)]}'
{
  "log": [
    {
      "commit": "7f092bb5fcc37296f16ed5dae464386ac2915c0c",
      "tree": "48efee1b69264eb82d285285a2117a320ff0774f",
      "parents": [
        "d41ea55c22574f74af7a1acbf42ff84ea318851a"
      ],
      "author": {
        "name": "Jan Svoboda",
        "email": "jan_svoboda@apple.com",
        "time": "Fri Jun 05 08:13:57 2026 -0700"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Fri Jun 05 08:15:59 2026 -0700"
      },
      "message": "[clang] Fix VFS creation crash with missing `DiagnosticConsumer` (#201397)\n\nFor convenience, the `CompilerInstance::createVirtualFileSystem()` API\nallows omitting the diagnostic consumer for clients that don\u0027t care\nabout missing overlay files and other VFS creation errors. However, even\nin that case, the temporary `DiagnosticsEngine` created internally\nwithin the function does need a consumer. This PR sets it up.\n\nrdar://176754115\n\nGitOrigin-RevId: 3dad31f753a9950ac1841dea1a9b03f32e8b6997\n"
    },
    {
      "commit": "d41ea55c22574f74af7a1acbf42ff84ea318851a",
      "tree": "9f6a64e988da0d57f768bfa9fd8b387a9f841a31",
      "parents": [
        "558923e2a7231ce80d0fa48ac87c74596d531886"
      ],
      "author": {
        "name": "Nick Sarnie",
        "email": "nick.sarnie@intel.com",
        "time": "Sat Jun 06 00:04:18 2026 +0900"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Fri Jun 05 08:06:20 2026 -0700"
      },
      "message": "[IR][CodeGen][mlir] Overload pointer argument for compressstore/expandload intrinsics (#200492)\n\nUpdate the `llvm.masked.expandload` and `llvm.masked.compressstore`\nintrinsics to have overloaded pointer arguments so other address spaces\ncan be used. This puts these intrinsics in line with other similar\nintrinsics, which already have overloaded pointer arguments.\n\nThe change may look large, but almost all of the changes are adding\n`.p0` to tests.\n\nThis is needed for targets that have non-zero default address space, the\nproblem was found when trying to compile the libc LIT tests for SPIRV,\nspecifically\n[this](https://github.com/llvm/llvm-project/blob/4ac26f45fa8f5c58a90effb903808cb0e908cf1c/libc/test/src/__support/CPP/simd_test.cpp)\none.\n\nCo-Authored-By: Claude Sonnet 4.5 \u003cnoreply@anthropic.com\u003e\n\n---------\n\nSigned-off-by: Nick Sarnie \u003cnick.sarnie@intel.com\u003e\nCo-authored-by: Claude Sonnet 4.5 \u003cnoreply@anthropic.com\u003e\nGitOrigin-RevId: 36ba42d09737872243f1968067fa835fbaecbe73\n"
    },
    {
      "commit": "558923e2a7231ce80d0fa48ac87c74596d531886",
      "tree": "4662f69baf7c6a15aed122bb4a4f0c652594bc55",
      "parents": [
        "09864a9f3f4b8fe88ead03d1a643c83c755e215b"
      ],
      "author": {
        "name": "Matt Arsenault",
        "email": "Matthew.Arsenault@amd.com",
        "time": "Fri Jun 05 15:31:30 2026 +0200"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Fri Jun 05 06:36:05 2026 -0700"
      },
      "message": "OpenMP: Accept amdgpu name in arch directive (#201789)\n\nAccept amdgpu as an alias for amdgcn as part of the general\ntrend of preferring the amdgpu name. This is so the name is\nconsistent in the future when the triple arch name changes.\n\nGitOrigin-RevId: 85da6a915345ce0a8614f8ff4531ba4efb592564\n"
    },
    {
      "commit": "09864a9f3f4b8fe88ead03d1a643c83c755e215b",
      "tree": "1eb137df23438b51e040c665ffabb1b3bdeba2a8",
      "parents": [
        "3593d16f9263b55bcfad8f3149e0ff56d8477d11"
      ],
      "author": {
        "name": "Timm Baeder",
        "email": "tbaeder@redhat.com",
        "time": "Fri Jun 05 15:06:01 2026 +0200"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Fri Jun 05 06:12:41 2026 -0700"
      },
      "message": "[clang][bytecode] Fix conditional operator scoping (#201777)\n\nThe code and the comment are correct, but they also applied to the\nconditional operator condition, which they shouldn\u0027t.\n\nGitOrigin-RevId: 07bd30557f2b00225d4eb7c125e74188b54f1faa\n"
    },
    {
      "commit": "3593d16f9263b55bcfad8f3149e0ff56d8477d11",
      "tree": "05ab660ca350d7034ff743763ea1659cacbfa781",
      "parents": [
        "7707540ed072845f074e285bc97eb2c065bdff91"
      ],
      "author": {
        "name": "lijinpei-amd",
        "email": "jinpli@amd.com",
        "time": "Fri Jun 05 20:50:40 2026 +0800"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Fri Jun 05 05:56:13 2026 -0700"
      },
      "message": "[Sema] Fix assertion crash on section conflict with a non-identifier decl (#200873)\n\nNamedDecl::getName() asserts the name is a simple identifier, so\nUnifySection crashed when the conflicting section decl had a special\nname such as a lambda\u0027s call operator.\n\nDrop the argument: `note_declared_at` has no format placeholder, so it\nwas dead code. The error already prints the section decl.\n\nFixes https://github.com/llvm/llvm-project/issues/192264\n\nGitOrigin-RevId: ddfa5eb8ae2e0cdd1c0542fff9c6671b8b3ff85c\n"
    },
    {
      "commit": "7707540ed072845f074e285bc97eb2c065bdff91",
      "tree": "93e4e2d87b9f24523360e6c39f4cbb9793c3a56a",
      "parents": [
        "b6d8e2e252df6c01b88a206612d3332db7189af1"
      ],
      "author": {
        "name": "Nikita Popov",
        "email": "npopov@redhat.com",
        "time": "Fri Jun 05 12:36:55 2026 +0200"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Fri Jun 05 03:40:49 2026 -0700"
      },
      "message": "[FunctionAttrs] Add support for nofree argument inference (#201591)\n\nUnlike Attributor, FunctionAttrs currently only supports inferring\nnofree on functions. This PR adds support for inferring it on arguments\nas well.\n\nI\u0027ve chosen to integrate this with the inference for\nreadnone/readonly/writeonly attributes, as the core logic for these is\nthe same.\n\nThere is a lot of test fallout. The only significant changes are in\nllvm/lib/Transforms/IPO/FunctionAttrs.cpp and\nllvm/test/Transforms/FunctionAttrs/nofree.ll.\n\nGitOrigin-RevId: 89905ff21441d8ed04b2bfafc57a8bb7c9402800\n"
    },
    {
      "commit": "b6d8e2e252df6c01b88a206612d3332db7189af1",
      "tree": "1db69542cf00b7fb52b59fea318b9852eaf2ce63",
      "parents": [
        "72def412ca9d316cb55c541b62ee55cc0339f7ea"
      ],
      "author": {
        "name": "Ryosuke Niwa",
        "email": "rniwa@webkit.org",
        "time": "Fri Jun 05 02:05:51 2026 -0700"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Fri Jun 05 02:23:01 2026 -0700"
      },
      "message": "[alpha.webkit.NoDeleteChecker] Allow no-delete default constructors (#201544)\n\nThis PR fixes the bug in TrivialFunctionAnalysis that it treats a\ndefault constructor without an explicit body / definition as not\n\"trivial\". Fixed the bug by allowing the function body to be missing\nwhen isThisDeclarationADefinition is true.\n\n---------\n\nCo-authored-by: Balazs Benics \u003cbenicsbalazs@gmail.com\u003e\nGitOrigin-RevId: 7d987f8a35a71a97754ccab804cb9966ece52e08\n"
    },
    {
      "commit": "72def412ca9d316cb55c541b62ee55cc0339f7ea",
      "tree": "7799e83dcae90d0b004acddfd173a2871c105023",
      "parents": [
        "c900389f3b613e5ffa8e402e76f25acb55e926c7"
      ],
      "author": {
        "name": "Kiva",
        "email": "zengtao@iscas.ac.cn",
        "time": "Fri Jun 05 16:10:34 2026 +0800"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Fri Jun 05 01:16:04 2026 -0700"
      },
      "message": "[RISCV][MC] Add experimental `Zvvmtls` and `Zvvmttls` support (#198229)\n\nThis patch adds experimental MC-layer support for the [RISC-V Integrated\nMatrix\nExtension](https://github.com/riscv/integrated-matrix-extension/releases/tag/riscv-isa-release-71c48b9-2026-05-17),\nspecifically the tile load/store extensions: `Zvvmtls` and `Zvvmttls`\n\nThis PR:\n\n- Adds the optional tile lambda operand syntax (`L1` through `L64`), and\nrelated asm operand.\n- Adds the `vmtl.v`, `vmts.v`, `vmttl.v` and `vmtts.v` instructions to\nthe MC\n- Modifies `parseMaskReg` to return `NoMatch` to allow overloaded\nmnemonics to continue matching alternative optional operands, such as\nparsing `vmtl.v v8, (a0), a1, L4` as the tile-lambda form instead of\nfailing by treating `L4` as a malformed mask operand. Real mask\nregisters missing .t, such as v0, still produce the existing diagnostic.\n\nGitOrigin-RevId: 0f18088dca8aa03124cac4d174e97174a9aa35be\n"
    },
    {
      "commit": "c900389f3b613e5ffa8e402e76f25acb55e926c7",
      "tree": "5ec74ee765bcb3ab40817ac6d5393c7518aff664",
      "parents": [
        "e313407b201d16a928793e4fa44dcd83ee5e650c"
      ],
      "author": {
        "name": "Tony Varghese",
        "email": "tonypalampalliyil@gmail.com",
        "time": "Fri Jun 05 12:14:59 2026 +0530"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Thu Jun 04 23:50:49 2026 -0700"
      },
      "message": "[NFC][clang] Format pragma comment kind StringSwitch (#201754)\n\nThis is a formatting-only change to the pragma comment kind\n`StringSwitch`.\n\nA follow-up change needs to add another pragma comment kind to this\nblock. Applying the formatter as part of that functional change would\nalso reformat the existing cases, making the functional diff harder to\nreview.\n\nNo functional change intended.\n\nCo-authored-by: Tony Varghese \u003ctony.varghese@ibm.com\u003e\nGitOrigin-RevId: 511d2e40ddeacf25f403b40ed73a41d1dea1b636\n"
    },
    {
      "commit": "e313407b201d16a928793e4fa44dcd83ee5e650c",
      "tree": "28baa62384206c8e829a2a1d2411996335825fbe",
      "parents": [
        "4003382745cd6287df6f1c992b0b98f605571ff9"
      ],
      "author": {
        "name": "Jiahao Guo",
        "email": "eoonguo@gmail.com",
        "time": "Fri Jun 05 12:32:13 2026 +0800"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Thu Jun 04 21:36:29 2026 -0700"
      },
      "message": "[CIR][AArch64] Lower NEON vslid intrinsics (#199415)\n\n### summary\n\npart of: https://github.com/llvm/llvm-project/issues/185382\n\nthis is a follow up of :\nhttps://github.com/llvm/llvm-project/pull/198309\n\nThis adds CIR lowering for the scalar AArch64 NEON shift-left-and-insert\nintrinsics `vslid_n_s64` and `vslid_n_u64`.\n\nThese ACLE builtins expose scalar `i64` interfaces, but the\ncorresponding LLVM intrinsic is vector-shaped:\n\n  ```llvm\n  llvm.aarch64.neon.vsli.v1i64(\u003c1 x i64\u003e, \u003c1 x i64\u003e, i32) -\u003e \u003c1 x i64\u003e\n  ```\n\nThe new lowering wraps the scalar 64-bit operands in one-lane CIR vector\ntypes, emits the aarch64.neon.vsli LLVM\nintrinsic call, and bitcasts the result back to the scalar return type.\n\nGitOrigin-RevId: dbe91f4a4b55a17197c52335db4b7a0aa9c69f6b\n"
    },
    {
      "commit": "4003382745cd6287df6f1c992b0b98f605571ff9",
      "tree": "4d942c305e8b146ee98e42e8197d3340eca64ec0",
      "parents": [
        "d01530622db45329f48d26c7cc51ada220c3e11d"
      ],
      "author": {
        "name": "adeshcom14",
        "email": "aadikane@amd.com",
        "time": "Fri Jun 05 09:45:00 2026 +0530"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Thu Jun 04 21:21:03 2026 -0700"
      },
      "message": "[AMDGPU] Track tensor load/store DMAs with asyncmark (#200775)\n\nWire existing variants of the tensor load-to/store-from LDS intrinsics\ninto the existing asyncmark/wait.asyncmark mechanism via TENSOR_CNT\nwaitcnt counter.\n\nFixes: LCOMPILER-1619\nGitOrigin-RevId: 6f233ceb01347d37bd943dde84e81710ac24bfdf\n"
