Replace SIMD assembly with initialized Node16::index + valgrind client request #77

Merged
andrew merged 2 commits from client-request-instead-of-asm into main 2026-08-04 00:07:21 +00:00
2 Commits
Author SHA1 Message Date
andrew 6d9810fe7f Poison Node48 reverseIndex slot on erase
CI / test (arm64, -DCMAKE_BUILD_TYPE=Debug -DMSAN_TOOLCHAIN_PATH=/opt/msan, 22, https://minio.weaselab.dev/public/aarch64/msan-toolchain-22.1.8.tar.zst, debug-arm64, ubuntu-latest-arm64) (push) Successful in 3m46s
CI / release (arm64, , ubuntu-latest-arm64) (push) Successful in 3m21s
CI / test (arm64, -DCMAKE_BUILD_TYPE=Debug -DMSAN_TOOLCHAIN_PATH=/opt/msan, 22, https://minio.weaselab.dev/public/aarch64/msan-toolchain-22.1.8.tar.zst, debug-arm64, ubuntu-latest-arm64) (pull_request) Successful in 3m50s
CI / release (arm64, , ubuntu-latest-arm64) (pull_request) Successful in 3m24s
CI / pre-commit (push) Successful in 2m5s
CI / test (amd64, -DCMAKE_BUILD_TYPE=Debug -DMSAN_TOOLCHAIN_PATH=/opt/msan, 21, https://minio.weaselab.dev/public/x86_64/msan-toolchain-21.1.8.tar.zst, debug, ubuntu-latest-amd64) (push) Successful in 3m56s
CI / test (amd64, -DCMAKE_CXX_FLAGS=-DUSE_64_BIT=1, 21, , 64-bit-versions, ubuntu-latest-amd64) (push) Successful in 3m23s
CI / test (amd64, -DCMAKE_C_COMPILER=gcc -DCMAKE_CXX_COMPILER=g++, 21, , gcc, ubuntu-latest-amd64) (push) Successful in 3m31s
CI / test (amd64, -DUSE_SIMD_FALLBACK=ON, 21, , simd-fallback, ubuntu-latest-amd64) (push) Successful in 3m24s
CI / release (amd64, -DMSAN_TOOLCHAIN_PATH=/opt/msan, ubuntu-latest-amd64) (push) Successful in 5m46s
CI / coverage (push) Successful in 3m55s
CI / pre-commit (pull_request) Successful in 2m8s
CI / test (amd64, -DCMAKE_BUILD_TYPE=Debug -DMSAN_TOOLCHAIN_PATH=/opt/msan, 21, https://minio.weaselab.dev/public/x86_64/msan-toolchain-21.1.8.tar.zst, debug, ubuntu-latest-amd64) (pull_request) Successful in 3m53s
CI / test (amd64, -DCMAKE_CXX_FLAGS=-DUSE_64_BIT=1, 21, , 64-bit-versions, ubuntu-latest-amd64) (pull_request) Successful in 3m21s
CI / test (amd64, -DCMAKE_C_COMPILER=gcc -DCMAKE_CXX_COMPILER=g++, 21, , gcc, ubuntu-latest-amd64) (pull_request) Successful in 3m20s
CI / test (amd64, -DUSE_SIMD_FALLBACK=ON, 21, , simd-fallback, ubuntu-latest-amd64) (pull_request) Successful in 3m20s
CI / release (amd64, -DMSAN_TOOLCHAIN_PATH=/opt/msan, ubuntu-latest-amd64) (pull_request) Successful in 5m46s
CI / coverage (pull_request) Successful in 4m5s
When a Node48 loses a child, the vacated reverseIndex slot would otherwise
remain defined, letting valgrind miss reads that rely on the stale value.
Mark it undefined, mirroring the existing poisoning of the vacated children
slot.
