)]}' { "log": [ { "commit": "55d60a30667f25e4eab32372b9990080aa20954c", "tree": "5de9c29d2831fd5ecbd8d4ac687e3cb507cde37e", "parents": [ "f105adfb93cfb08f9dc4b12e7492f3576f77fe9d" ], "author": { "name": "Hans Wennborg", "email": "hans@hanshq.net", "time": "Sat Apr 11 09:27:46 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Sat Apr 11 09:30:41 2026" }, "message": "[compiler-rt] Disable soft_rss_limit_mb_test.cpp (#191370)\n\nThe test has been failing flakily for a while; see PRs #170911, #171469,\n#188441.\n\nCo-authored-by: Vitaly Buka \u003cvitalybuka@gmail.com\u003e\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 6c4bd0b3ac6939e5d88f5e8de7679463c0604a9c\n" }, { "commit": "f105adfb93cfb08f9dc4b12e7492f3576f77fe9d", "tree": "9df211e9882f14e8ab1b150fa77513a04e2611c4", "parents": [ "0fd2aaa85ed426b97efc1aad5b2123fd227d6c7f" ], "author": { "name": "Vitaly Buka", "email": "vitalybuka@google.com", "time": "Fri Apr 10 20:23:15 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Fri Apr 10 20:25:34 2026" }, "message": "[NFC][compiler-rt][test] Fix indentation in lit.common.cfg.py (#191509)\n\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: d2786ecc1699a411d474f0631de0efaf17cd5935\n" }, { "commit": "0fd2aaa85ed426b97efc1aad5b2123fd227d6c7f", "tree": "ea8ee0a2d63304e674ca2a5cd191629101d5c7bb", "parents": [ "b7108959236d507abbfdcc923a2e75b3c658f701" ], "author": { "name": "Christopher Ferris", "email": "cferris1000@users.noreply.github.com", "time": "Fri Apr 10 17:50:34 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Fri Apr 10 17:55:40 2026" }, "message": "[scudo] Remove fill when realloc to smaller size. (#191321)\n\nIn the reallocate function, when there is a realloc smaller than the\ncurrent size, the code would attempt to fill in the bytes after the new\nsize. This doesn\u0027t really add any extra security and is mostly a waste\nof time, so skip it.\n\nRemove the test that verifies this functionality.\n\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: f0004d8143b346e0f5a13b4021f348165e27fb9a\n" }, { "commit": "b7108959236d507abbfdcc923a2e75b3c658f701", "tree": "7a8e17b3162ce4c0c3f099332969adba1dbd5517", "parents": [ "cabc3546290094fb966215b684f563a20e851638" ], "author": { "name": "David Justo", "email": "david.justo.1996@gmail.com", "time": "Fri Apr 10 17:34:00 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Fri Apr 10 17:35:40 2026" }, "message": "Disable MSVC-incompatible portions of `disable_container_overflow_checks` for MSVC (#191456)\n\n**Context:**\n\nThe test `disable_container_overflow_checks` recently started running on\nWindows, as per:\nhttps://github.com/llvm/llvm-project/pull/181721/changes\n\nAs a result, the MSVC ASan fork of LLVM ASan started executing this\ntest, which has been failing for 2 reasons.\n\n1) MSVC does not support the `__has_feature` syntax.\n2) The `__SANITIZER_DISABLE_CONTAINER_OVERFLOW__` macro is not supported\nin MSVC ASan (we have an equivalent in `_DISABLE_STL_ANNOTATION`)\nbecause `__SANITIZER_DISABLE_CONTAINER_OVERFLOW__` also invokes\nMSVC-incompatible syntax.\n\n**This PR** addresses these two failures.\n\nFor (1), we replace:\n\n```C++\n#if __has_feature(address_sanitizer)\n```\n\nwith\n\n```C++\n// compilers such as MSVC do not support `__has_feature`\n#ifndef __has_feature\n#define __has_feature(x) 0\n#endif\n\n// MSVC defines `__SANITIZE_ADDRESS__`\n#if __has_feature(address_sanitizer) || defined(__SANITIZE_ADDRESS__)\n```\n\nand for (2), we disable the tests for\n`__SANITIZER_DISABLE_CONTAINER_OVERFLOW__` for MSVC.\n\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 03bb8745a0e886934f93e5639637f42db2f2575f\n" }, { "commit": "cabc3546290094fb966215b684f563a20e851638", "tree": "95d9043ff43def3e821e6fe64f25e2727d9a6392", "parents": [ "db380e876fe3f842a701e57ea54388dbd49c54c4" ], "author": { "name": "google-yfyang", "email": "yfyang@google.com", "time": "Fri Apr 10 16:50:53 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Fri Apr 10 16:56:06 2026" }, "message": "[compiler-rt] Address dlvsym not found compilation error when targeting certain platforms (try 2) (#191458)\n\nThe previous attempt #191444 was incomplete.\n\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 05bfe73bf6b5311f01503d81996d1e714c6cce81\n" }, { "commit": "db380e876fe3f842a701e57ea54388dbd49c54c4", "tree": "f8f37d4935278f3f2f84660506f8bd8d33a45358", "parents": [ "f48b818001c44796138d23869d7bd7652875f483" ], "author": { "name": "google-yfyang", "email": "yfyang@google.com", "time": "Fri Apr 10 16:06:35 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Fri Apr 10 16:10:29 2026" }, "message": "[compiler-rt] Address dlvsym not found compilation error when targeting certain platforms (#191444)\n\n#191098 Add some unguarded dlvsym calls. This causes build issues when\ntargeting certain platforms such as iOS.\n\nThis change should restore the original behavior.\n\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: aca430027b0a308f6206dfbd7e66f57faf8a7700\n" }, { "commit": "f48b818001c44796138d23869d7bd7652875f483", "tree": "72f7c3f8aa425a2cee42a010ee2cb3ee168d645b", "parents": [ "5178c4e2979601c39e26c5d2768469f491bff475" ], "author": { "name": "Yaxun (Sam) Liu", "email": "yaxun.liu@amd.com", "time": "Fri Apr 10 13:30:07 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Fri Apr 10 13:35:58 2026" }, "message": "[compiler-rt] Expose shared DSO helpers for compiler-rt runtimes (#191098)\n\nThe motivation of this PR is to refactor and expose DSO helper functions\nso\nthey can be used by all compiler-rt libraries, including the profile\nlibrary,\nwithout duplicating dlopen/dlsym (non-Windows) or\nLoadLibrary/GetProcAddress\n(Windows) logic in each runtime.\n\nImplement the helpers in namespace __interception in\ninterception_linux.cpp for\nnon-Windows targets and interception_win.cpp for Windows, and use them\nfrom the\nexisting Linux interception path for RTLD_NEXT/RTLD_DEFAULT/dlvsym\nlookups.\n\nThis is NFC for existing libraries that already use interception\u0027s\npublic APIs;\nsanitizer and interception lit behavior is unchanged.\n\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 43d6e3874e67ac1f94482b43a98f595163e8cb6e\n" }, { "commit": "5178c4e2979601c39e26c5d2768469f491bff475", "tree": "d96a16770b9863183925c12f306c0a0cb24b9625", "parents": [ "ed91d78022a386fd2626867e26c61a4660e70d5d" ], "author": { "name": "Christopher Ferris", "email": "cferris1000@users.noreply.github.com", "time": "Thu Apr 09 19:02:11 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Thu Apr 09 19:06:11 2026" }, "message": "[scudo] Refactor the wrapper code into a single file. (#190857)\n\nAllow using a define to create a prefix rather than trying to use\nandroid/bionic checks.\n\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: b742ec91c0876fc5f9fb85aceecb67c22021b69d\n" }, { "commit": "ed91d78022a386fd2626867e26c61a4660e70d5d", "tree": "e0074557b5b799d590114ee65808780dd48d519d", "parents": [ "007abb52b74ee36af77a00eb293e2585aa42b561" ], "author": { "name": "Andrew Haberlandt", "email": "ahaberlandt@apple.com", "time": "Thu Apr 09 16:09:50 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Thu Apr 09 16:15:52 2026" }, "message": "[compiler-rt] [Darwin] Move macOS ASAN reservation above 512G (#191039)\n\nOn macOS, the first 512G may contain platform-specific reservations. To\nensure compatibility with these reservations, this changes ASAN to\nalways map shadow memory above 512G on macOS.\n\nrdar://174252720\n\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 857a98e50e279c0c1cf3658337dadb347f7ff5c4\n" }, { "commit": "007abb52b74ee36af77a00eb293e2585aa42b561", "tree": "f07df4ce4b7117e5d5c65307ae30e648b7f57567", "parents": [ "dd3d9fcad21efd524d52bc73304bddb39d558c7e" ], "author": { "name": "Brian Cain", "email": "brian.cain@oss.qualcomm.com", "time": "Thu Apr 09 04:43:56 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Thu Apr 09 04:45:18 2026" }, "message": "[compiler-rt] Enable sanitizers, XRay, etc for hexagon (#190300)\n\nEnable additional compiler-rt components in the hexagon-linux-clangrt\nstandalone cache file.\n\nAlso set CAN_TARGET_hexagon\u003d1 to bypass the try_compile() probe that\nfails during cross-compilation, and configure SANITIZER_CXX_ABI to use\nthe in-tree libc++ to avoid link failures.\n\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 3cc34fb973408b0cf1be082ccc78d5f153c25387\n" }, { "commit": "dd3d9fcad21efd524d52bc73304bddb39d558c7e", "tree": "0e1a96aae4021c8e2d55227b2bf71ced7ed2876d", "parents": [ "b91f2f99771f16f58f5b29524b41e171668576fe" ], "author": { "name": "Mads Marquart", "email": "mads@marquart.dk", "time": "Thu Apr 09 03:29:56 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Thu Apr 09 03:36:10 2026" }, "message": "[LSAN] Add extra suppressions on Apple Aarch64 (#117478)\n\n- _fetchInitializingClassList for startup code in dyld.\n- dyld4::RuntimeState::_instantiateTLVs for thread locals.\n\nFixes https://github.com/llvm/llvm-project/issues/115992, now it shows\nthe following instead (in a project using thread locals):\n```\n-----------------------------------------------------\nSuppressions used:\n count bytes template\n 3 120 *_fetchInitializingClassList*\n 1 104 *dyld4::RuntimeState::_instantiateTLVs*\n-----------------------------------------------------\n```\n\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 3b9f7913148d0f5d952ce85e8d8dd6ae4fec678b\n" }, { "commit": "b91f2f99771f16f58f5b29524b41e171668576fe", "tree": "da703b20bda056249c1fa4a61d09206b93f31ac9", "parents": [ "b917156f9bdf0b7f9bb88e056da32409f5d71630" ], "author": { "name": "Marian Buschsieweke", "email": "maribu@users.noreply.github.com", "time": "Thu Apr 09 03:11:15 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Thu Apr 09 03:14:05 2026" }, "message": "[compiler-rt] fix __sanitizer::struct_sock_fprog_sz availability (#183411)\n\n`struct sock_fprog` is not provided by glibc, but rather by the linux\nheaders (`#include \u003clinux/filter.h`). It seems that glibc due to an\nimplementation detail internally includes `\u003clinux/filter.h\u003e` somewhere\nother C libs (e.g. musl) do not, which previously caused build failures\nand let to disabling `struct sock_fprog` on non-glibc Linux systems.\n\nThis adds the missing include and provides it again for all Linux\nsystems regardless of C lib.