Merge branch 'main' into tosa_dialect_pad

Author: oscarandersson8218

.claude/skills/binary-size/SKILL.md

@@ -0,0 +1,72 @@
+---
+name: binary-size
+description: Analyze and reduce ExecuTorch binary size. Use when investigating binary size, running size tests, or optimizing the runtime for size-constrained deployments.
+---
+
+# Binary Size
+
+## Start from the `main` branch of executorch
+Ask the user where the executorch repo is.
+
+```bash
+git checkout main && git pull
+```
+
+## Build and measure baseline
+```bash
+conda activate executorch
+bash test/build_size_test.sh
+strip -o /tmp/size_test_stripped cmake-out/test/size_test
+strip -o /tmp/size_test_all_ops_stripped cmake-out/test/size_test_all_ops
+ls -la /tmp/size_test_stripped /tmp/size_test_all_ops_stripped
+```
+
+Produces two binaries:
+- `cmake-out/test/size_test` — ExecuTorch runtime without operator implementations
+- `cmake-out/test/size_test_all_ops` — ExecuTorch runtime with portable ops
+
+## Analyze with bloaty
+```bash
+bloaty cmake-out/test/size_test -d symbols -n 30   # by symbol
+bloaty cmake-out/test/size_test -d sections        # by ELF section
+bloaty <after> -- <before>                          # diff two builds
+nm -S <binary> | sort -k2 -rn | head -30           # symbol sizes
+strings <binary> | less                             # string literals in .rodata
+```
+
+Note: `bloaty -d compileunits` requires debug info (`-g`). The Release build does not include it.
+
+## Key build flags
+Set by `test/build_size_test.sh`:
+- `CMAKE_BUILD_TYPE=Release`
+- `EXECUTORCH_OPTIMIZE_SIZE=ON` — enables `-Os`, `-fno-exceptions`, `-fno-rtti`, unwind table suppression
+- `CXXFLAGS="-fno-exceptions -fno-rtti -Wall -Werror"`
+
+## Constraints
+- Use **CMake** to build (not Buck)
+- **C++17 minimum** language standard
+- Must build on **GCC 9** (CI uses `executorch-ubuntu-22.04-gcc9-nopytorch`) and **Clang 12** — avoid compiler-specific flags or pragmas without version guards
+- Do not regress existing functionality — run tests for modified files
+- Do not change build flags in `build_size_test.sh` for size reductions
+- Do not increase latency in the core runtime
+
+## Where to look for size reductions
+- `.text`: look for large functions, template bloat, duplicate instantiations
+- `.rodata`: verbose error messages, format strings, embedded file paths (`__FILE__`)
+- `.eh_frame`: should already be suppressed when `EXECUTORCH_OPTIMIZE_SIZE=ON`
+- Static init functions (`nm -S <binary> | grep GLOBAL__sub_I`): use `constexpr` constructors to constant-initialize static arrays
+- Logging strings: `ET_LOG_ENABLED=0` in Release eliminates format strings; ensure it propagates to consumers via `PUBLIC` compile definitions on cmake targets
+- Inline header functions: watch for compile-define mismatches between library and consumer TUs (e.g. `ET_LOG_ENABLED` set in library but not in consumer)
+
+## For each change
+1. Create a branch: `git checkout -b binary-size-<N>`
+2. Implement, rebuild, measure stripped sizes
+3. Create a separate PR — one logical change per PR
+4. Record results in `binary-size-<N>.md`:
+
+| Binary | This change (N vs N-1) | Cumulative (N vs main) |
+|---|---|---|
+| `size_test` (stripped) | -X | -Y |
+| `size_test_all_ops` (stripped) | -X | -Y |
+
+5. Update the CI size threshold in `.github/workflows/pull.yml` if sizes decrease

.flake8

@@ -74,6 +74,7 @@ exclude =
     ./build,
     ./configurations,
     ./docs,
+    ./exir/_serialize/generated/executorch_flatbuffer,
     ./third_party,
     *.pyi
 

.github/workflows/pull.yml

@@ -475,10 +475,8 @@ jobs:
         output=$(ls -la cmake-out/test/size_test)
         arr=($output)
         size=${arr[4]}
-        # threshold=48120 on devserver with gcc9
-        # todo(lfq): update once binary size is below 50kb.
-        # Note: using gcc9-nopytorch container with pinned nightly PyTorch
-        threshold="63785"
+        # Current CI size: 48008 (gcc9-nopytorch, 2026-03-06)
+        threshold="48500"
         if [[ "$size" -le "$threshold" ]]; then
           echo "Success $size <= $threshold"
         else
@@ -513,7 +511,8 @@ jobs:
         output=$(ls -la cmake-out/test/size_test)
         arr=($output)
         size=${arr[4]}
-        threshold="51752"
+        # Current CI size: 44160 (clang12, 2026-03-06)
+        threshold="45000"
         if [[ "$size" -le "$threshold" ]]; then
           echo "Success $size <= $threshold"
         else