    },
    {
      "commit": "d01530622db45329f48d26c7cc51ada220c3e11d",
      "tree": "18f8823a9520c8e1f432cb095695099e4b1f0e44",
      "parents": [
        "bb86921df2357355780a155eb3c3c6dd8ec8b544"
      ],
      "author": {
        "name": "Timm Baeder",
        "email": "tbaeder@redhat.com",
        "time": "Fri Jun 05 05:24:51 2026 +0200"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Thu Jun 04 20:30:50 2026 -0700"
      },
      "message": "[clang][bytecode] Get the right definition before compiling functions (#201105)\n\nThis broke libc++\u0027s\nstd/ranges/range.adaptors/range.concat/iterator/arithmetic.pass.cpp.\n\nThe (reduced via cvise but not enough) function looks like this:\n\n```c++\n  friend constexpr unsigned\n  operator-(const __iterator \u0026__x, const __iterator \u0026__y)\n    {\n      (void)-(__y - __x);\n      return 0;\n    }\n```\n\nWhen evaluating the binary operator for overflow, we will compile the\noperator- (_this_ function) to bytecode. At that point,\n::isThisDeclarationADefiniton() will return true and ::getDefiniton()\nreturns the function itself. However, all this is happening while the\nfunction is being instantiated, which means the function doesn\u0027t have a\nbody yet and the bytecode ends up being just a NoRet op. This will of\ncourse later fail.\n\nFix this by querying the body before trying to compile a function.\n\nUnfortunately I wasn\u0027t able to create a reproducer of reasonable size.\n\nGitOrigin-RevId: 2cc6b14b11010b8e537a5dfe627159ce96f5a690\n"
    },
    {
      "commit": "bb86921df2357355780a155eb3c3c6dd8ec8b544",
      "tree": "4449833c979af046a6fc5414ec015087db2bc3c7",
      "parents": [
        "2ba3d1e21419f1f67391a6cf2b1fb0dd85257cf7"
      ],
      "author": {
        "name": "Akira Hatanaka",
        "email": "ahatanak@gmail.com",
        "time": "Thu Jun 04 19:05:13 2026 -0700"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Thu Jun 04 19:10:53 2026 -0700"
      },
      "message": "[ExprConstant] Treat `\u0026*p` as not a dereference in C constant initializers (#201483)\n\nIn C, [C11 6.5.3.2p3] specifies that when the operand of unary `\u0026` is\nthe result of a unary `*` operator, neither operator is evaluated and\nthe result is as if both were omitted. So `\u0026*p` yields the pointer value\n`p` without performing a dereference, and forming it is well-defined\neven when `p` is null (e.g. `\u0026*(int *)0`).\n\nThe constant evaluator did not honor this: it evaluated the `*` as a\nreal lvalue access and diagnosed a null dereference as undefined\nbehavior. This went unnoticed for ordinary scalar initializers, which\nuse the relaxed `Expr::isConstantInitializer()` check, but a bit-field\ninitializer is evaluated via `EvaluateAsInt()` with `SE_NoSideEffects`,\nso the same expression was rejected there with \"initializer element is\nnot a compile-time constant\":\n\n```\n  struct S { long v : 8; };\n  const struct S s \u003d { .v \u003d (long)\u0026*(int *)0 };   // error\n  const long x     \u003d (long)\u0026*(int *)0;            // accepted\n```\n\nHandle `\u0026*p` in `PointerExprEvaluator::VisitUnaryAddrOf` (and the\ncorresponding `UO_AddrOf` case in the bytecode interpreter) by\nevaluating the pointer operand directly in C.\n\nFixes #197846.\n\nrdar://158774335\n\nGitOrigin-RevId: 4417a7501aae8a9b8d3ffee1d491540a0858cc29\n"
    },
    {
      "commit": "2ba3d1e21419f1f67391a6cf2b1fb0dd85257cf7",
      "tree": "4edf4259912cfac954a87eb3594237428b928e61",
      "parents": [
        "8adf81aacdb89177e719f96d66aee62ef2d711d7"
      ],
      "author": {
        "name": "Rashmi Mudduluru",
        "email": "r_mudduluru@apple.com",
        "time": "Thu Jun 04 16:14:46 2026 -0700"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Thu Jun 04 16:20:58 2026 -0700"
      },
      "message": "[Webkit Checkers][SaferCpp] Detect base-to-derived downcasts laundered through void* in MemoryUnsafeCastChecker (#200294)\n\nAdds a matcher for static_cast\u003cDerived*\u003e(static_cast\u003cvoid*\u003e(base)),\nwhich previously evaded detection because the outer cast\u0027s immediate\nsource expression is void*, not Base*.\n\nrdar://173770143\n\n---------\n\nCo-authored-by: Balázs Benics \u003cbenicsbalazs@gmail.com\u003e\nGitOrigin-RevId: 60bd8f8088eba2ede01b9d112fc6b927e9e99795\n"
    },
    {
      "commit": "8adf81aacdb89177e719f96d66aee62ef2d711d7",
      "tree": "29fc750554c49ecedf48ba64e159b0929d83b435",
      "parents": [
        "41d192d8b384480c170fc37de98625c42e36cb9d"
      ],
      "author": {
        "name": "Alex Crichton",
        "email": "alex@alexcrichton.com",
        "time": "Thu Jun 04 18:07:20 2026 -0500"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Thu Jun 04 16:11:56 2026 -0700"
      },
      "message": "[WebAssembly] Fix crash combining complex numbers and multivalue (#200514)\n\nThis fixes a crash in Clang when the `experimental-mv` ABI is used on\nWebAssembly targets in conjunction with complex numbers as arguments.\nThere\u0027s no strict definition for what the multivalue ABI is at this\ntime, so the main goal is to just not crash for now.\n\nCloses #70402\nCloses #153567\n\nGitOrigin-RevId: bb180dd616c61932a0000ff45cd5d69fc1195b87\n"
    },
    {
      "commit": "41d192d8b384480c170fc37de98625c42e36cb9d",
      "tree": "cb6d4b6aca69062f987b543aec7ccf03ba648821",
      "parents": [
        "e8dca4f3e100e4d9ffbcb03e4617e51ad47eca7e"
      ],
      "author": {
        "name": "Sam Elliott",
        "email": "aelliott@qti.qualcomm.com",
        "time": "Thu Jun 04 15:34:21 2026 -0700"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Thu Jun 04 15:35:32 2026 -0700"
      },
      "message": "[RISCV] Clang flags for controlling zilsd alignment (#181439)\n\nCalled `-mzilsd-word-align` and `-mzilsd-strict-align`. These interact\nwith scalar/strict alignment, in hopefully a reasonable way.\n\nThey cause errors on rv64, where zilsd is not available.\n\nGitOrigin-RevId: becc74f032080821ddceb04106ef7f30223b4913\n"
    },
    {
      "commit": "e8dca4f3e100e4d9ffbcb03e4617e51ad47eca7e",
      "tree": "499f2d70633191861ff4322828b87c0829b7e064",
      "parents": [
        "e500befdf2df67d13a90d4d88643adc0f16ca3d9"
      ],
      "author": {
        "name": "Farzon Lotfi",
        "email": "farzonlotfi@microsoft.com",
        "time": "Thu Jun 04 18:06:23 2026 -0400"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Thu Jun 04 15:11:24 2026 -0700"
      },
      "message": "[HLSL][CBuffer][Matrix] Honor row_major/column_major keyword in cbuffer layout (#201671)\n\nfixes #201668\n\nA per-declaration `row_major`/`column_major` keyword on a cbuffer matrix\nwas being dropped when building the cbuffer layout, so the layout struct\nand the buffer-layout copy fell back to the translation-unit\n`-fmatrix-memory-layout\u003d`\n\nNeeded to fix the desugar in two places:\n* HLSLBufferLayoutBuilder::layOutMatrix took a `const ConstantMatrixType\n*` and called ConvertTypeForMem(QualType(MT, 0)), discarding the sugar.\nIt now takes the sugared QualType.\n* SemaHLSL\u0027s host-layout struct construction called\ngetUnqualifiedDesugaredType() on each field, erasing the orientation\nattribute. A getHostLayoutFieldType() helper now keeps the sugared type\nfor constant matrices while desugaring everything else.\n\nGitOrigin-RevId: 72c2f7e0552ed468bbfb06f651b221f5c7b4b0bf\n"
    },
    {
      "commit": "e500befdf2df67d13a90d4d88643adc0f16ca3d9",
      "tree": "5b89e7f8148b29c7757f862a0028fbd7dafa1ec1",
      "parents": [
        "eb0e04a424718fc96662b9d163e55e9547710de0"
      ],
      "author": {
        "name": "Jan Svoboda",
        "email": "jan_svoboda@apple.com",
        "time": "Thu Jun 04 14:37:52 2026 -0700"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Thu Jun 04 14:40:44 2026 -0700"
      },
      "message": "Revert \"[clang][lex] Store `HeaderFileInfo` in a `DenseMap`\" (#201702)\n\nReverts llvm/llvm-project#200968\n\nThis is causing some non-determinism in PCM files in the\n`clang/test/Modules/rebuild.m` test.\n\nGitOrigin-RevId: adaa9e08b73829701b0b7ebc86d35be2602d3f2a\n"
    },
    {
      "commit": "eb0e04a424718fc96662b9d163e55e9547710de0",
      "tree": "c9e10db056418da562da743e5921e2469675c3ce",
      "parents": [
        "c8522d6aac8d3dafc9aa914dd92eb5c118ba617e"
      ],
      "author": {
        "name": "dpalermo",
        "email": "dan.palermo@amd.com",
        "time": "Thu Jun 04 16:21:25 2026 -0500"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Thu Jun 04 14:25:50 2026 -0700"
      },
      "message": "Revert \"[OpenMP] Use ext linkage for kernels handles and globals handles keep…\" (#201698)\n\nReverts llvm/llvm-project#200964\n\nThis patch breaks flang declare target on a common block\n\nGitOrigin-RevId: aa513cbc24a4dcfd3f98bff0e96fa85748080e75\n"
    },
    {
      "commit": "c8522d6aac8d3dafc9aa914dd92eb5c118ba617e",
      "tree": "5a1581e0daaad1f3115ef4642426492156afce2a",
      "parents": [
        "d12431dceab603dbbcb39562ab6f49b10c3552d0"
      ],
      "author": {
        "name": "Srividya Sundaram",
        "email": "srividya.sundaram@intel.com",
        "time": "Thu Jun 04 13:25:19 2026 -0700"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Thu Jun 04 13:31:10 2026 -0700"
      },
      "message": "[clang][doc][SYCL] Expand SYCL driver release notes (#200449)\n\n## Summary\n\nThe SYCL Support section in the release notes previously had a single\nentry for the default SPIR-V target change. This PR adds driver-level\nentries covering all notable SYCL changes landed upstream:\n\n- `-fsycl`, `-fsycl-device-only`, `-fsycl-host-only` flag introduction\n(#GH117268)\n- C++17 default enforcement and sub-C++17 rejection (#GH194014)\n- Runtime library rename from `libsycl.so` to `libLLVMSYCL.so` and\nautomatic runtime path / header include injection (#GH188770, #GH174877)\n- Windows SYCL runtime library linking support with CRT management\n(#GH194744)\n- Fix for `-nolibsycl` being silently ignored and spurious spirv-link\n  flags forwarded to `clang-sycl-linker` (#GH200252)\n\nAddresses the review comment on #200200 requesting that notable SYCL\ndriver and offload tool changes be documented alongside the front-end\nchanges. This PR covers only the Clang Driver changes.\n\n---------\n\nCo-authored-by: Nick Sarnie \u003cnick.sarnie@intel.com\u003e\nGitOrigin-RevId: 6e310bf37a199297a86f6338d94af14220bed969\n"
    },
    {
      "commit": "d12431dceab603dbbcb39562ab6f49b10c3552d0",
      "tree": "2d877d544a581b4e78b81595a826f264ad72ceb4",
      "parents": [
        "d2eaedf1406d2192c6e048963b582aee392ada06"
      ],
      "author": {
        "name": "Matheus Izvekov",
        "email": "mizvekov@gmail.com",
        "time": "Thu Jun 04 16:59:10 2026 -0300"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Thu Jun 04 13:01:27 2026 -0700"
      },
      "message": "[clang] fix assertion for DeducedAsPack DeducedTemplateSpecializationType (#201664)\n\nThis fixes a regression introduced in #186727, which was never released,\nso there are no release notes.\n\nFixes #200418\n\nGitOrigin-RevId: 889496c13bd4d3114e375bd0aeb1f161eff67c33\n"
    },
    {
      "commit": "d2eaedf1406d2192c6e048963b582aee392ada06",
      "tree": "783a0715370f792919297f5259c2c89d5a5a9701",