\n\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: dabd8cc2334917272fe73b70626fd3a5f9e63cf5\n" }, { "commit": "b917156f9bdf0b7f9bb88e056da32409f5d71630", "tree": "238a43d0820a075a46f509697e910611cf374f67", "parents": [ "da1f534a2132575113111ef9931144194c8d9aeb" ], "author": { "name": "Stefan Schulze Frielinghaus", "email": "stefansf@linux.ibm.com", "time": "Thu Apr 09 03:10:01 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Thu Apr 09 03:13:50 2026" }, "message": "[TSan] Fix determining static TLS blocks (#183106)\n\nRunning gcc test c-c++-common/tsan/tls_race.c on s390 we get:\n\nThreadSanitizer: CHECK failed: tsan_platform_linux.cpp:618 \"((thr_beg))\n\u003e\u003d ((tls_addr))\" (0x3ffaa35e140, 0x3ffaa35e250) (tid\u003d2419930)\n#0 __tsan::CheckUnwind() /devel/src/libsanitizer/tsan/tsan_rtl.cpp:696\n(libtsan.so.2+0x91b57)\n#1 __sanitizer::CheckFailed(char const*, int, char const*, unsigned long\nlong, unsigned long long)\n/devel/src/libsanitizer/sanitizer_common/sanitizer_termination.cpp:86\n(libtsan.so.2+0xd211b)\n#2 __tsan::ImitateTlsWrite(__tsan::ThreadState*, unsigned long, unsigned\nlong) /devel/src/libsanitizer/tsan/tsan_platform_linux.cpp:618\n(libtsan.so.2+0x8faa3)\n#3 __tsan::ThreadStart(__tsan::ThreadState*, unsigned int, unsigned long\nlong, __sanitizer::ThreadType)\n/devel/src/libsanitizer/tsan/tsan_rtl_thread.cpp:225\n(libtsan.so.2+0xaadb5)\n#4 __tsan_thread_start_func\n/devel/src/libsanitizer/tsan/tsan_interceptors_posix.cpp:1065\n(libtsan.so.2+0x3d34d)\n#5 start_thread \u003cnull\u003e (libc.so.6+0xae70d) (BuildId:\nd3b08de1b543c2d15d419bf861b3c2e4c01ac75b)\n#6 thread_start \u003cnull\u003e (libc.so.6+0x12d2ff) (BuildId:\nd3b08de1b543c2d15d419bf861b3c2e4c01ac75b)\n\nIn order to determine the static TLS blocks in GetStaticTlsBoundary we\niterate over the modules and try to find the largest range without a\ngap. Here we might have that modules are spaced exactly by the\nalignment. For example, for the failing test we have:\n\n(gdb) p/x ranges.data_[0]\n$1 \u003d {begin \u003d 0x3fff7f9e6b8, end \u003d 0x3fff7f9e740, align \u003d 0x8, tls_modid\n\u003d 0x3} (gdb) p/x ranges.data_[1]\n$2 \u003d {begin \u003d 0x3fff7f9e740, end \u003d 0x3fff7f9eed0, align \u003d 0x40,\ntls_modid \u003d 0x2} (gdb) p/x ranges.data_[2]\n$3 \u003d {begin \u003d 0x3fff7f9eed8, end \u003d 0x3fff7f9eef8, align \u003d 0x8, tls_modid\n\u003d 0x4} (gdb) p/x ranges.data_[3]\n$4 \u003d {begin \u003d 0x3fff7f9eefc, end \u003d 0x3fff7f9ef00, align \u003d 0x4, tls_modid\n\u003d 0x1}\n\nwhere ranges[3].begin \u003d\u003d ranges[2].end + ranges[3].align holds. Since in\nthe loop a strict inequality test is used we compute the wrong address\n\n(gdb) p/x *addr\n$5 \u003d 0x3fff7f9eefc\n\nwhereas 0x3fff7f9e6b8 is expected which is why we bail out in the\nsubsequent.\n\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: fc48fac95b8b1bad161c705dd4bf71300bd2b609\n" }, { "commit": "da1f534a2132575113111ef9931144194c8d9aeb", "tree": "af71125c09d509992070c46649455b7f283dc49d", "parents": [ "d9dca05ab45293a71f60ccae3b7c4846be6b853c" ], "author": { "name": "Andrew Haberlandt", "email": "ahaberlandt@apple.com", "time": "Tue Apr 07 17:18:55 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Tue Apr 07 17:21:03 2026" }, "message": "[sanitizer_common] Fix memory leak in ListOfModules (#190104)\n\nI found a rather old leak in `ListOfModules`. `ListOfModules` has a\nmember\n\n```\n InternalMmapVectorNoCtor\u003cLoadedModule\u003e modules_;\n```\n\nWhich has neither a constructor nor destructor. As a consequence, we\nneed to call `Destroy` on it when ListOfModules is destroyed, so that\nthe memory is released to the OS (via munmap).\n\nrdar://173906291\n\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 07e8db61f9e10a5825330d63c36022e849c96116\n" }, { "commit": "d9dca05ab45293a71f60ccae3b7c4846be6b853c", "tree": "badfcf49aaf42a68a477993e7991e198faba0fd4", "parents": [ "9b33af39f25a9864babe4a1fbd19fb9cb0dcc7b1" ], "author": { "name": "Joseph Huber", "email": "huberjn@outlook.com", "time": "Tue Apr 07 11:53:51 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Tue Apr 07 11:55:31 2026" }, "message": "[compiler-rt] Enable minimal ubsan for GPU targets (#188289)\n\nSummary:\nThis PR enabled the ubsan-minimal target for the GPU. This should enable\nthe trivial checks on the GPU-side. I had to make some changes to the\ncommon sanitizer requirements, but I think this is reasonable\nconsidering the common_sanitizer requires a fully hosted environment and\nall we needed were the headers.\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 46bef90e9c60ccc79d3fefa881270ac53a88226f\n" }, { "commit": "9b33af39f25a9864babe4a1fbd19fb9cb0dcc7b1", "tree": "3c3abaae5a5e9fa5969525c50dc58a19dd0cd80d", "parents": [ "9c2d763ac44091e4e5b293ffd9474685d8d4b845" ], "author": { "name": "Brian Cain", "email": "brian.cain@oss.qualcomm.com", "time": "Mon Apr 06 19:49:43 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Mon Apr 06 19:55:31 2026" }, "message": "[compiler-rt] Add hexagon to libFuzzer supported architectures (#190297)\n\nLibFuzzer builds successfully for Hexagon Linux.\n\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 2aa4100fa710ed83c5acd7505c27b4498f727c8e\n" }, { "commit": "9c2d763ac44091e4e5b293ffd9474685d8d4b845", "tree": "31fc6c793680eac4b6376f07beba2973c29306fd", "parents": [ "76287b5da8e155135536c8e3a67432d97d74fe3a" ], "author": { "name": "Joseph Huber", "email": "huberjn@outlook.com", "time": "Fri Apr 03 15:44:04 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Fri Apr 03 15:45:36 2026" }, "message": "[compiler-rt] Split the GPU.cmake cache file to AMDGPU and NVPTX (#190349)\n\nSummary:\nThese will have different functionality going forward. They should be\nsplit so we can more easily support things only feasible in AMDGPU.\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: d8ba56ce3f98871ae4e5782c4af2df4c98bebde7\n" }, { "commit": "76287b5da8e155135536c8e3a67432d97d74fe3a", "tree": "579d7dd43a94f461175f96b526bc75ac262c0635", "parents": [ "e5507743676ea1feeb1f2488a78c577bfb555dd6" ], "author": { "name": "Michael Kruse", "email": "llvm-project@meinersbur.de", "time": "Thu Apr 02 10:32:14 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Thu Apr 02 10:35:36 2026" }, "message": "[Runtimes] Introduce variables containing resource dir paths (#177953)\n\nIntroduce common infrastructure for runtimes that determines compiler\nresource path locations. These variables introduced are:\n\n * RUNTIMES_OUTPUT_RESOURCE_DIR\n * RUNTIMES_INSTALL_RESOURCE_PATH\n\nThat contain the location for the compiler resource path (typically\n`lib/clang/\u003cversion\u003e`) in the build tree and the install tree (the\nlatter relative to CMAKE_INSTALL_PREFIX).\n\nAdditionally, define\n\n * RUNTIMES_OUTPUT_RESOURCE_LIB_DIR\n * RUNTIMES_INSTALL_RESOURCE_LIB_PATH\n\nas for the location of clang/flang version-locked libraries (typically\n`lib${LLVM_LIBDIR_SUFFIX}/\u003ctarger-triple\u003e`, but also depends on `APPLE`\nand `LLVM_ENABLE_PER_TARGET_RUNTIME_DIR`). This code is moved from\nflang-rt and initially becomes its only user.\n\nRefactored out of #171610 as requested\n[here](https://github.com/llvm/llvm-project/pull/171610#discussion_r2687382481).\n\nExtracted `get_runtimes_target_libdir_common` from compiler-rt as\nrequested\n[here](https://github.com/llvm/llvm-project/pull/171610#discussion_r2689565634).\n\nAdded TODO comments to all runtimes as requested\n[here](https://github.com/llvm/llvm-project/pull/171610#issuecomment-3789598635).\n\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: afb80bddf147c134c5e3cf3f357f8d00bdfd526c\n" }, { "commit": "e5507743676ea1feeb1f2488a78c577bfb555dd6", "tree": "81c43d9a8a38461f3a1be90cbd8d2c12b8b28cc3", "parents": [ "6b5a632c21bb5d783511bfa9962f909422f9e1d9" ], "author": { "name": "Christopher Ferris", "email": "cferris1000@users.noreply.github.com", "time": "Wed Apr 01 23:44:31 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Wed Apr 01 23:50:30 2026" }, "message": "[scudo] Fix reallocate for MTE. (#190086)\n\nFor MTE, we can\u0027t use the whole size or we might trigger a segfault.\nTherefore, use the exact size when MTE is enabled or the exact usable\nsize parameter is true.\n\nAlso, optimize out the call to getUsableSize and use a simpler\ncalculation.\n\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 7c260d3966534031f9fe3a35b54887f6d979606d\n" }, { "commit": "6b5a632c21bb5d783511bfa9962f909422f9e1d9", "tree": "5dcc24605e49a8ce6175ba34bb0614933568a50b", "parents": [ "45f169ea615b1da27d8d5ce22280be9468978c69" ], "author": { "name": "Sadaf Ebrahimi", "email": "sadafebrahimi@google.com", "time": "Wed Apr 01 17:30:47 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Wed Apr 01 17:35:58 2026" }, "message": "[scudo] Ensure that reallocate copies everything\n\ngetUsableSize returns the actual capacity of the underlying block, which\nmay be larger than the size originally requested by the user. If the\nuser writes data into this extra space accessible via getUsableSize and\nsubsequently calls reallocate, the existing implementation only copies\nthe original requested number of bytes. This resulted in data loss for\nany information stored beyond the requested size but within the usable\nbounds.\n\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 7c4b6dee7401769a359d2922629f7cd0103d6c72\n" }, { "commit": "45f169ea615b1da27d8d5ce22280be9468978c69", "tree": "77a4b32058eb69afed21760ccfb022ade05da632", "parents": [ "3fc455ea68179d1c8e65bb182b119db6e93424bd" ], "author": { "name": "Andrew Marshall", "email": "planetmarshall@users.noreply.github.com", "time": "Wed Apr 01 10:00:53 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Wed Apr 01 10:05:48 2026" }, "message": "pass target triple to `check_assembler_flag` (#188521)\n\nTarget specific flags (Notably `-mimplict\u003dalways` for ARM) are not\nrecognized by the clang assembler unless the target is specified. This\nPR passes the value of `CMAKE_C_COMPILER_TARGET` to the assembler so\nthat target specific flags are recognized.\n\n## Previous behaviour\n\nWhen configuring builtins for an ARMv7 target:\n\n```\n-- Builtin supported architectures: armv7\n-- Checking for assembler flag -mimplicit-it\u003dalways\n-- Checking for assembler flag -mimplicit-it\u003dalways - Not accepted\n-- Checking for assembler flag -Wa,-mimplicit-it\u003dalways\n-- Checking for assembler flag -Wa,-mimplicit-it\u003dalways - Not accepted\nCMake Warning at CMakeLists.txt:462 (message):\n Don\u0027t know how to set the -mimplicit-it\u003dalways flag in this assembler; not\n including Arm optimized implementations\n```\n\n## New behaviour\n\n```\n-- Builtin supported architectures: armv7\n-- Checking for assembler flag -mimplicit-it\u003dalways\n-- Checking for assembler flag -mimplicit-it\u003dalways - Accepted\n```\n\n## Steps to reproduce:\n\n```\n$ cmake -S compiler-rt/lib/builtins -B build \\\n -DCMAKE_C_COMPILER\u003dclang \\\n -DCMAKE_C_COMPILER_TARGET\u003darmv7-unknown-linux-gnueabihf \\\n -DCOMPILER_RT_DEFAULT_TARGET_ONLY\u003dON\n```\n\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 97562e7e5cae5d6dc019b8f3275eb9639978c05f\n" }, { "commit": "3fc455ea68179d1c8e65bb182b119db6e93424bd", "tree": "7822688e4e932b0be11b548b93731062183b5f62", "parents": [ "ef6c3b137deaa8b8325c4d7ba582b6625345e37c" ], "author": { "name": "Hans Wennborg", "email": "hans@hanshq.net", "time": "Wed Apr 01 08:19:48 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Wed Apr 01 08:24:54 2026" }, "message": "[compiler-rt] Try bumping soft_rss_limit again (#188441)\n\nPrevious instance: #171469\n\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 21f94950a8411d0a10eaf3380ae44b411f65a0d3\n" }, { "commit": "ef6c3b137deaa8b8325c4d7ba582b6625345e37c", "tree": "3a95b968333f281f13f23c1a1540a3c16e8e51bd", "parents": [ "4faa19bd7af63994080fd7cde377a5948632bb32" ], "author": { "name": "jinge90", "email": "ge.jin@intel.com", "time": "Tue Mar 31 23:31:15 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Tue Mar 31 23:33:43 2026" }, "message": "[NFC][compiler-rt] Use __builtin_huge_val for CRT_INFINITY for double precision (#188197)\n\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 3f2a9c6f5486e4202ec01e61fbef0e7d094e2371\n" }, { "commit": "4faa19bd7af63994080fd7cde377a5948632bb32", "tree": "92f906f2b4b59b3a02c95b1aed0b27dfb6466518", "parents": [ "9f1df202a90516ffd65e5f97a8801326ddc5f03a" ], "author": { "name": "PiJoules", "email": "leonardchan@google.com", "time": "Tue Mar 31 22:04:54 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Tue Mar 31 22:10:28 2026" }, "message": "[compiler-rt][asan] Forward fix for free_aligned_sized_mismatch.cpp (#189760)\n\nMark this test as UNSUPPORTED for android since android\u0027s libc doesn\u0027t\nseem to support aligned_alloc.