)]}' { "log": [ { "commit": "d19905ff0c87d266fae76d50ace0f56712def70f", "tree": "30e65a28a5923491d4517527ebbc2b4d4495d4d4", "parents": [ "eeb33a920d4e1c56826ea6488193985341ebcdda" ], "author": { "name": "Sadaf Ebrahimi", "email": "sadafebrahimi@google.com", "time": "Tue Sep 01 22:04:58 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Tue Sep 01 22:14:09 2026" }, "message": "[scudo] Add MaxCacheResidentBytes to Secondary\n\nIntroduce DefaultMaxCacheResidentBytes in the cache configuration and\nMaxCacheResidentBytes at runtime to enforce an upper bound on total\nresident memory in the Secondary cache.\n\nWhen storing a block causes resident bytes to exceed the threshold,\nphysical pages from the oldest cached entries are released to the OS via\nreleaseAndZeroPagesToOS() while preserving their virtual address\nmappings in the cache.\n\nAlso adds Option::MaxCacheResidentBytes and M_CACHE_RESIDENT_BYTES_MAX\nto allow tuning this limit dynamically via mallopt().\n\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: b2c390e532f1a89c6946b65644c6cf05ca2b8b90\n" }, { "commit": "eeb33a920d4e1c56826ea6488193985341ebcdda", "tree": "03b1be8d04e5c9f94246034fa775573a20db3c17", "parents": [ "f3c1023259b9d6eaa4689467ddf10b3006a8fa31" ], "author": { "name": "Joseph Huber", "email": "huberjn@outlook.com", "time": "Tue Sep 01 20:21:08 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Tue Sep 01 20:26:29 2026" }, "message": "[compiler-rt] Fix undefined abort for ubsan_minimal GPU targets (#220314)\n\nSummary:\nWhen compiling with debug enabled this would call `abort()`. The minimal\nUBSan for GPUs is intentionally written to just depend on `printf`\nwhich most vendors support and aborting is done through hardware traps.\nI neglected to update this part because I had not done it with debugging\nenabled, just wrap this into a helper function.\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 1e5ad0192e93337f0f72fddb9421d4ecac63da53\n" }, { "commit": "f3c1023259b9d6eaa4689467ddf10b3006a8fa31", "tree": "837f65c0b0139d863c0a6ed88fe2c3950cb54c8c", "parents": [ "272630912b18345733ee92ccebd0e8af44723b65" ], "author": { "name": "Andrew Haberlandt", "email": "ahaberlandt@apple.com", "time": "Tue Sep 01 15:58:14 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Tue Sep 01 16:07:11 2026" }, "message": "[libFuzzer] [Darwin] Let the kernel kill fuzzers when OOM (#216836)\n\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 49d043f831b92a0b824472bbfa5ee99eb003b726\n" }, { "commit": "272630912b18345733ee92ccebd0e8af44723b65", "tree": "69854301242864fcb1b728a028ad69cd3a4f4403", "parents": [ "949185fe89d2c78514ddaa6b8874de9415036491" ], "author": { "name": "Justin T. Gibbs", "email": "gibbs@meta.com", "time": "Mon Aug 31 23:50:02 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Mon Aug 31 23:56:58 2026" }, "message": "[sanitizer_common] Add operator new chain-handling framework (#201151)\n\nGroundwork for #196388\n\n[sanitizer_common] Add operator new chain-handling framework\n\nImplement the operator new wrapper machinery required by\n[new.delete.single]/3+/4 in shared sanitizer_common files so every\nsanitizer can reuse it (avoids ~130 lines of duplication).\n\nsanitizer_new_handler.h provides three main templates in namespace\n__sanitizer (plus a NORETURN InvokeOnExhausted wrapper used internally):\n\n * RunNewHandlerChain\u003cAlloc\u003e(alloc)\n Runs std::get_new_handler() in a loop until either the\n allocation succeeds or the chain is exhausted (returns nullptr).\n\n * NewImplThrowing\u003cAlloc, OnExhausted\u003e(alloc, on_exhausted)\n Throwing operator new: runs the chain; on exhaustion either\n throws std::bad_alloc (AllocatorMayReturnNull()\u003dtrue opts into\n standards-conformant throw) or invokes on_exhausted (default;\n historical abort-on-OOM).\n\n * NewImplNothrow\u003cAlloc, OnExhausted\u003e(alloc, on_exhausted) noexcept\n Nothrow operator new: per [new.delete.single]/4 behaves as-if\n the throwing form is called within a try/catch converting any\n exceptions to a nullptr return.\n\nNOTE: Windows builds do not support exceptions, so the throwing\n std::bad_alloc behavior described above is converted to an\n invocation of on_exhausted(); a user new_handler that throws\n on Windows trips the noexcept on NewImplNothrow and aborts via\n std::terminate.\n\nNOTE: The framework auto-detects exception support via the standard\n __cpp_exceptions feature-test macro: when a consuming TU is\n compiled with -fno-exceptions, the throw/try-catch logic is\n compiled out and the framework collapses to the abort path\n (same shape as Windows). This means TUs do not have to be\n built with -fexceptions, and a -fno-exceptions adopter incurs\n no std::bad_alloc symbol dependency. When -fexceptions IS used,\n instantiating the throwing-form template introduces a runtime\n dependency on std::bad_alloc — adopters must link a C++ ABI\n library (libstdc++ / libc++abi) into the resulting runtime.\n\nsanitizer_new_operators.inc builds the eight standard\nOPERATOR_NEW_BODY* macros on top of these templates. A consuming\nsanitizer defines six ingredient macros (a stack-trace setup,\na report-OOM invocation, and four alloc helpers) and then includes the\nfile. The resulting OPERATOR_NEW_BODY / OPERATOR_NEW_BODY_NOTHROW /\nOPERATOR_NEW_BODY_ARRAY / ... / OPERATOR_NEW_BODY_ALIGN_ARRAY_NOTHROW\nmacros can then be used directly as the bodies of the eight operator new\noverrides.\n\nNFC. No consumer yet — this is a prerequisite for a follow-up that\nrestructures compiler-rt/lib/asan/asan_new_delete.cpp to use the\nshared framework.\n\nAssisted by: Claude Opus 4.7\n\n--------------------\nFurther context for the PR summary...\n\nHere\u0027s a breakdown of framework applicability and the work required to\ndeploy it.\n\n```\nSanitizer | Delta SLOC | aligned | OOM-null | adoption | notes\n-----------+------------+---------+----------+-----------------+----------------------\nasan | -43 | yes | yes | DONE | framework consumer\nhwasan | -50 | yes | no | easy + prereq | needs may_return_null\nmemprof | -25 | yes | no | easy + prereq | needs may_return_null\nmsan | -25 | yes | no | easy + prereq | needs may_return_null\ndfsan | -25 | yes | no | medium + prereq | needs may_return_null\nnsan | -25 | yes | no | medium + prereq | needs may_return_null\ntsan | -30 | yes | no | medium + prereq | user_alloc shim too\nlsan | - | no | no | medium | only 4 of 8 overloads\n\nDelta SLOC \u003d code size effect of applying the framework. All but asan are estimated.\naligned \u003d has std::align_val_t overloads\nOOM-null \u003d internal alloc returns nullptr on OOM (vs. abort)\nadoption \u003d effort to wire to the framework\nprereq \u003d per-sanitizer commit to plumb \"may_return_null\" behavior\n and enabling exceptions in the \"*_new_delete.cpp\" TU.\n dfsan/nsan also need a CXX slice defined in the build infrastructure.\n```\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 4bb34ea0473cd6207c79ebb81ce85f995d093d1a\n" }, { "commit": "949185fe89d2c78514ddaa6b8874de9415036491", "tree": "ec868093bf8443ad59cfab70f2304f63445e21d8", "parents": [ "a931360f2073eaff84177c9d9f7cf42393c45678" ], "author": { "name": "Tom Tromey", "email": "tromey@adacore.com", "time": "Mon Aug 31 14:12:00 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Mon Aug 31 14:16:36 2026" }, "message": "Remove two declarations without definitions (#219511)\n\nCommit 3dc4fd6d removed the definitions of struct_scc_modem_sz and\nstruct_scc_stat_sz, but left the declarations in a header. This patch\nremoves the leftovers.\n\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: d9d947ffadf1cebcb8cc14a18d7ed970332231ff\n" }, { "commit": "a931360f2073eaff84177c9d9f7cf42393c45678", "tree": "8aa63e1c582653ed22c75363fbd9196e24bfd626", "parents": [ "7ac469c198cb1c72bb7a3fe3a1e9420d6af74b69" ], "author": { "name": "Rainer Orth", "email": "ro@gcc.gnu.org", "time": "Sat Aug 29 08:28:40 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Sat Aug 29 08:31:57 2026" }, "message": "[fuzzer] Restrict merge-sigusr.test to Linux (#216702)\n\nThe `fuzzer/merge-sigusr.test` test causes the whole `ninja check-all`\nrun on NetBSD to be killed with `SIGUSR2`.\n\nIt turns out the test is highly Linux-specific in at least two ways:\n\n- The `setsid` command doesn\u0027t exist on any of Darwin, FreeBSD, and\nNetBSD.\n\n- `ps -o sess\u003d \u003cpid\u003e` is highly unportable, too:\n\n - FreeBSD `ps` doesn\u0027t have the `sess` keyword at all.\n\n- NetBSD `ps` does, but with different semantics: it\u0027s the session\npointer, not the session id as on Linux, which leads to randomly killing\nunrelated processes.\n\nTherefore this patch restricts the test to Linux instead of simply\nskipping it on Darwin and Windows.\n\nTested on `x86_64-pc-netbsd11.0`, `x86_64-pc-linux-gnu`, and\n`x86_64-pc-freebsd15.1`.\n\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: f2c31b8ffd99378e099720d7ddbf8078c650543a\n" }, { "commit": "7ac469c198cb1c72bb7a3fe3a1e9420d6af74b69", "tree": "ea827ad7636406bd448551f8ea0065f2a7f7b554", "parents": [ "ef796e117995e0083d3f9438c44d748c453147c4" ], "author": { "name": "Rainer Orth", "email": "ro@gcc.gnu.org", "time": "Sat Aug 29 08:26:18 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Sat Aug 29 08:31:31 2026" }, "message": "[sanitizer] Skip hanging tests on FreeBSD (#216703)\n\nThree sanitizer tests hang indefinitely on FreeBSD:\n\n```\nMemorySanitizer-X86_64 :: fork.cpp\nSanitizerCommon-tsan-x86_64-FreeBSD :: Posix/fork_threaded.c\nThreadSanitizer-x86_64 :: fork_multithreaded.cpp\n```\n\nAll of them loop and don\u0027t time out, so they need to be terminated\nmanually for `ninja check-all` to complete. To avoid this, this patch\nskips the affected tests.\n\nTested on `x86_64-pc-freebsd15.1`, `x86_64-pc-netbsd11.0` and\n`x86_64-pc-linux-gnu`.\n\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: cb2f67d1556c1232f99b4512a35129ec04f73c75\n" }, { "commit": "ef796e117995e0083d3f9438c44d748c453147c4", "tree": "7ebafc406bcc2e197279131bf556c9ce6ba0500f", "parents": [ "20678aa96a461d53afcd3bdb20eaf81f309dcf72" ], "author": { "name": "Sadaf Ebrahimi", "email": "sadafebrahimi@google.com", "time": "Thu Aug 27 21:49:46 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Thu Aug 27 21:57:07 2026" }, "message": "[scudo] Track maximum resident memory in cache (#219036)\n\nAdd tracking for the resident memory in MapAllocatorCache. Unreleased\ncache entries where Time !\u003d 0 are assumed to be resident, while entries\nthat have passed through releaseOlderThan where Time \u003d\u003d 0 have had their\nphysical pages discarded.\n\nTracking is performed by updating CurrentResidentBytes on store, remove,\nand release operations, recoding the peak in maxResidentBytes and\nreporting it in getStats.\n\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 5964066e909c5c443bd2f48a88dc54f263a4d47e\n" }, { "commit": "20678aa96a461d53afcd3bdb20eaf81f309dcf72", "tree": "cec1b0ad3ea0e59cd49a67df1eb73774f08df5bc", "parents": [ "39aeee4ba78b9b410d7d2ae91e7ca5f003115f5d" ], "author": { "name": "Bernhard Übelacker", "email": "bernhardu@mailbox.org", "time": "Thu Aug 27 11:50:55 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Thu Aug 27 11:57:27 2026" }, "message": "[win/asan] GetInstructionSize: Add some more instructions. (#219038)\n\nThese got reported in #96270.\n\n- Allow \"0F B6 44 24 XX : movzx eax, byte ptr [esp + XX]\" for x86 and\nx86_64.\n- Add \"0F B6 4C 24 XX : movzx ecx, byte ptr [esp + XX]\"\n- Add \"4F 8D 0C XX : lea r9, [...]\"\n- Add \"66 83 3A XX : cmp word ptr [rdx], XX\"\n- Add \"85 D2 : test edx, edx\"\n- Add \"8D 44 24 XX : lea eax, [esp + XX]\"\n- Add \"F3 0F 1E FA : endbr64\"\n- Add \"F3 0F 1E FB : endbr32\"\n\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 7fd0dd89cf63bd363b7884232202fa958ccc974a\n" }, { "commit": "39aeee4ba78b9b410d7d2ae91e7ca5f003115f5d", "tree": "807829d868d26cbde045599ff804a8a69c3d5cf5", "parents": [ "bfd30296d54cd1beb2527e80ea45ae653f52b78c" ], "author": { "name": "KAWASHIMA Takahiro", "email": "t-kawashima@fujitsu.com", "time": "Thu Aug 27 02:12:13 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Thu Aug 27 02:17:55 2026" }, "message": "[compiler-rt][cmake] Change orc-rt target names (#216901)\n\nWhen configured with `LLVM_ENABLE_RUNTIMES\u003dcompiler-rt;orc-rt`,\nfollowing two CMake errors occur.\n\n```\nCMake Error at llvm-project/llvm/cmake/modules/AddLLVM.cmake:2245 (add_custom_target):\n add_custom_target cannot create target \"check-orc-rt\" because another\n target with the same name already exists. The existing target is a custom\n target created in source directory\n \"llvm-project/compiler-rt/test/orc\". See documentation\n for policy CMP0002 for more details.\nCall Stack (most recent call first):\n llvm-project/llvm/cmake/modules/AddLLVM.cmake:2326 (add_lit_target)\n llvm-project/orc-rt/test/CMakeLists.txt:24 (add_lit_testsuite)\n\nCMake Error at llvm-project/orc-rt/test/unit/CMakeLists.txt:1 (add_custom_target):\n add_custom_target cannot create target \"OrcRTUnitTests\" because another\n target with the same name already exists. The existing target is a custom\n target created in source directory\n \"llvm-project/compiler-rt/lib/orc/tests\". See\n documentation for policy CMP0002 for more details.\n```\n\nBoth `compiler-rt/lib/orc/tests/CMakeLists.txt` and\n`orc-rt/test/unit/CMakeLists.txt` define targets named `OrcRTUnitTests`.\nBoth `compiler-rt/test/orc/CMakeLists.txt` and\n`orc-rt/test/CMakeLists.txt` define targets named `check-orc-rt`.\n\nThis commit changes target names in compiler-rt.