.github/workflows/trunk.yml

@@ -402,7 +402,7 @@ jobs:
         setup_script_args=""
         if [[ ${{ matrix.os}} == "bare_metal" ]]; then
           toolchain_prefix=arm-none-eabi-
-          threshold="110642" # 108 KiB
+          threshold="111000" # 111 KiB
           toolchain_cmake=examples/arm/ethos-u-setup/arm-none-eabi-gcc.cmake
         elif [[ ${{ matrix.os}} == "zephyr-preset" ]]; then
           setup_script_args="--target-toolchain zephyr"

.github/workflows/validate_flatbuffer_gen.yml

@@ -0,0 +1,35 @@
+name: Validate Flatbuffer Generation
+
+on:
+  workflow_dispatch:
+  pull_request:
+    paths:
+      - "schema/**"
+      - "exir/_serialize/generated/executorch_flatbuffer/**"
+
+jobs:
+  exir-flatbuffer:
+    runs-on: ubuntu-latest
+    steps:
+      - name: Checkout
+        uses: actions/checkout@v4
+
+      - name: setup python
+        uses: actions/setup-python@v5
+        with:
+          python-version: "3.12"
+
+      - name: install executorch  (with flatc)
+        run: ./install_executorch.sh -e
+
+      - name: Generate flatbuffer Python
+        run: python exir/_serialize/generate_program.py
+
+      - name: Validate executorch_flatbuffer is unchanged
+        run: |
+          git add -A exir/_serialize/generated/executorch_flatbuffer
+          if ! git diff --cached --quiet -- exir/_serialize/generated/executorch_flatbuffer; then
+            echo "Error: executorch_flatbuffer has uncommitted changes."
+            echo "Please run 'python exir/_serialize/generate_program.py' to regenerate the files and commit the changes."
+            exit 1
+          fi

.lintrunner.toml

@@ -8,6 +8,7 @@ exclude_patterns = [
     '**/third-party/**',
     '.github/scripts/**',
     'exir/serde/**',
+    'exir/_serialize/generated/executorch_flatbuffer/**',
 ]
 command = [
     'python',
@@ -39,6 +40,7 @@ exclude_patterns = [
     'third-party/**',
     '**/third-party/**',
     'exir/serde/**',
+    'exir/_serialize/generated/executorch_flatbuffer/**',
 ]
 command = [
     'python',

CLAUDE.md

@@ -6,6 +6,7 @@
 - `/building` - Build runners or C++ libs
 - `/profile` - Profile execution
 - `/cortex-m` - Build, test, or develop the Cortex-M backend
+- `/binary-size` - Analyze and reduce binary size
 
 Reference docs in `.claude/`: backends, runtime-api, quantization, llm-export, faq, tokenizers
 

CMakeLists.txt

@@ -221,6 +221,23 @@ endif()
 
 add_subdirectory(third-party)
 
+# Size-optimized builds disable exceptions, RTTI, and unwind tables.
+# ExecuTorch's runtime uses Result<T>/Error instead of exceptions, so
+# -fno-exceptions is safe. Suppressing .eh_frame via -fno-unwind-tables is safe
+# because exceptions are disabled. Placed after add_subdirectory(third-party) to
+# avoid leaking into third-party configure checks (e.g. gflags).
+if(EXECUTORCH_OPTIMIZE_SIZE
+   AND NOT WIN32
+   AND NOT MSVC
+)
+  set(CMAKE_CXX_FLAGS_RELEASE
+      "${CMAKE_CXX_FLAGS_RELEASE} -fno-exceptions -fno-rtti -fno-asynchronous-unwind-tables -fno-unwind-tables"
+  )
+  set(CMAKE_C_FLAGS_RELEASE
+      "${CMAKE_C_FLAGS_RELEASE} -fno-asynchronous-unwind-tables -fno-unwind-tables"
+  )
+endif()
+
 if(NOT DEFINED FXDIV_SOURCE_DIR)
   set(ORIGINAL_CMAKE_POSITION_INDEPENDENT_CODE_FLAG
       ${CMAKE_POSITION_INDEPENDENT_CODE}
@@ -477,6 +494,9 @@ target_include_directories(
 target_compile_definitions(
   executorch_core PUBLIC C10_USING_CUSTOM_GENERATED_MACROS
 )
+if(NOT EXECUTORCH_ENABLE_LOGGING)
+  target_compile_definitions(executorch_core PUBLIC ET_LOG_ENABLED=0)
+endif()
 target_compile_options(executorch_core PUBLIC ${_common_compile_options})
 if(MAX_KERNEL_NUM)
   target_compile_definitions(
@@ -525,6 +545,9 @@ add_library(executorch ${_executorch__srcs})
 target_link_libraries(executorch PRIVATE executorch_core)
 target_include_directories(executorch PUBLIC ${_common_include_directories})
 target_compile_definitions(executorch PUBLIC C10_USING_CUSTOM_GENERATED_MACROS)
+if(NOT EXECUTORCH_ENABLE_LOGGING)
+  target_compile_definitions(executorch PUBLIC ET_LOG_ENABLED=0)
+endif()
 target_compile_options(executorch PUBLIC ${_common_compile_options})
 executorch_target_link_options_shared_lib(executorch)
 