\n\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 911dc4139dce1999c72144a9656f65caffca555b\n" }, { "commit": "9f1df202a90516ffd65e5f97a8801326ddc5f03a", "tree": "c18eb0324a6fec387a4ccc74f10eea35b1491fa5", "parents": [ "cf7b82d41efd24c5cb048f1f8168f9caac4f1bed" ], "author": { "name": "Samuel Thibault", "email": "samuel.thibault@ens-lyon.org", "time": "Tue Mar 31 20:55:12 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Tue Mar 31 21:00:37 2026" }, "message": "[compiler-rt] Disable asan tests on GNU/Hurd (#169209)\n\nLike OpenBSD, asan does not support GNU/Hurd yet.\n\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: c1ebd2f1c07f1dc51df1c1a56891fe564cbadc19\n" }, { "commit": "cf7b82d41efd24c5cb048f1f8168f9caac4f1bed", "tree": "a92c854024cf2dbd857c68f5e6d5218d14709167", "parents": [ "f352341cf81261011fcd95aefd016772e7d405aa" ], "author": { "name": "PiJoules", "email": "leonardchan@google.com", "time": "Tue Mar 31 19:10:43 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Tue Mar 31 19:15:48 2026" }, "message": "[asan] Add size/alignment checks for free_[aligned_]sized (#189216)\n\nHistorically, alignment and size weren\u0027t taken into account when freeing\nallocations since `free` just takes a pointer. With `free_sized` and\n`free_aligned_sized`, we can do these size and alignment checks in asan\nnow. This adds a new report type specifically for these functions.\n\nChecking is hidden behind a new env flag `free_size_mismatch` which is\nenabled by default, but downstream users can opt out of it.\n\nThe bulk of this PR was generated by gemini but thoroughly reviewed and\nedited by me to the best of my ability.\n\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 78c5d68029c0507123de28ecbe242490ee487b16\n" }, { "commit": "f352341cf81261011fcd95aefd016772e7d405aa", "tree": "db1ce7cc3d766a8c1c8bf81faee3920776f1f803", "parents": [ "fddf09ba9610b710269fab27f755d1f94c03a3c7" ], "author": { "name": "Simon Tatham", "email": "simon.tatham@arm.com", "time": "Tue Mar 31 11:00:11 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Tue Mar 31 11:06:05 2026" }, "message": "[compiler-rt][ARM] Optimized double precision FP add/sub (#179921)\n\nThe one new assembly source file, `arm/adddf3.S`, implements both\naddition and subtraction via cross-branching after flipping signs, since\nboth operations must provide substantially the same logic. The new cmake\nproperties introduced in a prior commit are used to arrange that\nincluding `adddf3.S` supersedes the C versions of both addition and\nsubtraction, and also informs the test suite that both functions are\navailable to test.\n\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 0cecacd971a5471803b79f2b4a976ce75a2539b2\n" }, { "commit": "fddf09ba9610b710269fab27f755d1f94c03a3c7", "tree": "3aa70355ed7639c47d32ff22a8693765f1e0c986", "parents": [ "d85a6b4019d74bc707dddeee5eb39196a028a6ca" ], "author": { "name": "Roland McGrath", "email": "mcgrathr@google.com", "time": "Mon Mar 30 18:24:04 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Mon Mar 30 18:26:26 2026" }, "message": "[fuzzer] Use LIBCXX_ABI_UNSTABLE for hermetic libc++ (#189096)\n\nThis build of libc++ never interacts with any other, so\nit can always use the latest and best ABI.\n\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: bdf28a6d48112681afd24efe51a43e9807ff4fbb\n" }, { "commit": "d85a6b4019d74bc707dddeee5eb39196a028a6ca", "tree": "047812b1e320a9ebceb4dcb209dc4fd8854a1312", "parents": [ "9f10e4b8c42708cb44a57c4e2a83f10d87066247" ], "author": { "name": "Brian Cain", "email": "brian.cain@oss.qualcomm.com", "time": "Mon Mar 30 14:26:16 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Mon Mar 30 14:31:04 2026" }, "message": "[compiler-rt] Enable GWP-ASan for Hexagon (#188410)\n\nAdd Hexagon to ALL_GWP_ASAN_SUPPORTED_ARCH, add struct layout\nassertions.\n\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: a33acdb0b2dca00db08630050b10b09a7546235a\n" }, { "commit": "9f10e4b8c42708cb44a57c4e2a83f10d87066247", "tree": "58914b0d1dc2f8ddbe1a5d32c320a9562e27b21e", "parents": [ "1d86b1485db20454dd39c5ed5ee738b9103c8ed2" ], "author": { "name": "Brian Cain", "email": "brian.cain@oss.qualcomm.com", "time": "Mon Mar 30 14:25:14 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Mon Mar 30 14:30:50 2026" }, "message": "[Hexagon][XRay] Fix sled layout and trampoline to preserve retaddr (#188784)\n\nThe Hexagon XRay sled was 5 words (20 bytes) and the patched sequence\nclobbered r31 (the link register) via callr without saving it first.\nWhen the trampoline returned, the instrumented function\u0027s own allocframe\nwould then save the wrong return address, causing a crash or misrouted\nreturn.\n\nExpand the sled to 7 words (28 bytes) and wrap the call with\nallocframe(#0)/deallocframe so the caller\u0027s r31:30 are preserved across\nthe trampoline call.\n\nDetailed fixes:\n- HexagonAsmPrinter: emit 6 nop words after the jump (7 words total)\n- xray_hexagon.cpp: patch allocframe(#0) as first word, immext+r7 (func\nID), immext+r6 (trampoline), callr r6, deallocframe; write the first\nword last for atomicity\n- xray_trampoline_hexagon.S: complete rewrite -- properly load and\ndereference the global handler pointer, save/restore r0-r5 and r31, add\nstack frame with correct 8-byte alignment, add jumpr r31 to actually\nreturn from trampolines\n- xray_interface.cpp: update Hexagon cSledLength from 20 to 28\n- Update lit tests for 6-nop sled\n\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 5e7c66013bce8916d976355f5455e771d3bbcff4\n" }, { "commit": "1d86b1485db20454dd39c5ed5ee738b9103c8ed2", "tree": "3b691b7cffd57b21269c5b7b7c500caa4d8b6280", "parents": [ "26efb940ab2a9be8dbe83061947cfe044e749648" ], "author": { "name": "Simon Tatham", "email": "simon.tatham@arm.com", "time": "Mon Mar 30 14:11:00 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Mon Mar 30 14:15:35 2026" }, "message": "[compiler-rt][ARM] Rename endian.h to crt_endian.h (#189336)\n\nApparently on macOS there\u0027s a system header file also called\narm/endian.h, and another system header #includes it with \"\" rather than\n\u003c\u003e, so that this compiler-rt header accidentally shadows it. Worked\naround by prefixing \"crt\" to the name.\n\nNo changes are needed except the rename, because the planned functions\nthat use this header are still under review.\n\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: c7706be4bf06ca53ffcddcc96dacb4b6c4ac2782\n" }, { "commit": "26efb940ab2a9be8dbe83061947cfe044e749648", "tree": "d59aff632b73ef37b520fc66677bd1798ed717f9", "parents": [ "251b9df483e0c84aa3bdc353e8c9bb3d11158f9d" ], "author": { "name": "Brian Cain", "email": "brian.cain@oss.qualcomm.com", "time": "Sun Mar 29 18:01:39 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Sun Mar 29 18:02:46 2026" }, "message": "[compiler-rt][msan] Guard shmat interceptor w SANITIZER_INTERCEPT_SHMCTL (#189198)\n\nThe shmat interceptor calls REAL(shmctl), but shmctl is not intercepted\non all targets (e.g. 32-bit Linux with musl). Guard shmat behind\nSANITIZER_INTERCEPT_SHMCTL and use a MSAN_MAYBE_INTERCEPT pattern\nconsistent with other conditional interceptors.\n\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: a8cdc5a483efc90e4699df25027ae4916115dfed\n" }, { "commit": "251b9df483e0c84aa3bdc353e8c9bb3d11158f9d", "tree": "852696673db97a9c7867ff49676665431834e728", "parents": [ "0784dcca6673b744ada678779dc2c003ec232aec" ], "author": { "name": "Brian Cain", "email": "brian.cain@oss.qualcomm.com", "time": "Sun Mar 29 17:56:42 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Sun Mar 29 18:00:53 2026" }, "message": "[compiler-rt][msan] Add MSan support for Hexagon (Linux) (#189124)\n\nAdd the runtime infrastructure for MemorySanitizer on Hexagon Linux.\nHexagon is 32-bit, so the shadow memory layout uses a compact XOR-based\n mapping that fits within the lower 3GB of address space:\n\n 0x00000000 - 0x10000000 APP-1 (256MB, program text/data/heap)\n 0x10000000 - 0x20000000 ALLOCATOR (256MB)\n 0x20000000 - 0x40000000 SHADOW-1 (512MB, covers APP-1 + ALLOCATOR)\n 0x40000000 - 0x50000000 APP-2 (256MB, shared libs + stack)\n 0x60000000 - 0x70000000 SHADOW-2 (256MB, covers APP-2)\n 0x70000000 - 0x90000000 ORIGIN-1 (512MB)\n 0xB0000000 - 0xC0000000 ORIGIN-2 (256MB)\n\nMEM_TO_SHADOW uses XOR 0x20000000, and SHADOW_TO_ORIGIN adds 0x50000000.\n The dual-APP layout accommodates QEMU user-mode, which places shared\n libraries and the stack at 0x40000000.\n\n The allocator uses SizeClassAllocator32 with a 256MB region at\n 0x10000000, and kMaxAllowedMallocSize is set to 1GB consistent with\n other 32-bit targets.\n\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 44f1fa909953febfe019da533ec172da1348ac7b\n" }, { "commit": "0784dcca6673b744ada678779dc2c003ec232aec", "tree": "2aed5ebc86efdf6e409c3c02e36dfc08a7696ba8", "parents": [ "b51d4b3c64d234baac6a8ca4ce7e2966c474cdad" ], "author": { "name": "Brian Cain", "email": "brian.cain@oss.qualcomm.com", "time": "Sun Mar 29 03:14:52 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Sun Mar 29 18:00:41 2026" }, "message": "[compiler-rt][msan] Fix 32-bit overflow in CheckMemoryLayoutSanity (#189199)\n\nUse start + (end - start) / 2 instead of (start + end) / 2 to compute\nthe midpoint address. The original expression overflows when start + end\nexceeds UPTR_MAX, which happens on 32-bit targets whose memory layout\nincludes regions above 0x80000000.\n\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 670de1f5229a2b75a9711edd677b18eed60a587a\n" }, { "commit": "b51d4b3c64d234baac6a8ca4ce7e2966c474cdad", "tree": "10818ab3dfe8e97c73bad54be42614a2b6caef51", "parents": [ "721abfc9951e42ca53ad32604a8bbdce00bdd94b" ], "author": { "name": "Brian Cain", "email": "brian.cain@oss.qualcomm.com", "time": "Sun Mar 29 03:13:32 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Sun Mar 29 03:15:51 2026" }, "message": "[compiler-rt][sanitizer] Add struct_rlimit64_sz for musl (#189197)\n\nOn musl, rlimit64 is an alias for rlimit rather than a distinct type\nprovided by glibc. Add a SANITIZER_MUSL elif branch so that\nstruct_rlimit64_sz is defined for musl-based Linux targets.\n\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 89d57d03bfa0369fdc5d6825085291f3eb604ea8\n" }, { "commit": "721abfc9951e42ca53ad32604a8bbdce00bdd94b", "tree": "ce8c7ab0f9d0391d2f9a55c6a1e956ee24d8f47a", "parents": [ "00305d47c745d00b4f245518599068a00334e67d" ], "author": { "name": "PiJoules", "email": "leonardchan@google.com", "time": "Fri Mar 27 23:16:52 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Fri Mar 27 23:19:40 2026" }, "message": "[compiler-rt] Add interceptors for free_[aligned_]sized for asan+hwasan (#189109)\n\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: a5fa4dba6e2e01c4e8c4571301048c392facaed5\n" }, { "commit": "00305d47c745d00b4f245518599068a00334e67d", "tree": "5092682e4e345413c88adfd0d8ba86d995550f4a", "parents": [ "6fcac4c2c35597c5ed115b3461b9a73e052821a6" ], "author": { "name": "Luke Wren", "email": "wren6991@gmail.com", "time": "Fri Mar 27 22:47:57 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Fri Mar 27 22:50:30 2026" }, "message": "[RISCV] Allocate feature bits for Zifencei and Zmmul (#143306)\n\nAs proposed in\nhttps://github.com/riscv-non-isa/riscv-c-api-doc/pull/110.\n\nNo real compiler-rt implementation as Linux does not list these\nextensions in hwprobe.\n\nSigned-off-by: Luke Wren \u003cwren6991@gmail.com\u003e\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: efba01ae12703f1c6d6283a392ba852ac06a4f90\n" }, { "commit": "6fcac4c2c35597c5ed115b3461b9a73e052821a6", "tree": "b792995ad53f90ba270f1985877705ea9f8be83d", "parents": [ "292d1ee7b076e0353d4f40854d76fee21766e6b1" ], "author": { "name": "Joseph Huber", "email": "huberjn@outlook.com", "time": "Fri Mar 27 19:28:25 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Fri Mar 27 19:30:55 2026" }, "message": "[compiler-rt] Add PTX feature specifically when CUDA is not available (#189083)\n\nSummary:\nPeople need to be able to build this without a CUDA installation.