\n\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 86440cce63c88ca5491320ae72d43db9964c83f2\n" }, { "commit": "bfd30296d54cd1beb2527e80ea45ae653f52b78c", "tree": "724dbb5fafd3b954afe9525e854a12eaf56b6491", "parents": [ "e960037e0e0239cdd3b022587ec22e57db9b2e07" ], "author": { "name": "Justin T. Gibbs", "email": "gibbs@meta.com", "time": "Tue Aug 25 20:30:22 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Tue Aug 25 20:38:44 2026" }, "message": "[asan] Build the asan_new_delete.cpp C++ slice with -frtti (#213077)\n\nRequired by #196388\n\nIn preparation for asan\u0027s operator new throwing `std::bad_alloc` on OOM,\ncompile the C++ slice (asan_new_delete.cpp, the only asan TU that\nthrows) with -frtti instead of the -fno-rtti inherited from ASAN_CFLAGS.\n\nThrowing `std::bad_alloc` needs the type_info for std::bad_alloc. Under\n-fno-rtti clang cannot reference the C++ ABI library\u0027s canonical\n_ZTISt9bad_alloc and instead emits a private, hidden per-image copy of\nit. On Mach-O that copy does not coalesce with the definition in\nlibc++abi, so the type_info the runtime throws with differs from the one\nthe user\u0027s `catch (const std::bad_alloc\u0026)` matches against, and the\nprogram terminates. On ELF the copy is weak with default visibility, so\nthe linker coalesces every copy to one and the addresses match, but that\ncoalescing is not guaranteed (e.g. `-Bsymbolic` defeats it).\n\nBuilding this one TU with -frtti makes clang reference the external\ncanonical type_info, so the runtime\u0027s throw and the user\u0027s catch agree\non every platform. The change lives inside the existing exceptions\ncarve-out, so it only takes effect when the throwing path is compiled\nin, and it adds no new library dependency (the exception throw already\nrequires the C++ ABI library).\n\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 7d2e3e7e5b0b0ab6588fb7ec05e4e321518eff09\n" }, { "commit": "e960037e0e0239cdd3b022587ec22e57db9b2e07", "tree": "287062ad2cab522beafc6533c4d7314cd6ae7c2f", "parents": [ "d33e78aae8f72732f351c9a32b5ba59e9f032197" ], "author": { "name": "Yaxun (Sam) Liu", "email": "yaxun.liu@amd.com", "time": "Tue Aug 25 14:04:01 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Tue Aug 25 14:08:34 2026" }, "message": "[compiler-rt] Add version info to Windows runtime DLLs (#216408)\n\nWindows runtime DLLs built by compiler-rt currently do not include\nVERSIONINFO metadata. This leaves fields such as FileVersion,\nProductName, and ProductVersion empty in Windows file properties.\n\nLLVM and Clang already use windows_version_resource.rc to add this\nmetadata to Windows executables and shared libraries. compiler-rt uses a\nseparate CMake helper for runtime libraries, so its DLLs do not inherit\nthat support.\n\nAdd compiler-rt support for attaching the shared LLVM Windows version\nresource to MSVC shared runtime targets. The resource file is copied to\na target-specific file before being added as a source, so each DLL can\nuse its own source properties.\n\nUse LLVMVersion.cmake to fill in the version fields. This gives runtime\nDLLs such as clang_rt.asan_dynamic-x86_64.dll FileVersion and\nProductVersion values matching the LLVM version.\n\nThis only applies to shared runtime targets. Static and object runtime\nlibraries are unchanged.\n\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 9e30a36318c07d1875576d4cb9beeb06489d2a31\n" }, { "commit": "d33e78aae8f72732f351c9a32b5ba59e9f032197", "tree": "41ec4c539ad206c2e8f3982655e8fb34c27b9f8a", "parents": [ "867c4c4b71e800c8f9e6696e116ce5c287401674" ], "author": { "name": "oltolm", "email": "oleg.tolmatcev@gmail.com", "time": "Fri Aug 21 18:12:47 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Fri Aug 21 18:18:13 2026" }, "message": "[compiler-rt] Adjust ASan Windows tests for mingw-w64 (#216585)\n\nAdd MinGW-specific configurations to the fuse-lld and unsymbolized\ntests. Use the Clang driver and GNU-compatible LLD options instead of\nthe MSVC-oriented lld-link invocation, and account for differences in\nthe resulting symbolization output.\n\nThis change was split out of\nhttps://github.com/llvm/llvm-project/pull/209902, which adds GCC support\nto the sanitizer runtimes on Windows.\n\n---------\n\nCo-authored-by: Hannes Domani \u003cssbssa@yahoo.de\u003e\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 94f216a9cf0e69db996fd573b764cc00d88eba02\n" }, { "commit": "867c4c4b71e800c8f9e6696e116ce5c287401674", "tree": "ea8a4f24cf8424553deffd37f7ece44e15c5842a", "parents": [ "bc8ca28ca027445abe49d31fac7e853656245740" ], "author": { "name": "oltolm", "email": "oleg.tolmatcev@gmail.com", "time": "Fri Aug 21 18:12:34 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Fri Aug 21 18:17:46 2026" }, "message": "[compiler-rt] Add libbacktrace support on Windows (#216584)\n\nAllow the libbacktrace symbolizer on non-POSIX targets when libbacktrace\nis supported and does not use malloc.\n\nFor Windows runtimes built with GCC, register libbacktrace before the\nDbgHelp symbolizer instead of searching for llvm-symbolizer. Keep the\nexisting symbolizer selection for Clang builds.\n\nThis change was split out of\nhttps://github.com/llvm/llvm-project/pull/209902, which adds GCC support\nto the sanitizer runtimes on Windows.\n\nCo-authored-by: Hannes Domani \u003cssbssa@yahoo.de\u003e\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 8ada288c70876f89d2c4c5caf525ef9719e33804\n" }, { "commit": "bc8ca28ca027445abe49d31fac7e853656245740", "tree": "e2f6cc51d3a301be684071be8a808ebd8a47334d", "parents": [ "90ba498e85aa559e8d202fcdbbeb6a4e24aa720c" ], "author": { "name": "oltolm", "email": "oleg.tolmatcev@gmail.com", "time": "Fri Aug 21 18:12:16 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Fri Aug 21 18:17:13 2026" }, "message": "[NFC][compiler-rt] Refactor Windows IN_SECTION declarations (#216582)\n\nThis refactors all uses of __declspec(allocate()) on Windows to use an\nequivalent newly defined macro, IN_SECTION. Future work\n(https://github.com/llvm/llvm-project/pull/209902) will generalize\nIN_SECTION to work for GNU.\n\nCo-authored-by: Hannes Domani \u003cssbssa@yahoo.de\u003e\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 78b2bf730da12e4095e64200d7c4c3c3968e2c6c\n" }, { "commit": "90ba498e85aa559e8d202fcdbbeb6a4e24aa720c", "tree": "9d8cb6d949a574c2080de33c99af40a97bae7d7e", "parents": [ "fd21a429362ea9b04d97129a9fe5163f4c9d1ae2" ], "author": { "name": "bernhardu", "email": "bernhardu@mailbox.org", "time": "Fri Aug 21 13:12:45 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Fri Aug 21 13:17:26 2026" }, "message": "[win/asan] GetInstructionSize: Add some more instructions. (#214707)\n\nBelow instructions can be found in wine-11.14 Debian packages.\n\n```\n8D 45 XX : lea eax, [ebp + XX]\n66 0f ef c0 : pxor xmm0, xmm0\n41 B9 XX XX XX XX : mov r9d, XX XX XX XX\n45 84 c9 : test r9b,r9b\n```\n\nThis patch also adds `80 3A XX` to tests (followup of 3ca5b7dc4f) and\nfixes the sorting of test lists.\n\n```\n\u003d\u003d484\u003d\u003dinterception_win: unhandled instruction at 0x7bcb2264: 8d 45 f8 83 ec 18 8b 5d\n ntdll.dll`RtlExitUserThread:\n 0x7bcb2264 \u003c+4\u003e: 8d 45 f8 lea eax, [ebp - 0x8]\n\n\u003d\u003d484\u003d\u003dinterception_win: unhandled instruction at 0x7a5f2fe1: 66 0f ef c0 89 e5 53 83\n ucrtbase.dll`memset:\n 0x7a5f2fe1 \u003c+1\u003e: 66 0f ef c0 pxor xmm0, xmm0\n\n\u003d\u003d760\u003d\u003dinterception_win: unhandled instruction at 0x6fffffbe6b25: 41 b9 04 00 00 00 ba 0c\n ntdll.dll`RtlExitUserThread:\n 0x6fffffbe6b25 \u003c+5\u003e: 41 b9 04 00 00 00 movl $0x4, %r9d\n\n\u003d\u003d760\u003d\u003dinterception_win: unhandled instruction at 0x6fffffc342a4: 45 84 c9 74 45 0f 1f 80\n ntdll.dll`strpbrk:\n 0x6fffffc342a4 \u003c+4\u003e: 45 84 c9 testb %r9b, %r9b\n```\n\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 5af1c55f938ff025718129f64e8e41136496ff05\n" }, { "commit": "fd21a429362ea9b04d97129a9fe5163f4c9d1ae2", "tree": "ab63e2c29491d23c090693cec38e8ec52b63fa89", "parents": [ "4b58de477d5b6039e23e8372d40c0793a6d2289c" ], "author": { "name": "George Burgess IV", "email": "george.burgess.iv@gmail.com", "time": "Wed Aug 19 17:35:37 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Wed Aug 19 17:41:59 2026" }, "message": "[profile] hoist length check before use of pointer (#217109)\n\nIf a truncated profile is passed into this endpoint (say, 1 byte), we\nunconditionally do `((__llvm_profile_header\n*)ProfileData)-\u003eBinaryIdsSize`, which reaches outside of the `char`\narray passed to this function.\n\nHoisting the check prevents the out-of-bounds read.\n\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 9932f190f42e1a12c3cafea23938241abf6337ca\n" }, { "commit": "4b58de477d5b6039e23e8372d40c0793a6d2289c", "tree": "808d91b8b50dc40d0e464df8892dad020e50226f", "parents": [ "6798025b387ae1f7c485bbb1bc3e565d1b87bd98" ], "author": { "name": "flovent", "email": "flbven@protonmail.com", "time": "Wed Aug 19 13:29:30 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Wed Aug 19 13:32:53 2026" }, "message": "[ASan] Correctly handle vectorized pointer sub/cmp for `invalid-pointer-pair` (#213546)\n\nBefore this PR, asan will treat vector operands just like pointer and\npass it to `__sanitizer_ptr_sub/__sanitizer_ptr_cmp(i64, i64)`,\nwhich leads to assertion failure because it doesn\u0027t matches the needed\nparameter type.\n\nThis PR extracts vector\u0027s elements and creates runtime call for each\npair of them.\n\nCloses #212453 .\n\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 297fe1e500a8dbb21a70ed5ace5f9c4b68c21e15\n" }, { "commit": "6798025b387ae1f7c485bbb1bc3e565d1b87bd98", "tree": "bd4a5d626425de5fe94ad367c5cacb11b157df93", "parents": [ "66e09cd4803d25d7325727daca3d35fa199a5823" ], "author": { "name": "jinge90", "email": "ge.jin@intel.com", "time": "Wed Aug 19 07:41:54 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Wed Aug 19 07:47:27 2026" }, "message": "[compiler-rt] Add clang_rt library path to LIB env var for profile lit (#217189)\n\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: ba30b230ce829f642aec8601b7752f10ec2c761b\n" }, { "commit": "66e09cd4803d25d7325727daca3d35fa199a5823", "tree": "f422419188fc893401332986208771689bbe1bdb", "parents": [ "7556d2bc30f2a006f67e5c4b3d7a6a10176abf22" ], "author": { "name": "Sadaf Ebrahimi", "email": "sadafebrahimi@google.com", "time": "Tue Aug 18 20:53:17 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Tue Aug 18 20:56:06 2026" }, "message": "[scudo] Fix MTE remapping for cached allocations\n\nWhen MTE is enabled, force page-alignment for all secondary allocations.\nSince the allocation position can shift when reusing a cached block, the\nnew header page must be remapped with MAP_MEMTAG to enable MTE and the\nold header page must be remapped without MAP_MEMTAG to disable MTE, as\nit may now be a part of the user playload. This is necessary because\nPROT_MTE is sticky and not cleared by the PROT_NONE mapping used during\ncaching.\n\nCo-authored-by: Christopher Ferris \u003ccferris1000@users.noreply.github.com\u003e\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 95bde4372e3f98a60da6f98f2bc29d435e9cef70\n" }, { "commit": "7556d2bc30f2a006f67e5c4b3d7a6a10176abf22", "tree": "143ff0e8bfe660e00cad6dfa5bd650d914b84dce", "parents": [ "09d9d65f0ffdec7e096f3dc1f8d0a068374b30af" ], "author": { "name": "Thurston Dang", "email": "thurston@google.com", "time": "Tue Aug 18 18:31:32 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Tue Aug 18 18:38:18 2026" }, "message": "[asan][test] Disable invalid-pointer-pairs-vector-extract.cpp (#217088)\n\nThis test (introduced in\nhttps://github.com/llvm/llvm-project/pull/216532) has been XPASS\u0027ing on\nS390X (reported in\nhttps://github.com/llvm/llvm-project/pull/216532#issuecomment-5327707172).\n\nThere has already been a prior fix-forward (marking Solaris as\nunsupported: https://github.com/llvm/llvm-project/pull/216606). Since\nhttps://github.com/llvm/llvm-project/pull/213546 is expected to make the\ntest pass on all platforms anyway, rather than playing whack-a-mole with\nmarking targets unsupported, this patch disables the test.\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: b21c2e248ec12993f81306290523910471367060\n" }, { "commit": "09d9d65f0ffdec7e096f3dc1f8d0a068374b30af", "tree": "6aa74c2fe0df96efb75c80e92ef5695d4a05fbb9", "parents": [ "c81d79311f1fe335871354455e30ed9e5ba616cd" ], "author": { "name": "Chris Apple", "email": "cja-private@pm.me", "time": "Tue Aug 18 17:56:03 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Tue Aug 18 18:04:58 2026" }, "message": "[rtsan] Fix sanitizer_common/TestCases/dlsym_alloc.c (#216579)\n\nThis test crashes on `free(NULL)` when DlsymAlloc is being used. This is\nbecause `DlysmAlloc::Free` seems to crash on linux when it\u0027s called with\n`nullptr`.\n\nThe vast majority of other sanitizers guard against it on the sanitizer\nside in this way:\n\nhttps://github.com/llvm/llvm-project/blob/65e0fe689bb7c1a6516644c27c25c3fe3c8b1ac7/compiler-rt/lib/tsan/rtl/tsan_interceptors_posix.cpp#L728\n\nhttps://github.com/llvm/llvm-project/blob/65e0fe689bb7c1a6516644c27c25c3fe3c8b1ac7/compiler-rt/lib/lsan/lsan_interceptors.cpp#L78\n\nFor example.\n\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: e93d6090dc1e8d4122512a3bab66929579e47bd4\n" }, { "commit": "c81d79311f1fe335871354455e30ed9e5ba616cd", "tree": "4c6c7eb47528e06e8fbfd44e08ee5246dd7de002", "parents": [ "36e41d114451498cc26b37b50fbbbff5d1b26587" ], "author": { "name": "Andrew Haberlandt", "email": "ahaberlandt@apple.com", "time": "Tue Aug 18 16:43:41 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Tue Aug 18 16:49:53 2026" }, "message": "[compiler-rt] [Darwin] Bump min-deployment targets to match libc++ (#216855)\n\nFixes the following warning when compiling with newer SDKs:\n\n```\n/....../usr/include/c++/v1/__configuration/availability.h:204:4: warning: \"The selected platform is no longer supported by libc++.