      "parents": [
        "1e494bb78bcbfab159b94aab542b2bb5ea0e4e75"
      ],
      "author": {
        "name": "Matheus Izvekov",
        "email": "mizvekov@gmail.com",
        "time": "Thu Jun 04 15:29:45 2026 -0300"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Thu Jun 04 11:32:38 2026 -0700"
      },
      "message": "[clang] Reland: fix getTemplateInstantiationArgs (#201373)\n\nRelands #199528\n\nThis implements a new strategy for collecting the template arguments, by\nrelying on the qualifiers and template parameter lists to navigate the\ntemplate\ncontext of out-of-line definitions.\n\nThis greatly simplifies the signature of that function, by removing a\nbunch\nof workarounds, and simpliffying a couple that weren\u0027t removed yet.\n\nSince this now relies on qualifiers and template parameter lists,\nthis patch expends most of its effort making sure these are placed,\ntransformed and propagated to template instantiations.\n\nAlso makes the explicit specialization AST nodes stop abusing the\ntemplate\nparameter lists by storing it\u0027s own template parameter list, creating a\ndedicated field for them, similar to partial specializations.\n\nGitOrigin-RevId: bcdb732df219fc5de7c16b9712d52aa56b22fed0\n"
    },
    {
      "commit": "1e494bb78bcbfab159b94aab542b2bb5ea0e4e75",
      "tree": "e0ad55daba7a9f63d1358a20e32b1a4af258fef7",
      "parents": [
        "d3947430d5f2725b7872c8dabf38368e980cc138"
      ],
      "author": {
        "name": "Arthur Eubanks",
        "email": "aeubanks@google.com",
        "time": "Thu Jun 04 10:54:45 2026 -0700"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Thu Jun 04 10:57:22 2026 -0700"
      },
      "message": "[clang] Add -fcrash-diagnostics-tar for tarball of crash reproducer files (#201643)\n\nMakes it easier to move around crash diagnostics.\n\nReland of #198838 with crash-diagnostics-tar.c and\ncrash-report-crashfile.m fixed.\n\nGitOrigin-RevId: 990811766057d912deab1a4aca9d61c73f136b2e\n"
    },
    {
      "commit": "d3947430d5f2725b7872c8dabf38368e980cc138",
      "tree": "0780838ca8b7f0215ffe219f84c906a279f9c07d",
      "parents": [
        "eca68ed17f94611539667011463dae833c58e3bc"
      ],
      "author": {
        "name": "Jan Svoboda",
        "email": "jan_svoboda@apple.com",
        "time": "Thu Jun 04 10:49:49 2026 -0700"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Thu Jun 04 10:51:46 2026 -0700"
      },
      "message": "[clang][lex] Store `HeaderFileInfo` in a `DenseMap` (#200968)\n\nCalling `FileManager::GetUniqueIDMapping()` during modular builds gets\nvery expensive if the `FileManager` has seen lots of files. This\nfunction is used in two places in the `ASTWriter` to look up\n`HeaderFileInfo` in `HeaderSearch`.\n\nThis PR changes the storage of `HeaderFileInfo` from\n`FileEntry::getUID()`-indexed `std::vector\u003cT\u003e` to\n`llvm::DenseMap\u003cFileEntryRef, T\u003e`, improving scanning performance by\n~2.5%.\n\nGitOrigin-RevId: f357a470de89660b6a887abc95f6a15aae205e64\n"
    },
    {
      "commit": "eca68ed17f94611539667011463dae833c58e3bc",
      "tree": "67a1dd6a16ed3ecc88a3671a58f2dd7fa26f42d2",
      "parents": [
        "19c840ee172e43835776c9c8cdfff55a061cb34c"
      ],
      "author": {
        "name": "Andy Kaylor",
        "email": "akaylor@nvidia.com",
        "time": "Thu Jun 04 10:02:33 2026 -0700"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Thu Jun 04 10:09:00 2026 -0700"
      },
      "message": "[CIR] Separate floating-point bin ops from integer bin ops (#201462)\n\nThis change creates new FP-specific binary operations and updates the\nexisting binary operations that previously accepted any arithmetic type\nto only allow integer and vector-of-integer types.\n\nThis change is being done to prepare for extended floating-point\nhandling such as strict FP semantics and fast-math handling. It also\nsimplifies the handling of integer overflow flags.\n\nAssisted-by: Cursor / claude-opus-4.8\nGitOrigin-RevId: a8267ccddd8324733eb5ac2663b6a0ca1d904c10\n"
    },
    {
      "commit": "19c840ee172e43835776c9c8cdfff55a061cb34c",
      "tree": "21ab20c67c50c74c249ca40c78b06d1734d0c61c",
      "parents": [
        "1901c3a17c58359b255e499c4e36dc4629c44189"
      ],
      "author": {
        "name": "Arthur Eubanks",
        "email": "aeubanks@google.com",
        "time": "Thu Jun 04 09:47:03 2026 -0700"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Thu Jun 04 09:51:09 2026 -0700"
      },
      "message": "Revert \"[clang] Add -fcrash-diagnostics-tar for tarball of crash reproducer files\" (#201622)\n\nReverts llvm/llvm-project#198838\n\nTest failing at\nhttps://lab.llvm.org/buildbot/#/builders/190/builds/43494\n\nGitOrigin-RevId: 69f7aeb52e71ebb7d264bc9e613bc4bc90cb0c47\n"
    },
    {
      "commit": "1901c3a17c58359b255e499c4e36dc4629c44189",
      "tree": "3fa5790e7ceaafaf7d8e81b68445e5b7182f0085",
      "parents": [
        "76cc227fdd43e6f6e06ec54dade687926dc34b84"
      ],
      "author": {
        "name": "Arthur Eubanks",
        "email": "aeubanks@google.com",
        "time": "Thu Jun 04 09:01:01 2026 -0700"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Thu Jun 04 09:06:26 2026 -0700"
      },
      "message": "[clang] Add -fcrash-diagnostics-tar for tarball of crash reproducer files (#198838)\n\nMakes it easier to move around crash diagnostics.\n\nGitOrigin-RevId: e44d060dfd617e8a5e37b61adcb38ed4ac5d782d\n"
    },
    {
      "commit": "76cc227fdd43e6f6e06ec54dade687926dc34b84",
      "tree": "21ab20c67c50c74c249ca40c78b06d1734d0c61c",
      "parents": [
        "94594e1af4cd8e122669f2e457423ee1696e50e2"
      ],
      "author": {
        "name": "Tom Honermann",
        "email": "tom.honermann@intel.com",
        "time": "Thu Jun 04 11:15:30 2026 -0400"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Thu Jun 04 08:21:12 2026 -0700"
      },
      "message": "[Clang] Correct diagnostic notes for C++11 range-based for statements with invalid iterator types (#201461)\n\nPreviously, diagnostic notes issued for errors encountered due to invalid\niterator types in C++11 range-based for statements reported the range type\nas the iterator type instead of the invalid iterator type.  Now fixed.\n\nGitOrigin-RevId: 816e2924786a0d7d300aa9a2ff80e5a2a082e6fa\n"
    },
    {
      "commit": "94594e1af4cd8e122669f2e457423ee1696e50e2",
      "tree": "a40d411bc273734abdb497678221d33996094541",
      "parents": [
        "7ebea8d8aa83f926fea73a53b87d184405d160c5"
      ],
      "author": {
        "name": "Tony Guillot",
        "email": "tony.guillot@protonmail.com",
        "time": "Thu Jun 04 16:49:40 2026 +0200"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Thu Jun 04 08:20:56 2026 -0700"
      },
      "message": "[Clang][Docs] Documented sentinel attribute (#196088)\n\nThe documentation of the sentinel attribute was missing, this PR\ndocuments the behavior of the sentinel attribute.\n\nGitOrigin-RevId: 8858ddd8a4735ae120855beb23b3e97dae1887ac\n"
    },
    {
      "commit": "7ebea8d8aa83f926fea73a53b87d184405d160c5",
      "tree": "27b5c6c8f496b3114cbbc51cddb1b98bccc49089",
      "parents": [
        "0b01d299c897d168d286f32bbe469936156f6b41"
      ],
      "author": {
        "name": "Kartik Ohlan",
        "email": "kartik7ohlan@gmail.com",
        "time": "Thu Jun 04 07:49:18 2026 -0700"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Thu Jun 04 07:50:54 2026 -0700"
      },
      "message": "[CIR] Maximum across vector (IEEE754) (#199779)\n\nPart of https://github.com/llvm/llvm-project/issues/185382\n\nMove the test cases to\n\n[intrinsics.c](https://github.com/llvm/llvmproject/pull/clang/test/CodeGen/AArch64/neon/intrinsics.c)\nRemoved the test cases from\n\n[neon-intrinsics.c](https://github.com/llvm/llvmproject/pull/clang/test/CodeGen/AArch64/neon/intrinsics.c)\n\nRemoved [neon-across.c](clang/test/CodeGen/AArch64/neon-across.c)\n\n---------\n\nCo-authored-by: Andrzej Warzyński \u003candrzej.warzynski@gmail.com\u003e\nGitOrigin-RevId: 6530683b31d8b9a3cb7e04eae7c59abeb223faae\n"
    },
    {
      "commit": "0b01d299c897d168d286f32bbe469936156f6b41",
      "tree": "a4236a70a8d695f168509018c0ccbab717b116f1",
      "parents": [
        "1eb3a9e4fd8648a73dd2eb551fca9f5ca6fb3c49"
      ],
      "author": {
        "name": "Kelvin Li",
        "email": "kli@ca.ibm.com",
        "time": "Thu Jun 04 10:38:30 2026 -0400"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Thu Jun 04 07:40:51 2026 -0700"
      },
      "message": "[clang][driver] Rename ClangExecutable and getClangProgramPath (NFC) (#200814)\n\nThis patch is to rename ClangExecutable to DriverExecutable and\ngetClangProgramPath to getDriverProgramPath. This makes the\nname more neutral and less confusing when used in flang.\n\nGitOrigin-RevId: 2e9f45a7020ba1bb99409e61d17c20cfe60e2e23\n"
    },
    {
      "commit": "1eb3a9e4fd8648a73dd2eb551fca9f5ca6fb3c49",
      "tree": "e14edfcab0ee71de5f6dbd00474a98b299d0a9c0",
      "parents": [
        "3821bcedc02aa82c54e06f4dfdb51924f758c266"
      ],
      "author": {
        "name": "Ariel-Burton",
        "email": "arielburton@yahoo.com",
        "time": "Thu Jun 04 09:28:14 2026 -0400"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Thu Jun 04 06:31:15 2026 -0700"
      },
      "message": "Fix clang/test/DebugInfo/CXX/line.cpp (#198401)\n\nThis PR twaeks the clang/test/DebugInfo/line.cpp test to pass on z/OS.\n\nThe reason the test was failing is that the RUN lines which specify\n-triple %itanium_abi_triple expands to s390x-ibm-zos when run on z/OS.\nThe IR that is emitted for this triple does not match the patterns\nexpected by the test.\n\nThis PR tweaks the patterns in the CHECK lines so that the test also\npasses on z/OS.\n\nGitOrigin-RevId: 07e081dcd784f3286ed9aa1ac549beb61066ce97\n"
    },
    {
      "commit": "3821bcedc02aa82c54e06f4dfdb51924f758c266",
      "tree": "5afca176ef0a413b35bb87bfd8375063d4053a8b",
      "parents": [
        "8b7b1811a023d950e0702c0169a6a368fdea98b9"
      ],
      "author": {
        "name": "Vassil Vassilev",
        "email": "v.g.vassilev@gmail.com",
        "time": "Thu Jun 04 16:16:10 2026 +0300"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Thu Jun 04 06:21:28 2026 -0700"
      },
      "message": "[clang-repl] Fix Value::setRawBits unit confusion and right-size raw storage. (#200886)\n\nValue::setRawBits had inconsistent units: the default value and the size\nassert treated the parameter as bytes (sizeof(Storage)), while the\nmemcpy treated it as bits (NBits / 8). A caller passing the natural byte\ncount (e.g. sizeof(long long)) ended up copying only sizeof(T)/8 bytes\n-- one byte for an 8-byte payload, leaving the rest stale. The one\nin-tree caller compensated by multiplying by 8, hiding the bug.\n\nRename the parameter to NBytes and drop the / 8 so the API name,\ndefault, assert, and memcpy all agree on bytes. Update the caller in\nInterpreterValuePrinter.cpp to pass ElemSize directly.\n\nRight-size the Storage::m_RawBits array while we are here: it was\nsizeof(long double) * 8 bytes, which reads like a bit/byte confusion\nsince the widest typed member of the union is long double itself. The\noversized array made sizeof(Value) ~144 bytes on x86_64 instead of ~40,\nbloating every copy/move of a Value.\n\nAdd a regression test exercising setRawBits with both an explicit byte\ncount and the default argument. Pre-fix the test fails for both: the\nexplicit-count branch copies 1 byte instead of 8, and the default branch\ncopies sizeof(Storage)/8 bytes instead of the full union width.\n\nGitOrigin-RevId: 3a1d8a8db2645a2a62b4f13916e2ba6e639a7fad\n"