backends/arm/scripts/pre-commit

@@ -1,5 +1,5 @@
 #!/bin/bash
-# Copyright 2025 Arm Limited and/or its affiliates.
+# Copyright 2025-2026 Arm Limited and/or its affiliates.
 #
 # This source code is licensed under the BSD-style license found in the
 # LICENSE file in the root directory of this source tree.
@@ -9,5 +9,14 @@ git rev-list --format=%s --max-count=1 HEAD | grep -q WIP && exit 0
 
 # Check 2: lintunner on latest patch.
 lintrunner -a --revision 'HEAD^' --skip MYPY
-commit_files=$(git diff-tree --no-commit-id --name-only --diff-filter=M HEAD -r)
-git add $commit_files || true
+
+# Re-stage only paths that were already staged before lintrunner ran.
+# This preserves lint fixes for files the user intended to commit, and avoids
+# accidentally staging unrelated working-tree changes during commit --amend.
+staged_files=$(git diff --cached --name-only --diff-filter=ACMR)
+while IFS= read -r path; do
+    [ -z "${path}" ] && continue
+    if [ -f "${path}" ]; then
+        git add -- "${path}" || true
+    fi
+done <<< "${staged_files}"

backends/arm/scripts/pre-push

@@ -99,34 +99,52 @@ for COMMIT in ${COMMITS}; do
     # most of the time.
     echo -e "${INFO} License check"
 
+    # Check only files with content changes.
+    # Skip pure moves/renames (R100) where file content is unchanged.
+    license_files=()
+    while IFS=$'\t' read -r status path1 path2; do
+        case "$status" in
+            A|M)
+                license_files+=("$path1")
+                ;;
+            R100)
+                ;;
+            R*)
+                license_files+=("$path2")
+                ;;
+        esac
+    done < <(git diff-tree --no-commit-id --name-status --diff-filter=AMR -r -M ${COMMIT})
+
     current_year=$(date +%Y)
     failed_license_check=false
-    for commited_file in $commit_files; do
+    for committed_file in "${license_files[@]}"; do
         # Skip files with certain extensions
-        case "$commited_file" in
+        case "$committed_file" in
             *.md|*.md.in|*.json|*.yml|*.yaml|*.cmake|*.patch|.gitignore)
-                echo -e "${INFO} Skipping license check for ${commited_file} (excluded extension)"
+                echo -e "${INFO} Skipping license check for ${committed_file} (excluded extension)"
                 continue
                 ;;
         esac
 
-        file_header=$(head "$commited_file")
+        file_header=$(head "$committed_file")
         if ! echo "$file_header" | grep -qi "Arm"; then
-            echo -e "${WARNING} No Arm copyright header in ${commited_file}"\
+            echo -e "${WARNING} No Arm copyright header in ${committed_file}"\
             " (skipping license year check)"
             continue
         fi
 
         if ! echo "$file_header" | grep -q "$current_year Arm"; then
-            echo -e "${ERROR} Header in $commited_file did not contain"\
+            echo -e "${ERROR} Header in $committed_file did not contain"\
             "'$current_year Arm'"
             failed_license_check=true
         else
-            echo -e "${SUCCESS} $commited_file passed license check"
+            echo -e "${SUCCESS} $committed_file passed license check"
         fi
     done
 
-    if [[ $failed_license_check == true ]]; then
+    if [[ ${#license_files[@]} -eq 0 ]]; then
+        echo -e "${INFO} No files with content changes to check"
+    elif [[ $failed_license_check == true ]]; then
         FAILED=1
     else
         echo -e "${SUCCESS} All files passed license check"
@@ -181,8 +199,8 @@ for COMMIT in ${COMMITS}; do
 
     # Determine whether all modifications are under backends/arm or examples/arm
     only_arm_changes=true
-    for commited_file in $commit_files; do
-        if [[ $commited_file != backends/arm/* && $commited_file != examples/arm/* ]]; then
+    for committed_file in $commit_files; do
+        if [[ $committed_file != backends/arm/* && $committed_file != examples/arm/* ]]; then
             only_arm_changes=false
             break
         fi