ip, now, state_len, 64);
+        assert!(
+            start < state_len,
+            "start offset must never exceed state length when scan limit is larger"
+        );
+    }
+}
+
+#[test]
+fn light_fuzz_scan_offset_budget_never_exceeds_effective_window() {
+    let shared = ProxySharedState::new();
+    let mut seed = 0x5A41_5356_4C32_3236u64;
+    let base = Instant::now();
+
+    for _ in 0..4096 {
+        seed ^= seed << 7;
+        seed ^= seed >> 9;
+        seed ^= seed << 8;
+
+        let ip = IpAddr::V4(Ipv4Addr::new(
+            (seed >> 24) as u8,
+            (seed >> 16) as u8,
+            (seed >> 8) as u8,
+            seed as u8,
+        ));
+        let state_len = ((seed >> 8) as usize % 131_072).saturating_add(1);
+        let scan_limit = ((seed >> 32) as usize % 512).saturating_add(1);
+        let now = base + Duration::from_nanos(seed & 0xffff);
+        let start =
+            auth_probe_scan_start_offset_in(shared.as_ref(), ip, now, state_len, scan_limit);
+
+        assert!(
+            start < state_len,
+            "scan offset must stay inside state length"
+        );
+    }
+}

src/proxy/tests/handshake_auth_probe_scan_offset_stress_tests.rs

@@ -0,0 +1,127 @@
+use super::*;
+use std::collections::HashSet;
+use std::net::{IpAddr, Ipv4Addr};
+use std::time::{Duration, Instant};
+
+#[test]
+fn positive_same_ip_moving_time_yields_diverse_scan_offsets() {
+    let shared = ProxySharedState::new();
+    let ip = IpAddr::V4(Ipv4Addr::new(198, 51, 100, 77));
+    let base = Instant::now();
+    let mut uniq = HashSet::new();
+
+    for i in 0..512u64 {
+        let now = base + Duration::from_nanos(i);
+        let offset = auth_probe_scan_start_offset_in(shared.as_ref(), ip, now, 65_536, 16);
+        uniq.insert(offset);
+    }
+
+    assert!(
+        uniq.len() >= 256,
+        "offset randomization collapsed unexpectedly for same-ip moving-time samples (uniq={})",
+        uniq.len()
+    );
+}
+
+#[test]
+fn adversarial_many_ips_same_time_spreads_offsets_without_bias_collapse() {
+    let shared = ProxySharedState::new();
+    let now = Instant::now();
+    let mut uniq = HashSet::new();
+
+    for i in 0..1024u32 {
+        let ip = IpAddr::V4(Ipv4Addr::new(
+            (i >> 16) as u8,
+            (i >> 8) as u8,
+            i as u8,
+            (255 - (i as u8)),
+        ));
+        uniq.insert(auth_probe_scan_start_offset_in(
+            shared.as_ref(),
+            ip,
+            now,
+            65_536,
+            16,
+        ));
+    }
+
+    assert!(
+        uniq.len() >= 512,
+        "scan offset distribution collapsed unexpectedly across adversarial peer set (uniq={})",
+        uniq.len()
+    );
+}
+
+#[tokio::test(flavor = "multi_thread", worker_threads = 4)]
+async fn stress_parallel_failure_churn_under_saturation_remains_capped_and_live() {
+    let shared = ProxySharedState::new();
+    clear_auth_probe_state_for_testing_in_shared(shared.as_ref());
+
+    let start = Instant::now();
+    let mut workers = Vec::new();
+    for worker in 0..8u8 {
+        let shared = shared.clone();
+        workers.push(tokio::spawn(async move {
+            for i in 0..8192u32 {
+                let ip = IpAddr::V4(Ipv4Addr::new(
+                    10,
+                    worker,
+                    ((i >> 8) & 0xff) as u8,
+                    (i & 0xff) as u8,
+                ));
+                auth_probe_record_failure_in(
+                    shared.as_ref(),
+                    ip,
+                    start + Duration::from_micros((i % 128) as u64),
+                );
+            }
+        }));
+    }
+
+    for worker in workers {
+        worker.await.expect("saturation worker must not panic");
+    }
+
+    assert!(
+        auth_probe_state_for_testing_in_shared(shared.as_ref()).len()
+            <= AUTH_PROBE_TRACK_MAX_ENTRIES,
+        "state must remain hard-capped under parallel saturation churn"
+    );
+
+    let probe = IpAddr::V4(Ipv4Addr::new(10, 4, 1, 1));
+    let _ = auth_probe_should_apply_preauth_throttle_in(
+        shared.as_ref(),
+        probe,
+        start + Duration::from_millis(1),
+    );
+}
+
+#[test]
+fn light_fuzz_scan_offset_stays_within_window_for_randomized_inputs() {
+    let shared = ProxySharedState::new();
+    let mut seed = 0xA55A_1357_2468_9BDFu64;
+    let base = Instant::now();
+
+    for _ in 0..8192 {
+        seed ^= seed << 7;
+        seed ^= seed >> 9;
+        seed ^= seed << 8;
+
+        let ip = IpAddr::V4(Ipv4Addr::new(
+            (seed >> 24) as u8,
+            (seed >> 16) as u8,
+            (seed >> 8) as u8,
+            seed as u8,
+        ));
+        let state_len = ((seed >> 8) as usize % 200_000).saturating_add(1);
+        let scan_limit = ((seed >> 40) as usize % 1024).saturating_add(1);
+        let now = base + Duration::from_nanos(seed & 0x1fff);
+
+        let offset =
+            auth_probe_scan_start_offset_in(shared.as_ref(), ip, now, state_len, scan_limit);
+        assert!(
+            offset < state_len,
+            "scan offset must always remain inside state length"
+        );
+    }
+}