    },
    {
      "commit": "8b7b1811a023d950e0702c0169a6a368fdea98b9",
      "tree": "5e93981aff69e1094ecf7f0442110244228951e4",
      "parents": [
        "07603f194a742b5fe249e11d5da5c8254480556b"
      ],
      "author": {
        "name": "Timm Baeder",
        "email": "tbaeder@redhat.com",
        "time": "Thu Jun 04 14:13:01 2026 +0200"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Thu Jun 04 05:16:20 2026 -0700"
      },
      "message": "[clang][bytecode] Fix a diagnostic difference with extern variables (#201369)\n\nIf the extern variable is constexpr of of non-array type, we should\ndiagnose it as missing an initializer. Otherwise, we diagose a read of\nnon-constexpr variable.\n\nGitOrigin-RevId: 03127a03860b9d8cb440fe8f51c00647f45eb8be\n"
    },
    {
      "commit": "07603f194a742b5fe249e11d5da5c8254480556b",
      "tree": "055da36c24543e348f5ba6f00f3768168018a076",
      "parents": [
        "85b4e3fef78b2e1ce3255f8bc3c573867a44358a"
      ],
      "author": {
        "name": "Matt Arsenault",
        "email": "Matthew.Arsenault@amd.com",
        "time": "Thu Jun 04 14:08:50 2026 +0200"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Thu Jun 04 05:11:26 2026 -0700"
      },
      "message": "clang: Remove hacky OpenMP handling for appending bound arch (#201555)\n\nUse the same path as CUDA/HIP and stop doing a hacky search through\nthe arguments looking for -march.\n\nGitOrigin-RevId: 36a1b5156e94f2c07ee8eeabe710ab99fc1a5489\n"
    },
    {
      "commit": "85b4e3fef78b2e1ce3255f8bc3c573867a44358a",
      "tree": "db3084e7b4518eee8271e452bb08be0de2bbe13d",
      "parents": [
        "8a1f93993be545013b378cde919dd9b3e32fca35"
      ],
      "author": {
        "name": "Ryosuke Niwa",
        "email": "rniwa@webkit.org",
        "time": "Thu Jun 04 03:33:53 2026 -0700"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Thu Jun 04 03:36:26 2026 -0700"
      },
      "message": "[alpha.webkit.UncountedCallArgsChecker] Protect the const member getter\u0027s this argument (#201147)\n\nThis PR fixes a bug that when a const member variable getter is\ndetected, we don\u0027t check if its object argument is kept alive for the\nduration of the function call.\n\nGitOrigin-RevId: 7a008c0ad92a434e3bd419f4bb1c94ce85c56844\n"
    },
    {
      "commit": "8a1f93993be545013b378cde919dd9b3e32fca35",
      "tree": "f95bf41daed7c0cd461ad7a71374a16d2057f74f",
      "parents": [
        "c671838eafc5dc5193d8612d73cca5c4f5237ab5"
      ],
      "author": {
        "name": "Timm Baeder",
        "email": "tbaeder@redhat.com",
        "time": "Thu Jun 04 11:00:54 2026 +0200"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Thu Jun 04 02:06:19 2026 -0700"
      },
      "message": "[clang][bytecode] Reject erroneous vector conversions (#201368)\n\nGitOrigin-RevId: 1960c7a9fdf2fc613a148fbc078173adddd6ca13\n"
    },
    {
      "commit": "c671838eafc5dc5193d8612d73cca5c4f5237ab5",
      "tree": "43a212988b98c00ea995aa601f9ea57614ff475b",
      "parents": [
        "1bde5d863d16d9369cc33480bc07183f3a666b26"
      ],
      "author": {
        "name": "Liao Chunyu",
        "email": "chunyu@iscas.ac.cn",
        "time": "Thu Jun 04 16:45:05 2026 +0800"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Thu Jun 04 01:51:03 2026 -0700"
      },
      "message": "[RISCV][MC] Support experimental Zvdota Family instructions (#195069)\n\nSpec:\nhttps://github.com/aswaterman/riscv-misc/blob/main/isa/ldot-bdot/ldot-bdot.adoc\n\n---------\n\nCo-authored-by: Brandon Wu \u003csongwu0813@gmail.com\u003e\nCo-authored-by: Craig Topper \u003ccraig.topper@sifive.com\u003e\nGitOrigin-RevId: 7da29bc529d74d327804ad25e49e0cdeccfed263\n"
    },
    {
      "commit": "1bde5d863d16d9369cc33480bc07183f3a666b26",
      "tree": "b1e3a44f25027766d28ccad0ae43c495caf7aeb1",
      "parents": [
        "5cb8579ce0cf9421a35685c118789ba4388760f7"
      ],
      "author": {
        "name": "paperchalice",
        "email": "liujunchang97@outlook.com",
        "time": "Thu Jun 04 16:45:02 2026 +0800"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Thu Jun 04 01:50:47 2026 -0700"
      },
      "message": "[IR] Add fast-math support to {u,s}itofp (#198470)\n\n- `{u,s}itofp` are floating point typed values.\n- CodeGen part (foldFPToIntToFP in DAGCombiner) needs `nsz` to fold\npattern (uitofp (fptoui x)) -\u003e (trunc x).\n- LLVM has intrinsic variants of `{u,s}itofp`, which already support\nfast-math flags.\n\nNow optimization flags require 9 bits in bitcode, fast-math flags of\n`uitofp` are stored in high 8 bits.\nVPlan part may need some extra work, it assumes optimization flags from\ndifferent categories are disjoint.\n\nGitOrigin-RevId: b66b10e852ea3bcb4a35ff83c781bf875ed50870\n"
    },
    {
      "commit": "5cb8579ce0cf9421a35685c118789ba4388760f7",
      "tree": "d00e7e1e30d5d1c512a943477f923f1e1d7c5264",
      "parents": [
        "4e0cf378439d24dce1d2dc3d724fdba2066bba26"
      ],
      "author": {
        "name": "Amr Hesham",
        "email": "amr96@programmer.net",
        "time": "Thu Jun 04 10:21:50 2026 +0200"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Thu Jun 04 01:28:53 2026 -0700"
      },
      "message": "[CIR] Fix Complex Casting with Atomic qualifier (#201424)\n\nApply Complex Casting with Atomic qualifier fixes from PRs #172210 and\n#172163\n\nIssue https://github.com/llvm/llvm-project/issues/192331\n\nGitOrigin-RevId: b6599484337145cd03cdc42b30de08939e3f6947\n"
    },
    {
      "commit": "4e0cf378439d24dce1d2dc3d724fdba2066bba26",
      "tree": "0ab0b55b810542bb7c4e1d7ce225ee63e508d9cf",
      "parents": [
        "d0355fc5a2efa7b9f1fa5098af85d4650ac6a049"
      ],
      "author": {
        "name": "Nikita Popov",
        "email": "npopov@redhat.com",
        "time": "Thu Jun 04 09:26:12 2026 +0200"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Thu Jun 04 00:30:24 2026 -0700"
      },
      "message": "[CaptureTracking] Volatile operations only capture address (#201316)\n\nThe fact that a volatile access was performed on a certain address is an\nobservable effect in the abstract machine, so volatile operations\ncapture the address of the accessed pointer. However, they do not\ncapture the provenance. This matches what we document in LangRef.\n\nWhile I\u0027m pretty sure that this models the semantics correctly, I\u0027m\nslightly concerned that we might be using the provenance capture here to\npaper over some other issue, though nothing specific comes to mind (and\nthe test changes don\u0027t show anything problematic).\n\nGitOrigin-RevId: f02dd4afbac15705016a9a7962d87161524b66cc\n"
    },
    {
      "commit": "d0355fc5a2efa7b9f1fa5098af85d4650ac6a049",
      "tree": "8f1352be5fe6bcadcc5279bb7a92e7f4afda6dd1",
      "parents": [
        "f94d57c576cd7d926dbcc4199b4edc29d3e33018"
      ],
      "author": {
        "name": "ShashwathiNavada",
        "email": "shashwathinavada@gmail.com",
        "time": "Thu Jun 04 12:35:57 2026 +0530"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Thu Jun 04 00:11:45 2026 -0700"
      },
      "message": "[Flang] Fix -frelaxed-c-loc-checks being ignored when using the driver (#200733)\n\n`-frelaxed-c-loc-checks` worked correctly when passed directly to -fc1,\nbut was silently ignored when using the driver (e.g., flang -c\n-frelaxed-c-loc-checks), causing the flag to go unused. This patch fixes\nit by adding `OPT_relaxed_c_loc` to the `addAllArgs` call in Flang.cpp\nAlso extend the existing test with a driver-mode RUN line to cover this\npath.\n\nGitOrigin-RevId: 364a8837a9d67135d050fe594ba51e319fbab938\n"
    },
    {
      "commit": "f94d57c576cd7d926dbcc4199b4edc29d3e33018",
      "tree": "0e517b0785687dcf54d4943da685d4251f1a7b02",
      "parents": [
        "2d9c83d1e9ee4f15655466aedfa71d24f6acadb3"
      ],
      "author": {
        "name": "Timm Baeder",
        "email": "tbaeder@redhat.com",
        "time": "Thu Jun 04 05:45:34 2026 +0200"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Wed Jun 03 20:50:26 2026 -0700"
      },
      "message": "[clang][bytecode] Fix a crash with covariant return types (#201354)\n\n`Context::collectBaseOffset()` will assert if the passed-in classes are\nthe same.\n\nGitOrigin-RevId: a53507fb5c108f0fd71270bf6498fa0f283b27c1\n"
    },
    {
      "commit": "2d9c83d1e9ee4f15655466aedfa71d24f6acadb3",
      "tree": "a1429d4b2e33823ab74300aac630573861de7ec7",
      "parents": [
        "f3ae146b4c478e17ceaee5f6b501ede9901055d3"
      ],
      "author": {
        "name": "Johannes Doerfert",
        "email": "jdoerfert.llvm@gmail.com",
        "time": "Wed Jun 03 19:14:30 2026 -0700"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Wed Jun 03 19:15:46 2026 -0700"
      },
      "message": "[OpenMP] Use ext linkage for kernels handles and globals handles keep… (#200964)\n\n… linkage\n\nHost handles are now emmitted with external linkage to clash if two\nkernels with the same name are registered. This could have happen right\nnow and silently corrupt the program, but it can happen more easily once\nwe allow users to name their kernels.\n\nIn the same patch we make global variable handles retain the linkage of\nthe global variable, forcing clashes for external ones and continue to\nsupport weak use cases.\n\n---------\n\nCo-authored-by: Shilei Tian \u003ci@tianshilei.me\u003e\nGitOrigin-RevId: 9263a42920ea470ff7377fa3ca3b3759aeb86fa1\n"
    },
    {
      "commit": "f3ae146b4c478e17ceaee5f6b501ede9901055d3",
      "tree": "f18fefee6eaa9c72a5cc3e797f7fbc3e582c0204",
      "parents": [
        "4f3ee6f5b88599fd831ad60be424f3e043776d33"
      ],
      "author": {
        "name": "Joseph Huber",
        "email": "huberjn@outlook.com",
        "time": "Wed Jun 03 18:46:09 2026 -0500"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Wed Jun 03 16:49:20 2026 -0700"
      },
      "message": "[Clang] Accept `-flto\u003dnone` similar to `-fno-lto` (#201460)\n\nSummary:\nPersonal preference, but I would like to be able to set all the LTO\nkinds from the string, and it feels a little odd to need to use\n`-fno-lto` to override the mode.\nGitOrigin-RevId: 96a50e0ee0959dd3095b0d3c561fab0c5f5fe94c\n"
    },
    {
      "commit": "4f3ee6f5b88599fd831ad60be424f3e043776d33",
      "tree": "0828d75b3d539d93bdfc82c191ebc67e19f18f8d",
      "parents": [
        "16641ec718ce2af44c0eb5a1f628d3a5b8381580"
      ],
      "author": {
        "name": "Michael Spencer",
        "email": "bigcheesegs@gmail.com",
        "time": "Wed Jun 03 15:33:22 2026 -0700"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Wed Jun 03 15:36:11 2026 -0700"
      },
      "message": "[libclang] Add clang_ModuleCache_pruneWithCallback (#199789)\n\nclang_ModuleCache_pruneWithCallback takes a callback that is invoked for\neach PCM that gets pruned. This is to support build systems that would\nlike to clean up additional data when PCMs are removed from disk.\n\nGitOrigin-RevId: 5f975bb6afec710e4ef47c3b90a299b2212c6ec0\n"
    },
    {