\n\nLong term we should bump up the minimum version as I\u0027m pretty sure every\narchitecture before this has been deprecated by NVIDIA.\n\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 871d675c52ca4e2fad51aa90dd065d4f12229f06\n" }, { "commit": "292d1ee7b076e0353d4f40854d76fee21766e6b1", "tree": "f0be52dddc7806aea65aff596597eee9672c893f", "parents": [ "af1b293bfbbae36bf2efa3eb7ed5431a828889ee" ], "author": { "name": "Aiden Grossman", "email": "aidengrossman@google.com", "time": "Fri Mar 27 19:15:46 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Fri Mar 27 19:17:31 2026" }, "message": "[Scudo] Disable ScudoCombinedTests.NewType (#189070)\n\nThis is failing in some configurations on AArch64 Linux. Given there are\na lot of follow-up commits that makes this hard to revert, just disable\nit for now pending future investigation.\n\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: df6d6c9cd14113637cae8d4f51f9a05a09e7dd50\n" }, { "commit": "af1b293bfbbae36bf2efa3eb7ed5431a828889ee", "tree": "481e9bae0d97db29c0dd761071005996970ecd31", "parents": [ "b1899f796f64b22b6cba1816c6069641627068e9" ], "author": { "name": "Joseph Huber", "email": "huberjn@outlook.com", "time": "Fri Mar 27 17:07:42 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Fri Mar 27 17:10:28 2026" }, "message": "[compiler-rt] Fix irrelevant warning on the builtins target (#189055)\n\nSummary:\nCurrently, building through runtimes will yield this warning:\n```\n CMake Warning at compiler-rt/cmake/Modules/CompilerRTUtils.cmake:335 (message):\n LLVMTestingSupport not found in LLVM_AVAILABLE_LIBS\n Call Stack (most recent call first)\n```\n\nThis is due to the fact that the builtins target does not go through the\ns tandard runtimes patch and sets them as BUILDTREE_ONLY so they do not\nshow up. These are not used in this case, so just guard the condition to\nsuppress the warning.\n\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: f52797c54d4ef9269ac1da63388d794e5c3d4c0a\n" }, { "commit": "b1899f796f64b22b6cba1816c6069641627068e9", "tree": "b11d1be3018897c17570589417f855dfc37b3471", "parents": [ "385923496d3f1707c6ad8c9ac88fe1b8b07c6ab4" ], "author": { "name": "Brian Cain", "email": "brian.cain@oss.qualcomm.com", "time": "Fri Mar 27 14:32:42 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Fri Mar 27 14:35:21 2026" }, "message": "[XRay] Always register constructor(0) alongside .preinit_array (#188788)\n\nOn musl-based systems the dynamic linker does not process\nDT_PREINIT_ARRAY, so the .preinit_array entry alone never calls\n__xray_init(). Without initialization, the global XRay Flags struct is\nzero-initialized and flags()-\u003exray_mode is NULL. When the basic-mode or\nFDR-mode static initializers run from .init_array and call\ninternal_strcmp(flags()-\u003exray_mode, ...), they dereference NULL and\ncrash.\n\nFix this by always registering a constructor(0) in addition to the\n.preinit_array entry. On glibc where .preinit_array works, __xray_init()\nwill have already run and the constructor returns immediately (the\nfunction is idempotent). On musl, the constructor ensures __xray_init()\nruns before other .init_array entries that depend on XRay flags being\ninitialized.\n\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 3bff1a81c1b44e097b61dbf57090f61267baafe8\n" }, { "commit": "385923496d3f1707c6ad8c9ac88fe1b8b07c6ab4", "tree": "8801fd6373504ee237120e0253e68ac8721a6cf4", "parents": [ "f262f486bddff61f7fab53554e3bda5eb466969a" ], "author": { "name": "Dan Blackwell", "email": "dan_blackwell@apple.com", "time": "Fri Mar 27 10:13:01 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Fri Mar 27 10:15:59 2026" }, "message": "[NFCI][sanitizer_common] Realign #ifdefs in sanitizer_internal_defs.h (#186861)\n\nCurrently it is very hard to tell these nested ifdefs apart. This patch\nfixes that, while trying to be as light-touch as possible.\n\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 6a994417d85116c4ae4f21f15c35f944944443fa\n" }, { "commit": "f262f486bddff61f7fab53554e3bda5eb466969a", "tree": "4c725d94ba12771287401127379996111d934c8e", "parents": [ "d4fa79b40ac7283be590602a2a36e44362edebcf" ], "author": { "name": "Dan Blackwell", "email": "dan_blackwell@apple.com", "time": "Fri Mar 27 10:10:09 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Fri Mar 27 10:15:49 2026" }, "message": "[ASan][test-only] Remove superfluous guards in stack_container_dynamic_lib.c (#188469)\n\nAs noted in https://github.com/llvm/llvm-project/pull/188406 comments,\nthe documentation recommends guarding only with\n__has_feature(address_sanitizer). This patch updates the test to follow\nthe same pattern by removing the\n__SANITIZER_DISABLE_CONTAINER_OVERFLOW__ checks. Having this macro\ndefined results in the common_interface_defs.h header defining the\ncontiguous container functions as no-ops anyway.\n\nThis is a followup to https://github.com/llvm/llvm-project/pull/188406.\n\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 015994b7f9bce814c8851b7dd6d1718fc8dbb514\n" }, { "commit": "d4fa79b40ac7283be590602a2a36e44362edebcf", "tree": "70a6c580e72871b363f3b83e53434ce990e42d67", "parents": [ "1514ddd8c712de827f40335233df72d61eceada3" ], "author": { "name": "Christopher Ferris", "email": "cferris1000@users.noreply.github.com", "time": "Thu Mar 26 19:32:56 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Thu Mar 26 19:35:51 2026" }, "message": "[scudo] Add Last entry to ReleaseToOS enum. (#188645)\n\nThis allows static asserts to be set in tracing code that might use the\nReleaseToOS values as indexes.\n\nThis would have caused a compile failure instead of a runtime crash when\nI added the use of a new ReleaseToOS value.\n\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 6b3556a6ea2fda31be3a615119d1c8d89452ce36\n" }, { "commit": "1514ddd8c712de827f40335233df72d61eceada3", "tree": "4b948075c9493001a6f8261fa1bd1361b0377f64", "parents": [ "4ac7fcf4b6b648f39fc186555d012780d8240106" ], "author": { "name": "Joseph Huber", "email": "huberjn@outlook.com", "time": "Thu Mar 26 15:17:43 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Thu Mar 26 15:20:06 2026" }, "message": "[compiler-rt] Rework profile data handling for GPU targets (#187136)\n\nSummary:\nCurrently, the GPU iterates through all of the present symbols and\ncopies them by prefix. This is inefficient as it requires a lot of small\nhigh-latency data transfers rather than a few large ones. Additionally,\nwe force every single profiling symbol to have protected visibility.\nThis means potentially hundreds of unnecessary symbols in the symbol\ntable.\n\nThis PR changes the interface to move towards the start / stop section\nhandling. AMDGPU supports this natively as an ELF target, so we need\nlittle changes. Instead of overriding visibility, we use a single table\nto define the bounds that we can obtain with one contiguous load.\n\nUsing a table interface should also work for the in-progress HIP\nimplementation for this, as it wraps the start / stop sections into\nstandard void pointers which will be inside of an already mapped region\nof memory, so they should be accessible from the HIP API.\n\nNVPTX is more difficult as it is an ELF platform without this support. I\nhave hooked up the \u0027Other\u0027 handling to work around this, but even then\nit\u0027s a bit of a stretch. I could remove this support here, but I wanted\nto demonstrate that we can share the ABI. However, NVPTX will only work\nif we force LTO and change the backend to emit variables in the same\n\nTL;DR, we now do this:\n```c\nstruct { start1, stop1, start2, stop2, start3, stop3, version; } device;\nstruct host \u003d DtoH(lookup(\"device\"));\ncounters \u003d DtoH(host.stop - host.start)\nversion \u003d DtoH(host.version);\n```\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: ffd6a13b5ffed8bb2e314db29be784381f984ae1\n" }, { "commit": "4ac7fcf4b6b648f39fc186555d012780d8240106", "tree": "dbc1054199e8525974e13433f7929d7bc46c5c9c", "parents": [ "6c49aa534ca3e309f55583c3e7ba9d88e35daeaa" ], "author": { "name": "Dan Blackwell", "email": "dan_blackwell@apple.com", "time": "Thu Mar 26 13:31:27 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Thu Mar 26 13:36:10 2026" }, "message": "[ASan] Make __SANITIZER_DISABLE_CONTAINER_OVERFLOW__ bypass definitions return values (#188237)\n\nCurrently all of these functions are empty bodies; this means that when\nincluding this header and compiling with\n__SANITIZER_DISABLE_CONTAINER_OVERFLOW__ defined, warnings are emitted\nabout the missing return values for those functions that do return\nvalues.\n\nThis patch returns success values for all those check functions with\nnon-void return types. Note: these were originally added in\nhttps://github.com/llvm/llvm-project/pull/163468.\n\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: bbccfa5654fe9202219e0811b1f518a1a039ddb7\n" }, { "commit": "6c49aa534ca3e309f55583c3e7ba9d88e35daeaa", "tree": "2e2d4cfd1851d0ee476d386533e90cc9e44373e9", "parents": [ "e57628c55305f07bc49de2e6e3644f96d1bccbf0" ], "author": { "name": "Simon Tatham", "email": "simon.tatham@arm.com", "time": "Thu Mar 26 09:10:15 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Thu Mar 26 09:15:28 2026" }, "message": "[compiler-rt][ARM] Double-precision FP support functions (#179920)\n\nThis commit adds C helper functions `dnan2`, `dnorm2` and `dunder` for\nhandling the less critical edge cases of double-precision arithmetic,\nsimilar to `fnan2`, `fnorm2` and `funder` that were added in commit\nf7e652127772e93.\n\nIt also adds a header file that defines some register aliases for\nhandling double-precision numbers in AArch32 software floating point in\nan endianness-independent way, by providing aliases `xh` and `xl` for\nthe high and low words of the first double-precision function argument,\nregardless of which of them is in r0 and which in r1, and similarly `yh`\nand `yl` for the second argument in r2/r3.\n\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 80831832e03f8184fb24ffd08ccba59d4e57ec0e\n" }, { "commit": "e57628c55305f07bc49de2e6e3644f96d1bccbf0", "tree": "fe68b100ae36207b30584f3f8868f4b7c0a58703", "parents": [ "9b87f644b1ca99e4ee2fce3f4d2217b3c3d60fed" ], "author": { "name": "Jake Egan", "email": "Jake.egan@ibm.com", "time": "Thu Mar 26 06:13:47 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Thu Mar 26 06:15:46 2026" }, "message": "[sanitizer_common] Implement address sanitizer on AIX: stack unwinding (#138188)\n\nImplement unwind.h-based stack unwinding routines on AIX.\n\nIssue: https://github.com/llvm/llvm-project/issues/138916\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 25904ac91554f0a94fbe2fc8715a16e17a84ce2b\n" }, { "commit": "9b87f644b1ca99e4ee2fce3f4d2217b3c3d60fed", "tree": "5055d3ede1cef3971f5edece4912867dd56bd4f0", "parents": [ "838a7243c3fff8ecab030383d202b81bc9fdc0e4" ], "author": { "name": "Christopher Ferris", "email": "cferris1000@users.noreply.github.com", "time": "Wed Mar 25 17:48:22 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Wed Mar 25 17:51:14 2026" }, "message": "[scudo] Remove some android wrapper checking. (#188339)\n\nTests for Android specific behavior don\u0027t really belong here since it is\naffected by the config which is not necessarily the same on Android.\nThere are already tests that the config options and flag options work\nproperly. Android wrapper tests belong to Android.