\" [-W#warnings]\n 204 | # warning \"The selected platform is no longer supported by libc++.\"\n | ^\n```\n\nrdar://175395925\n\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: bc2e50bf7baf1c52370f8d8cdfee4bba2180d27a\n" }, { "commit": "36e41d114451498cc26b37b50fbbbff5d1b26587", "tree": "197c8b6ba413edd1850d02c1cb0f1a10f38aaaae", "parents": [ "0ccff08262472b88b5d5d193342d84ba4f6fa3fa" ], "author": { "name": "Ilya Leoshkevich", "email": "iii@linux.ibm.com", "time": "Tue Aug 18 11:47:11 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Tue Aug 18 11:50:58 2026" }, "message": "[JITLink] Allocate frames in SystemZ atexit tests (#215482)\n\nSystemZ JITLink atexit tests crash, because they call functions without\nfirst allocating a frame.\n\nFix by allocating one. While at it, switch to the more conventional\nslots for %r14 and %r15.\n\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 05cfc88431991c27e83eddd701692f9aebd53c1f\n" }, { "commit": "0ccff08262472b88b5d5d193342d84ba4f6fa3fa", "tree": "3cf6898daf737cd0fab6839eea22db349d89686b", "parents": [ "278d36c3d2d645a159405728c96d434bf9337064" ], "author": { "name": "Thurston Dang", "email": "thurston@google.com", "time": "Mon Aug 17 21:15:31 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Mon Aug 17 21:21:00 2026" }, "message": "[asan][test] Fix invalid-pointer-pairs-vector-extract.cpp breakage after #210729 (#216827)\n\ncompiler-rt/test/asan/TestCases/invalid-pointer-pairs-vector-extract.cpp\n(introduced in #216532) started failing (or rather, unexpectedly\npassing) after #210729 (e.g.,\nhttps://lab.llvm.org/buildbot/#/builders/66/builds/35645,\nhttps://lab.llvm.org/buildbot/#/builders/51/builds/42365), because it\nchanges the codegen for pointer subtraction if -fwrapv-pointer is not\nset.\n\nThis patch fixes the test case by adding -fwrapv-pointer.\n\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 9c139023cf2dd79cd1870766ba2118cb87f9bcf6\n" }, { "commit": "278d36c3d2d645a159405728c96d434bf9337064", "tree": "bfe57f7941fbb3a1c42bdb30ef5936cb4a97bedc", "parents": [ "e3963728983a66133592dac0a346e3e602baf63a" ], "author": { "name": "Rainer Orth", "email": "ro@gcc.gnu.org", "time": "Mon Aug 17 16:30:39 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Mon Aug 17 16:35:09 2026" }, "message": "[sanitizer] Skip hanging tests on NetBSD (#216712)\n\nSeveral sanitizer tests hang indefinitely on NetBSD:\n\n```\nMemorySanitizer-Unit :: ./Msan-x86_64-Test\nMemorySanitizer-Unit :: ./Msan-x86_64-with-call-Test\nMemorySanitizer-X86_64 :: zero_alloc.cpp\nThreadSanitizer-x86_64 :: signal_cond.cpp\nlibFuzzer-x86_64-default-NetBSD :: reload.test\n```\n\nAll of them loop and don\u0027t time out, so they need to be terminated\nmanually for `ninja check-all` to complete. To avoid this, this patch\nskips the affected tests or subtests. Unfortunately, the Msan unit tests\nstill hang on exit in `__cxa_finalize` even if all subtests are skipped\nwith `llvm-lit -gtest_filter\u003d-*`.\n\nTested on `x86_64-pc-netbsd11.0`, `x86_64-pc-freebsd15.1`, and\n`x86_64-pc-linux-gnu`.\n\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 782f1130177b2c792dc37f0923d21d187a4c150b\n" }, { "commit": "e3963728983a66133592dac0a346e3e602baf63a", "tree": "7340ba3249ebacb6b1691895c9141fed17138485", "parents": [ "1738b42e9c1b34ecaf2ddde937db76a2bffeac65" ], "author": { "name": "oltolm", "email": "oleg.tolmatcev@gmail.com", "time": "Mon Aug 17 00:43:56 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Mon Aug 17 00:46:16 2026" }, "message": "[compiler-rt] Clean up Windows includes in sanitizer runtimes (#216586)\n\nExtracted from https://github.com/llvm/llvm-project/pull/209902.\nIt\u0027s one of the patches that enables ASan and UBSan in GCC for Windows.\n\nCo-authored-by: Hannes Domani \u003cssbssa@yahoo.de\u003e\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 391dd417e0864e2d1b50550e041e80b6a153a528\n" }, { "commit": "1738b42e9c1b34ecaf2ddde937db76a2bffeac65", "tree": "9972ccb6bfe97553fb31c8ec752b5f9bcb09688c", "parents": [ "c8b2785f7af22a5cf9af9435c6bf4c8d32d37597" ], "author": { "name": "Thurston Dang", "email": "thurston@google.com", "time": "Sun Aug 16 22:42:50 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Sun Aug 16 22:46:04 2026" }, "message": "[asan][test] Exclude invalid-pointer-pairs-vector-extract.cpp from Solaris (#216606)\n\nIntended to fix buildbot breakage on Solaris (XPASS), as reported in\nhttps://github.com/llvm/llvm-project/pull/216532#issuecomment-5309297518\n\nThis change is only a transient step for the test, since the upcoming\nhttps://github.com/llvm/llvm-project/pull/213546 is expected to make the\ntest pass on all platforms.\n\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: c926bfd8069d46406603405ed01eb7e2d8362077\n" }, { "commit": "c8b2785f7af22a5cf9af9435c6bf4c8d32d37597", "tree": "c244366a4f27ebef10b6b5f3622988e0408a5601", "parents": [ "d66020cef8abc3154a71418239003c91d3bfff74" ], "author": { "name": "flovent", "email": "flbven@protonmail.com", "time": "Sun Aug 16 11:27:38 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Sun Aug 16 11:30:50 2026" }, "message": "[ASan] Add xfail test for PR #213546 (#216532)\n\nAdd xfail test for PR #213546\n\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 2d70006c2a40ed35e6622c6cb38fa579d76c7777\n" }, { "commit": "d66020cef8abc3154a71418239003c91d3bfff74", "tree": "f8ff4345d5e5627bd59ed5f28aa16201f799c3a2", "parents": [ "06f928b9590feccfceca8199f547bbf4e3aff744" ], "author": { "name": "flovent", "email": "flbven@protonmail.com", "time": "Sun Aug 16 03:04:32 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Sun Aug 16 03:05:42 2026" }, "message": "[ASan] Register assert lit feature in `lit.cfg.py` (#216529)\n\nAdd this feature to test a functional change related to asserts, see\nhttps://github.com/llvm/llvm-project/pull/213546#discussion_r3790705838.\n\nAnd this will allow other people to add `// REQUIRES: asserts` in\ntestcases in the future too.\n\nI follow the same pattern already used in `flang/test/lit.cfg.py` and\nother configs.\n\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: bfb662baf51dfc6e5c75b98c0f913cdbcbec96ad\n" }, { "commit": "06f928b9590feccfceca8199f547bbf4e3aff744", "tree": "2e5fd460c4a3f1199bc4ab8e396924718ad2d75a", "parents": [ "b86fdb4266bfe98e40947e22e8687aac772213df" ], "author": { "name": "Chris Apple", "email": "cja-private@pm.me", "time": "Sat Aug 15 14:21:20 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Sat Aug 15 14:25:53 2026" }, "message": "[sanitizer-common][rtsan] Fix log and stack printing for rtsan (#215707)\n\nFixes:\n* sanitizer_common/TestCases/suffix-log-path_test.c\n* sanitizer_common/TestCases/print-stack-trace.cpp\n\nThese two tests have some overlap, which is why they are fixed in the\nsame commit. Functionally, this means:\n* RTSan now exports the common `__sanitizer_print_stack_trace` API, from\nwhat I understand from the comment on the master version of this, this\nis for debuggers to call\n* RTSan can now report out to a log file using the common utilities\n\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 227ec3e87fd013ad1963827144ff8f6e9d762510\n" }, { "commit": "b86fdb4266bfe98e40947e22e8687aac772213df", "tree": "a57f0057220afe24ed54c33642829368ed1106f0", "parents": [ "c52233c1331c74b358ea439441f9a3fa71cd8563" ], "author": { "name": "jinge90", "email": "ge.jin@intel.com", "time": "Fri Aug 14 14:51:40 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Fri Aug 14 14:55:51 2026" }, "message": "[compiler-rt] Fix clang_rt.builtins spirv64 bitcode install component (#216276)\n\nInstall component for libclang_rt.builtins static archive library for\nspirv64 is clang_rt.builtins-spirv64, libclang_rt.builtins bitcode\nlibrary install should use same component as static archive library.\n\nSigned-off-by: jinge90 \u003cge.jin@intel.com\u003e\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: d0f98d63f6ea03009f8105619ab0e922dd334564\n" }, { "commit": "c52233c1331c74b358ea439441f9a3fa71cd8563", "tree": "335bc738f79cc1a4b0497ebf7e7f66e2da1e4bc0", "parents": [ "6ef009d66ab01be5450567c00d8ee4ce750bf884" ], "author": { "name": "jinge90", "email": "ge.jin@intel.com", "time": "Thu Aug 13 21:31:20 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Thu Aug 13 21:35:27 2026" }, "message": "[compiler-rt] Emit wrapper function for __eqsf2, __eqdf2 for SPIR-V (#215957)\n\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: f395eb5beec8b6868582c75d620a5b0962bf7da7\n" }, { "commit": "6ef009d66ab01be5450567c00d8ee4ce750bf884", "tree": "61ffb49be20235c50127e9777992d0376b8730f4", "parents": [ "6c2fdc89a3580fc75aed2b7d12c9534e00b7a707" ], "author": { "name": "jinge90", "email": "ge.jin@intel.com", "time": "Thu Aug 13 07:24:19 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Thu Aug 13 07:25:34 2026" }, "message": "[compiler-rt] Partially enable building compiler-rt unit tests for spirv64 target (#215717)\n\nCurrently, compilation/linking error will happen for all compiler-rt\nbuiltin unit test when building them for spirv64 target, there are a\nbunch of problem to resolve, this patch is the 1st one aiming to resolve\nfollowing 2 problems in building phase:\n1. fatal error: error in backend:\nSPV_EXT_relaxed_printf_string_address_space is required because printf\nuses a format string not in constant address space.\n2. -lc/-lm flags are not recognized by spirv64 toolchain.\n\nFor 1, unit tests are plain C source files and use printf, the format\nstring is a C-string which is not decorated by any address space, so\nneed to apply SPV_EXT_relaxed_printf_string_address_space to allow it\nfor spirv64.\nFor 2, spirv64 toolchain has not supported to use static archive\nlibrary, so this -lc/-lm is not accepted by clang. So, we need to use\nbitcode libc in device compilation phase via -mlink-builtin-bitcode.\nCurrently, we only link libc since libm spirv64 library has not been\nlanded in community. In order to build compiler-rt tests for spirv64, we\nneed to enable LLVM libc when building llvm-project.\n\nUsing following command to build llvm-project:\ncmake ../llvm -G Ninja -DLLVM_ENABLE_PROJECTS\u003d\"clang\"\n-DRUNTIMES_spirv64-unknown-unknown_LLVM_ENABLE_RUNTIMES\u003d\"compiler-rt;libc\"\n-DLLVM_RUNTIME_TARGETS\u003d\"default;spirv64-unknown-unknown\"\n-DRUNTIMES_spirv64-unknown-unknown_COMPILER_RT_INCLUDE_TESTS\u003dON\n-DRUNTIMES_spirv64-unknown-unknown_COMPILER_RT_EMULATOR\u003d/bin/true\nninja all\nninja check-compiler-rt-spirv64-unknown-unknown\nBefore applying the PR, the building result is:\nTotal Discovered Tests: 270\n Unsupported: 120 (44.44%)\n Passed : 1 (0.37%)\n Failed : 149 (55.19%)\nAfter applying the PR, the building result is:\nTotal Discovered Tests: 270\n Unsupported: 120 (44.44%)\n Passed : 119 (44.07%)\n Failed : 31 (11.48%)\n\n---------\n\nSigned-off-by: jinge90 \u003cge.jin@intel.com\u003e\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 50828033d5bb75b5bc75f695ece5fdcf2df76d1b\n" }, { "commit": "6c2fdc89a3580fc75aed2b7d12c9534e00b7a707", "tree": "1b89212717dfde37f8c69caf59beb0d641c794f6", "parents": [ "0404e0e0ba2d15bba03b6e3ced1cc1d6e8d07b86" ], "author": { "name": "Duo Wang", "email": "duo_wang@apple.com", "time": "Wed Aug 12 19:28:59 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Wed Aug 12 19:30:54 2026" }, "message": "[compiler-rt][Darwin] Drop i386 slice for builtins on modern SDKs (#215388)\n\nThe SDK \u003e\u003d 10.15 i386-removal in darwin_test_archs was gated by `if(NOT\nTEST_COMPILE_ONLY)`, so it never fired for the builtins config, which\nsets TEST_COMPILE_ONLY\u003dOn. `-arch i386` still compiles fine on current\nSDKs — ld is the only stage that rejects it — so the compile-only probe\nkept i386 in DARWIN_osx_BUILTIN_ARCHS, and downstream `strip` on the\nresulting fat archives fails on Xcode 27 with \"ld: linking for i386 is\nno longer supported\".\n\nHoist the check above the guard so it applies to both the runtime and\nbuiltin arch probes.\n\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 7c372d9f7a6e89f80350c895ac295c537055a2aa\n" }, { "commit": "0404e0e0ba2d15bba03b6e3ced1cc1d6e8d07b86", "tree": "744796f66cc37a604fe708267c44b520de6a45ca", "parents": [ "a6b30d73349b5b38721025e9dc19e0203f343e55" ], "author": { "name": "Christopher Ferris", "email": "cferris1000@users.noreply.github.com", "time": "Tue Aug 11 16:02:16 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Tue Aug 11 16:05:10 2026" }, "message": "[scudo] Remove malloc_enable_child from API. (#215439)\n\nThe malloc_enable_child does not need to be a public API, instead move\nthe function to be static and reference it that way.\n\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: a558267da71fe93b7ff6fa9c7abcbd9b4eb87444\n" }, { "commit": "a6b30d73349b5b38721025e9dc19e0203f343e55", "tree": "3c131101e8f9b3b35b8ddc0404e992dfc10bbde8", "parents": [ "d266864150e00567f307701c65d627cecb4395d4" ], "author": { "name": "Joseph Huber", "email": "huberjn@outlook.com", "time": "Mon Aug 10 21:23:07 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Mon Aug 10 21:26:06 2026" }, "message": "[compiler-rt] Use `fprintf` for proper targets (#215362)\n\nSummary:\nNeeds fprintf to flush properly for sanitization.