      "commit": "16641ec718ce2af44c0eb5a1f628d3a5b8381580",
      "tree": "482c671b25a116ec06b3943c07a67e9b80299501",
      "parents": [
        "2d643045f4bd8ed147b4982dba6a0c4b6fc3f679"
      ],
      "author": {
        "name": "Ryosuke Niwa",
        "email": "rniwa@webkit.org",
        "time": "Wed Jun 03 15:18:42 2026 -0700"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Wed Jun 03 15:21:08 2026 -0700"
      },
      "message": "[alpha.webkit.NoDeleteChecker] Treat a r-value smart pointer return value as no-delete. (#200912)\n\nSkip the checkin g of the destructor of T in ExprWithCleanups /\nCXXBindTemporaryExpr when returning a value using r-value reference\nsince such a construct never invokes delete.\n\n---------\n\nCo-authored-by: Balazs Benics \u003cbenicsbalazs@gmail.com\u003e\nGitOrigin-RevId: 9ae0d890a9236a94f23d5c3ff84ea08d7a670f03\n"
    },
    {
      "commit": "2d643045f4bd8ed147b4982dba6a0c4b6fc3f679",
      "tree": "b2e0fb13964df5e34feb1c972749194d5ec5240b",
      "parents": [
        "4fed21bb98628fe3e0ead91d350fefceccc2fac4"
      ],
      "author": {
        "name": "mygitljf",
        "email": "101249452+mygitljf@users.noreply.github.com",
        "time": "Thu Jun 04 04:51:12 2026 +0800"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Wed Jun 03 13:56:06 2026 -0700"
      },
      "message": "[clang-format][Objective-C] Fix assertion crash on stray \u0027-\u0027/\u0027+\u0027 in @interfa… (#199104)\n\nBefore calling `parseObjCMethod()`, it now checks whether the next token\nis a `(` or an identifier; if not, it simply skips ahead, thereby\npreventing the assertion failure and subsequent crash.\nFixes #199075\n\nGitOrigin-RevId: 0a4f19e4769eae755cdad1a4af1d73988762dd2c\n"
    },
    {
      "commit": "4fed21bb98628fe3e0ead91d350fefceccc2fac4",
      "tree": "c510f9f015f8abcf6d422d93749dff8346ba091a",
      "parents": [
        "a3e92b21ce445ea757d63a0c1218910a22b46c67"
      ],
      "author": {
        "name": "adams381",
        "email": "adams@nvidia.com",
        "time": "Wed Jun 03 13:27:41 2026 -0500"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Wed Jun 03 11:31:14 2026 -0700"
      },
      "message": "[CIR][NFC] Remove unused DataLayout in array lowering (#201408)\n\nThe per-element loop in CIRAttrToValue::visitCirAttr(ConstArrayAttr) constructed an mlir::DataLayout for every element of an ArrayAttr-backed constant array and never used it.  Constructing a DataLayout walks the parent ops to gather the layout spec, so for an N-element array this was N redundant constructions on the lowering path.  Removing the dead local is NFC -- the generated IR is unchanged.\n\nSplit out of #198427, where it was flagged as an unrelated drive-by.\n\nGitOrigin-RevId: 4169b17acfe53c79e0e170ebb8249ef749e061a2\n"
    },
    {
      "commit": "a3e92b21ce445ea757d63a0c1218910a22b46c67",
      "tree": "c083872237a112175be1b220ffa73b898727e18a",
      "parents": [
        "5a54ba59da62c2131cf276f3cf6c07f22bb23e35"
      ],
      "author": {
        "name": "Hans Wennborg",
        "email": "hans@hanshq.net",
        "time": "Wed Jun 03 20:05:47 2026 +0200"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Wed Jun 03 11:11:16 2026 -0700"
      },
      "message": "Revert \"[PGO][AMDGPU] Add basic HIP offload PGO support (#177665)\" (#201416)\n\nThis broke profiling builds on Windows by switching the profile library\nto link against the dynamic CRT; see discussion on the PR.\n\nThere were already a number of issues reported and fixed after this PR.\nRather than piling on the fixes (and this one may need some work),\nrevert back to green for now to let the project recover.\n\nThis reverts commit 5db13643f4b7038db0ca304d9f8900122502935c.\n\nAdditionally, this reverts the followup PRs in\n635e120fb87304924508a7a204574727e3c37363,\n2766733764f4bdf8399d48c0225e9c64bdfa95f7,\n4c33844b4b560f24b2a3a0cc689d73510cb01ad5, and\n5eca8b67ff3c5f371141d54b6e4544a0ebe77fdb:\n\n\"[PGO][HIP] Stop pulling ROCm.o into every PGO host link (#200101)\"\n\"[compiler-rt][profile] Add COMPILER_RT_BUILD_PROFILE_ROCM option\n(#200127)\"\n\"[PGO][HIP] Skip ROCm interceptor in profile-only compiler-rt builds\n(#200111)\"\n\"[PGO][HIP] Fix profile-only Windows link by gating ROCm interceptor\nmacro (#200859)\"\n\nGitOrigin-RevId: 6cfa1a01a0f737ed1d54963810057fb1bd67a274\n"
    },
    {
      "commit": "5a54ba59da62c2131cf276f3cf6c07f22bb23e35",
      "tree": "05c908a0f5e44f64efadbc960180b0ac2ad4772d",
      "parents": [
        "3549d85d40388cf507fafbba6c50f3bb3f083120"
      ],
      "author": {
        "name": "Brian Cain",
        "email": "brian.cain@oss.qualcomm.com",
        "time": "Wed Jun 03 11:37:49 2026 -0500"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Wed Jun 03 09:40:46 2026 -0700"
      },
      "message": "[Hexagon] Add cmake caches for cross-toolchain distribution build (#201207)\n\nAdds and extends the\nclang/cmake/caches/hexagon-unknown-linux-musl-clang* files to drive a\nfull install-distribution build: host tools, per-target builtins (Linux\nand baremetal), and runtimes for hexagon-unknown-linux-musl.\n\nGitOrigin-RevId: 3a1420effe06748cb9bd1b2ef746bd5be25fad52\n"
    },
    {
      "commit": "3549d85d40388cf507fafbba6c50f3bb3f083120",
      "tree": "a1a858dee51333d9fa62dac37c2f273110e64db5",
      "parents": [
        "3656dfeb8d5ce7349b04fcab5d2dcfb240eb8304"
      ],
      "author": {
        "name": "Andre Kuhlenschmidt",
        "email": "andre.kuhlenschmidt@gmail.com",
        "time": "Wed Jun 03 09:17:39 2026 -0700"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Wed Jun 03 09:20:47 2026 -0700"
      },
      "message": "[flang][openacc] change option feature flag names to address comments. (#200037)\n\n- Fixes FeatureFlags and CLI flags to use OpenACC instead of ACC.\n- Offline comments further refined the name to be\nOpenACCDefaultNoneScalarsStrict\n- Which changes the semantics requiring the user to opt out of the\ndefault behavior, instead of opt-in to an extension.\n- Also makes the CLI flag OptOut instead of OptIn to match the cli\nbehavior.\n- Gets rid of the unneeded warning flag since FeatureFlags have both a\ndisabled bit and a warning bit.\n- Updates old test for these changes and adds a new test to document all\nthe different cli configurations.\n\nGitOrigin-RevId: 15375d84799ec260d101aa35670f61fa456b6a4e\n"
    },
    {
      "commit": "3656dfeb8d5ce7349b04fcab5d2dcfb240eb8304",
      "tree": "9aef25857cbb53374fc03b26319a931deabd4b1b",
      "parents": [
        "6f469f09ad7455ae6166f3695e85924e7e1890fd"
      ],
      "author": {
        "name": "Harald van Dijk",
        "email": "hdijk@accesssoftek.com",
        "time": "Wed Jun 03 16:28:00 2026 +0100"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Wed Jun 03 08:31:58 2026 -0700"
      },
      "message": "[HIP] Fix another test for --no-lto (#201382)\n\nGitOrigin-RevId: 0cea23a93700e1a454369f617525f628df84a064\n"
    },
    {
      "commit": "6f469f09ad7455ae6166f3695e85924e7e1890fd",
      "tree": "f6358be34b204db7804c7161e5752fb3a1f674e2",
      "parents": [
        "b03364306297402db676a937905d8c969ac461b9"
      ],
      "author": {
        "name": "Andre Kuhlenschmidt",
        "email": "andre.kuhlenschmidt@gmail.com",
        "time": "Wed Jun 03 07:50:30 2026 -0700"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Wed Jun 03 07:56:49 2026 -0700"
      },
      "message": "[flang][openacc] add extension which accepts multiple names in a OpenACC routine directive (#200296)\n\nThis PR adds an extension which allows one or more function names in a\nsingle named routine directive. This is treated as multiple named\nroutine directives with the same clauses. The bind clause is forbidden.\nThe empty list of names isn\u0027t excepted. Routine clauses are stable under\nunparsing.\n\nThis PR tests Parsing, Unparsing, Semantics, and Lowering.\n\nGitOrigin-RevId: fdf8fb6c13deb88c2e179a4eed0c66900dcd92d8\n"
    },
    {
      "commit": "b03364306297402db676a937905d8c969ac461b9",
      "tree": "3cd8f249cc9294ea8074de65ad55d3c365b0fae4",
      "parents": [
        "3d3cfe94310f7a205b16fc3856a042f67b96b6cf"
      ],
      "author": {
        "name": "Yair Ben Avraham",
        "email": "yairba@protonmail.com",
        "time": "Wed Jun 03 17:45:01 2026 +0300"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Wed Jun 03 07:51:27 2026 -0700"
      },
      "message": "[CIR][AArch64] Lower vfmaq_lane_v and vfma_laneq_v (#197084)\n\nLower BI__builtin_neon_vfmaq_lane_v and BI__builtin_neon_vfma_laneq_v in\nCIR.\n\nThis handles the covered vfmaq_lane_* and vfma_laneq_* ACLE wrappers by\nbitcasting operands to the expected types, selecting the requested lane\nfrom the lane source operand, and emitting fma through\nemitCallMaybeConstrainedBuiltin.\n\nFor vfmaq_lane_v, the selected lane is splatted with emitNeonSplat.\nFor vfma_laneq_v, the lane is selected from the wider lane source; the\nf64 case extracts the scalar lane before emitting scalar fma.\n\nNeighboring scalar lane/laneq wrappers and other out-of-scope forms\nremain explicit NYI cases.\n\nTests are moved into the existing CIR-enabled fused multiply files under\nclang/test/CodeGen/AArch64/neon/, reusing upstream LLVM checks where\npossible and adding CIR coverage for lane selection and fma lowering.\n\nPart of #185382\nFollow-up to #195602\n\nGitOrigin-RevId: 4a8c5d231227d7c6c61160e5be25f4626cdabbb8\n"
    },
    {
      "commit": "3d3cfe94310f7a205b16fc3856a042f67b96b6cf",
      "tree": "a7d084b21f59d3189ae17a33d27e74e10cc8542d",
      "parents": [
        "198aa0428d2198c49d67bfa3fff62242bdd993ce"
      ],
      "author": {
        "name": "Jameson Nash",
        "email": "vtjnash@gmail.com",
        "time": "Wed Jun 03 10:41:23 2026 -0400"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Wed Jun 03 07:51:08 2026 -0700"
      },
      "message": "Reapply \"[clang] remove lots of \"innocuous\" addrspacecasts\"  (#200427)\n\nReapply #197745, with an additional commit to undo a small part of the\nfirst commit, pending further analysis of alternatives to that part of it.\n\nIn particular, make the `agg.tmp` (CreateAggTemp) values keep using the\ndeclared expression type of the RValue. This is indeed probably\nsensible, since it lets Sema influence this via the expression type,\nthough it runs into some issues where some expression types (notably for\nany load) haven\u0027t had one of the equivalent functions\ngetNonLValueExprType/DeduceAutoType/getUnqualifiedType called on them,\nso they are bringing along additional annotation baggage which doesn\u0027t\napply the the rvalue when turned into a temporary (see comments in\ngetNonLValueExprType for relevant part of C++ standard). This in turn is\nalso rarely relevant, since inventing temporaries aren\u0027t often allowed\nin this part of the pipeline (it\u0027d require a move constructor) so the\nLValue Dest already provides the type and the RValue type is ignored.\nHowever, it does affect a single test (which loads a global but ignores\nthe result) and so this adds an extra `getUnqualifiedType` call to\n`CodeGenFunction::EmitAnyExpr` so that load doesn\u0027t propagate the\nqualifiers from source to temporary. This might be a common issue with\nload-like expressions (for example, there\u0027s a `getUnqualifiedType` in\nAtomicExpr code when computing the appropriate codegen for the result,\nso it also didn\u0027t trust Sema to have created this correctly--and indeed\nit seems that Sema expects CG to wait until after\nDefaultLvalueConversion before it is does anything with the result type)\n\nGitOrigin-RevId: 9606c9ff18f0d90a639b132a7f69ebc8a8f61c99\n"