\n\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: a3260f658242ff8613f9a852ef95c6f1e644c95a\n" }, { "commit": "838a7243c3fff8ecab030383d202b81bc9fdc0e4", "tree": "ae66369857c2722c101a36dba6947f8a0665f9a3", "parents": [ "0ba0fee05ed512c647586982bcd0e7d899a16289" ], "author": { "name": "Fabio D\u0027Urso", "email": "fdurso@google.com", "time": "Wed Mar 25 10:46:52 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Wed Mar 25 10:51:08 2026" }, "message": "[scudo] Use portable TEST_SKIP macro (#188045)\n\nWhich expands to ZXTEST_SKIP on Fuchsia.\n\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 1d2f14f625cf3fd2025e45ea584440a8e7e0cc26\n" }, { "commit": "0ba0fee05ed512c647586982bcd0e7d899a16289", "tree": "0f431cc79c4f97e432e4e0dca01c592a6fb189e3", "parents": [ "706faaccd4af2de0211e5eb013ff7f8d6b0fa663" ], "author": { "name": "Andrew Haberlandt", "email": "ahaberlandt@apple.com", "time": "Wed Mar 25 06:49:14 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Wed Mar 25 06:50:24 2026" }, "message": "[compiler-rt] CRT builtins tests should not run on mac/windows under LLVM_ENABLE_RUNTIMES (#187835)\n\n#171941 got the builtins tests running under LLVM_ENABLE_RUNTIMES by\ntesting the builtins as part of the runtimes build.\n\nAs a consequence, CMake in `lib/builtins/` is no longer visible when\nconfiguring the tests (but `test/builtins/` is). This means that the\n`cmake_dependent_option` from `lib/builtins/` is not accounted for by\nthe tests, allowing COMPILER_RT_BUILD_CRT to be YES when\nCOMPILER_RT_HAS_CRT is NO. As a consequence, the CRT tests are running\non platforms where COMPILER_RT_HAS_CRT is false (#176892).\n\nhttps://github.com/llvm/llvm-project/blob/367da15a11c52886c50e7f020cb4de59fe6d07ca/compiler-rt/lib/builtins/CMakeLists.txt#L1106-L1108\n\nAlthough the long-term solution could be to split both the builtins (and\ntheir tests) out of compiler-rt into a top-level directory with shared\noptions, this works around the issue for the moment by checking both\nCOMPILER_RT_HAS_CRT and COMPILER_RT_BUILD_CRT before enabling the \"crt\"\nfeature.\n\nFixes #176892\n\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: a5a7f6266ef05aa0093a779d6f2fc8e513a352e6\n" }, { "commit": "706faaccd4af2de0211e5eb013ff7f8d6b0fa663", "tree": "691bacdfd3140e64246502dcb0fdf53cedbfe1ef", "parents": [ "21b19cf94c02ae093a731d6fc763f88adc010330" ], "author": { "name": "Vitaly Buka", "email": "vitalybuka@google.com", "time": "Wed Mar 25 05:28:46 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Wed Mar 25 05:30:38 2026" }, "message": "[asan] Convert __SANITIZER_DISABLE_CONTAINER_OVERFLOW__ tests to C (#188406)\n\nAs-is tests do not pass on android with older C++ headers.\nThere is nothing C++ specific in tests.\n\nFollowup to #181721.\n\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 3b04d26abd50bb50f8ffa889439a30f2b778d9a1\n" }, { "commit": "21b19cf94c02ae093a731d6fc763f88adc010330", "tree": "8a656c0999a93cf3b79a1aae648fd4cc5640391f", "parents": [ "ce8678569038bfe1824b109a6ffba1f310e6a6a3" ], "author": { "name": "Vitaly Buka", "email": "vitalybuka@google.com", "time": "Tue Mar 24 19:50:35 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Wed Mar 25 01:25:50 2026" }, "message": "[compiler-rt] Suppress unused variable report in emutls\n\nPull Request: https://github.com/llvm/llvm-project/pull/188329\n\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: cb125346509a10b48d32788a8242b3722e5113d9\n" }, { "commit": "ce8678569038bfe1824b109a6ffba1f310e6a6a3", "tree": "d6a4f5fe09541ee179fbb412195d418e3e056b2b", "parents": [ "bb7645f5e11c9c1d719a890fcb09ccfaaa14580f" ], "author": { "name": "Vitaly Buka", "email": "vitalybuka@google.com", "time": "Tue Mar 24 18:23:37 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Tue Mar 24 18:25:38 2026" }, "message": "[libFuzzer] Fix -Wunused-variable when building with NDEBUG (#188301)\n\nThe variable `FuzzerInitIsRunning` is only used within `assert()`.\nFollow up to #178342.\n\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 9fd8bc0f0eedd3d6ca2c71f90de768a7759ab44c\n" }, { "commit": "bb7645f5e11c9c1d719a890fcb09ccfaaa14580f", "tree": "a01b93db2db32fde884702812e15e84ed931935e", "parents": [ "741bad5403e3244786e1da0e3f0e706023e72017" ], "author": { "name": "Joseph Huber", "email": "huberjn@outlook.com", "time": "Tue Mar 24 16:44:24 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Tue Mar 24 16:45:26 2026" }, "message": "[compiler-rt] Support unit tests for the GPU build (#187895)\n\nSummary:\nThis PR enables the basic unit tests for builtins to be run on the GPU\narchitectures. Other targets like profiling are supported, but the\nhost-device natures will make it more difficult to adequately unit\ntest. It may be be possible to do basic tests there, to simply verify\nthat\ncounters are present and in the proper format for when they are copied\nto the host.\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: e73d8f8fc59ae2108baf863a6c8c8cfddbbca537\n" }, { "commit": "741bad5403e3244786e1da0e3f0e706023e72017", "tree": "bdf93f69ecb88176fb8749084a60c10228c2e361", "parents": [ "3f13273a685bc932e3b580c0707e011c9dfb97d5" ], "author": { "name": "Daniil Kovalev", "email": "dkovalev@accesssoftek.com", "time": "Tue Mar 24 09:24:19 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Tue Mar 24 09:25:57 2026" }, "message": "[PAC][compiler-rt] Use `__ptrauth` qualifier instead of builtins (#188109)\n\nSince #100830 has landed, we no longer need to rely on builtins\n\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 34875331d956862d6f5b750c0bcf026d40564671\n" }, { "commit": "3f13273a685bc932e3b580c0707e011c9dfb97d5", "tree": "a3ccad6b72ff3aef2d5fe3873957880c99c8a139", "parents": [ "cb6a351a238424067a0b019388fa88de116042ab" ], "author": { "name": "Dan Blackwell", "email": "dan_blackwell@apple.com", "time": "Tue Mar 24 08:23:15 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Tue Mar 24 08:25:46 2026" }, "message": "[ASan] Enable __SANITIZER_DISABLE_CONTAINER_OVERFLOW__ tests (#181721)\n\nNow that the corresponding libcxx change has landed, these tests should\nbe passing on some platforms.\n\nThis patch re-enables them for all platforms, so that we can see which\nbots these do not work on and mark them unsupported accordingly.\n\nrdar://167946476\n\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 96fb2942cb522b05f773f7492c1c125c7aef620c\n" }, { "commit": "cb6a351a238424067a0b019388fa88de116042ab", "tree": "18cb5b8cfc06743a938ab95dce9168d191b31cd8", "parents": [ "28c233d2e119c0a260083be9247c85ccd981b18d" ], "author": { "name": "Midhunesh", "email": "midhunesh.p@ibm.com", "time": "Tue Mar 24 00:48:25 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Tue Mar 24 00:51:03 2026" }, "message": "[compiler-rt] Error only if sanitizer builds are enabled (#187944)\n\nThis fix is to error out only if sanitizer builds are enabled as opposed\nto forcing flang to use an option or install the development package.\n\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 6a8a1008b56c8a3532b19551292e71216f0ea03e\n" }, { "commit": "28c233d2e119c0a260083be9247c85ccd981b18d", "tree": "f1549f72a1bcea951a274c9ecabcf19204554223", "parents": [ "0a638b1a5e41834d0b438b925e53c70d3108f34a" ], "author": { "name": "Simon Tatham", "email": "simon.tatham@arm.com", "time": "Mon Mar 23 16:01:12 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Mon Mar 23 16:05:36 2026" }, "message": "[compiler-rt][ARM] cmake properties for complicated builtin sources (#179919)\n\nIn the builtins library, most functions have a portable C implementation\n(e.g. `mulsf3.c`), and platforms might provide an optimized assembler\nimplementation (e.g. `arm/mulsf3.S`). The cmake script automatically\nexcludes the C source file corresponding to each assembly source file it\nincludes. Additionally, each source file name is automatically\ntranslated into a flag that lit tests can query, with a name like\n`librt_has_mulsf3`, to indicate that a function is available to be\ntested.\n\nIn future commits I plan to introduce cases where a single .S file\nprovides more than one function (so that they can share code easily),\nand therefore, must supersede more than one existing source file.\n\nI\u0027ve introduced the `crt_supersedes` cmake property, which you can set\non a .S file to name a list of .c files that it should supersede. Also,\nthe `crt_provides` property can be set on any source file to indicate a\nlist of functions it makes available for testing, in addition to the one\nimplied by its name.\n\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 44df4116c87c023264c2a0551f808f031489e8f9\n" }, { "commit": "0a638b1a5e41834d0b438b925e53c70d3108f34a", "tree": "b38101fe773869b495d31c49600347c78771dfc9", "parents": [ "2ab760b3c391c8175bf0b9557c756a62ce4e727c" ], "author": { "name": "Michael Jones", "email": "michaelrj@google.com", "time": "Fri Mar 20 23:46:38 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Fri Mar 20 23:50:27 2026" }, "message": "[compiler-rt] Add bitmask to fix warning (#187812)\n\nAfter #186881 was merged the gcc libc bots started complaining about the\nconversion from u8 to 2 bit integer being unsafe (see:\nhttps://lab.llvm.org/buildbot/#/builders/131/builds/42788). This PR\nadds a bitmask that fixes the warning.\n\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 6891a6ef0e2466ffb8a921d7f58934ef464f137a\n" }, { "commit": "2ab760b3c391c8175bf0b9557c756a62ce4e727c", "tree": "61d32fa0476728f80bc6d8defdfb7f4319fd678c", "parents": [ "e6de493dc05969ead81d7ebab65d009a74d2cce2" ], "author": { "name": "Christopher Ferris", "email": "cferris1000@users.noreply.github.com", "time": "Fri Mar 20 22:32:16 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Fri Mar 20 22:33:28 2026" }, "message": "[scudo] Make the default for size/align checks to not die. (#187799)\n\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 8cc0124508a55de25ccb2422ec9a2a95f1bc8020\n" }, { "commit": "e6de493dc05969ead81d7ebab65d009a74d2cce2", "tree": "e9f62adefa56c9418685cec27746dac35af04d85", "parents": [ "0212f0888c75e9fb9f48586b33bc1d981a43dad6" ], "author": { "name": "Christopher Ferris", "email": "cferris1000@users.noreply.github.com", "time": "Fri Mar 20 20:26:35 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Fri Mar 20 20:30:36 2026" }, "message": "[scudo] Add free_sized and free_aligned_sized (#186881)\n\nAdd one new flag, dealloc_align_mismatch that turns on/off alignment\nchecks. Add three new config parameters, one for deallocate type\nmismatch (such as abort on new/free if true), one for checking if the\nsize parameter matches on dealloc and one for checking if the alignment\nis correct on a dealloc.\n\nAdd extra flags to be passed for to indicate to do an align/size check.\n\nUpdate report functions to better indicate the errors. Add unit tests\nfor all of these.\n\nThis is based on these upstream cls by jcking:\n\nhttps://github.com/llvm/llvm-project/pull/147735\nhttps://github.com/llvm/llvm-project/pull/146556\n\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 1b44e34b18e8823c858036087407c5e5976cec06\n" }, { "commit": "0212f0888c75e9fb9f48586b33bc1d981a43dad6", "tree": "b4c86b71434c1a25d677881e7d6b3449af7ea90c", "parents": [ "84bfdcd752cb00ccc8087aeb584d3c08e9da96ba" ], "author": { "name": "Roman Vinogradov", "email": "roman.vinogradov@sap.com", "time": "Fri Mar 20 17:05:19 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Fri Mar 20 17:09:28 2026" }, "message": "[ASan] Fix missed poisoned suffix in first granule in __asan_region_is_poisoned (#187466)\n\nAlign beg address down instead of up in __asan_region_is_poisoned(), so\nthe shadow scan includes the first granule. This fixes a false negative\nwhen first granule has an unpoisoned prefix and poisoned suffix.