\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: aaf177eceb12f35d43c088b9d185be909eb963b4\n" }, { "commit": "d266864150e00567f307701c65d627cecb4395d4", "tree": "07c175c411182827240dfd968af21568f4054e39", "parents": [ "287a9e550d89873f0035921aee885dc85c3bb899" ], "author": { "name": "Joseph Huber", "email": "huberjn@outlook.com", "time": "Mon Aug 10 17:41:00 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Mon Aug 10 17:43:55 2026" }, "message": "[compiler-rt] Fix GPU builds picking up incorrect headers (#215339)\n\nSummary:\nThis started failing after a change to AMDGPU\u0027s default types.\nPreviously we were not excluding the system includes, which means that\nwe were picking up glibc headers on an incompatible target. Make this\nexplicit. means we don\u0027t get the other system headers. This is\nunnecessary for now, but later one we will explicitly link it up to the\nLLVM libc headers once needed.\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 4548e6b80fb07c6e63dad48266064330953f90d3\n" }, { "commit": "287a9e550d89873f0035921aee885dc85c3bb899", "tree": "a93aa39e576b5c78babdd2ee69c28c7818a1836b", "parents": [ "fb536f2276a39e3eee46c8c94a9ba67de8306d94" ], "author": { "name": "Chris Apple", "email": "cja-private@pm.me", "time": "Sun Aug 09 17:36:53 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Sun Aug 09 17:40:33 2026" }, "message": "[sanitizer-common][rtsan] Mark unsupported tests (#213404)\n\nFrom #213170, turning that PR into separate commits as suggested.\n\nTo enable rtsan in sanitizer common, first disable the tests that are\nnot trivially fixable (or will remain permanently unsupported).\n\nQuoting from that other review, assuming all of these are merged and\nrtsan is enabled for sanitizer-common\n---\nDarwin:\nNew tests that run and pass: 249\nUNSUPPORTED: 26\nTotal: 275\nUNSUPPORTED percentage: 9.5%\n\nLinux is about the same, few additional unsupported:\nNew tests that run and pass: 249\nUNSUPPORTED: 40\nTotal: 289\nUNSUPPORTED percentage 13.8%\n\nAgain, some of these UNSUPPORTED are temporary, and we\u0027ll go back\nthrough and clean them up (or at least discuss suport) so I expect this\npercentage to go down over time\n\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 3d8868224e00d2f3821380c7df985c61fd5dfab3\n" }, { "commit": "fb536f2276a39e3eee46c8c94a9ba67de8306d94", "tree": "f6ff77bfa09d44ad2512850ec468179c1f381e78", "parents": [ "f8d2608c6a9e5089f4c017038e19b07626bfbe5a" ], "author": { "name": "hulxv", "email": "hulxxv@gmail.com", "time": "Sat Aug 08 21:24:13 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Sat Aug 08 21:25:37 2026" }, "message": "[compiler-rt][builtins] libc-backed int-to-double/float builtins (#209900)\n\nAdd libc-backed integer to double/float conversion builtins\n\nPart of #197824\n\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: f9b0ba3fdae3df4e6e1559b243d6bc786bc90807\n" }, { "commit": "f8d2608c6a9e5089f4c017038e19b07626bfbe5a", "tree": "58e6562f193e3cb082c39d3bef33849228f0b1ed", "parents": [ "634b5943f777b9755e131bc006c3b7d129d61f49" ], "author": { "name": "hulxv", "email": "hulxxv@gmail.com", "time": "Sat Aug 08 20:48:11 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Sat Aug 08 20:50:48 2026" }, "message": "[compiler-rt][builtins] add libc-backed double/float-to-int builtins (#207543)\n\nAdd libc-backed double/float to integer conversion builtins\n\nPart of #197824\n\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 2abc0e5685096d507e2967c3d5ad069a824f1a61\n" }, { "commit": "634b5943f777b9755e131bc006c3b7d129d61f49", "tree": "be848300887111ef2b42033dc3f743a49a3742f1", "parents": [ "3ec69a1ce2be5a7fe5117ef94db7709e52c8736f" ], "author": { "name": "Jin Huang", "email": "jinhuang1102@gmail.com", "time": "Sat Aug 08 13:53:05 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Sat Aug 08 13:55:47 2026" }, "message": "[compiler-rt][pgo] Use mock coroutine header in instrprof-coroutine-profile.cpp (#214593)\n\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 35854f82e5ce4e2bdea12cad413d30945e6e6b8a\n" }, { "commit": "3ec69a1ce2be5a7fe5117ef94db7709e52c8736f", "tree": "eea1a822e2397dde9065963fb4a0ef695e08ba46", "parents": [ "3478ebcc365572a3bcdeff6c2d8e355e944907aa" ], "author": { "name": "Jakub Jakacki", "email": "jakub.jakacki@intel.com", "time": "Thu Aug 06 21:24:42 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Thu Aug 06 21:30:57 2026" }, "message": "Rename Compiler-RT target to compiler-rt in CMake files (#214364)\n\nVisual Studio 2026 has trouble generating valid solutions for projects\nwith mismatching directory names.\nThe change renames the Compiler-RT target to match the directory name\n\"compiler-rt\" in all CMake files.\n\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 36e1616725cb56bd4b9fee255d85294407b213cc\n" }, { "commit": "3478ebcc365572a3bcdeff6c2d8e355e944907aa", "tree": "ddfdf4ca2e8f979dc4dc829a23c164c005ae6fe0", "parents": [ "7578fa6369a1e76324164f33c4d20d668c98a6b5" ], "author": { "name": "Christopher Ferris", "email": "cferris1000@users.noreply.github.com", "time": "Thu Aug 06 19:07:22 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Thu Aug 06 19:10:58 2026" }, "message": "[scudo] Add check to verify locking is correct. (#214377)\n\nAdded checks that at the end of lockSlow, the lock is held and when\ndoing an unlock, the lock is held.\n\nThis found a bug in primary32.h:unmapTestOnly where the lock is not\nheld, so fixed that.\n\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 38ca11fed5ec1448f16b717c2676562c7f99b969\n" }, { "commit": "7578fa6369a1e76324164f33c4d20d668c98a6b5", "tree": "05c097bd6f7635598b323ee927298a07f825a1a1", "parents": [ "bc24702387f6092b4e2da6860fbb5a60fb91757d" ], "author": { "name": "Trevor Gross", "email": "tg@trevorgross.com", "time": "Thu Aug 06 13:08:27 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Thu Aug 06 13:10:38 2026" }, "message": "[compiler-rt] Disable executable stack on `aeabi_u{read,write}*.S` (#214465)\n\nThese were missing `NO_EXEC_STACK_DIRECTIVE` to add `.note.GNU-stack`;\nwithout it, a binary including any of these files will have the stack\nmarked executable. Add the directive here, matching other similar files.\n\nSymtab diff before:\n\n$ clang compiler-rt/lib/builtins/arm/aeabi_uread4.S\n--target\u003darm-unknown-linux-gnueabi -c\n $ llvm-readelf aeabi_uread4.o -S\n There are 5 section headers, starting at offset 0xe4:\n\n Section Headers:\n[Nr] Name Type Address Off Size ES Flg Lk Inf Al\n[ 0] NULL 00000000 000000 000000 00 0 0 0\n[ 1] .strtab STRTAB 00000000 0000a8 000039 00 0 0 1\n[ 2] .text PROGBITS 00000000 000034 000020 00 AX 0 0 4\n[ 3] .ARM.attributes ARM_ATTRIBUTES 00000000 000054 000022 00 0 0 1\n[ 4] .symtab SYMTAB 00000000 000078 000030 10 1 2 4\n\nAfter:\n\n$ clang compiler-rt/lib/builtins/arm/aeabi_uread4.S\n--target\u003darm-unknown-linux-gnueabi -c\n $ llvm-readelf aeabi_uread4.o -S\n There are 6 section headers, starting at offset 0xf4:\n\n Section Headers:\n[Nr] Name Type Address Off Size ES Flg Lk Inf Al\n[ 0] NULL 00000000 000000 000000 00 0 0 0\n[ 1] .strtab STRTAB 00000000 0000a8 000049 00 0 0 1\n[ 2] .text PROGBITS 00000000 000034 000020 00 AX 0 0 4\n[ 3] .note.GNU-stack PROGBITS 00000000 000054 000000 00 0 0 1\n[ 4] .ARM.attributes ARM_ATTRIBUTES 00000000 000054 000022 00 0 0 1\n[ 5] .symtab SYMTAB 00000000 000078 000030 10 1 2 4\n\nFixes: 39413af931a7 (\"[Compiler-rt] Implement AEABI Unaligned Read/Write\n Helpers in compiler-rt (#167913)\")\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 849c51e082b0958524246a0a880f46d468a55147\n" }, { "commit": "bc24702387f6092b4e2da6860fbb5a60fb91757d", "tree": "6dc54f29fb7a4f276f6b602d05696bda4cc4a564", "parents": [ "74a75f454eeb829249834034afd4ccbbb3f0a5a2" ], "author": { "name": "Jin Huang", "email": "jinhuang1102@gmail.com", "time": "Thu Aug 06 07:33:40 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Thu Aug 06 07:35:31 2026" }, "message": "[compiler-rt][pgo] Add profile instrumentation test for coroutines (#213801)\n\nThis PR adds a compiler-rt test (`instrprof-coroutine-profile.cpp`) to\nverify PGO profile counter generation and ingestion for C++20\ncoroutines.\n\nDuring investigating the iFDO profile ingested for coroutine pass, we\nfound that the profile data contains entry counts for the original\ncoroutine function (`foo`), but lacks the entry counters for the split\n`foo.resume` function.\n\nAfter inspecting\n[PassBuilderPipelines.cpp](https://github.com/llvm/llvm-project/blob/f7b7ec8d5542dcede1ae607aa74dfe36d7a3c530/llvm/lib/Passes/PassBuilderPipelines.cpp#L1273)\nconfirms that `PGOInstrumentationGen` runs before `CoroSplitPass`.\nTherefore,\n- 1) the iFDO profile correctly ingests and associates profile counters\nwith coroutine function `foo`.\n- 2) Separate function entry counters for `foo.resume` are not expected\nin the iFDO profile.\n\nAdded `compiler-rt/test/profile/instrprof-coroutine-profile.cpp` and use\n`llvm-profdata show` verify coroutine profile data. The IR dump after\n`PGOInstrumentationGen` confirms that the function entry count is\ntracking by counter[6] in the 8-element counter for `foo`.\n\n```llvm\ncoro.alloc: ; preds \u003d %entry\n call void @llvm.instrprof.increment(ptr @__profn__Z3fooi, i64 650973721613012168, i32 8, i32 6)\n %2 \u003d call i64 @llvm.coro.size.i64()\n %call \u003d call noalias noundef nonnull ptr @_Znwm(i64 noundef %2) #14\n br label %coro.init\n\ncoro.init: ; preds \u003d %coro.alloc, %entry\n```\n\nCo-authored-by: Jin Huang \u003cjingold@google.com\u003e\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 2f4043eb92f789541b483dec8c794ecd67d167c1\n" }, { "commit": "74a75f454eeb829249834034afd4ccbbb3f0a5a2", "tree": "ce6b972cd6b27c350f3a4950c9eee79a4f0c2040", "parents": [ "ad5f058c6aa4933528402466c6f3b916867a13be" ], "author": { "name": "jinge90", "email": "ge.jin@intel.com", "time": "Wed Aug 05 15:40:46 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Wed Aug 05 15:45:24 2026" }, "message": "[compiler-rt] Create a combined compiler-rt builtin bitcode for SPIRV64 target (#214149)\n\nThis PR re-lands the compiler-rt builtin part from\nhttps://github.com/llvm/llvm-project/pull/196656\n\nSigned-off-by: jinge90 \u003cge.jin@intel.com\u003e\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: f6eb7c5ca11aa0d6c2463f0e7c83171362e16fd9\n" }, { "commit": "ad5f058c6aa4933528402466c6f3b916867a13be", "tree": "389ac454771ec8e0f3c45e84d64ee8069c1e8d4e", "parents": [ "33e2c4c3de2b4d522e3e9fae179d4d192e5f9f67" ], "author": { "name": "hulxv", "email": "hulxxv@gmail.com", "time": "Wed Aug 05 00:38:04 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Wed Aug 05 00:40:09 2026" }, "message": "[compiler-rt][cmake] filter libc-backed builtins superseded by assembly (#213481)\n\napply\nhttps://github.com/llvm/llvm-project/pull/209900#discussion_r3615454574\n\nfix the duplication that causes issues like this:\n```\n lit-23: /__w/llvm-project/llvm-project/compiler-rt/test/builtins/Unit/lit.cfg.py:186: fatal: builtins_source_features contains duplicates: [\u0027librt_has_floatdidf\u0027, \u0027librt_has_floatdisf\u0027, \u0027librt_has_floatundidf\u0027, \u0027librt_has_floatundisf\u0027]\nFAILED: compiler-rt/test/builtins/CMakeFiles/check-builtins /__w/llvm-project/llvm-project/build/compiler-rt/test/builtins/CMakeFiles/check-builtins\ncd /__w/llvm-project/llvm-project/build/compiler-rt/test/builtins \u0026\u0026 /usr/bin/python3 /usr/bin/lit-23 -sv --show-xfail --show-unsupported /__w/llvm-project/llvm-project/build/compiler-rt/test/builtins/TestCases /__w/llvm-project/llvm-project/build/compiler-rt/test/builtins/Unit/X86_64LinuxConfig\nninja: build stopped: subcommand failed.\n```\n\nPart of #197824\n\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: af2a0e9c8897060b4442e7d096459aad922f37d6\n" }, { "commit": "33e2c4c3de2b4d522e3e9fae179d4d192e5f9f67", "tree": "8f0a29a86b150bc1641ee3e6b954318b60cf4279", "parents": [ "9ee0925ae75737364674af7f294b99734439beb6" ], "author": { "name": "Vitaly Buka", "email": "vitalybuka@google.com", "time": "Tue Aug 04 19:27:26 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Tue Aug 04 19:32:00 2026" }, "message": "[NFC][libFuzzer] Remove unused GetModuleName helper in TracePC (#214032)\n\nFixing build bots after\nhttps://github.com/llvm/llvm-project/pull/203084.\n\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 70253af8709617b860865ce8d74204718e8a1975\n" }, { "commit": "9ee0925ae75737364674af7f294b99734439beb6", "tree": "ac3aa7ee09bb748875e1315a03a89e5de9573ccc", "parents": [ "0c35537bf46e7ae5946f335e66b684a26536465f" ], "author": { "name": "Daniel Scherzer", "email": "daniel.e.scherzer@gmail.com", "time": "Fri Jul 31 20:43:07 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Fri Jul 31 20:45:55 2026" }, "message": "[compiler-rt] [docs] Clean up explanation of `__fixunsdfdi` (#212624)\n\nReplace \"is compiling into\" with \"results in\"\n\n---------\n\nCo-authored-by: hulxv \u003chulxxv@gmail.com\u003e\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 6512cd96dd1b2e637666e1e6dfc84c1efacbe40b\n" }, { "commit": "0c35537bf46e7ae5946f335e66b684a26536465f", "tree": "fcba5e3fac768db9741ff9f9979febc3ac12d00f", "parents": [ "2449d9b3a35795926f010f2b4399e9d85061c8ea" ], "author": { "name": "Jakob Koschel", "email": "jakobkoschel@google.com", "time": "Fri Jul 31 08:56:15 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Fri Jul 31 09:00:42 2026" }, "message": "[ubsan] Report specific CFI checks in UBSan summaries (#213224)\n\nInstead of reporting generic `cfi-bad-type` in UBSan summaries for CFI\nfailures, report the specific CFI check kind (e.g., `cfi-vcall`,\n`cfi-nvcall`, `cfi-icall`, `cfi-mfcall`, `cfi-derived-cast`,\n`cfi-unrelated-cast`).