src/proxy/tests/handshake_baseline_invariant_tests.rs

@@ -0,0 +1,237 @@
+use super::*;
+use crate::crypto::sha256_hmac;
+use crate::stats::ReplayChecker;
+use std::net::{IpAddr, Ipv4Addr, SocketAddr};
+use std::time::{Duration, Instant};
+use tokio::time::timeout;
+
+fn test_config_with_secret_hex(secret_hex: &str) -> ProxyConfig {
+    let mut cfg = ProxyConfig::default();
+    cfg.access.users.clear();
+    cfg.access
+        .users
+        .insert("user".to_string(), secret_hex.to_string());
+    cfg.access.ignore_time_skew = true;
+    cfg.censorship.mask = true;
+    cfg
+}
+
+fn make_valid_tls_handshake(secret: &[u8], timestamp: u32) -> Vec<u8> {
+    let session_id_len: usize = 32;
+    let len = tls::TLS_DIGEST_POS + tls::TLS_DIGEST_LEN + 1 + session_id_len;
+    let mut handshake = vec![0x42u8; len];
+
+    handshake[tls::TLS_DIGEST_POS + tls::TLS_DIGEST_LEN] = session_id_len as u8;
+    handshake[tls::TLS_DIGEST_POS..tls::TLS_DIGEST_POS + tls::TLS_DIGEST_LEN].fill(0);
+
+    let computed = sha256_hmac(secret, &handshake);
+    let mut digest = computed;
+    let ts = timestamp.to_le_bytes();
+    for i in 0..4 {
+        digest[28 + i] ^= ts[i];
+    }
+
+    handshake[tls::TLS_DIGEST_POS..tls::TLS_DIGEST_POS + tls::TLS_DIGEST_LEN]
+        .copy_from_slice(&digest);
+    handshake
+}
+
+#[tokio::test]
+async fn handshake_baseline_probe_always_falls_back_to_masking() {
+    let shared = ProxySharedState::new();
+    clear_auth_probe_state_for_testing_in_shared(shared.as_ref());
+
+    let cfg = test_config_with_secret_hex("11111111111111111111111111111111");
+    let replay_checker = ReplayChecker::new(64, Duration::from_secs(60));
+    let rng = SecureRandom::new();
+    let peer: SocketAddr = "198.51.100.210:44321".parse().unwrap();
+
+    let probe = b"not-a-tls-clienthello";
+    let res = handle_tls_handshake(
+        probe,
+        tokio::io::empty(),
+        tokio::io::sink(),
+        peer,
+        &cfg,
+        &replay_checker,
+        &rng,
+        None,
+    )
+    .await;
+
+    assert!(matches!(res, HandshakeResult::BadClient { .. }));
+}
+
+#[tokio::test]
+async fn handshake_baseline_invalid_secret_triggers_fallback_not_error_response() {
+    let shared = ProxySharedState::new();
+    clear_auth_probe_state_for_testing_in_shared(shared.as_ref());
+
+    let good_secret = [0x22u8; 16];
+    let bad_cfg = test_config_with_secret_hex("33333333333333333333333333333333");
+    let replay_checker = ReplayChecker::new(64, Duration::from_secs(60));
+    let rng = SecureRandom::new();
+    let peer: SocketAddr = "198.51.100.211:44322".parse().unwrap();
+
+    let handshake = make_valid_tls_handshake(&good_secret, 0);
+    let res = handle_tls_handshake(
+        &handshake,
+        tokio::io::empty(),
+        tokio::io::sink(),
+        peer,
+        &bad_cfg,
+        &replay_checker,
+        &rng,
+        None,
+    )
+    .await;
+
+    assert!(matches!(res, HandshakeResult::BadClient { .. }));
+}
+
+#[tokio::test]
+async fn handshake_baseline_auth_probe_streak_increments_per_ip() {
+    let shared = ProxySharedState::new();
+    clear_auth_probe_state_for_testing_in_shared(shared.as_ref());
+
+    let cfg = test_config_with_secret_hex("44444444444444444444444444444444");
+    let replay_checker = ReplayChecker::new(64, Duration::from_secs(60));
+    let rng = SecureRandom::new();
+
+    let peer: SocketAddr = "203.0.113.10:5555".parse().unwrap();
+    let untouched_ip = IpAddr::V4(Ipv4Addr::new(203, 0, 113, 11));
+    let bad_probe = b"\x16\x03\x01\x00";
+
+    for expected in 1..=3 {
+        let res = handle_tls_handshake_with_shared(
+            bad_probe,
+            tokio::io::empty(),
+            tokio::io::sink(),
+            peer,
+            &cfg,
+            &replay_checker,
+            &rng,
+            None,
+            shared.as_ref(),
+        )
+        .await;