\n\nAdd test that covers this scenario.\n\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: ca54948d0b7a914211ae3d8d95e309358576c3f1\n" }, { "commit": "84bfdcd752cb00ccc8087aeb584d3c08e9da96ba", "tree": "14215d617eb305f919f14ceafe8b7d7c71d9e1d8", "parents": [ "59066a662511e1f9959b07a32c88906d7212b6a7" ], "author": { "name": "Joseph Huber", "email": "huberjn@outlook.com", "time": "Thu Mar 19 15:51:48 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Thu Mar 19 15:56:26 2026" }, "message": "[compiler-rt] Define GPU specific handling of profiling functions (#185763)\n\nSummary:\nThe changes in https://www.github.com/llvm/llvm-project/pull/185552\nallowed us to\nstart building the standard `libclang_rt.profile.a` for GPU targets.\nThis PR expands this by adding an optimized GPU routine for counter\nincrement and removing the special-case handling of these functions in\nthe OpenMP runtime.\n\nVast majority of these functions are boilerplate, but we should be able\nto do more interesting things with this in the future, like value or\nmemory profiling.\n\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: d18a784d410617fbb3f9d2bf6d30fcd84b3533db\n" }, { "commit": "59066a662511e1f9959b07a32c88906d7212b6a7", "tree": "0fb0ab14337e17318d21bcef972880ed17a2ecc0", "parents": [ "a6484a8fb26b15c56b2e0fb069ea0f1cd9ffa706" ], "author": { "name": "Ross Burton", "email": "ross.burton@arm.com", "time": "Thu Mar 19 00:48:08 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Thu Mar 19 00:50:40 2026" }, "message": "[compiler-rt] Unify python shebangs (#187285)\n\nAs per PEP-0394[1], there is no real concensus over what binary names\nPython has, specifically \u0027python\u0027 could be Python 3, Python 2, or not\nexist.\n\nHowever, everyone has a python3 interpreter and the scripts are all\nwritten for Python 3. Unify the shebangs so that the ~50% of shebangs\nthat use python now use python3.\n\n[1] https://peps.python.org/pep-0394/\n\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 015e3d2b20924e966345d3c493fb16f2fa3160b4\n" }, { "commit": "a6484a8fb26b15c56b2e0fb069ea0f1cd9ffa706", "tree": "70786e17a8bd89843e1ccac868c718bcc852cc7a", "parents": [ "c09cee4577254cc33ee4a7699bbe3d7ae3ea730a" ], "author": { "name": "Roman Vinogradov", "email": "roman.vinogradov@sap.com", "time": "Wed Mar 18 16:43:19 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Wed Mar 18 16:45:40 2026" }, "message": "[ASan] Fix overflow and last byte handling in __asan_region_is_poisoned (#183900)\n\n__asan_region_is_poisoned() uses an exclusive end address\n(end \u003d beg + size) to validate the region [beg, end) and to compute\nthe aligned inner shadow region. This causes correctness issue\nnear memory range upper boundary and could trigger address space\noverflow on 32-bit targets.\n\n1. Incorrect handling of the last byte of a memory range\n\n The implementation checks AddrIsInMem(end) instead of the last\n application byte (end - 1). For regions ending at the last byte\n of Low/Mid/HighMem (e.g. __asan_region_is_poisoned(kHighMemEnd, 1)),\n this returns end (kHighMemEnd + 1) instead of the original\n pointer. This behavior is inconsistent with the function’s\n semantics and with __asan_address_is_poisoned().\n\n2) address space overflow and invalid shadow range\n\nIf a region ends at the top of the virtual address space (kHighMemEnd),\n e.g. on 32-bit targets, end \u003d beg + size could wrap to 0.\n This violated the invariant beg \u003c end and could trigger\n the CHECK failure.\n\n Additionally, overflow in RoundUpTo alignment computations\n for aligned_b could produce an invalid shadow region spanning\n LowShadow to HighShadow across ShadowGap, leading mem_is_zero()\n to access unmapped memory and crash.\n\nFix by switching to an inclusive last byte:\n\n last \u003d beg + size - 1\n\nAll checks are now performed on beg and last. The aligned inner\nshadow region is also computed from [beg, last]. Additional guard\nfor aligned_b prevents the mapping to shadow if aligned_b is wrapped\n(in this case the aligned inner region is also empty and doesn\u0027t\nrequire the shadow scan via mem_is_zero()).\n\nThis fixes incorrect return values at memory range ends and\nprevents overflow related crashes on 32-bit targets.\n\nTest is extended to cover these boundary cases.\n\n---------\n\nCo-authored-by: Vitaly Buka \u003cvitalybuka@gmail.com\u003e\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 2caba086abeb8f10b4d723b3d2cc205f2da7bbc3\n" }, { "commit": "c09cee4577254cc33ee4a7699bbe3d7ae3ea730a", "tree": "efe5486214600132b87511cbad6aea5cf8db9d15", "parents": [ "573ea4ab862e838d5c8512f654799d14deca07e7" ], "author": { "name": "Midhunesh", "email": "midhunesh.p@ibm.com", "time": "Wed Mar 18 14:22:22 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Wed Mar 18 14:25:47 2026" }, "message": "[compiler-rt] Update runtime build script to detect RPC XDR header for AIX (#186977)\n\n`sanitizer_common` and its tests depend on the RPC XDR header for layout\ncompatibility. When this header is absent from a CI or build\nenvironment, changes that silently break the expected struct layout go\nundetected, since there is nothing to fail the build.\nThe default is opt-in — error on missing header is on by default for AIX\n(where the dependency is known and the package is `bos.net.nfs.adt`) and\noff by default elsewhere.\n\nChanges:\n\n1. On AIX, checks for `tirpc/rpc/xdr.h`; on all other platforms, checks\nfor `rpc/xdr.h`\n2. Introduces `COMPILER_RT_REQUIRE_RPC_XDR_H` CMake option (default ON\non AIX, OFF elsewhere) that, when set, turns a missing header into a\nfatal configuration error with an actionable message\n3. Drive-by fix: Normalizes `HAVE_RPC_XDR_H` to 0 when the header is\nabsent, for consistent downstream `if()/#cmakedefine` behavior\n\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: d783723a584a1cab30c3a92ca247abb3401fc6da\n" }, { "commit": "573ea4ab862e838d5c8512f654799d14deca07e7", "tree": "f26b41ad4a179c625934357c199b664d04e5008f", "parents": [ "b93edfeca73cf0d8537cf1c84ede88a9768c7bab" ], "author": { "name": "cherrymui", "email": "lunaria21@gmail.com", "time": "Mon Mar 16 18:49:56 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Mon Mar 16 18:51:46 2026" }, "message": "[sanitizer_common] Define SANITIZER_WEAK_IMPORT for Go race detector (#186525)\n\nCurrently, when building the Go race detector (when SANITIZER_GO\nis set), SANITIZER_WEAK_IMPORT is no-op. It is perfectly fine to\ndefine SANITIZER_WEAK_IMPORT for Go just like other cases. That\nwill tell the Go linker to treat _dyld_get_dyld_header as a weak\nimport.\n\nPerhaps SANITIZER_WEAK_ATTRIBUTE can also be defined for Go. That\nwould be a separate patch.\n\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 1098b95dac5d07534da4d5e4a6396e0881150ef5\n" }, { "commit": "b93edfeca73cf0d8537cf1c84ede88a9768c7bab", "tree": "995360b76da8b53ce5c1882850e76261be9dab40", "parents": [ "ea94b34eeb6351bda6a3185cc5393263a7c5a1d4" ], "author": { "name": "Vitaly Buka", "email": "vitalybuka@google.com", "time": "Sat Mar 14 00:33:28 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Sat Mar 14 00:35:26 2026" }, "message": "[NFC][sanitizer] Accept ETIMEDOUT in getpwnam_r_invalid_user.cpp (#186538)\n\nOn some systems, looking up an result in a timeout.\n\nError here is not a sign of compiler-rt issue.\n\nFixes flakiness on\nhttps://lab.llvm.org/buildbot/#/builders/sanitizer-ppc64le-linux\n\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 5a87cba9294c4d5ddb4a800f21e6e96fcc83cefa\n" }, { "commit": "ea94b34eeb6351bda6a3185cc5393263a7c5a1d4", "tree": "d2c5f61fa7275be599b3cb828f6750cb2e32799d", "parents": [ "c8ee6566881410a159d66e99081779d28124fd3f" ], "author": { "name": "Brian Cain", "email": "brian.cain@oss.qualcomm.com", "time": "Fri Mar 13 01:27:11 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Fri Mar 13 01:30:43 2026" }, "message": "[compiler-rt] Add ASan/UBSan runtime support for Hexagon Linux (#183982)\n\nAdd the architecture-specific pieces needed for the ASan and UBSan\nsanitizer runtimes to build and run on hexagon-unknown-linux-musl.\n\nWithout this patch, building sanitizer runtimes for Hexagon Linux fails\nwith:\n\n sanitizer_linux.cpp: error: member access into incomplete type\n \u0027struct stat64\u0027\n\nbecause musl libc does not provide struct stat64. This patch routes\nHexagon through the statx() syscall path (like LoongArch) to avoid the\nstat64 dependency entirely.\n\nChanges:\n\n* asan_mapping.h: Add ASAN_SHADOW_OFFSET_CONST (0x20000000) for Hexagon\nwith shadow layout documentation.\n* sanitizer_linux.cpp: Implement internal_clone() for Hexagon using\ninline assembly (trap0 syscall, generic clone argument order: flags,\nstack, ptid, ctid, tls). Route Hexagon through the statx() path for stat\noperations since musl lacks struct stat64.\n* sanitizer_linux.h: Add Hexagon to the internal_clone() declaration\nguard.\n* sanitizer_stoptheworld_linux_libcdep.cpp: Add Hexagon to the\nStopTheWorld architecture guard with register definitions.\n* sanitizer_asm.h: Define ASM_TAIL_CALL as \u0027jump\u0027 for Hexagon.\n* CMakeLists.txt: Add -fno-emulated-tls for Hexagon targets. Hexagon\nLinux uses native TLS via the UGP register; emulated TLS produces broken\nsanitizer runtimes with unresolvable __emutls references.\n\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: cef418ec4b8bbc92700e84a873f2b4c1985e966e\n" }, { "commit": "c8ee6566881410a159d66e99081779d28124fd3f", "tree": "2ed60c111df1519b15560fe3a377e8de38c71634", "parents": [ "ba616e8b55173e8d6362af87fa22ee601cceb1c0" ], "author": { "name": "Christopher Ferris", "email": "cferris1000@users.noreply.github.com", "time": "Thu Mar 12 20:19:33 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Thu Mar 12 20:25:40 2026" }, "message": "[scudo] Display flags on secondary cache entries. (#185786)\n\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: dae2923901fa456d6b5a39426ceb67185835a427\n" }, { "commit": "ba616e8b55173e8d6362af87fa22ee601cceb1c0", "tree": "ffd81667dab07b129989de75d9447de6079a5e78", "parents": [ "b40bcbfe9118507084e2529ce5a2e561081be587" ], "author": { "name": "DylanFleming-arm", "email": "85629460+DylanFleming-arm@users.noreply.github.com", "time": "Thu Mar 12 14:14:04 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Thu Mar 12 14:15:37 2026" }, "message": "[libc] Remove unused AOR_v20.02 directory (#185951)\n\nAs far as I am aware, AOR is no longer used anywhere within LLVM, as\nmost of the required code has since been ported to elsewhere within the\nproject.\n\nRemoves the entire directory, and updates some now outdated comments.\n\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: d682325961bad9217264201843f38c2b881af80a\n" }, { "commit": "b40bcbfe9118507084e2529ce5a2e561081be587", "tree": "a316444fc6c49d47d5079cd3cc2adabe1abcebec", "parents": [ "e1e57b877616abd748cdd63bb78e913e26f5bdfa" ], "author": { "name": "Keith Packard", "email": "keithp@keithp.com", "time": "Thu Mar 12 10:37:48 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Thu Mar 12 10:40:24 2026" }, "message": "compiler-rt/arm: Check for overflow when adding float denorms (#185245)\n\nWhen the sum of two sub-normal values is not also subnormal, we need to\nset the exponent to one.\n\nTest case:\n\nstatic volatile float x \u003d 0x1.362b4p-127;\nstatic volatile float x2 \u003d 0x1.362b4p-127 * 2;\n\nint\nmain (void)\n{\n\tprintf(\"x %a x2 %a x + x %a\\n\", x, x2, x + x);\n\treturn x2 \u003d\u003d x + x ? 