\n\nThis is done by splitting CFIBadType into specific error types in\nubsan_checks.inc, and updating the handlers to report the appropriate\nErrorType based on the check kind. The suppression flag name for all of\nthem remains `cfi` to maintain backward compatibility.\n\nAlso replaces cfi-bad-type expectations in the existing tests.\n\nAssisted-by: Automated tooling, human reviewed.\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 6f9c189f738d47c7ad80f6bdf928fea441bfa8d6\n" }, { "commit": "2449d9b3a35795926f010f2b4399e9d85061c8ea", "tree": "393bdc5bb2ba9aed4f8d122e0b08e65596595bf5", "parents": [ "2da3d90e811c58607e1fcd7d5f7e41508010e548" ], "author": { "name": "Jakob Koschel", "email": "jakobkoschel@google.com", "time": "Fri Jul 31 07:24:44 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Fri Jul 31 07:25:38 2026" }, "message": "[ubsan] Test CFI summaries globally (#206434)\n\nAlways enable report_error_type\u003d1 and print_summary\u003d1 for CFI tests, and\nupdate existing tests to check for cfi-bad-type. This allows us to\nverify summary output comprehensively across the test suite.\n\nThis is a prerequisite for the change to the UBSan summary for CFI in\nhttps://github.com/llvm/llvm-project/pull/203341.\n\nAssisted-by: Automated tooling, human reviewed.\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: bcf20310bdafd37253c6c849cd9bd3ba5b79752f\n" }, { "commit": "2da3d90e811c58607e1fcd7d5f7e41508010e548", "tree": "2d853e3123d72e85d4bc11ef0ee436e4125fd0a4", "parents": [ "c9e8a87a5def666108fb842939c9f46a36bf25df" ], "author": { "name": "hulxv", "email": "hulxxv@gmail.com", "time": "Fri Jul 31 00:50:13 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Fri Jul 31 00:55:42 2026" }, "message": "[compiler-rt][builtins] libc-backed floating-point extend/trunct builtins (#209984)\n\nAdd libc-backed floating-point extend/truncate builtins\n\nPart of #197824\n\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 0879c1092c66242027065634210692cc06359fee\n" }, { "commit": "c9e8a87a5def666108fb842939c9f46a36bf25df", "tree": "82d6d2a3eb1e407df1bb26f144135983a76d927e", "parents": [ "49199be78295abfd8547c823017c3d1eabae997a" ], "author": { "name": "Thurston Dang", "email": "thurston@google.com", "time": "Thu Jul 30 21:59:14 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Thu Jul 30 22:00:24 2026" }, "message": "[msan][test] Disable UB cases in vararg_shadow.cpp (#204246)\n\nThis patch comments out the tests for char and float, which have\nundefined behavior: `warning: second argument to \u0027va_arg\u0027 is of\npromotable type \u0027char\u0027/\u0027float\u0027; this va_arg has undefined behavior\nbecause arguments will be promoted to \u0027int\u0027/\u0027double\u0027 [-Wvarargs]`.\n\nThe non-UB way to rewrite the tests is to use `va_arg()` with int and\ndouble types (when the original variables are char and float), but there\nare already test cases for int and double.\n\nNote also that the float test makes an assumption that\n`cast\u003cfloat\u003e(cast\u003cdouble\u003e(float))` will maintain the original shadow,\nwhich MSan need not guarantee. In particular,\nhttps://github.com/llvm/llvm-project/pull/204197 breaks that assumption\n(which led to a buildbot breakage and revert).\n\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 56a48d712006ff7d9a09eef6a86b78d4b2eec3aa\n" }, { "commit": "49199be78295abfd8547c823017c3d1eabae997a", "tree": "366963d358ec6ab445a8dbbfa29f63e7a27d3a14", "parents": [ "9dced0e25824a13b671590a033a95f628445830b" ], "author": { "name": "Michael Jones", "email": "michaelrj@google.com", "time": "Thu Jul 30 19:52:08 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Thu Jul 30 19:55:57 2026" }, "message": "[scudo] Fix GCC build after 212604 (#212904)\n\nChange #212604 introduced an issue when building with gcc where it\ncomplains there are \"too few template-parameter-lists\" (see log:\n\nhttps://lab.llvm.org/buildbot/#/builders/131/builds/51269/steps/6/logs/stdio\n)\nThis PR fixes the build.\n\nAssisted-by: Automated tooling, human reviewed.\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: aae5b3f2a92cc1c08a4bd65ba397fba5e37272fe\n" }, { "commit": "9dced0e25824a13b671590a033a95f628445830b", "tree": "899d97e8b40bc7cc735fb0a59ed4b5cb0d579ec2", "parents": [ "d385b98e482c47ea6774280aaf405c1fc4631c5c" ], "author": { "name": "Rainer Orth", "email": "ro@gcc.gnu.org", "time": "Thu Jul 30 19:11:41 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Thu Jul 30 19:15:34 2026" }, "message": "[tsan] Fix buildgo.sh on FreeBSD/NetBSD (#213047)\n\nThis patch fixes the remaining compile errors when running `ninja\ncheck-all` on FreeBSD and NetBSD:\n\n- `sanitizer_linux.cpp` doesn\u0027t compile on NetBSD, so this patch adds a\ncast:\n ```\nsanitizer_common/sanitizer_linux.cpp:2492:10: error: format specifies\ntype \u0027unsigned long long\u0027 but the argument has type \u0027__greg_t\u0027 (aka\n\u0027unsigned long\u0027) [-Werror,-Wformat]\n ```\n\n- `tsan_interface_atomic.cpp` doesn\u0027t compile on both FreeBSD and\nNetBSD, so this patch disables the warning:\n ```\ntsan_interface_atomic.cpp:353:12: error: unused function template\n\u0027NoTsanAtomic\u0027 [-Werror,-Wunused-template]\ntsan_interface_atomic.cpp:358:12: error: unused function template\n\u0027Atomic\u0027 [-Werror,-Wunused-template]\ntsan_interface_atomic.cpp:389:12: error: unused function template\n\u0027NoTsanAtomic\u0027 [-Werror,-Wunused-template]\ntsan_interface_atomic.cpp:394:12: error: unused function template\n\u0027Atomic\u0027 [-Werror,-Wunused-template]\ntsan_interface_atomic.cpp:401:12: error: unused function template\n\u0027NoTsanAtomic\u0027 [-Werror,-Wunused-template]\ntsan_interface_atomic.cpp:406:12: error: unused function template\n\u0027Atomic\u0027 [-Werror,-Wunused-template]\n ```\n\n- `sanitizer_linux_libcdep.cpp` doesn\u0027t compile on NetBSD:\n ```\nsanitizer_linux_libcdep.cpp:527:27: error: use of undeclared identifier\n\u0027__lwp_getprivate_fast\u0027; did you mean \u0027_lwp_getprivate\u0027?\n ```\nThis is the same issue already handled in `tsan_platform_linux.cpp`: to\nexpose the `__lwp_getprivate_fast` definition in `\u003cmachine/mcontext.h\u003e`,\n`_RTLD_SOURCE` needs to be defined before `\u003cmachine/mcontext.h\u003e` is\nincluded (indirectly from `\u003csignal.h\u003e` in this case). It also needs to\ninclude `\u003csys/types.h\u003e` to get a definition of the `__aligned` macro.\n\nTested on `amd64-pc-freebsd15.1` and `amd64-pc-netbsd10.1`.\n\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 6e92d6d0577cbe102bd48f396301ea5cd0ebf247\n" }, { "commit": "d385b98e482c47ea6774280aaf405c1fc4631c5c", "tree": "efddcfb7a5eaf1b9be8002c09a89f5aa9f372cdf", "parents": [ "6fc45ef4ec7ae33901e3fa7ff6dc452a857f20e4" ], "author": { "name": "Christopher Ferris", "email": "cferris1000@users.noreply.github.com", "time": "Wed Jul 29 22:02:32 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Wed Jul 29 22:05:43 2026" }, "message": "[scudo] Fix all deadlocks with condition code variables. (#212604)\n\nAdded a ScopedFLLock that automatically handles notify calls if there\nare waiters. Replace all ScopedLock of FLLock with this new\nScopedFLLock. Updated the enable function to properly notify FLLock\nwaiters if necessary.\n\nAdded a new function for re-enabling in the child process after a fork\nto clear the variables used by the condition code.\n\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 738cbf1883083525a76a6eeb30671ea778661bd3\n" }, { "commit": "6fc45ef4ec7ae33901e3fa7ff6dc452a857f20e4", "tree": "67043fdde95b9ad5c2fe24fc9faff6a62e774927", "parents": [ "4b0944a24e859ff6c4e32dce33a40129db35e94e" ], "author": { "name": "Nikita Taranov", "email": "nikita.taranov@clickhouse.com", "time": "Wed Jul 29 17:52:36 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Wed Jul 29 17:55:51 2026" }, "message": "[XRay][test] Raise fdr-mode.cpp unwrite threshold to deflake it (#211595)\n\nIt was observed failing on `sanitizer-aarch64-linux` ([build\n40750](https://lab.llvm.org/buildbot/#/builders/51/builds/40750)).\n\nThe failure is in the `UNWRITE` run\n(`XRAY_FDR_OPTIONS\u003d\"func_duration_threshold_us\u003d5000\"`), which asserts\nthat every short function record is \"unwritten\" (rewound) except\narg1-logging records. In the failing output an `fA()` enter/exit pair\nsurvived:\n\n```\n\u003cstdin\u003e:9: - { func-id: 3, function: \u0027fA()\u0027, cpu: 0, thread: 1909237, kind: function-enter, tsc: 1784775492023431847 }\n\u003cstdin\u003e:10: - { func-id: 3, function: \u0027fA()\u0027, cpu: 0, thread: 1909237, kind: function-exit, tsc: 1784775492029462345 }\n\u003cstdin\u003e:11: - { func-id: 4, function: \u0027fArg(int)\u0027, args: [ 1 ], ..., kind: function-enter-arg, ... }\n```\n\nso `UNWRITE-NEXT` (and `UNWRITE-NOT: function-enter`) no longer hold.\n\nFDR mode only rewinds a function\u0027s enter record when its exit lands\nwithin the threshold (`XRayFDRController::functionExit`: `TSC -\nLastFunctionEntryTSC \u003c CycleThreshold`).\n\n```\n1784775492029462345 - 1784775492023431847 \u003d 6,030,498 cycles / 1 GHz \u003d 6.03 ms \u003e 5 ms threshold\n```\n\ni.e. the thread was preempted for ~6 ms mid-function, so the record was\n(correctly, by the runtime\u0027s own rule) not unwritten.\n\nThe proposed fix is to raise the `UNWRITE` run\u0027s\n`func_duration_threshold_us` from `5000` to `100000`, so realistic\nscheduling jitter cannot push a short function over the threshold. This\nmirrors the accepted fix for the sibling test in\n[PR #186611](https://github.com/llvm/llvm-project/pull/186611)\n(`basic-filtering.cpp`), which flaked the same way.\n\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 130dccc2c3192d9f49c2e85bcd047ed67a9fa9e1\n" }, { "commit": "4b0944a24e859ff6c4e32dce33a40129db35e94e", "tree": "48f5d279045578f4bab2c6e23ffa744727fca8d7", "parents": [ "cdd58e48ab09d8cbc0df73af2c4c2cd57b0098d9" ], "author": { "name": "hulxv", "email": "hulxxv@gmail.com", "time": "Wed Jul 29 00:38:54 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Wed Jul 29 00:40:38 2026" }, "message": "[compiler-rt][builtins] add libc-backed arithmetic builtins (#207092)\n\nAdd libc-backed arithmetic builtins\n\nDepends on #197950\n\nPart of #197824\n\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: a13822ff14b6e4082caa553b230beb1a7252cb8e\n" }, { "commit": "cdd58e48ab09d8cbc0df73af2c4c2cd57b0098d9", "tree": "498fc3eeeaf818fb4d6b51b3b98c46d766583d90", "parents": [ "9d3c9e033e689560f5192a46754b38889dca9489" ], "author": { "name": "Srividya Sundaram", "email": "srividya.sundaram@intel.com", "time": "Tue Jul 28 18:11:57 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Tue Jul 28 18:17:36 2026" }, "message": "Revert \"[Driver][SYCL] Add compile-time device library linking for SPIR-V targets\" (#212550)\n\nReverts llvm/llvm-project#196656\nThe CI does not yet build compiler-rt with SPIR-V support, so\nlibclang_rt.builtins.bc is absent on that bot. This change made the\nmissing file a hard error, breaking all libsycl functional tests that\ncompile with -fsycl.\n\nRe-landing after CI is updated to build compiler-rt for SPIR-V.\n\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 3919897e1b1b2c4f4327e9a8a9eebfb91bc75616\n" }, { "commit": "9d3c9e033e689560f5192a46754b38889dca9489", "tree": "4a9ad5035f6db6430c3e57d564799cb1f6994882", "parents": [ "9ee960589aadf779d42272087227d680aa3d7bd9" ], "author": { "name": "Jinseok Kim", "email": "always.starving0@gmail.com", "time": "Tue Jul 28 17:52:17 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Tue Jul 28 17:55:59 2026" }, "message": "[asan][NFC] Fix comments in asan_poisoning.cpp (#212449)\n\nRemove duplicated wording and fix typos in comments.\n\nSigned-off-by: Jinseok Kim \u003calways.starving0@gmail.com\u003e\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: d4d007b581bef93d6b7cee4bf9d621804826a9d6\n" }, { "commit": "9ee960589aadf779d42272087227d680aa3d7bd9", "tree": "a334bee8d561f7444247458dfd3d405f2652ec62", "parents": [ "4a08706fffe66ede459c36064ff267dd03c5e854" ], "author": { "name": "Srividya Sundaram", "email": "srividya.sundaram@intel.com", "time": "Tue Jul 28 14:31:09 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Tue Jul 28 14:34:40 2026" }, "message": "[Driver][SYCL] Add compile-time device library linking for SPIR-V targets (#196656)\n\nThis PR implements compile-time device library linking for SYCL\noffloading to SPIR-V targets, using `libclang_rt.builtins.bc` - an\nin-tree compiler-rt artifact produced alongside the existing\n`libclang_rt.builtins.a` for `SPIRV64`.\n\n## Motivation\n\nSYCL device compilations targeting SPIR-V need access to compiler\nbuiltins (integer arithmetic, floating-point helpers, etc.) at compile\ntime so the compiler can optimize across user code and builtins, inline\naggressively, and eliminate dead code. This PR lays the foundation by\nwiring up the first in-tree device library : `libclang_rt.builtins.bc` -\nusing the same `-mlink-builtin-bitcode` mechanism already used by\n`libclc` and `HIP`.\n\n## Changes\n\n**compiler-rt (compiler-rt/lib/builtins/CMakeLists.txt)**\nFor the `spirv64` target, after building\n`libclang_rt.builtins-spirv64.a`, runs `llvm-link` to merge all bitcode\nobjects into a single `libclang_rt.builtins.bc`. This file is installed\nto `\u003cResourceDir\u003e/lib/spirv64-unknown-unknown/` alongside the static\narchive. The `-mlink-builtin-bitcode` flag requires raw LLVM IR and\ncannot unpack a static archive directly, hence the extra step.