2026-08-03 17:38:21 -04:00
andrew bae2b22e95 Replace SIMD assembly with initialized Node16::index + valgrind client request
CI / test (arm64, -DCMAKE_BUILD_TYPE=Debug -DMSAN_TOOLCHAIN_PATH=/opt/msan, 22, https://minio.weaselab.dev/public/aarch64/msan-toolchain-22.1.8.tar.zst, debug-arm64, ubuntu-latest-arm64) (push) Successful in 3m44s
CI / release (arm64, , ubuntu-latest-arm64) (push) Successful in 3m19s
CI / test (arm64, -DCMAKE_BUILD_TYPE=Debug -DMSAN_TOOLCHAIN_PATH=/opt/msan, 22, https://minio.weaselab.dev/public/aarch64/msan-toolchain-22.1.8.tar.zst, debug-arm64, ubuntu-latest-arm64) (pull_request) Successful in 3m47s
CI / release (arm64, , ubuntu-latest-arm64) (pull_request) Successful in 3m26s
CI / pre-commit (push) Successful in 2m1s
CI / test (amd64, -DCMAKE_BUILD_TYPE=Debug -DMSAN_TOOLCHAIN_PATH=/opt/msan, 21, https://minio.weaselab.dev/public/x86_64/msan-toolchain-21.1.8.tar.zst, debug, ubuntu-latest-amd64) (push) Successful in 3m51s
CI / test (amd64, -DCMAKE_CXX_FLAGS=-DUSE_64_BIT=1, 21, , 64-bit-versions, ubuntu-latest-amd64) (push) Successful in 3m21s
CI / test (amd64, -DCMAKE_C_COMPILER=gcc -DCMAKE_CXX_COMPILER=g++, 21, , gcc, ubuntu-latest-amd64) (push) Successful in 3m15s
CI / test (amd64, -DUSE_SIMD_FALLBACK=ON, 21, , simd-fallback, ubuntu-latest-amd64) (push) Successful in 3m18s
CI / release (amd64, -DMSAN_TOOLCHAIN_PATH=/opt/msan, ubuntu-latest-amd64) (push) Successful in 5m43s
CI / coverage (push) Successful in 4m8s
CI / pre-commit (pull_request) Successful in 2m16s
CI / test (amd64, -DCMAKE_BUILD_TYPE=Debug -DMSAN_TOOLCHAIN_PATH=/opt/msan, 21, https://minio.weaselab.dev/public/x86_64/msan-toolchain-21.1.8.tar.zst, debug, ubuntu-latest-amd64) (pull_request) Successful in 4m25s
CI / test (amd64, -DCMAKE_CXX_FLAGS=-DUSE_64_BIT=1, 21, , 64-bit-versions, ubuntu-latest-amd64) (pull_request) Successful in 3m49s
CI / test (amd64, -DCMAKE_C_COMPILER=gcc -DCMAKE_CXX_COMPILER=g++, 21, , gcc, ubuntu-latest-amd64) (pull_request) Successful in 3m51s
CI / test (amd64, -DUSE_SIMD_FALLBACK=ON, 21, , simd-fallback, ubuntu-latest-amd64) (pull_request) Successful in 3m55s
CI / release (amd64, -DMSAN_TOOLCHAIN_PATH=/opt/msan, ubuntu-latest-amd64) (pull_request) Successful in 6m36s
CI / coverage (pull_request) Successful in 4m20s
Commit 6fed133 moved SIMD operations on Node16::index into file-level
assembly (later extended to aarch64) to avoid reading indeterminate bytes
(UB in C++). Revert that in favor of a simpler, architecture-independent
fix:

- Restore the original HAS_AVX / HAS_ARM_NEON intrinsics at the affected
  call sites, keeping the newer conflictMask16 helper and scalar fallback.
- memset Node16::index in both copy constructors (from Node3 and Node48) so
  the [numChildren, 16) tail is always initialized, making the full 16-byte
  SIMD loads well-defined. This also fixes the same latent UB in the NEON
  path and the Node16->Node16 full-array memcpy.
- Issue a VALGRIND_MAKE_MEM_UNDEFINED client request on the tail so valgrind
  still flags any read that relies on the unused slots rather than numChildren.
- Drop simd.h, simd_x86_64.S, simd_aarch64.S, and the CMake ASM wiring.
2026-08-03 17:32:41 -04:00