0 : 1;\n}\n\nSigned-off-by: Keith Packard \u003ckeithp@keithp.com\u003e\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: ab6bb1bab69233128e859282b969fbda3cc5dfed\n" }, { "commit": "e1e57b877616abd748cdd63bb78e913e26f5bdfa", "tree": "88c0f07f7a9f913f62f006aa83e7d46f47b46de8", "parents": [ "b2341a254b6762d5e279b9811157e7d0a5643e7f" ], "author": { "name": "Nico Weber", "email": "thakis@chromium.org", "time": "Wed Mar 11 17:42:05 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Wed Mar 11 17:45:52 2026" }, "message": "Revert \"[CFI] Expand test to include minimal runtime\" (#185935)\n\nReverts llvm/llvm-project#183646\n\nBreaks tests if ubsan_minimal runtime isn\u0027t built, see:\nhttps://github.com/llvm/llvm-project/pull/183646#issuecomment-3994252478\n\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: d371f42a47f1ee71a2db65e19de5626a2e7bb2b9\n" }, { "commit": "b2341a254b6762d5e279b9811157e7d0a5643e7f", "tree": "c843ad419a0a7c278ab45d0fbc476f611b2c5178", "parents": [ "90e2a1b0c63a66af44919d1a7b97fb03078f5c62" ], "author": { "name": "Alex Crichton", "email": "alex@alexcrichton.com", "time": "Wed Mar 11 04:55:06 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Wed Mar 11 05:00:13 2026" }, "message": "[WebAssembly] Remove `__c_longjmp` from compiler-rt (#185798)\n\nThis is similar to #185770 where it removes an\nexception-handling-related symbol from `compiler-rt` in favor of having\ndefinitions elsewhere. The compiler-rt library is linked into all shared\nobjects, for example, which can result in duplicate definitions of a\nsymbol where this tag wants to have one unique definition. The intention\nbehind this commit is to defer the definition of this symbol to\ndownstream libraries, such as the definition of `longjmp` itself. An\nexample of this is WebAssembly/wasi-libc#772 where the responsibility of\ndefining this symbol now lies with wasi-libc.\n\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 7a43f770a0a493d9eb31101251de32fa62cb7ab1\n" }, { "commit": "90e2a1b0c63a66af44919d1a7b97fb03078f5c62", "tree": "c1e95dd2a3b25415e1bbfb31139f9a2461cd2c1c", "parents": [ "5a8af2774bbdb6f26ad9229580ad90ea2e563d85" ], "author": { "name": "Alex Crichton", "email": "alex@alexcrichton.com", "time": "Wed Mar 11 00:43:11 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Wed Mar 11 00:45:59 2026" }, "message": "[WebAssembly] Move __cpp_exception to libunwind (#185770)\n\nThe `__cpp_exception` symbol is now defined in libunwind instead of\ncompiler-rt. This is moved for a few reasons, but the primary reason is\nthat compiler-rt is linked duplicate-ly into all shared objects meaning\nthat it\u0027s not suitable for define-once symbols such as\n`__cpp_exception`. By moving the definition to the user of the symbol,\nlibunwind itself, that guarantees that the symbol should be defined\nexactly once and only when appropriate. A secondary reason for this\nmovement is that it avoids the need to compile compiler-rt twice: once\nwith exception and once without, and instead the same build can be used\nfor both exceptions-and-not.\n\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: a3f28233fc860d77fa70ad57a93a67963fa20a0c\n" }, { "commit": "5a8af2774bbdb6f26ad9229580ad90ea2e563d85", "tree": "5e46c8982c1ca3472fdeec27a1a63253e7b226ec", "parents": [ "c8a567684659fe05ede3a220965b94fd71c4a7cb" ], "author": { "name": "Joseph Huber", "email": "huberjn@outlook.com", "time": "Tue Mar 10 18:45:18 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Tue Mar 10 18:50:42 2026" }, "message": "[copmiler-rt] Initial support for building profile library on the GPU (#185552)\n\nSummary:\nAs suggested in https://github.com/llvm/llvm-project/pull/177665, we\nshould build a GPU version of the compiler-rt profile library instead of\nwriting it in-line in the lowering. This PR does not define anything GPU\nspecific, it simply re-uses the baremetal handling. Later PRs will\nprevent the GPU specific handling we would want to do to optimize\ncounter handling on the GPU.\n\nNote that this will require using the cache file, or setting these\noptions\nmanually for existing users. Hopefully if people are using the cache\nfile\nas they should it won\u0027t break anything.\n\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: fd069a46bf77227e058475508c679f3a004d7b36\n" }, { "commit": "c8a567684659fe05ede3a220965b94fd71c4a7cb", "tree": "3fc7b7e59688c0c580c3e59951bdd6d14dd3790a", "parents": [ "ec616cf086689eb669ebb93c34bc757756c87531" ], "author": { "name": "Christopher Ferris", "email": "cferris1000@users.noreply.github.com", "time": "Mon Mar 09 19:50:32 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Mon Mar 09 19:53:42 2026" }, "message": "[scudo] Use a fixed format for the milliseconds in latest release. (#185097)\n\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 3e9808befb70a73593d6d98cf744018b4ecab89c\n" }, { "commit": "ec616cf086689eb669ebb93c34bc757756c87531", "tree": "2a4e274f199c5b0db8aa530af08a754314712bc5", "parents": [ "338a5c004c774a8927899b1f1c0c25a82d14510f" ], "author": { "name": "Fabio D\u0027Urso", "email": "fdurso@google.com", "time": "Mon Mar 09 17:48:53 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Mon Mar 09 17:51:01 2026" }, "message": "[scudo] Add missing class name specifier in mem_map_fuchsia (#185389)\n\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: ae8f614a0cb63081487506227e373c05bfc45e0c\n" }, { "commit": "338a5c004c774a8927899b1f1c0c25a82d14510f", "tree": "f295a0afd3d1fdf9fb5424eb2995bace3e6d47b5", "parents": [ "3cac82da8c6809e6a4e4db3a38eb71a4255cf4a6" ], "author": { "name": "Matthew Nagy", "email": "matthew.nagy@sony.com", "time": "Mon Mar 09 15:42:25 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Mon Mar 09 15:43:34 2026" }, "message": "halt_on_error flag for TySan and docs (#182479)\n\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 2a9372ff021546a9b3310c2fb435a0cb4c6c744d\n" }, { "commit": "3cac82da8c6809e6a4e4db3a38eb71a4255cf4a6", "tree": "17a8efe6e96fa9a803975781216205a1bfbad438", "parents": [ "44ae0d2a969bde917724fb0d8856ffd7b8ad0b2b" ], "author": { "name": "Rainer Orth", "email": "ro@gcc.gnu.org", "time": "Sat Mar 07 07:30:33 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Sat Mar 07 07:35:11 2026" }, "message": "[ubsan_minimal] Build on Solaris (#184976)\n\nTwo tests currently `FAIL` on Solaris/amd64 and Solaris/sparcv9:\n\n```\n SafeStack-Standalone-i386 :: overflow.c\n SafeStack-Standalone-x86_64 :: overflow.c\n```\n\nThis happens because `libclang_rt.ubsan_minimal.a` isn\u0027t built on\nSolaris although it\u0027s required with `-fsanitize-minimal-runtime`.\n\nThis patch fixes this.\n\nTested on `amd64-pc-solaris2.11` and `sparcv9-sun-solaris2.11`.\n\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: d6b33bda134f933f5ec1b7a948441ab4e98bbb8d\n" }, { "commit": "44ae0d2a969bde917724fb0d8856ffd7b8ad0b2b", "tree": "e758fca0a1e5945c9dc6d4e0dd086af9f2e433b7", "parents": [ "f285bd77250d18d262ff5c3cbb425acce7757c6c" ], "author": { "name": "Nico Weber", "email": "thakis@chromium.org", "time": "Fri Mar 06 20:33:28 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Fri Mar 06 20:35:57 2026" }, "message": "builtins: Make cmake formatting self-consistent aftr #183871\n\nNo behavior change.\n\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 4d53c42f55c06e9dbae48723e366f391ccd9312f\n" }, { "commit": "f285bd77250d18d262ff5c3cbb425acce7757c6c", "tree": "1ca05a2f6db566eed79b27c8f7710032b566cc68", "parents": [ "6ac361a10e7fc6879a2ce5e30bd15cc46fdb537e" ], "author": { "name": "Saleem Abdulrasool", "email": "compnerd@compnerd.org", "time": "Fri Mar 06 19:12:13 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Fri Mar 06 19:15:37 2026" }, "message": "builtins: adjust FP80 source management (#183871)\n\nWe would previously include the FP80 sources into the Windows build if\nwe built with the GNU driver rather than the `cl` driver.\n\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 57f1ec6e0a4ec3cd4a485ca1ec85e3efe541933b\n" }, { "commit": "6ac361a10e7fc6879a2ce5e30bd15cc46fdb537e", "tree": "d29848678587a95755875c5c3cbb17f9a3c19b32", "parents": [ "568f6a925307495d6ba838a68d0dfe7ab375c946" ], "author": { "name": "Kito Cheng", "email": "kito.cheng@sifive.com", "time": "Fri Mar 06 11:37:11 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Fri Mar 06 11:40:52 2026" }, "message": "[compiler-rt][RISCV] Initialize length only when __init_riscv_feature… (#115449)\n\n…_bits success\n\nThat could give us a simple way to detect `__init_riscv_feature_bits`\ngot fail.\n\nSee also https://github.com/riscv-non-isa/riscv-c-api-doc/pull/95\n\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: d64a9fe7814d4061aaaf067fef5b1393f4f1529e\n" }, { "commit": "568f6a925307495d6ba838a68d0dfe7ab375c946", "tree": "c65ac47cdab5a77de59850b2d6684dc9bbfad49d", "parents": [ "a18fb0c5f8c7b5fc6996dfc77b3fb3df92ff549d" ], "author": { "name": "Sadaf Ebrahimi", "email": "sadafebrahimi@google.com", "time": "Thu Mar 05 18:38:06 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Thu Mar 05 18:41:45 2026" }, "message": "[scudo] Move getResidentPages function (#183138)\n\nMoving getResidentPages out of linux.cpp and adding it to MemMapBase\n\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: ea65e8593e21b055dfb1f5f74ea7a5618a6fa275\n" }, { "commit": "a18fb0c5f8c7b5fc6996dfc77b3fb3df92ff549d", "tree": "1dcbc68308be1cb6194d72be4aeee933fa6b98da", "parents": [ "df58f59ff0760c6c373547c9443742f38bcb7a23" ], "author": { "name": "Sean Fertile", "email": "sd.fertile@gmail.com", "time": "Tue Mar 03 20:27:20 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Tue Mar 03 20:30:42 2026" }, "message": "Disable leak sanitizer test on ppc. (#184414)\n\nTest is flaky and fails due to machine load on the build bots. Disable\nuntil we can split the build bots over more machines.\n\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: df1a53ae242418f5ac22adb5bb2178d3f931565f\n" }, { "commit": "df58f59ff0760c6c373547c9443742f38bcb7a23", "tree": "0f09188b9b1ea813774f07f7a36aaee55f95fdbe", "parents": [ "0cd354e178fec2af0c516235f945ceca8611b337" ], "author": { "name": "PiJoules", "email": "leonardchan@google.com", "time": "Tue Mar 03 18:48:12 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Tue Mar 03 18:54:41 2026" }, "message": "[sanitizer][Fuchsia] Define interceptor for reallocarray on Fuchsia (#184410)\n\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: f42b8a18d9043f36f173184d1bf1c28dc6a5be2f\n" }, { "commit": "0cd354e178fec2af0c516235f945ceca8611b337", "tree": "2340229a1bb847a9ff4ca8db0a86f734cac31eb5", "parents": [ "64c252741e3d8d9906473b1d09733cff93b91f50" ], "author": { "name": "Vitaly Buka", "email": "vitalybuka@google.com", "time": "Tue Mar 03 05:07:19 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Tue Mar 03 05:10:49 2026" }, "message": "[CFI] Expand test to include minimal runtime (#183646)\n\n`ubsan_minimal` contains some CFI tests, but it would be nice have one\non CFI side.\n\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: eba4a76597dd1d655794cb5732ace534b58bd97d\n" }, { "commit": "64c252741e3d8d9906473b1d09733cff93b91f50", "tree": "b10c2b8f8a7c58e7242fb1f36afc9bdfa05e23fe", "parents": [ "344e193355ce1055000e04001cdddf3f16a147fa" ], "author": { "name": "Simon Tatham", "email": "simon.tatham@arm.com", "time": "Mon Mar 02 09:30:19 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Mon Mar 02 09:35:46 2026" }, "message": "[compiler-rt][ARM] Fix conditions for strict-mode FP testing (#183507)\n\nOn sufficiently old versions of the Arm architecture, the optimized FP\nroutines are not enabled. So commit a84ee1416b6c179 should not have\nenabled the extra-strict tests that go with them.