\n\n**Clang Driver (clang/lib/Driver/ToolChains/SYCL.cpp)**\n`SYCLToolChain::getDeviceLibs()` now constructs the path to\n`libclang_rt.builtins.bc` under the clang resource directory and returns\nit for `-mlink-builtin-bitcode` injection during `spirv64` device\ncompilation. If the file is absent (e.g. compiler-rt was not built with\nspirv64 support), a specific diagnostic is emitted. Linking is\nsuppressed when `--no-offloadlib` is passed.\n\n**Diagnostics (clang/include/clang/Basic/DiagnosticDriverKinds.td)**\nAdded `err_drv_no_compiler_rt_builtins_bc`, following the\n`err_drv_libclc_not_found` community pattern, with actionable guidance\non how to build compiler-rt with SPIR-V support.\n\n### Test Infrastructure\n\n1. `clang/test/Driver/sycl-device-lib-spirv64.cpp` - new test verifying\n-mlink-builtin-bitcode is passed for spirv64, suppressed with\n--no-offloadlib, and that a missing .bc produces the correct diagnostic.\n2. `clang/test/Driver/Inputs/spirv64-sycl/` - dummy resource-dir fixture\nused by driver tests.\n3. All existing SYCL driver tests updated to pass -`resource-dir\n%S/Inputs/spirv64-sycl `on -fsycl RUN lines, making them independent of\nthe installed clang resource directory on CI runners.\n\n## Behavior\n\n### Default\n\n```bash\nclang -fsycl myprogram.cpp\n# Device compilation includes:\n# \"-mlink-builtin-bitcode\" \"\u003cResourceDir\u003e/lib/spirv64-unknown-unknown/libclang_rt.builtins.bc\"\n```\n\nAutomatically links device libraries from `\u003cclang-install\u003e/lib/` during\ndevice compilation.\n\n### Without device libraries\n```bash\nclang -fsycl --no-offloadlib myprogram.cpp\n# -mlink-builtin-bitcode is suppressed\n```\nMissing .bc (compiler-rt not built with spirv64 support)\n```bash\nerror: no compiler-rt builtins bitcode library \u0027...\u0027 found in the clang resource directory;\nbuild compiler-rt with SPIR-V support or pass \u0027--no-offloadlib\u0027 to compile without it ```\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 5a650306e4e1069d194d8fcfeee821d7e424f9b3\n" }, { "commit": "4a08706fffe66ede459c36064ff267dd03c5e854", "tree": "9bdb21b04dda0c3395cf010707bac34b5901c9b1", "parents": [ "1c2e0e2f28b09adc0a4b5921bae4191de23e4759" ], "author": { "name": "Christopher Ferris", "email": "cferris1000@users.noreply.github.com", "time": "Mon Jul 27 17:49:24 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Mon Jul 27 17:51:08 2026" }, "message": "[scudo] Test to verify fork and allocate does not deadlock. (#211975)\n\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: ad791f26ef495dbc2c3c6c283518872697210898\n" }, { "commit": "1c2e0e2f28b09adc0a4b5921bae4191de23e4759", "tree": "a23c9132068d2dbebe6ffe9669fa4444503c83de", "parents": [ "e1c9323385b40780f574f440a8a9ad83855712b0" ], "author": { "name": "Sadaf Ebrahimi", "email": "sadafebrahimi@google.com", "time": "Mon Jul 27 15:39:00 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Mon Jul 27 15:40:29 2026" }, "message": "[scudo] Add unmap and eviction stats to Secondary\n\nTrack and print stats on the number of memory unmaps and cache evictions\nin the MapAllocator. These statistics track total unmaps, as well as\nunmaps triggered by exceeding the maximum cache entry size or the\nmaximum number of entries in the cache.\n\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 0d70ca384090b709eb651fe4ea43398d591e4711\n" }, { "commit": "e1c9323385b40780f574f440a8a9ad83855712b0", "tree": "c48070c45101e395b6edab4020c61f52d662c643", "parents": [ "2d62de9efaa4eb5aebfbda260c34b888111317f9" ], "author": { "name": "Rainer Orth", "email": "ro@gcc.gnu.org", "time": "Mon Jul 27 07:54:12 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Mon Jul 27 07:56:29 2026" }, "message": "[sanitizer_common][tests] Skip MemoryMapping::ParseUnixMemoryProfile … (#212114)\n\n…etc. on NetBSD\n\nThe `Sanitizer-x86_64-Test` test doesn\u0027t link on NetBSD/amd64 10.1:\n\n```\n/usr/bin/ld: /usr/bin/ld: DWARF error: invalid or unhandled FORM value: 0x25\nSANITIZER_TEST_OBJECTS.sanitizer_procmaps_test.cpp.x86_64.o: in function `__sanitizer::MemoryMapping_ParseUnixMemoryProfile_Test::TestBody()\u0027:\nsanitizer_procmaps_test.cpp:(.text._ZN11__sanitizer41MemoryMapping_ParseUnixMemoryProfile_Test8TestBodyEv+0x52): undefined reference to `__sanitizer::ParseUnixMemoryProfile(void (*)(unsigned long, unsigned long, bool, unsigned long*), unsigned long*, char*, unsigned long)\u0027\n/usr/bin/ld: SANITIZER_TEST_OBJECTS.sanitizer_procmaps_test.cpp.x86_64.o: in function `__sanitizer::MemoryMapping_ParseUnixMemoryProfileTruncated_Test::TestBody()\u0027:\nsanitizer_procmaps_test.cpp:(.text._ZN11__sanitizer50MemoryMapping_ParseUnixMemoryProfileTruncated_Test8TestBodyEv+0x17a): undefined reference to `__sanitizer::ParseUnixMemoryProfile(void (*)(unsigned long, unsigned long, bool, unsigned long*), unsigned long*, char*, unsigned long)\u0027\n\n```\n\nFixed by skipping the affected subtest.\n\nTested on `amd64-pc-netbsd10.1` and `x86_64-pc-linux-gnu`.\n\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 0cc7159125eacd3abdff94eb7e634c1a8f889d1e\n" }, { "commit": "2d62de9efaa4eb5aebfbda260c34b888111317f9", "tree": "ae27c85e70f5c8e3c604a771d79d843ac2511e69", "parents": [ "268a1483b6fa220f7c0cffe4e280767605cb9a7a" ], "author": { "name": "Yaxun (Sam) Liu", "email": "yaxun.liu@amd.com", "time": "Sat Jul 25 13:50:07 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Sat Jul 25 13:55:49 2026" }, "message": "[PGO][HIP] Support hipModuleLoad in offload PGO (#211875)\n\nOffload PGO finds profile sections by inspecting the image passed to\nhipModuleLoadData. hipModuleLoad only provides a file name, so its\ndevice\nprofile counters were not collected.\n\nAfter a successful hipModuleLoad, read the code object with the existing\nprofile file-buffer helper and register it through the same path used by\nin-memory module loads.\n\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: ae5e065a1128ce8e206da6dc39ae54036f85e7d2\n" }, { "commit": "268a1483b6fa220f7c0cffe4e280767605cb9a7a", "tree": "51103095045a2e06939c90a48ecae42f32da79ac", "parents": [ "e4275cd7221fceef65b9dbd7d619d5c37776796c" ], "author": { "name": "c4x64", "email": "bloodmeridian.prabhas@gmail.com", "time": "Fri Jul 24 21:13:11 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Fri Jul 24 21:15:42 2026" }, "message": "[tsan][ppc64] Fix copy-paste bug in __sigsetjmp OPD TOC loading (#210589)\n\nIn `__sigsetjmp`\u0027s big-endian OPD path, the TOC was loaded from\n`_setjmp`\u0027s OPD entry instead of `__sigsetjmp`\u0027s own entry, due to a\ncopy-paste error.\n\n```diff\n- addis r2,r2,_setjmp-1b@ha\n- addi r2,r2,_setjmp-1b@l\n+ addis r2,r2,__sigsetjmp-1b@ha\n+ addi r2,r2,__sigsetjmp-1b@l\n```\n\nThis caused an incorrect TOC pointer to be loaded on big-endian PPC64\nwhen `__sigsetjmp`/sigsetjmp was intercepted by TSan. The corresponding\n`_setjmp` path (which uses label `0b`) correctly references its own\nsymbol for the OPD lookup.\n\nCo-authored-by: pmr \u003cprabhas@pmrs-MacBook.local\u003e\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 8ba8c16cde64c0c9c33e73fea8d75a19a748cede\n" }, { "commit": "e4275cd7221fceef65b9dbd7d619d5c37776796c", "tree": "089f61ed28051e2170bca8dce079237f32a1df02", "parents": [ "70ee990da50fdf4837433425ef7f8ea4a9e55874" ], "author": { "name": "Justin Cady", "email": "desk@justincady.com", "time": "Fri Jul 24 17:17:34 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Fri Jul 24 17:21:10 2026" }, "message": "[compiler-rt] FreeBSD kernel headers in baremetal libprofile (#211278)\n\nAdd conditional FreeBSD kernel headers when building baremetal profile\nlibrary. This is unconnected to a CMake target; it requires building\nlibprofile out of band with `-DCOMPILER_RT_PROFILE_BAREMETAL\u003d1`\nand `-I/usr/src/sys/sys`.\n\nThe headers replaced in this diff are not available when building for\nFreeBSD kernel space. This is the first issue I hit when experimenting\nwith using the baremetal libprofile like this, so there may be other\nchanges required as I make progress.\n\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: aafd116191c66625e75ca8efac0281e7c52e39c5\n" }, { "commit": "70ee990da50fdf4837433425ef7f8ea4a9e55874", "tree": "9a3bcd938404f7e972135aef18f27ae4e7b1029d", "parents": [ "f5ea1d95d8406bc0f26ce6b9ee5bd20187002357" ], "author": { "name": "Ian.han", "email": "jackinwhat@gmail.com", "time": "Thu Jul 23 17:50:20 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Thu Jul 23 17:55:40 2026" }, "message": "[sanitizer_common] Don\u0027t crash in fopen64 interceptor when path is NULL (#211468)\n\n`fopen` was fixed to tolerate a NULL `path` in 2015 (1d1be3dd8822), and\n`freopen`/`freopen64` carry the same `if (path)` guard. `fopen64` was\nmissed, so `fopen64(NULL, mode)` dereferences NULL inside the\ninterceptor and crashes under sanitizers, even though real `fopen64`\nwould just return NULL/EFAULT.\n\nAdd the missing `if (path)` guard, plus a regression test mirroring\n`fopen_nullptr.c`. Since `fopen64` is only intercepted on glibc\n(`SANITIZER_INTERCEPT_FOPEN64`), the test is placed under `Linux/` and\ngated with `// REQUIRES: glibc`.\n\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 8f64a4806ffbf577a91390750c2701de2c6496af\n" }, { "commit": "f5ea1d95d8406bc0f26ce6b9ee5bd20187002357", "tree": "a5570b9191fc044b5e31a64f31293e5454019dde", "parents": [ "f39acf7e52deac52290149e672d5c74eb596dcae" ], "author": { "name": "Patrik Dokoupil", "email": "61749375+pdokoupil@users.noreply.github.com", "time": "Thu Jul 23 04:49:55 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Thu Jul 23 04:55:38 2026" }, "message": "[compiler-rt][sanitizer_common] Size readlink/getsockopt post-hook unpoison by bytes written (#209209)\n\n# About\n\n`readlink`/`readlinkat` return the number of bytes placed in the buffer\nin res and do not NUL-terminate it, and getsockopt writes *optlen bytes\nof binary option data. The post-hooks instead sized their POST_WRITE\n(MSan unpoison) with internal_strlen(buf) + 1, which reads past what the\nkernel wrote -- over- unpoisoning the uninitialized tail (masking real\nbugs) and, on a buffer with no NUL, reading out of bounds inside the\nruntime. For binary option data an early zero byte instead\nunder-unpoisons.\n\nSize the unpoison by the actual written length, matching the\ncorresponding libc interceptors (readlink unpoisons res bytes, and\ngetsockopt unpoisons *optlen bytes).\n\nSeems to be present since the file\u0027s 2013 import.\n\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 73a7638c0d832182dd7df912115d80be87567cb4\n" }, { "commit": "f39acf7e52deac52290149e672d5c74eb596dcae", "tree": "48d616848e458d2083ab0ceed6e9b8130ffcedc5", "parents": [ "7b86f93d42ebf714651f5fa54afd82c0131d6aee" ], "author": { "name": "Haowei", "email": "haowei@google.com", "time": "Wed Jul 22 06:23:13 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Wed Jul 22 06:26:03 2026" }, "message": "Remove debug leftovers (#211068)\n\na04f87fdd9b59c6089bf731c669c6f149208363c merged with some debug print\nleftovers. This patch removes them.\n\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 91406e8dff2b1578ff4dd22593ea1a8039b074a4\n" }, { "commit": "7b86f93d42ebf714651f5fa54afd82c0131d6aee", "tree": "23c0b5037e619b79a0a302c74e9ec513255e0bbb", "parents": [ "ddf6253b3226ab64a6f86a34b82e21256fc698ea" ], "author": { "name": "Chris Apple", "email": "cja-private@pm.me", "time": "Wed Jul 22 01:51:17 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Wed Jul 22 01:55:57 2026" }, "message": "[compiler-rt][rtsan] Loosen requirements of halt_on_error test (#210734)\n\nSome systems may call other intercepted functions during the course of a\n`malloc`. This would result in some error stack such as:\n\n```\n ERROR malloc\n ERROR pthread_mutex_lock\n ERROR pthread_mutex_unlock\n\n ERROR free\n ERROR pthread_mutex_lock\n ERROR pthread_mutex_unlock\n```\n\nWe support this as RTSan, but this test would be overly specific on\nthose platforms. As written, this requires \"only malloc, then free\nimmediately after\". The change makes this \"at least malloc is called,\nand then later at least free is called\". This still keeps the spirit of\nthe check the same but allows for more (valid) interpretation in\nbetween.\n\nNOTE:\nIn the future, we may want to consider doing something similar here:\n\nhttps://github.com/llvm/llvm-project/blob/40c4fea67f49b841bd064624219efceab91b65e0/compiler-rt/test/rtsan/stack_suppressions.cpp#L65\n\nThis specifies ONLY 7 suppressions should happen here which is extremely\nspecific. However I did not change this because it\u0027s a little trickier.\nWe don\u0027t really have a test for this stack suppressions method anywhere\nelse. We could do something like \"at least 7 things should be\nsuppressed\", but it\u0027s messier so I left that for a future exercise.\n\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 0cf295574bbef856646e8321e4e798128b84b73b\n" }, { "commit": "ddf6253b3226ab64a6f86a34b82e21256fc698ea", "tree": "e081ec78d298a2658bc7fac067a7a1a334687333", "parents": [ "c45c8e2b2dd6905c88ad01250b3bcaae596b0e6a" ], "author": { "name": "Dan Blackwell", "email": "dan_blackwell@apple.com", "time": "Tue Jul 21 13:08:51 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Tue Jul 21 13:10:51 2026" }, "message": "[Darwin][TSan] Fix race in mach_vm_deallocate interceptor (#210716)\n\nI have seen an issue whereby the meta store in MetaMap::AllocBlock lands\nin the unmapped gap left during the tsan::MetaMap::ResetRange call;\nwhere the range is first unmapped, and then remapped with MAP_FIXED.\nThis occurred under the mach_vm_deallocate interceptor, because it first\ndoes the deallocate call, and only then does it call UnmapShadow -\nleaving a gap where a fresh slab for malloc can land, only to have the\nmeta region for that range mapped out from under it, resulting in a bad\naccess fault.