    },
    {
      "commit": "198aa0428d2198c49d67bfa3fff62242bdd993ce",
      "tree": "ff0637235b7736108432b72c7e5749e8c3332585",
      "parents": [
        "5008e038dc6a455ab0d95eca1dfec893c5dca6fc"
      ],
      "author": {
        "name": "Farzon Lotfi",
        "email": "farzonlotfi@microsoft.com",
        "time": "Wed Jun 03 10:34:53 2026 -0400"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Wed Jun 03 07:42:31 2026 -0700"
      },
      "message": "[Matrix][HLSL] Add codegen support for Matrix Layout keywords (#198887)\n\nfixes #192262\n\n- Wrap Matrix Type in a row or column major layout attribute\n- Add Helper to know which Matrix Layout to apply in codegen or check\nfor in Sema\n- Remove the Decl Atribute and only store on the type.\n\nAssisted by Claud Opus 4.7\n\nGitOrigin-RevId: 3a9c6cdb5d99cbcbb98c8d3562c409ddc5822deb\n"
    },
    {
      "commit": "5008e038dc6a455ab0d95eca1dfec893c5dca6fc",
      "tree": "f195a373213b782263fe3a1ca773dec9b94b3995",
      "parents": [
        "c30aa7888116d3c600cf6d2bfa9a046d2cb2e89d"
      ],
      "author": {
        "name": "Joseph Huber",
        "email": "huberjn@outlook.com",
        "time": "Wed Jun 03 09:21:51 2026 -0500"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Wed Jun 03 07:25:52 2026 -0700"
      },
      "message": "[HIP] Fix test for --no-lto (#201367)\n\nGitOrigin-RevId: 7eff475c4998039878563b987121c7fcc002b7b3\n"
    },
    {
      "commit": "c30aa7888116d3c600cf6d2bfa9a046d2cb2e89d",
      "tree": "57eb4a86eacab26e6da7c99f12d6843c6d08f55b",
      "parents": [
        "3ced54ea63d0adcc39b7156b120ff1d6f5d8814f"
      ],
      "author": {
        "name": "Joseph Huber",
        "email": "huberjn@outlook.com",
        "time": "Wed Jun 03 08:50:52 2026 -0500"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Wed Jun 03 06:56:29 2026 -0700"
      },
      "message": "[Clang] Fix leftover use of old LTO path (#201360)\n\nSummary:\nThis was accidentally missed when I merged the refactor because it\nshowed up after I made the PR and didn\u0027t have any merge conflicts I\nnoticed.\nGitOrigin-RevId: a2369b9743391351b1dd4a5c38fb61e295ff7088\n"
    },
    {
      "commit": "3ced54ea63d0adcc39b7156b120ff1d6f5d8814f",
      "tree": "f4c8452590599b6ab53eec26c031cb027bd03374",
      "parents": [
        "af1cc909bc4e9d9d973bf09c97ca66ecfe79107c"
      ],
      "author": {
        "name": "Joseph Huber",
        "email": "huberjn@outlook.com",
        "time": "Wed Jun 03 08:26:16 2026 -0500"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Wed Jun 03 06:30:44 2026 -0700"
      },
      "message": "[Clang] Rework LTO mode selection to be a Toolchain property (#201155)\n\nSummary:\nCurrently, the LTO mode is a property of the Driver, which makes sense\nbecause it is used to set up phases. However, we currently have `-flto`\nand `-foffload-lto`, which is a split that doesn\u0027t fully work with the\nfull context of a heterogenous compilation as it is \u0027all-or-nothing\u0027.\n\nThis PR seeks to be mostly NFC for now, just moving the queries to a\nper-toolchain interface rather than the static driver mode setting we\nhave right now. The *single* use of this before ToolChains are created\nis for the Webassembly toolchain to set an include path. This is now\njust a direct check on the flag, which is consistent. In the future they\ncould shift to fat LTO objects as well.\n\nThe main goal for the PR is to allow the GPU / Offloading toolchains to\nspecify their \"real\" LTO behavior. Right now SPIR-V and AMDGCN both\ndefault to LTO, but rather than re-use the LTO handling we hack through\nthe driver phases to override it. Allowing this split would let us\nheavily simplify this logic.\n\nCo-authored-by: Cursor \u003ccursoragent@cursor.com\u003e\n\n---------\n\nCo-authored-by: Cursor \u003ccursoragent@cursor.com\u003e\nGitOrigin-RevId: 70d62e1e10907e1bd6dd0e92968acbf6fe58af37\n"
    },
    {
      "commit": "af1cc909bc4e9d9d973bf09c97ca66ecfe79107c",
      "tree": "871d13ce73867be62562b47e4bc618bcdea6ae96",
      "parents": [
        "40f5e9331aab528be0cc5e9b6d33a8a1a76a64b4"
      ],
      "author": {
        "name": "Yanzuo Liu",
        "email": "zwuis@outlook.com",
        "time": "Wed Jun 03 20:56:33 2026 +0800"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Wed Jun 03 05:59:00 2026 -0700"
      },
      "message": "[clang][NFC] Introduce `LangOptions::isCompatibleWith(ClangABI)` (#201067)\n\nThis slightly improves readability and reduces the probability of\noff-by-one errors.\n\nGitOrigin-RevId: 27fd70f849f417db18ac581457f0419332a7a1fa\n"
    },
    {
      "commit": "40f5e9331aab528be0cc5e9b6d33a8a1a76a64b4",
      "tree": "8eba9934d48aa63f734279be9f0d617b2bf85c09",
      "parents": [
        "cd6ce7d17621ed9cafe9ed9e1bc7da351d14421c"
      ],
      "author": {
        "name": "Chuanqi Xu",
        "email": "yedeng.yd@linux.alibaba.com",
        "time": "Wed Jun 03 17:19:41 2026 +0800"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Wed Jun 03 02:26:12 2026 -0700"
      },
      "message": "[C++20] [Modules] Ask for definition before calling isInlineDefinitionExternallyVisible (#201291)\n\nClose https://github.com/llvm/llvm-project/issues/195905\n\nThe root cause of the problem is, after decl merging, one of the\ndestructor for `base` (see test case) is chosen, and its body is not\ndeserialized for performance as we alreday see its body in current TU.\nAnd then, the code call\nFunctionDecl::isInlineDefinitionExternallyVisible from the destructor\nwithout a serialized body, but\nFunctionDecl::isInlineDefinitionExternallyVisible\u0027s implementation\nrequires to see a body. Then assertion failed.\n\nDecl merging and function decl merging is common. We can\u0027t always make\nsure we find the definition without checking. It is much more\nfundamental. So the PR doesn\u0027t try to touch the decl merging mechanism\nor how we handle the result of lookups.\n\nWe tried to fix the issue at the calling point to make sure we have a\ndefinition for isInlineDefinitionExternallyVisible. Other use of\nisInlineDefinitionExternallyVisible does the similar thing.\n\nGitOrigin-RevId: 74f1c9417e916aed314fd429ce2faee49668ffa6\n"
    },
    {
      "commit": "cd6ce7d17621ed9cafe9ed9e1bc7da351d14421c",
      "tree": "dec36f72bf0b9f87a96d3dc28066be8d2a503460",
      "parents": [
        "aae1652c30b75351ec43e5f18db2fa884337e05f"
      ],
      "author": {
        "name": "Martin Wehking",
        "email": "martin.wehking@arm.com",
        "time": "Wed Jun 03 09:33:22 2026 +0100"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Wed Jun 03 01:36:10 2026 -0700"
      },
      "message": "[AArch64] Add 9.7 CVT data processing intrinsics (#186807)\n\nAdd Clang intrinsics\nsvcvtt_f16_s8, _f32_s16, _f64_s32, _f16_u8, _f32_u16, _f64_u32\nsvcvtb_f16_s8, _f32_s16, _f64_s32, _f16_u8, _f32_u16, _f64_u32\n\n+ lowering to AARCH64 Instrs. SCVTF, SCVTFLT, UCVTF, UCVTFLT\n\nand Clang instrinsics:\nsvcvtn_s8[_f16_x2], _s32[_f64_x2], _u8[_f16_x2], _u16[_f32_x2],\n_u32[_f64_x2]\n\n+ lowering to AARCH64 Instrs. FCVTZSN, FCVTZUN\n\nThe Clang intrinsics are guarded by the sve2.3 and sme2.3 feature flags.\n\nACLE Patch:\nhttps://github.com/ARM-software/acle/pull/428\n\nThe patch reuses `IsReductionQV` for resolving the overload of\nintrinsics.\nThis naming is misleading and needs changed\n\nGitOrigin-RevId: 74bf9b5666a6450ef87b130070b5df3da365183d\n"
    },
    {
      "commit": "aae1652c30b75351ec43e5f18db2fa884337e05f",
      "tree": "b6224c1f7db54e838d9094fa16f5e7fee0237ac7",
      "parents": [
        "3308d91f15cf286e6779931cc5cbb35d0e9ff5c1"
      ],
      "author": {
        "name": "Wenju He",
        "email": "wenju.he@intel.com",
        "time": "Wed Jun 03 16:02:20 2026 +0800"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Wed Jun 03 01:06:15 2026 -0700"
      },
      "message": "Revert \"[OpenCL] Allow mesa3d OS in spirv32 and spirv64 targets\" (#200727)\n\nReverts llvm/llvm-project#197148\n\nlibclc will use generic target triple spirv32[64]-unknown-unknown for use in mesa.\n\nGitOrigin-RevId: 6b5bb08d3aa3f498b7bd9f6bb43376719bddfa60\n"
    },
    {
      "commit": "3308d91f15cf286e6779931cc5cbb35d0e9ff5c1",
      "tree": "50927f0ce1fa686e74bca0f9f6811e7e6ffa533f",
      "parents": [
        "9163c46be05b3104c8606f5e9be347f2ec145c8a"
      ],
      "author": {
        "name": "Steffen Larsen",
        "email": "sholstla@amd.com",
        "time": "Wed Jun 03 05:45:35 2026 +0200"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Tue Jun 02 20:51:12 2026 -0700"
      },
      "message": "[clang][CUDA] Avoid ambiguity in host/device template specializations (#201049)\n\nThis commit changes SemaOverload to resolve an otherwise diagnosed\nambiguity between addresses of template specializations of functions\nthat are overloaded for both device and host. Similar to how it works\nfor non-templated function overloads, these changes prioritizes the\nspecializations that corresponds to the target of the owning function,\ni.e. if compiling for host, the address of the host specialization takes\nprecedence over the device specialization and vice versa.\n\nFixes https://github.com/llvm/llvm-project/issues/199299\n\n---------\n\nSigned-off-by: Steffen Holst Larsen \u003csholstla@amd.com\u003e\nGitOrigin-RevId: 6d7962db9397eefe3a287115d413fd4c562e4178\n"
    },
    {
      "commit": "9163c46be05b3104c8606f5e9be347f2ec145c8a",
      "tree": "83d890898f7f3ecd1315a04112016f7e3bceb4bf",
      "parents": [
        "64055fe7bd180847186384bb67d6a4bfd508bd0d"
      ],
      "author": {
        "name": "Boyao Wang",
        "email": "wangboyao@bytedance.com",
        "time": "Wed Jun 03 10:07:32 2026 +0800"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Tue Jun 02 19:11:00 2026 -0700"
      },
      "message": "[RISCV] Remove experimental XRivosVizip support (#200761)\n\nRemove experimental XRivosVizip support which will not be maintained by\nRVIOS any more.\n\nGitOrigin-RevId: 53938ba7f3631583f79f1b7c1cf8a5d2347ba4d1\n"
    },
    {
      "commit": "64055fe7bd180847186384bb67d6a4bfd508bd0d",
      "tree": "dfa17a409e4b15a26e3dba4b1e127b8eed1a977c",
      "parents": [
        "1412de353bb5853754744cb3fa8c59c645f7cc26"
      ],
      "author": {
        "name": "Andy Kaylor",
        "email": "akaylor@nvidia.com",
        "time": "Tue Jun 02 18:24:23 2026 -0700"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Tue Jun 02 18:26:15 2026 -0700"
      },
      "message": "[CIR] Implement destruction of TLS and static global references (#200227)\n\nThis implements destruction of lifetime-extended reference temporaries\nused to initialize TLS or static duration reference variables.\n\nAssisted-by: Cursor / claude-opus-4.7\nGitOrigin-RevId: 19c7fdb3c3c68a9d6861627c85cd0251b7854f6f\n"
    },
    {
      "commit": "1412de353bb5853754744cb3fa8c59c645f7cc26",
      "tree": "81bd5746afc7d178e4f7ae3dc4c3bfe67047d3f6",
      "parents": [