+        assert!(matches!(res, HandshakeResult::BadClient { .. }));
+        assert_eq!(
+            auth_probe_fail_streak_for_testing_in_shared(shared.as_ref(), peer.ip()),
+            Some(expected)
+        );
+        assert_eq!(
+            auth_probe_fail_streak_for_testing_in_shared(shared.as_ref(), untouched_ip),
+            None
+        );
+    }
+}
+
+#[test]
+fn handshake_baseline_saturation_fires_at_compile_time_threshold() {
+    let shared = ProxySharedState::new();
+    clear_auth_probe_state_for_testing_in_shared(shared.as_ref());
+
+    let ip = IpAddr::V4(Ipv4Addr::new(198, 51, 100, 33));
+    let now = Instant::now();
+
+    for _ in 0..AUTH_PROBE_BACKOFF_START_FAILS.saturating_sub(1) {
+        auth_probe_record_failure_in(shared.as_ref(), ip, now);
+    }
+    assert!(!auth_probe_is_throttled_in(shared.as_ref(), ip, now));
+
+    auth_probe_record_failure_in(shared.as_ref(), ip, now);
+    assert!(auth_probe_is_throttled_in(shared.as_ref(), ip, now));
+}
+
+#[test]
+fn handshake_baseline_repeated_probes_streak_monotonic() {
+    let shared = ProxySharedState::new();
+    clear_auth_probe_state_for_testing_in_shared(shared.as_ref());
+
+    let ip = IpAddr::V4(Ipv4Addr::new(203, 0, 113, 42));
+    let now = Instant::now();
+    let mut prev = 0u32;
+
+    for _ in 0..100 {
+        auth_probe_record_failure_in(shared.as_ref(), ip, now);
+        let current =
+            auth_probe_fail_streak_for_testing_in_shared(shared.as_ref(), ip).unwrap_or(0);
+        assert!(current >= prev, "streak must be monotonic");
+        prev = current;
+    }
+}
+
+#[test]
+fn handshake_baseline_throttled_ip_incurs_backoff_delay() {
+    let shared = ProxySharedState::new();
+    clear_auth_probe_state_for_testing_in_shared(shared.as_ref());
+
+    let ip = IpAddr::V4(Ipv4Addr::new(198, 51, 100, 44));
+    let now = Instant::now();
+
+    for _ in 0..AUTH_PROBE_BACKOFF_START_FAILS {
+        auth_probe_record_failure_in(shared.as_ref(), ip, now);
+    }
+
+    let delay = auth_probe_backoff(AUTH_PROBE_BACKOFF_START_FAILS);
+    assert!(delay >= Duration::from_millis(AUTH_PROBE_BACKOFF_BASE_MS));
+
+    let before_expiry = now + delay.saturating_sub(Duration::from_millis(1));
+    let after_expiry = now + delay + Duration::from_millis(1);
+
+    assert!(auth_probe_is_throttled_in(
+        shared.as_ref(),
+        ip,
+        before_expiry
+    ));
+    assert!(!auth_probe_is_throttled_in(
+        shared.as_ref(),
+        ip,
+        after_expiry
+    ));
+}
+
+#[tokio::test]
+async fn handshake_baseline_malformed_probe_frames_fail_closed_to_masking() {
+    let shared = ProxySharedState::new();
+    clear_auth_probe_state_for_testing_in_shared(shared.as_ref());
+
+    let cfg = test_config_with_secret_hex("55555555555555555555555555555555");
+    let replay_checker = ReplayChecker::new(64, Duration::from_secs(60));
+    let rng = SecureRandom::new();
+    let peer: SocketAddr = "198.51.100.212:44323".parse().unwrap();
+
+    let corpus: Vec<Vec<u8>> = vec![
+        vec![0x16, 0x03, 0x01],
+        vec![0x16, 0x03, 0x01, 0xFF, 0xFF],
+        vec![0x00; 128],
+        (0..64u8).collect(),
+    ];
+
+    for probe in corpus {
+        let res = timeout(
+            Duration::from_millis(250),
+            handle_tls_handshake(
+                &probe,
+                tokio::io::empty(),
+                tokio::io::sink(),
+                peer,
+                &cfg,
+                &replay_checker,
+                &rng,
+                None,
+            ),
+        )
+        .await
+        .expect("malformed probe handling must complete in bounded time");
+
+        assert!(
+            matches!(
+                res,
+                HandshakeResult::BadClient { .. } | HandshakeResult::Error(_)
+            ),
+            "malformed probe must fail closed"
+        );
+    }
+}