\n\nIt is worth noting that there is a comment above the IsValidMmapRange\nfunction (which gets called during UnmapShadow), that describes this\nexact issue in the context of munmap.\n\nrdar://179172055\n\nAssisted by: Claude\n\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 1d0b606198b502a307da0f688262a1044c38f316\n" }, { "commit": "c45c8e2b2dd6905c88ad01250b3bcaae596b0e6a", "tree": "345595837cb701d850f46854caece3f56fbb8921", "parents": [ "07a7546236d966629cabdaefd7f107e21a87f52f" ], "author": { "name": "Simi Pallipurath", "email": "simi.pallipurath@arm.com", "time": "Tue Jul 21 09:18:01 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Tue Jul 21 09:20:51 2026" }, "message": "[compiler-rt][test] Skip page-size feature detection under emulators. (#210683)\n\nThis avoids adding a page-size-* lit feature when compiler-rt tests are\nconfigured to run through an emulator.\n\nPerviously this used to work because failure to detect the page size\nwould fall back to 4096. Now after this change\nhttps://github.com/llvm/llvm-project/pull/209175, the code tries to\nexecute config.python_executable through the emulator. For\nemulator-based test configurations, this is inappropriate. The function\nnow returns None when an emulator is configured, and the lit feature is\nonly added when a concrete page size is available.\n\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 46fb353e5d48097b0056ec3eddf77d6ce41bf226\n" }, { "commit": "07a7546236d966629cabdaefd7f107e21a87f52f", "tree": "d214f2123ff14b79a06f9321c9e1a345a1a84df9", "parents": [ "57f33802683ef36d46b42ce67e0429d38c21a056" ], "author": { "name": "Chris Apple", "email": "cja-private@pm.me", "time": "Mon Jul 20 20:32:11 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Mon Jul 20 20:35:43 2026" }, "message": "[rtsan] Clean up rtsan_opts default variable (#210803)\n\nFollow on from https://github.com/llvm/llvm-project/pull/210395.\n\nSome systems didn\u0027t like this empty variable when they appended to it\nlater (and the original variable that we had as a dummy was unsupported\nby rtsan)\n\nhttps://github.com/llvm/llvm-project/pull/210395/changes#diff-5f39176422b87f314a683584d1cabe670b4c1e3942d65bb2b5ce4d0623c40560L7\n\nChange this to another inconsequential value so we can append\n`:other_options` to it easily later in the file\n\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: f36452f40e849d8e4947689a3bd8252ad5bbbde5\n" }, { "commit": "57f33802683ef36d46b42ce67e0429d38c21a056", "tree": "6547496eddce08491788fa61c37c94e9d36b10f0", "parents": [ "0a8d3d671299436b58bc0bd33c7b537770fba0a2" ], "author": { "name": "Chris Apple", "email": "cja-private@pm.me", "time": "Mon Jul 20 20:07:07 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Mon Jul 20 20:09:27 2026" }, "message": "[NFC] [rtsan] Cleanup env for lit tests (#210395)\n\nThis matches the style of our other tests, and may make it easier for\nothers to maintain downstream if they override the `env_rtsan_opts`\nvariable with special platform specific options.\n\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: d903936a03eba0eb1f96b96ec0a956de0bfaaf47\n" }, { "commit": "0a8d3d671299436b58bc0bd33c7b537770fba0a2", "tree": "1be018f7a6bdc1afc43c0fde0314e28987e6887c", "parents": [ "73bfec87df77af3efe907b7b7d59b07c5a45f62e" ], "author": { "name": "Christopher Ferris", "email": "cferris1000@users.noreply.github.com", "time": "Mon Jul 20 18:08:53 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Mon Jul 20 18:10:45 2026" }, "message": "[scudo] Remove dead code and other small cleanups. (#210454)\n\nSome dead code from the error info refactor is still around, so delete\nthat. Also cleanup spelling and spacing.\n\nModified the tests to use TestOnly functions so it\u0027s clear what is\nactually used.\n\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: f010423d3dfac774c6157686b6f14fb8d357c715\n" }, { "commit": "73bfec87df77af3efe907b7b7d59b07c5a45f62e", "tree": "51f858a570f684c59370f1522ed5c2810cfd3f8a", "parents": [ "a282d2dd39fbdee7de35732a37f527390b829f57" ], "author": { "name": "Haowei", "email": "haowei@google.com", "time": "Mon Jul 20 17:47:59 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Mon Jul 20 17:50:16 2026" }, "message": "Reapply \"Skipping host target exports during cross-compilation\" (#210… (#210496)\n\nThis reverts commit 42eb7db188f23625872b6f1d722979a3100d170e. This\nreland 3fbb037d03e7ff1ef1cd9156363e895af7caaf98 (#209922).\n\nUnder CMake 4+, calling add_library(... SHARED IMPORTED) on a target\nplatform that lacks dynamic linking support triggers a fatal error. This\nbecomes an issue when building LLVM and runtimes for baremetal targets\nlike armv6m-none-eabi. This patch adds the option\n\"LLVM_OMIT_EXPORTS_FROM_CONFIG\" in LLVM. When used in sub builds like\nLLVM runtimes, it makes CMake to skip including the LLVM and Clang\nexports. This mitigates the CMake 4 errors on baremetal runtimes.\n\n42eb7db188f23625872b6f1d722979a3100d170e originally\napply this flag on all runtimes build and causing issues on runtimes\nlike intel-sycl-gpu, amdgpu-offload-ubuntu-22-cmake-build-only.\nThis PR only set this flag when the target platform lacks shared library\nsupport.\n\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: a04f87fdd9b59c6089bf731c669c6f149208363c\n" }, { "commit": "a282d2dd39fbdee7de35732a37f527390b829f57", "tree": "ad0b149168428c167cef3184b38d733531e58ef8", "parents": [ "0bc2bddce14a53d9184c230b49fa807bf1e6f3f6" ], "author": { "name": "Andrew Haberlandt", "email": "ahaberlandt@apple.com", "time": "Mon Jul 20 17:18:07 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Mon Jul 20 17:20:27 2026" }, "message": "[compiler-rt] Move sanitizer ignorelists to sanitizer-ignorelists target (#200518)\n\nThis adds a `sanitizer-ignorelists` target to compiler-rt, allowing the\nignorelists to be built independently of the sanitizer runtimes. The\nignorelists are still built by default, but this change is not quite\nNFC: when any of the sanitizers are built, all of the ignorelists will\nbe included (rather than the individual runtime ignorelists being built\nwith their respective runtimes).\n\nFor _why_ we want to be able to build these independently of the\nruntimes, see\nhttps://github.com/llvm/llvm-project/pull/195690#issuecomment-4384573223\n\nI did not move the dfsan abilist to this new target because it is not as\nstraightforward as a copying a static file.\n\nAssisted-by: Claude\n\nrdar://175962064\n\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 83ba8addf60668a971453eed207f9bfdaa379f97\n" }, { "commit": "0bc2bddce14a53d9184c230b49fa807bf1e6f3f6", "tree": "9eed80006cf3dc97031a11b712a24bf9340fd154", "parents": [ "3f50fe0309e4f6f8f00440f067b4c296826df0a5" ], "author": { "name": "Haowei", "email": "haowei@google.com", "time": "Sat Jul 18 00:17:48 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Sat Jul 18 00:20:50 2026" }, "message": "Revert \"Skipping host target exports during cross-compilation\" (#210468)\n\nReverts llvm/llvm-project#209922\nIt broke runtimes build on amdgpu-amd-amdhsa\n\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 42eb7db188f23625872b6f1d722979a3100d170e\n" }, { "commit": "3f50fe0309e4f6f8f00440f067b4c296826df0a5", "tree": "d55bfa30b86cf4a60c8b6f6a29cf89d160bf17ef", "parents": [ "133c088462bc66a8294d3a8328ed60716b975ed1" ], "author": { "name": "Haowei", "email": "haowei@google.com", "time": "Fri Jul 17 23:08:42 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Fri Jul 17 23:10:36 2026" }, "message": "Adding an option to Skip LLVM and Clang exports (#209922)\n\nUnder CMake 4+, calling add_library(... SHARED IMPORTED) on a\ntarget platform that lacks dynamic linking support triggers a fatal\nerror. This becomes an issue when building LLVM and runtimes for\nbaremetal targets like armv6m-none-eabi. This patch\nadds the option \"LLVM_OMIT_EXPORTS_FROM_CONFIG\" in LLVM.\nWhen used in sub builds like LLVM runtimes, it makes CMake to skip\nincluding the LLVM and Clang exports. This mitigates the CMake 4\nerrors on baremetal runtimes.\n\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 3fbb037d03e7ff1ef1cd9156363e895af7caaf98\n" }, { "commit": "133c088462bc66a8294d3a8328ed60716b975ed1", "tree": "9eed80006cf3dc97031a11b712a24bf9340fd154", "parents": [ "855c4b80cc60c8252f2af4ca24745de9e43ea3fc" ], "author": { "name": "Reid Kleckner", "email": "rkleckner@nvidia.com", "time": "Fri Jul 17 18:56:59 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Fri Jul 17 19:01:15 2026" }, "message": "[docs] Finish misc MyST migration (#210209)\n\nTracking issue: #201242\nSee the [migration guide] for more information.\n\n[migration guide]:\nhttps://llvm.org/docs/SphinxQuickstartTemplate.html#markdown-migration-guidelines\n\nThis is a stacked PR based on #210208, which is the standalone rename\ncommit for this miscellaneous Markdown migration batch.\n\nThis was prepared with rst2myst plus LLM-assisted cleanup. LLDB man\npages are intentionally left as reStructuredText because LLDB man-page\nbuilding should not take on Markdown parser dependencies.\n\nUnfortunately, many of these miscellaneous docs do not have working\nstandalone doc build targets, so I was not able to render HTML for all\nof them. I think the scope is small enough that this is reviewable by\nhand.\n\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 93b1240309309d5c2f8f94527ff7c7bd5df359af\n" }, { "commit": "855c4b80cc60c8252f2af4ca24745de9e43ea3fc", "tree": "8db979fdc43e8e0499981f471dc1708585778b7f", "parents": [ "9ccc42330d732269a53b3fd59c6186d87a320be2" ], "author": { "name": "Reid Kleckner", "email": "rkleckner@nvidia.com", "time": "Fri Jul 17 17:48:13 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Fri Jul 17 17:50:38 2026" }, "message": "[docs] Rename misc rst docs to md (#210208)\n\nTracking issue: #201242\nSee the [migration guide] for more information.\n\n[migration guide]:\nhttps://llvm.org/docs/SphinxQuickstartTemplate.html#markdown-migration-guidelines\n\nThis is the initial straight rename commit.\n\nMany of these miscellaneous .rst files are not actually part of a real\ndocs build, since docs/conf.py does not exist for many of them.\n\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: c42ce32264f94a8800f4250d7b716cca55f77df6\n" }, { "commit": "9ccc42330d732269a53b3fd59c6186d87a320be2", "tree": "0d94134c6456cd4e16691c3c9940e83eaf17d612", "parents": [ "a38bbdd0d4185b640553dc405bdfbb8e223c91e6" ], "author": { "name": "Aiden Grossman", "email": "aidengrossman@google.com", "time": "Fri Jul 17 16:59:29 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Fri Jul 17 17:00:43 2026" }, "message": "[compiler-rt] Fix tests after #205331 (#210370)\n\nNew optimizations changed some compiler-rt tests. Update the test\nexpectations based on this.\n\nFixes #210355.\n\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: fd55bbd6022ccb8d854f580b56c39177e80c256e\n" }, { "commit": "a38bbdd0d4185b640553dc405bdfbb8e223c91e6", "tree": "f34be9901fc55cc075c75f767d3504680ff1bf10", "parents": [ "85aac854430553d78d236e0b6ba82abb8eed3df4" ], "author": { "name": "Zachary Henkel", "email": "zachary.henkel@gmail.com", "time": "Fri Jul 17 15:47:03 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Fri Jul 17 15:50:37 2026" }, "message": "Replace Python\u0027s deprecated tempfile.mktemp() with secure temp APIs (#210123)\n\nPython\u0027s tempfile.mktemp() is deprecated and insecure: it returns a path\nwithout creating the file, leaving a TOCTOU/symlink window before the\nfile is opened. Replace every use, choosing the temp API by lifetime:\n\n- compiler-rt android_common.py adb(): tempfile.TemporaryFile\n(anonymous, auto-deleted), read back via seek(0); drops manual\nclose()/unlink().\n- compiler-rt android_common.py pull_from_device():\ntempfile.TemporaryDirectory so \"adb pull\" writes into a private,\nauto-cleaned directory.\n- lldb examples delta.py / gdbremote.py start_gdb_log():\ntempfile.NamedTemporaryFile(delete\u003dFalse); the log must outlive the call\nfor a later stop_gdb_log, so it is intentionally not auto-deleted.\n\nReported by internal static analysis and contributed upstream for the\nbenefit of the wider community.\n\nAssisted-by: GitHub Copilot CLI (Claude Opus 4.8)\n\nI do not have commit access and will need someone to submit on my\nbehalf.\n\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: bcb459015e12fb03e4ae3d56393883b2974f96fb\n" }, { "commit": "85aac854430553d78d236e0b6ba82abb8eed3df4", "tree": "5f5c55ef3ebca5020e867348f21105d53d22fe90", "parents": [ "409c99dc9f153bd9964ab127fc70fc425e935407" ], "author": { "name": "maflcko", "email": "6399679+maflcko@users.noreply.github.com", "time": "Fri Jul 17 05:50:28 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Fri Jul 17 05:55:46 2026" }, "message": "[compiler-rt][test] Use configured Python for lit page-size probe (#209175)\n\nUse the same Python executable configured for the test suite, rather\nthan relying on python3 being on PATH.\n\nCo-authored-by: MarcoFalke \u003c*~\u003d`\u0027#}+{/-|\u0026$^_@721217.xyz\u003e\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 30bc042395a6cf0aa1e5de4fcc625d5b8c611767\n" }, { "commit": "409c99dc9f153bd9964ab127fc70fc425e935407", "tree": "a855ff76d3c3e72a17768a1723fab98be5cb2c6c", "parents": [ "42990d2bcf74f95cdf43b5bc4d1caf8faefaee56" ], "author": { "name": "maflcko", "email": "6399679+maflcko@users.noreply.github.com", "time": "Fri Jul 17 05:42:16 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Fri Jul 17 05:44:49 2026" }, "message": "[compiler-rt][ubsan] Fix suppressions for inlined functions (#206735)\n\nAlign suppression PC handling with getCallerLocation(), so function\nsuppressions identify the inlined function that caused a report.