        "3b86f01a4551bb79b673806ee57d7514b731d210"
      ],
      "author": {
        "name": "Andy Kaylor",
        "email": "akaylor@nvidia.com",
        "time": "Tue Jun 02 18:23:57 2026 -0700"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Tue Jun 02 18:25:48 2026 -0700"
      },
      "message": "[CIR] Fix insertion point tracking for switch with cleanups (#201210)\n\nWe had some problems where we would incorrectly maintain the insertion\npoint for switch statements that contained cleanup scopes. This resulted\nin cir.scope statements without a terminator, tripping a verification\nerror.\n\nThis change adds a RunCleanupsScope RAII object for the switch statement\nand adds a check inside popCleanup() to avoid moving the insertion point\nto the point after the now-closed cleanup scope if the insertion point\nhad previously been somewhere other than inside the cleanup scope.\n\nAssisted-by: Cursor / claude-opus-4.8\nGitOrigin-RevId: 45bddca21c65106b41fa47a485a30c863e75da53\n"
    },
    {
      "commit": "3b86f01a4551bb79b673806ee57d7514b731d210",
      "tree": "3e4532b006e35e51b91678fb80d8a6620331d55b",
      "parents": [
        "de2acd5799e05c4e86c02c5260208ea19477c474"
      ],
      "author": {
        "name": "adams381",
        "email": "adams@nvidia.com",
        "time": "Tue Jun 02 20:11:37 2026 -0500"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Tue Jun 02 18:16:09 2026 -0700"
      },
      "message": "[CIR] Coerce Direct args and returns in CallConvLowering (#195879)\n\nFourth PR in the split of #192119/#192124. Implements the\nDirect-with-coercion path in CallConvLowering.\n\nEvery Direct argument or return whose ABI type differs from its source\ntype is now coerced through a store/reload roundtrip via an entry-block\nalloca, mirroring classic codegen\u0027s CreateCoercedLoad/CreateCoercedStore.\nThe temporary alloca uses max(srcAlign, dstAlign) from the DataLayout and\nis hoisted into the entry block so it composes with HoistAllocas\nregardless of pipeline order. When the coerced type is larger than the\nsource -- e.g. a 12-byte aggregate returned as { i64, i64 } -- the slot is\nsized to the larger type and accessed through a source-typed view for the\nstore and a destination-typed view for the load, so neither side\nover-reads.\n\nCallConvLowering is split into three phases (function-definition\ncoercion, call-site rewriting, and Ignore cleanup) because in-place\nblock-argument type changes from Direct-with-coerce otherwise confused the\nearlier ordering, synthesizing alloca/load chains for call-site uses of\nIgnore\u0027d arguments that were about to be dropped.\n\nCo-authored-by: Cursor \u003ccursoragent@cursor.com\u003e\nGitOrigin-RevId: f132e926290fe1cc69946addfb90d61b56c9402a\n"
    },
    {
      "commit": "de2acd5799e05c4e86c02c5260208ea19477c474",
      "tree": "9fe1d01a73e5c051af24ddca60f47be026575367",
      "parents": [
        "da0605e4eec93380a1622390819fa3e30ac6fdb0"
      ],
      "author": {
        "name": "Yury Plyakhin",
        "email": "yury.plyakhin@intel.com",
        "time": "Tue Jun 02 18:02:30 2026 -0700"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Tue Jun 02 18:06:41 2026 -0700"
      },
      "message": "[clang-sycl-linker][test] Improve dry-run mode and tighten test coverage (#200513)\n\n- Rework `--dry-run` in `clang-sycl-linker` so it skips all real output\n    (writing bitcode, executing tools, etc.).\n  - The `link:`, `sycl-module-split:`, and a new `sycl-bundle:` summary\n    line are now gated on `-v` alone.\n  - Tighten `sycl-bundle:` checks in `basic.ll`, `split-mode.ll`, and\n    `triple.ll` to pin kind, triple, and arch (instead of just kind),\n    and add `-NOT: {{.+}}` after fully-covered dry-run check groups.\n  - replace the `clang-sycl-linker` + `llvm-objdump --offloading`\n    round-trip with a single `--dry-run -v` invocation.\n- add dedicated `non-dry-run` mode test to verify code paths not exposed\nin `dry-run`.\n\n    Assisted by Claude.\n\nGitOrigin-RevId: 0e40e9e341cf78662ab468fb8cdd6cac7e58ad1f\n"
    },
    {
      "commit": "da0605e4eec93380a1622390819fa3e30ac6fdb0",
      "tree": "8ecef0746b72ce7c0fbc00934ae65da14b79937b",
      "parents": [
        "c35cd5ab83a5556817b695ad68ab2b58c4921a1c"
      ],
      "author": {
        "name": "Erich Keane",
        "email": "ekeane@nvidia.com",
        "time": "Tue Jun 02 16:51:15 2026 -0700"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Tue Jun 02 16:56:05 2026 -0700"
      },
      "message": "[CIR] Implement lowering for const-emitted global compound literals (#201152)\n\nThis came up in a test suite as a NYI, it is just emitting a\nconstant-backing literal for an initializer. These are specific to C, as\nglobal compound literals have static storage duration in C. This patch,\njust like classic codgen, just creates a \u0027.compoundliteral\u0027 object as\nbacking for these variables, and lets us create references to them.\n\n---------\n\nCo-authored-by: Andy Kaylor \u003cakaylor@nvidia.com\u003e\nGitOrigin-RevId: 678891b586ab0f3b97c83987eb2df0d2d5743ffb\n"
    },
    {
      "commit": "c35cd5ab83a5556817b695ad68ab2b58c4921a1c",
      "tree": "a2a18328d887c4574a3321d916c071484bd6e3a4",
      "parents": [
        "cf25df254a6e772b377e314b82791ed41f7a9049"
      ],
      "author": {
        "name": "Yury Plyakhin",
        "email": "yury.plyakhin@intel.com",
        "time": "Tue Jun 02 16:12:42 2026 -0700"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Tue Jun 02 16:16:15 2026 -0700"
      },
      "message": "[clang-linker-wrapper] Drop SYCL dry-run stub-image special case (#201222)\n\nRemove the `DryRun` branch in `bundleSYCL` that emitted a stub\n`OffloadBinary`. SYCL goes through the same empty-buffer path as other\noffload kinds, so the special case is no longer needed.\n\nUpdate `linker-wrapper-image.c` to expect the resulting `[0 x i8]\nzeroinitializer` constant and a size of `0` in the register/unregister\ncalls.\n\nAssisted by Claude.\n\nGitOrigin-RevId: d59899763f0df2bf16fca98ab11421646957602a\n"
    },
    {
      "commit": "cf25df254a6e772b377e314b82791ed41f7a9049",
      "tree": "7fe9b099cd70446cf58a192f0d93a2a8774efcf6",
      "parents": [
        "e46466c53f52743ad12a0c640a2d5744b3f420b1"
      ],
      "author": {
        "name": "adams381",
        "email": "adams@nvidia.com",
        "time": "Tue Jun 02 18:10:31 2026 -0500"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Tue Jun 02 16:16:00 2026 -0700"
      },
      "message": "[CIR] Set ExternalWeakLinkage on weak/weak_import function declarations (#198422)\n\nClassic CodeGen\u0027s `SetFunctionAttributes` calls `setLinkageForGV` to force `ExternalWeakLinkage` on `__attribute__((weak))` and Darwin `weak_import` declarations.  CIR had no equivalent: weak function declarations were emitted with `ExternalLinkage` instead of `ExternalWeakLinkage`.\n\nThis adds `setLinkageForFunction` — the same weak/external-weak logic as `setLinkageForGV` — and calls it from `setFunctionAttributes`.  The underlying crash on inline forward declarations (the original motivation) is already fixed by #195257; what remains is this linkage gap.\n\n`inline-forward-decl.c` covers `__attribute__((weak))` on an inline forward declaration; `func-linkage-weak-import.c` covers Darwin `weak_import` (→ `extern_weak` in CIR and LLVM).\n\nGitOrigin-RevId: 26e44e62fdaf0467e153d2f83251e2d48eb1f10e\n"
    },
    {
      "commit": "e46466c53f52743ad12a0c640a2d5744b3f420b1",
      "tree": "6c6b1259f6359cb19982098a6ac71d2484cef0b7",
      "parents": [
        "913c12213e549376924e71b5f6c593ef8f5a5f73"
      ],
      "author": {
        "name": "Mircea Trofin",
        "email": "mtrofin@google.com",
        "time": "Tue Jun 02 14:46:36 2026 -0700"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Tue Jun 02 14:50:44 2026 -0700"
      },
      "message": "Reverting PR #184065 and #200323 to address some interplay with CFI (#201194)\n\nThere is a relation between CFI and ThinLTO GUIDs that still needs to be\ndisentangled first. Note that we leave the `MD_unique_id` in\n`FixedMetadataKinds.def` to avoid needing to re-number it later. Plus\nthe metadata string (\"guid\") itself is used by ctxprof.\n\nGitOrigin-RevId: 87d6941017aa6022ce00baa9f1dfae5a49b113f4\n"
    },
    {
      "commit": "913c12213e549376924e71b5f6c593ef8f5a5f73",
      "tree": "fb23dcf092c8373213261f27a1cf9c916800b7f5",
      "parents": [
        "2779c7e97119194ba451e2dfa316952b68003f15"
      ],
      "author": {
        "name": "Ian Anderson",
        "email": "iana@apple.com",
        "time": "Tue Jun 02 13:20:13 2026 -0700"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Tue Jun 02 13:26:14 2026 -0700"
      },
      "message": "[clang][driver][darwin] Hold onto full triples in Darwin SDKPlatformInfo (#200896)\n\nThe architecture can be relevant when determining if an SDK supports a\nparticular triple. Record the full triples in SDKPlatformInfo instead of\nall of the non-architecture triple components.\n\nAssisted-by: Claude Code\n\nrdar://172876443\n\nGitOrigin-RevId: b89bb06afd069aa1b5e9f05ab692b3e6b41318c0\n"
    },
    {
      "commit": "2779c7e97119194ba451e2dfa316952b68003f15",
      "tree": "812748315b605fef167301b66cff0b9298c5a35e",
      "parents": [
        "66156a011bda298c7afbb6e35dd05a4b2c3ed3ed"
      ],
      "author": {
        "name": "Yaxun (Sam) Liu",
        "email": "yaxun.liu@amd.com",
        "time": "Tue Jun 02 16:10:16 2026 -0400"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Tue Jun 02 13:15:46 2026 -0700"
      },
      "message": "[HIP][AMDGPU] Use non-LTO pipeline for non-RDC in the linker wrapper (#201135)\n\nNon-RDC HIP does not need LTO, but the new offload driver compiles all\nAMDGPU device code through the LTO pipeline. That makes non-RDC builds\npay\nfull LTO codegen cost for no benefit.\n\nFix this in clang-linker-wrapper instead of the driver, so device\ncodegen\nstill runs in the wrapper\u0027s parallel device-link step (kept fast by\n--offload-jobs) rather than being serialized back in the driver.\n\nThe driver passes a new --no-lto flag for the non-RDC fat-binary job\n(unless the user asked for -foffload-lto). With it, the wrapper drops\n-flto\nand, because the device images are bitcode stored in object-extension\nfiles, also passes -x ir so clang actually compiles them (cc1 -emit-obj)\ninstead of handing the bitcode to lld for an LTO link.\n\nGitOrigin-RevId: eaff67c9838f98d221aa563c9e5d4e714d636674\n"
    },
    {
      "commit": "66156a011bda298c7afbb6e35dd05a4b2c3ed3ed",
      "tree": "ac8435cbdc8d709dea475f524d6cecfb59a3bda4",
      "parents": [
        "23965f9d81b3300429286607e9954d4aaaf19620"
      ],
      "author": {
        "name": "Vicky Nguyen",
        "email": "vicky.trucviennguyen@gmail.com",
        "time": "Tue Jun 02 11:51:29 2026 -0700"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Tue Jun 02 11:56:20 2026 -0700"
      },
      "message": "[CIR][AArch64] Upstream vector-shift-right-and-accumulate NEON builtins (#200630)\n\nRelated to https://github.com/llvm/llvm-project/issues/185382\n\nCIR lowering for vector-shift-right-and-accumulate\n(https://arm-software.github.io/acle/neon_intrinsics/advsimd.html#vector-shift-right-and-accumulate)\n\nPort tests from `clang/test/CodeGen/AArch64/neon_intrinsics.c` to\n`clang/test/CodeGen/AArch64/neon/intrinsics.c`\n\nGitOrigin-RevId: e21e5489588c2944d5071b52bdbca684a72c756f\n"