src/proxy/tests/handshake_fuzz_security_tests.rs

@@ -67,16 +67,10 @@ fn test_config_with_secret_hex(secret_hex: &str) -> ProxyConfig {
     cfg
 }
 
-fn auth_probe_test_guard() -> MutexGuard<'static, ()> {
-    auth_probe_test_lock()
-        .lock()
-        .unwrap_or_else(|poisoned| poisoned.into_inner())
-}
-
 #[tokio::test]
 async fn mtproto_handshake_duplicate_digest_is_replayed_on_second_attempt() {
-    let _guard = auth_probe_test_guard();
-    clear_auth_probe_state_for_testing();
+    let shared = ProxySharedState::new();
+    clear_auth_probe_state_for_testing_in_shared(shared.as_ref());
 
     let secret_hex = "11223344556677889900aabbccddeeff";
     let base = make_valid_mtproto_handshake(secret_hex, ProtoTag::Secure, 2);
@@ -110,13 +104,13 @@ async fn mtproto_handshake_duplicate_digest_is_replayed_on_second_attempt() {
     .await;
     assert!(matches!(second, HandshakeResult::BadClient { .. }));
 
-    clear_auth_probe_state_for_testing();
+    clear_auth_probe_state_for_testing_in_shared(shared.as_ref());
 }
 
 #[tokio::test]
 async fn mtproto_handshake_fuzz_corpus_never_panics_and_stays_fail_closed() {
-    let _guard = auth_probe_test_guard();
-    clear_auth_probe_state_for_testing();
+    let shared = ProxySharedState::new();
+    clear_auth_probe_state_for_testing_in_shared(shared.as_ref());
 
     let secret_hex = "00112233445566778899aabbccddeeff";
     let base = make_valid_mtproto_handshake(secret_hex, ProtoTag::Secure, 1);
@@ -178,13 +172,13 @@ async fn mtproto_handshake_fuzz_corpus_never_panics_and_stays_fail_closed() {
         );
     }
 
-    clear_auth_probe_state_for_testing();
+    clear_auth_probe_state_for_testing_in_shared(shared.as_ref());
 }
 
 #[tokio::test]
 async fn mtproto_handshake_mixed_corpus_never_panics_and_exact_duplicates_are_rejected() {
-    let _guard = auth_probe_test_guard();
-    clear_auth_probe_state_for_testing();
+    let shared = ProxySharedState::new();
+    clear_auth_probe_state_for_testing_in_shared(shared.as_ref());
 
     let secret_hex = "99887766554433221100ffeeddccbbaa";
     let base = make_valid_mtproto_handshake(secret_hex, ProtoTag::Secure, 4);
@@ -274,5 +268,5 @@ async fn mtproto_handshake_mixed_corpus_never_panics_and_exact_duplicates_are_re
         );
     }
 
-    clear_auth_probe_state_for_testing();
+    clear_auth_probe_state_for_testing_in_shared(shared.as_ref());
 }