\n\nIntentionally match only the innermost inline frame.\n\nExtend suppressions-nested-calls.c coverage to -O1.\n\nFixes https://github.com/llvm/llvm-project/issues/209501\n\n---------\n\nCo-authored-by: MarcoFalke \u003c*~\u003d`\u0027#}+{/-|\u0026$^_@721217.xyz\u003e\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: a376783e2ed3a44f63991c97513711faad09525b\n" }, { "commit": "42990d2bcf74f95cdf43b5bc4d1caf8faefaee56", "tree": "f6821016e146c9e0f57044d9c83ed4699004b994", "parents": [ "3f5d3686912785b3ec05dfb625593422a1f26e7d" ], "author": { "name": "John Paul Jepko", "email": "john.jepko@ericsson.com", "time": "Thu Jul 16 21:01:38 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Thu Jul 16 21:05:58 2026" }, "message": "[Clang][Sema] Reland -Wstringop-overread with computeDependence fix (#208012)\n\nReland #183004 and follow-up #205201 (reverted in #207840) which adds `-Wstringop-overread` to warn when memory functions `memcpy`, `memmove`, `memcmp`, and other related builtins read more bytes than the source buffer size.\n\nThe revert was due to a crash when a memory function is called on `\u0026x` where `x` is a static constexpr local inside a template. The original PR exposed a bug in `computeDependence(UnaryOperator*)` where only the value-dependence bit is set for this case, when the instantiation-dependence bit should also be set. This is the root cause behind the crash.\n\nFix `computeDependence` to properly set both value and instantiation-dependence bits, and add a regression for this crash via `warn-stringop-overread.c`.\n\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: aaa4ebd2ad174920ddce67c4641292a44788e879\n" }, { "commit": "3f5d3686912785b3ec05dfb625593422a1f26e7d", "tree": "253fcef80665313dbceb0cd5b314dac49abb7388", "parents": [ "2745a35b2ab5c648a854e561f8e6e08ddb73435a" ], "author": { "name": "Mauri de Souza Meneguzzo", "email": "mauri870@gmail.com", "time": "Thu Jul 16 20:44:53 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Thu Jul 16 20:50:27 2026" }, "message": "[tsan] Enable __TSAN_HAS_INT128 on s390x and mips64 (#197319)\n\nThe s390x exclusion was added in\nhttps://github.com/llvm/llvm-project/commit/b17673816d7f65e07015489993b22049e36b04db\n(https://reviews.llvm.org/D105629)\nciting lack of hardware 128-bit atomics. s390x does have the CDSG\ninstruction,\nbut `alignof(__int128) \u003d\u003d 8` on the s390x psABI while CDSG requires\n16-byte\nalignment, so Clang does not define\n`__GCC_HAVE_SYNC_COMPARE_AND_SWAP_16` for\nthis target. The mips64 exclusion has been present since\n`__TSAN_HAS_INT128` was\nintroduced in\nhttps://github.com/llvm/llvm-project/commit/06bbca9ec4bbc5a103c94c4973797ea4ddf944b5\n(https://reviews.llvm.org/D18543) with no documented\nrationale. Clang similarly does not define\n`__GCC_HAVE_SYNC_COMPARE_AND_SWAP_16`\nfor mips64.\n\nBoth exclusions are unnecessary. When\n`__GCC_HAVE_SYNC_COMPARE_AND_SWAP_16` is\nnot defined, all a128 operations use the SpinMutex-based emulation path,\nwhich\nserializes 128-bit accesses under a global lock without requiring any\nhardware\nCAS16 support.\n\nVerified with Clang (targets: s390x-linux-gnu, mips64-linux-gnuabi64,\nmips64el-linux-gnuabi64): all three produce the full set of 12\n`__tsan_atomic128_*` symbols with zero `__sync_*_16 / __atomic_*_16`\nlibcall references.\n\nRuntime testing has not been performed as I lack access to these\narchitectures.\n\nRelates https://github.com/llvm/llvm-project/pull/196833\nFor https://github.com/golang/go/issues/61236\n\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: c96a176dd15690ed8181c6764b26d0303008a40f\n" }, { "commit": "2745a35b2ab5c648a854e561f8e6e08ddb73435a", "tree": "e0f3453b538399fb6bb73c4cef12d3442ff57b5b", "parents": [ "a868992a2a5e1be65719d5954e9f7f5aa61833a3" ], "author": { "name": "David CARLIER", "email": "devnexen@gmail.com", "time": "Thu Jul 16 19:24:34 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Thu Jul 16 19:25:41 2026" }, "message": "[compiler-rt][memprof] clear histogram tail granule on allocation. (#208911)\n\nClearShadow rounded the size down, leaving the partial tail counter\n(ceil(size/8)-th) uncleared, so a recycled chunk leaked a stale count\ninto the last histogram bucket. Round up instead.\n\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: c0b8c59b65afd268f286f56ef71f6433ac47dfd1\n" }, { "commit": "a868992a2a5e1be65719d5954e9f7f5aa61833a3", "tree": "524aac22ccdc096fa7233c134176a456187ad2ba", "parents": [ "79d9d7fae549af6b33a69b49953a4c8bc35687c7" ], "author": { "name": "Mariusz Borsa", "email": "wrotki@msn.com", "time": "Thu Jul 16 18:50:31 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Thu Jul 16 18:55:52 2026" }, "message": "[Sanitizers][Darwin] Add atos-canary test (#209338)\n\nWe\u0027ve observed tests using atos symbolizer ocasionally failing for\nseveral test runs on x86_64 machines, then stabilizing again. We were\nunable to reproduce\nthis with local builds, so it\u0027s unclear what the root cause is.\n\nTo help investigate similar failures in the future, we\u0027re adding a test\nexercising atos itself. Should atos start to fail again, this test\nshould fail too, and provide way more info helping debugging than other\ntests do\n\nCo-authored-by: Claude Opus 4.8 (1M context) \u003cnoreply@anthropic.com\u003e\n\n---------\n\nCo-authored-by: Mariusz Borsa \u003cm_borsa@apple.com\u003e\nCo-authored-by: Claude Opus 4.8 (1M context) \u003cnoreply@anthropic.com\u003e\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: e1fa984ae0a87cf6de3b656496ced466b35995c0\n" }, { "commit": "79d9d7fae549af6b33a69b49953a4c8bc35687c7", "tree": "2a411d0bf159a64dde7fe24dd8abdc1e35b2a00b", "parents": [ "e5a75f5d38f4ec7916c424950e90c867934d3a4b" ], "author": { "name": "David CARLIER", "email": "devnexen@gmail.com", "time": "Thu Jul 16 18:49:48 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Thu Jul 16 18:55:42 2026" }, "message": "[sanitizer_common] fix shadowed ret hiding symbolizer read failures. (#209971)\n\nThe inner `bool ret` from ReadFromFile shadowed the outer one that is\nreturned, so a dead/closed symbolizer (0-byte read or read error) was\nreported as success: SendCommand never restarted the crashed process,\nand the addr2line path aborted on a CHECK over an empty buffer.\nRegression from acfeb1a6c244. Add a regression test that feeds an EOF fd\nthrough ReadFromSymbolizer and expects failure.\n\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 6563d2500c82deb27c7dab9bc17334de777f8cbc\n" }, { "commit": "e5a75f5d38f4ec7916c424950e90c867934d3a4b", "tree": "b77f0ca29a36f7328174d7ec535706b9719352a2", "parents": [ "e6c7fa970fbaad0cac22fcfb1f8dac4d8a144900" ], "author": { "name": "Mauri de Souza Meneguzzo", "email": "mauri870@gmail.com", "time": "Thu Jul 16 18:24:52 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Thu Jul 16 18:30:48 2026" }, "message": "[tsan][go]: add a128 store, load, compare_exchange (#196833)\n\nThis commit adds a128 load, store and compare_exchange to the Go\nsanitizer functions. This is needed for a new Go atomic type that\noperate on 16-byte values.\n\nThe SANITIZER_GO macro was gating a128 and is now lifted. Based on the\n__TSAN_HAS_INT128 macro definition, the only architecture excluded by it\nthat intersects with [Go race support\n](https://go.dev/doc/articles/race_detector#Requirements) is s390x.\n\nFor https://github.com/golang/go/issues/61236\n\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 5d99b0a9a79dd72ae45b88a7815edaf600808698\n" }, { "commit": "e6c7fa970fbaad0cac22fcfb1f8dac4d8a144900", "tree": "40b707f9c413bbfd08683a38c3e4c648e74efa01", "parents": [ "c3d8ee219a56b36348ba6fb4ca25778cc0865d3c" ], "author": { "name": "Fei Peng", "email": "pengfei.02@bytedance.com", "time": "Thu Jul 16 01:02:13 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Thu Jul 16 01:06:01 2026" }, "message": "[tsan] Add 47-bit VMA mapping for linux/aarch64 (#205949)\n\nAdd 47-bit VMA mapping to support TSan on the latest Android\nemulator(16kb page size and 47 va bits).\n\nFixes https://github.com/llvm/llvm-project/issues/151931\n\nAndroid kernel config:\nhttps://android.googlesource.com/kernel/build/+/refs/heads/main-kernel-2026/kleaf/impl/defconfig/arm64_16k_defconfig#6\n\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: f020acfd395d9094303e78c807bc168a46a783fa\n" }, { "commit": "c3d8ee219a56b36348ba6fb4ca25778cc0865d3c", "tree": "0134cdcede1edc46dad6a2671e82e247bc8d51ce", "parents": [ "9dff91ff24b1f04d285c8fd267a02d3d7e5b4a5f" ], "author": { "name": "Wenju He", "email": "wenju.he@intel.com", "time": "Wed Jul 15 19:57:32 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Wed Jul 15 20:00:05 2026" }, "message": "[compiler-rt] Fix asan_interceptors.cpp build warning -Wunused-template (#209750)\n\nmmap_interceptor/munmap_interceptor template functions are instantiated\nin sanitizer_common_interceptors.inc only under `#if\nSANITIZER_INTERCEPT_MMAP`. Fix `unused function template\n\u0027mmap_interceptor\u0027` warning on Windows, which is error under -Werror.\n\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 835ae525713e27aa98c46469792cee980669c6ca\n" }, { "commit": "9dff91ff24b1f04d285c8fd267a02d3d7e5b4a5f", "tree": "ca611459be8fdf22dd8f6037986357f26a487db9", "parents": [ "b2201a2cd8a614e91b4a0add2c1c9d021a2d9a6d" ], "author": { "name": "Patrik Dokoupil", "email": "61749375+pdokoupil@users.noreply.github.com", "time": "Wed Jul 15 19:54:08 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Wed Jul 15 19:59:53 2026" }, "message": "[compiler-rt][sanitizer_common] Fix setgroups syscall pre-hook to check its input (#209207)\n\n# About\n\n`PRE_SYSCALL(setgroups)`/`setgroups16` called `POST_WRITE` on their\n`grouplist` argument, but grouplist is a read-only input the kernel\nnever writes. Under MSan, `POST_WRITE` maps to `__msan_unpoison`, so\nannotating a raw setgroups(2) via the documented\n`__sanitizer_syscall_pre_setgroups` hook silently marked an\nuninitialized group list as initialized -- before the syscall even ran,\nand even on failure -- masking a real use-of-uninitialized-value bug.\nUnder TSan it recorded a write range where a read belongs.\n\nUse PRE_READ, matching every other input-only argument in this file\n(e.g. sched_setparam). Add an MSan regression test.\n\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 23ead65310da314ae04a1c30137f279beaac1a4f\n" }, { "commit": "b2201a2cd8a614e91b4a0add2c1c9d021a2d9a6d", "tree": "891bce51844ebf005c466a4ded85b185d479b722", "parents": [ "54fab4595a422cab70a6a639c335d32be8921649" ], "author": { "name": "Vy Nguyen", "email": "1762963+oontvoo@users.noreply.github.com", "time": "Wed Jul 15 19:53:15 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Wed Jul 15 19:54:50 2026" }, "message": "[codegen]Ensure __builtin_trap() has an unreachable (#197789)\n\n```\n void test2() {\n __builtin_trap();\n}\n```\n\nPreviously, this would generate this:\n\n```\n; Function Attrs: mustprogress noinline nounwind optnone uwtable\ndefine dso_local void @test2()() #2 !dbg !14 {\n call void @llvm.trap(), !dbg !15\n ret void, !dbg !16\n}\n```\n\nNow we\u0027ll have `unreachable` after the trap\n\n```\n; Function Attrs: noinline nounwind optnone uwtable\ndefine dso_local void @test2() #0 {\n call void @llvm.trap()\n unreachable\n}\n\n```\n\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 0f494fc6576600f79bbcdb5314fbed6c13da7395\n" }, { "commit": "54fab4595a422cab70a6a639c335d32be8921649", "tree": "fbd498562766c684a28cc9ca495529c1a5c8b8f1", "parents": [ "a07c611ffd73474b1c7d891997d15ba986b8396d" ], "author": { "name": "Kyungtak Woo", "email": "kevinwkt@google.com", "time": "Wed Jul 15 18:57:28 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Wed Jul 15 19:02:38 2026" }, "message": "[compiler-rt][tsan] Fix -Wunused-template in tsan_interface_atomic.cpp (NFC) (#209621)\n\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: b74c800578a06bba7b3bacc690cbe493759836a2\n" }, { "commit": "a07c611ffd73474b1c7d891997d15ba986b8396d", "tree": "483c6ec4b63e1c0956282b3b4727f1727c0be303", "parents": [ "3007de23975cbe786b95dadf88c83a679ac5cb72" ], "author": { "name": "Joseph Huber", "email": "huberjn@outlook.com", "time": "Wed Jul 15 15:24:06 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Wed Jul 15 15:26:20 2026" }, "message": "[compiler-rt] Fix amdgcn -\u003e amdgpu conversion for compiler-rt (#209765)\n\nSummary:\nWe are apparently transitioning to using `amdgpu` everywhere, so these\nneed to be updated.\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 2cd090afbe1eddfe8f781d5ccae391e1f69e8ee4\n" }, { "commit": "3007de23975cbe786b95dadf88c83a679ac5cb72", "tree": "8f551a7d839a03aa78aa6c8833bd38f6eba0328c", "parents": [ "d662304d76d33ddab220b1bbfaa8efd664486463" ], "author": { "name": "Aiden Grossman", "email": "aidengrossman@google.com", "time": "Wed Jul 15 03:30:59 2026" }, "committer": { "name": "Copybara-Service", "email": "copybara-worker@google.com", "time": "Wed Jul 15 03:34:16 2026" }, "message": "[compiler-rt] Drop support for using the external shell\n\nThe lit external shell is going to be removed soon, so drop support for\nusing it to run the compiler-rt tests.\n\nReviewers: petrhosek, ilovepi, thurstond\n\nPull Request: https://github.com/llvm/llvm-project/pull/209626\n\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 00112b0ac171c360ae60d2f7ffabb7f22d62cdd0\n" } ], "next": "d662304d76d33ddab220b1bbfaa8efd664486463" }