\n\nAlso in that commit, I wrote a comment saying I was setting two separate\ncompile-time definitions (-DCOMPILER_RT_ARM_OPTIMIZED_FP and\n-DCOMPILER_RT_ARM_OPTIMIZED_FP_THUMB1), and then didn\u0027t actually do it!\nThis caused the strict mulsf3 tests to be wrongly disabled in Thumb2.\n\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 36c6c689dc3141b557638808a073b71104e82c35\n" }, { "commit": "344e193355ce1055000e04001cdddf3f16a147fa", "tree": "ae964bee164e935074d9c996c3ddd64d69607614", "parents": [ "36b7ba36ef48cbb211135173c1285063df3f73c3" ], "author": { "name": "Benjamin Stott", "email": "Benjamin.Stott@sony.com", "time": "Sun Mar 01 06:25:12 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Mon Mar 02 09:35:35 2026" }, "message": "[UBSan] Wrap Location variants in anonymous union (#168866)\n\n(Addresses the FIXME)\n\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: d412b04a883c7d6212edf1d678349689cb7d723b\n" }, { "commit": "36b7ba36ef48cbb211135173c1285063df3f73c3", "tree": "2534f55b5d4ad4ae85baa79a76ee42c77cb7eb3e", "parents": [ "5737c66d4f85c258f6ff309da068a8f1be3e99d6" ], "author": { "name": "MacGyver Codilla", "email": "77024043+39otsu@users.noreply.github.com", "time": "Sun Mar 01 06:23:28 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Sun Mar 01 06:26:18 2026" }, "message": "[ASan] Mark recent integration tests as accordingly for MSVC (#135889)\n\nBoth of these tests will cause an unsuccessful pass when using msvc.\n\n`shadowed-stack-serialization.cpp` - XFAIL due to the metadata not being\ngenerated.\n`fakeframe-right-redzone.cpp` - UNSUPPORTED due to the optimization\nlimitations of the msvc compiler.\n\n---------\n\nCo-authored-by: MacGyver Codilla \u003cmcodilla@microsoft.com\u003e\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 10b1b7857b0564c447054ed5f451729230280f54\n" }, { "commit": "5737c66d4f85c258f6ff309da068a8f1be3e99d6", "tree": "25150554d7171edd8d329fa78b012f4bad8af4c7", "parents": [ "bc0a39e5b30af5cc47dd0df097a9b817e80ee3aa" ], "author": { "name": "Reagan", "email": "xbjfk.github@gmail.com", "time": "Sun Mar 01 06:22:27 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Sun Mar 01 06:26:02 2026" }, "message": "[compiler-rt][CMake] Fix build when specifying --stdlib\u003d (with 2 dashes) (#136111)\n\nYou can pass the stdlib argument either as -stdlib and --stdlib - the\nprevious regex did not account for this however - which caused the build\nto fail, as a --stdlib argument would be replaced with a single dash,\ncausing clang to assume reading from stdin and the build to fail:\n\nclang++: error: -E or -x required when input is from standard input\nclang++: error: cannot specify -o when generating multiple output files\n\nThe files\n[libcxxabi/CMakeLists.txt](https://github.com/llvm/llvm-project/blob/bf6986f9f09f79da38006a83c339226c429bb686/libcxxabi/CMakeLists.txt#L261)\nand\n[libunwind/CMakeLists.txt](https://github.com/llvm/llvm-project/blob/bf6986f9f09f79da38006a83c339226c429bb686/libunwind/CMakeLists.txt#L257)\naccount for this by removing --stdlib first.\n\nCo-authored-by: Vitaly Buka \u003cvitalybuka@google.com\u003e\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: b2ce908a48e0ab6b0404742f38e5e5d0ee51cc2f\n" }, { "commit": "bc0a39e5b30af5cc47dd0df097a9b817e80ee3aa", "tree": "38f34b541d8e27f0337454a567b0a8f99588435b", "parents": [ "d397657ad6ff072611c9c13f7aa982c097fcf1cd" ], "author": { "name": "Vitaly Buka", "email": "vitalybuka@google.com", "time": "Fri Feb 27 03:35:58 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Fri Feb 27 03:40:58 2026" }, "message": "[SafeStack] Allow -fsanitize-minimal-runtime with -fsanitize\u003dsafestack (#183644)\n\nSafeStack does not require a full sanitizer runtime, so it should be\ncompatible\nwith the minimal runtime flag.\n\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: b354b206d3bef08d254713c027e235c0251eebdb\n" }, { "commit": "d397657ad6ff072611c9c13f7aa982c097fcf1cd", "tree": "b60b5160f50340fb0906ccd45486889b76d51863", "parents": [ "aa5fc9a691e815794980b5310129832992c2e287" ], "author": { "name": "Alexey Samsonov", "email": "vonosmas@gmail.com", "time": "Fri Feb 27 00:14:15 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Fri Feb 27 00:15:30 2026" }, "message": "[scudo] Add reallocarray C wrapper. (#183385)\n\n`reallocarray()` is a POSIX extension to C standard which wraps\n`realloc` function and adds `calloc`-like overflow detection. It is\navailable in glibc and some other standard library implementations. Add\n`reallocarray` to the list of Scudo C wrappers, so that the code that\ndepends on `reallocarray` presence will continue to work.\n\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 7c022af37ef2bc9deb60c6be93b7925028d39e8a\n" }, { "commit": "aa5fc9a691e815794980b5310129832992c2e287", "tree": "039b14a64e0ede4d587642d687dd2c7bc4ca032a", "parents": [ "ecffa351c7a4c9e792a9a940c7c7bb3c4523eebe" ], "author": { "name": "Tamas Kaman", "email": "tamas.kaman@arm.com", "time": "Thu Feb 26 23:27:11 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Thu Feb 26 23:31:04 2026" }, "message": "[scudo] Change header tagging for the secondary allocator (#182487)\n\nWhen secondary allocator allocates a new chunk, the allocation is\nprepended with a chunk header (common with the primary allocator)\nand large header (only used for secondary).\nOnly the headers are tagged, the data is not, and the headers are\ntagged individually as different tags are used for them.\nIn the current implementation while tagging the large header the unused\narea is tagged with it, so the allocator can tag up to a page size (in\nworst case), which is costly and does not bring security benefit (as the\narea is unused).\nWith the current fix we can get rid of around 97-98% of the tagging for\nthe secondary allocator, measured with random benchmarks.\n\nCo-authored-by: Christopher Ferris \u003ccferris1000@users.noreply.github.com\u003e\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: e55945556a1e99a7d6618735873d513a0239b2a0\n" }, { "commit": "ecffa351c7a4c9e792a9a940c7c7bb3c4523eebe", "tree": "a77d6031cd12035c6520b4745f20151199e25930", "parents": [ "efa78c10673ca0ea0ab5c6c6165c4173117e6ed3" ], "author": { "name": "Dan Blackwell", "email": "dan_blackwell@apple.com", "time": "Wed Feb 25 18:55:41 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Wed Feb 25 19:00:32 2026" }, "message": "[compiler-rt][Fuzzer] Relax reduce_inputs.test to account for non-determinism (#182495)\n\nI have seen that very occasionally this test is failing on a bot, with\nonly 3 files in the corpus. After running the test in a loop 4000+\ntimes, I witnessed this same failure twice.\n\nIn both cases the first corpus member was some string not containing a\n\u0027F\u0027; the second corpus member was \u0027F[\u0027 or \u0027FZ\u0027; and the final corpus\nmember \u0027FUZ\u0027.\n\nIn a normal run there is an intermediate corpus member \u0027FU.\u0027 - so this\ntest is failing in very rare cases where the fuzzer gets lucky and\nmatches 2 branch conditions in one mutation.\n\nThis patch allows the FileCheck condition to match 3 or 4 corpus files.\nIt may be possible for the fuzzer to reach the target in 2 files, but I\nthink that if that is possible, it will be exceptionally rare.\n\nrdar://170440934\n\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: ed1a1eda0ff9c5d59b13551c0447793e92d375bd\n" }, { "commit": "efa78c10673ca0ea0ab5c6c6165c4173117e6ed3", "tree": "ff996ad85e759b6712611555e5b4e84a235f0f51", "parents": [ "25e7ac56471209b5e7a55d87639f1f6b888ff51e" ], "author": { "name": "PiJoules", "email": "leonardchan@google.com", "time": "Tue Feb 24 21:56:35 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Tue Feb 24 22:01:48 2026" }, "message": "Reapply \"[ASan][Fuchsia] Have Fuchsia use a dynamic shadow start\" (#182972) (#183154)\n\nThis reverts commit 19daed352f218f866ebcbbf3f35b14918588d8a1 but keeping\nthe original cmake in compiler-rt/lib/asan/CMakeLists.txt as is. Prior\nwe were attempting to use clang-cl for building ASM with all the normal\nflags but extra flags are needed to force clang-cl to be used as an\nassembler. For now, it\u0027s just easier to not touch any of the cmake\nmachinery and omit building the assembly for windows builds, but we\nstill keep the `#if` in the asm file to avoid having a fuchsia-specific\nexclusion.\n\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: ac45132c2f64ff6b189a5cd15c2b4e257433d243\n" }, { "commit": "25e7ac56471209b5e7a55d87639f1f6b888ff51e", "tree": "68eb40c87c69d9c3475fbef95bb5d14a200e69fd", "parents": [ "8417d605c60626408e3671ea979466952e6cf1c8" ], "author": { "name": "Andrew Haberlandt", "email": "ahaberlandt@apple.com", "time": "Tue Feb 24 16:29:36 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Tue Feb 24 16:35:43 2026" }, "message": "[sanitizer_common] [Darwin] Adopt _dyld_get_dyld_header (#182943)\n\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 2e7d07a33725a82ecfc514e27f047ece3ff13d4c\n" }, { "commit": "8417d605c60626408e3671ea979466952e6cf1c8", "tree": "89e1e997fac8d2f7203438bb71baa4abd08f8a76", "parents": [ "4b3abf0f808b772bfc2656689131f6c58ecc9f1c" ], "author": { "name": "Simon Tatham", "email": "simon.tatham@arm.com", "time": "Tue Feb 24 09:08:06 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Tue Feb 24 09:11:09 2026" }, "message": "[compiler-rt][ARM] Enable strict mode in divsf3/mulsf3 tests (#179918)\n\nCommit 5efce7392f3f6cc added optimized AArch32 assembly versions of\nmulsf3 and divsf3, with more thorough tests. The new tests included test\ncases specific to Arm\u0027s particular NaN handling rules, which are\ndisabled on most platforms, but were intended to be enabled for Arm.\n\nUnfortunately, they were not enabled under any circumstances, because I\nmade a mistake in `test/builtins/CMakeLists.txt`: the command-line `-D`\noption that should have enabled them was added to the cflags list too\nearly, before the list was reinitialized from scratch. So it never ended\nup on the command line.\n\nAlso, the test file mulsf3.S only even _tried_ to enable strict mode in\nThumb1, even though the Arm/Thumb2 implementation would also have met\nits requirements.\n\nBecause the strict-mode tests weren\u0027t enabled, I didn\u0027t notice that they\nwould also have failed absolutely everything, because they checked the\nresults using the wrong sense of comparison! I used `\u003d\u003d`, but that\ncomparison was supposed to be a drop-in replacement for\n`compareResultF`, which returns zero for equality. Changed the tests to\nuse `!\u003d`.\n\nFinally, I\u0027ve also added a macro to each test so that it records the\nsource line number of each failing test case. That way, when a test\nfails, you can find it in the test source more easily, without having to\nsearch for the hex numbers mentioned in the failure message.\n\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: a84ee1416b6c179a3d45939a6490f5268ee601bf\n" }, { "commit": "4b3abf0f808b772bfc2656689131f6c58ecc9f1c", "tree": "2962b79bb736a7602e716a2738b3c8b1adab4de5", "parents": [ "f0b7ec7780a7a8cd070536cc49dfcdafbfd3c11e" ], "author": { "name": "Paul Kirth", "email": "paulkirth@google.com", "time": "Tue Feb 24 01:00:03 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Tue Feb 24 01:08:04 2026" }, "message": "Revert \"[ASan][Fuchsia] Have Fuchsia use a dynamic shadow start\" (#182972)\n\nReverts llvm/llvm-project#180880\n\nThis is breaking Fuchsia\u0027s CI. something in the CMake needs to be\nadjusted. Reverting on the author\u0027s request.\n\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 19daed352f218f866ebcbbf3f35b14918588d8a1\n" } ], "next": "f0b7ec7780a7a8cd070536cc49dfcdafbfd3c11e" }