    },
    {
      "commit": "23965f9d81b3300429286607e9954d4aaaf19620",
      "tree": "b5a349477c432d6e2890e83f3dfa94d685002fa2",
      "parents": [
        "eacab9846f8f681c7098cc8be81f0b2b8121a7a9"
      ],
      "author": {
        "name": "Abid Qadeer",
        "email": "haqadeer@amd.com",
        "time": "Tue Jun 02 18:53:30 2026 +0100"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Tue Jun 02 10:55:42 2026 -0700"
      },
      "message": "[flang] Support -ffunction-sections and -fdata-sections. (#199731)\n\nWire the flags through the driver, frontend, and TargetMachine, and add\ndriver and codegen lit tests.\n\nFixes https://github.com/llvm/llvm-project/issues/163550\n\n---------\n\nCo-authored-by: Cursor \u003ccursoragent@cursor.com\u003e\nGitOrigin-RevId: fe2f7e40994574dd5d7efa4da8be5d9237905416\n"
    },
    {
      "commit": "eacab9846f8f681c7098cc8be81f0b2b8121a7a9",
      "tree": "eccb8b52e7aa6debd71f2ce3280cd67d4627bc05",
      "parents": [
        "cb34be98fad6c1ef5441b1620b22dc22dacd08af"
      ],
      "author": {
        "name": "Michael Buch",
        "email": "michaelbuch12@gmail.com",
        "time": "Tue Jun 02 18:34:13 2026 +0100"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Tue Jun 02 10:36:01 2026 -0700"
      },
      "message": "[clang] Treat unnamed bitfields as padding in `__builtin_clear_padding` (#201102)\n\nCurrently Clang\u0027s implementation of `__builtin_clear_padding` diverges\nfrom GCC in its treatment of unnamed bitfields.\n\nGCC treats them as padding (which seems correct since they can\u0027t be\nnamed and wouldn\u0027t be part of the value representation of an object,\nthough I\u0027m not sure what the standard has to say about this).\n\n[Gobdolt](https://godbolt.org/z/e9Mo91dhh)\n\nGCC trunk:\n```\npre-clear bytes:   ff ff ff ff\npost-clear bytes:  01 00 00 80\n```\n\nClang trunk:\n```\npre-clear bytes:   ff ff ff ff\npost-clear bytes:  ff ff ff ff\n```\n\nNote how we cleared the padding with GCC.\n\nThis patch skips marking unnamed bitfields as \"occupied\". The rest of\nthe machinery works out-of-the-box.\n\nGitOrigin-RevId: 6dfef47047def67fc141248eb8e7e2a67e4307e7\n"
    },
    {
      "commit": "cb34be98fad6c1ef5441b1620b22dc22dacd08af",
      "tree": "7989a337d889fecda3ff1a0a14b606d7dcd494ec",
      "parents": [
        "77b1b3b0cbbfb3a1f87c705fc6e47eccb611b81b"
      ],
      "author": {
        "name": "Zibi Sarbinowski",
        "email": "zibi@ca.ibm.com",
        "time": "Tue Jun 02 13:32:56 2026 -0400"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Tue Jun 02 10:35:47 2026 -0700"
      },
      "message": "[SystemZ][z/OS] Add XPLink ABI on z/OS (#188501)\n\nThis patch introduces full support for the XPLINK calling convention on\nz/OS within LLVM. XPLINK is documented in the Language Environment\nVendor Interface ([Chapter\n22](https://www.ibm.com/docs/en/zos/3.2.0?topic\u003dapplications-call-linkage-convention-amode-64))\nand in the IBM Redbook [XPLink: OS/390 Extra Performance\nLinkage](http://www.redbooks.ibm.com/abstracts/sg245991.html?Open\n). It defines a high‑performance linkage model used by the z/OS Language\nEnvironment (LE) and by the XL C/C++ compiler.\nThis work succeeds and replaces the earlier effort in [PR\n101024](https://github.com/llvm/llvm-project/pull/101024). The present\nimplementation restructures the original approach, significantly expands\ntest coverage, and aligns more closely with both the architectural\nrequirements of XPLINK and the de‑facto ABI behaviour of XL C/C++.\n\nGitOrigin-RevId: d9cf2e1bc5aaf2eaacc36c765b63429d42ac0c60\n"
    },
    {
      "commit": "77b1b3b0cbbfb3a1f87c705fc6e47eccb611b81b",
      "tree": "ac7a153ff2de2cca152a63286dcb4d418fdff6be",
      "parents": [
        "5df25caefedf24355ed2e391914853e5935fdba7"
      ],
      "author": {
        "name": "Jan Svoboda",
        "email": "jan_svoboda@apple.com",
        "time": "Tue Jun 02 10:21:09 2026 -0700"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Tue Jun 02 10:25:51 2026 -0700"
      },
      "message": "[clang] Remove `FileSystemStatCache` (#198411)\n\nThis mechanism is not used anywhere, and can be easily reimplemented in\nterms of a VFS if needed.\n\nGitOrigin-RevId: bc056ea2c2948d51455f6cf75975e4980b8b908f\n"
    },
    {
      "commit": "5df25caefedf24355ed2e391914853e5935fdba7",
      "tree": "668ec8356bb5da98b1918425eaca0f95f8ec0b32",
      "parents": [
        "01bdf9fe868f7d55c4326b4e1c8dc268ebbd9288"
      ],
      "author": {
        "name": "Rainer Orth",
        "email": "ro@gcc.gnu.org",
        "time": "Tue Jun 02 19:09:49 2026 +0200"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Tue Jun 02 10:16:15 2026 -0700"
      },
      "message": "[Driver] Switch clang++ to default Solaris 11.4 compilation environment (#201063)\n\n`clang++` has long followed `g++`\u0027s lead predefining `_XOPEN_SOURCE\u003d600`\non Solaris. As detailed in [Switch g++ to default Solaris 11.4\ncompilation\nenvironment](https://gcc.gnu.org/pipermail/gcc-patches/2026-May/716990.html),\nthis is no longer necessary in Solaris 11.4.\n\nTested on `amd64-pc-solaris2.11` and `sparcv9-sun-solaris2.11`.\n\nGitOrigin-RevId: 3658dafe2c26c3e1242ff0505b9e8f95f0f5abf1\n"
    },
    {
      "commit": "01bdf9fe868f7d55c4326b4e1c8dc268ebbd9288",
      "tree": "7337f6ca187edb1541fc8d11ade655495bf97b77",
      "parents": [
        "c2654f8f04b85f7b8b2fd9e7a20e025867b3d716"
      ],
      "author": {
        "name": "Jake Daly",
        "email": "48452250+jakemdaly@users.noreply.github.com",
        "time": "Tue Jun 02 09:29:24 2026 -0700"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Tue Jun 02 09:31:23 2026 -0700"
      },
      "message": "[clang][NFC] Bump the maximum number of Sema diagnostics (#200948)\n\nThe number of Sema diagnostics in DiagnosticSemaKinds.td has reached the\n5000 limit. This PR increases the max limit to 6000.\n\nGitOrigin-RevId: 22ef7baaf6dc409a353af42ff52599b2746107ca\n"
    },
    {
      "commit": "c2654f8f04b85f7b8b2fd9e7a20e025867b3d716",
      "tree": "ae5a717cb22c20d53fbdd8d4e824e58d4fc3d84b",
      "parents": [
        "d8ec8a1a445149c7c054664b245c753699c6f84a"
      ],
      "author": {
        "name": "Utkarsh Saxena",
        "email": "usx@google.com",
        "time": "Tue Jun 02 17:16:27 2026 +0200"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Tue Jun 02 08:21:10 2026 -0700"
      },
      "message": "[LifetimeSafety] Refactor buildOriginFlowChain to use PImpl pattern (#201071)\n\nMove `buildOriginFlowChain` into `AnalysisImpl`, removing `FactManager`\nparameter\n\nGitOrigin-RevId: 3dfd4f784d7d0837252482f28ed2c0ca99a109b0\n"
    },
    {
      "commit": "d8ec8a1a445149c7c054664b245c753699c6f84a",
      "tree": "399b392a8fac7f0891fb1bf5e475e339b2cee459",
      "parents": [
        "6f80c46bed639681a710d13dc7161922572a0ae7"
      ],
      "author": {
        "name": "Andrzej Warzyński",
        "email": "andrzej.warzynski@arm.com",
        "time": "Tue Jun 02 15:52:17 2026 +0100"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Tue Jun 02 07:55:42 2026 -0700"
      },
      "message": "[cir] Refine cir::CastOp semantics for int \u003c-\u003e float casts (#200005)\n\nInt-to-float and float-to-int casts in cir::CastOp are lowered directly\nto their LLVM equivalents. Update the verifier to reflect this semantics\nand ensure that, for vector casts, the source and destination vectors have\nthe same length.\n\nThis lets the CIR verifier reject invalid casts earlier, instead of relying\non errors reported later at the LLVM IR level.\n\nGitOrigin-RevId: ade0e1a49c2118a98b0b34b3136661747bc5aaa0\n"
    },
    {
      "commit": "6f80c46bed639681a710d13dc7161922572a0ae7",
      "tree": "8976289ae3f4752659fb757491bbb0d44cd56520",
      "parents": [
        "1a5195c57804e976fef55b7319a023c63c203c3f"
      ],
      "author": {
        "name": "Timm Baeder",
        "email": "tbaeder@redhat.com",
        "time": "Tue Jun 02 16:47:56 2026 +0200"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Tue Jun 02 07:51:01 2026 -0700"
      },
      "message": "[clang][ExprConst] Support DesignatedInitUpdateExpr of array type (#201000)\n\nI think this should work.\n\nGitOrigin-RevId: 5d9e965c3f9983278308a00c282810e0f04754cc\n"
    },
    {
      "commit": "1a5195c57804e976fef55b7319a023c63c203c3f",
      "tree": "f06979fa48d7cb58c0905c987ee47ac304f39133",
      "parents": [
        "e0c7d5498a03de0f1ecc3412f359dc064645b90e"
      ],
      "author": {
        "name": "Timm Baeder",
        "email": "tbaeder@redhat.com",
        "time": "Tue Jun 02 16:16:47 2026 +0200"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Tue Jun 02 07:21:06 2026 -0700"
      },
      "message": "[clang][bytecode] Variadic ctors also start lifetime (#201113)\n\nThis is the same thing we do in the non-variadic `Call()`.\n\nGitOrigin-RevId: 16853f627616581293a8561b5c00dad33056c75f\n"
    },
    {
      "commit": "e0c7d5498a03de0f1ecc3412f359dc064645b90e",
      "tree": "bb246c824ab6c191e77aa5096f38772a7d720d35",
      "parents": [
        "e1b4647a08be205c653112fe38de83bac0140bb8"
      ],
      "author": {
        "name": "Christian Sigg",
        "email": "csigg@google.com",
        "time": "Tue Jun 02 16:00:40 2026 +0200"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Tue Jun 02 07:06:04 2026 -0700"
      },
      "message": "[clang] Fix lit test after f341dab for envs that cannot write to file (#201121)\n\nGitOrigin-RevId: fda34a97729f254d1a184d341683132c4ec65489\n"
    },
    {
      "commit": "e1b4647a08be205c653112fe38de83bac0140bb8",
      "tree": "114a57034adfee0cbb82ed8af1f8348fb70071f8",
      "parents": [
        "6e70a75a468f662502fd85d55c69e9713b930d36"
      ],
      "author": {
        "name": "Hans Wennborg",
        "email": "hans@hanshq.net",
        "time": "Tue Jun 02 13:43:23 2026 +0200"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Tue Jun 02 04:46:24 2026 -0700"
      },
      "message": "Revert \"[clang] fix getTemplateInstantiationArgs\" (#201093)\n\nThis caused bootstrap builds on macOS to fail with\n\n  error: definition with same mangled name ... as another definition\n\nin JSONGenerator.cpp. See comments on the original PR.\n\nReverts llvm/llvm-project#199528\n\nGitOrigin-RevId: b29352f7ea1d46fa7d90900d7a279851b6de9f74\n"
    },
    {
      "commit": "6e70a75a468f662502fd85d55c69e9713b930d36",
      "tree": "ad4b176989b1ed92dc5b219e007658d0386f1218",
      "parents": [
        "6db82a2a25a7436cb2a1dc9acdfd18da2614ad26"
      ],
      "author": {
        "name": "Lukacma",
        "email": "Marian.Lukac@arm.com",
        "time": "Tue Jun 02 12:17:02 2026 +0100"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Tue Jun 02 04:21:08 2026 -0700"
      },
      "message": "[AArch64] Add intrinsic support for Fdot instr. (#189987)\n\nThis patch adds intrinsics for new NEON Fdot instruction variants. The\nimplementation is based on the\n[proposal](https://github.com/ARM-software/acle/pull/428) and adds these\nACLE builtins:\n\n```\nfloat32x2_t vdot_f32_f16(float32x2_t r, float16x4_t a, float16x4_t b);\nfloat32x4_t vdotq_f32_f16(float32x4_t r, float16x8_t a, float16x8_t b);\n\nfloat32x2_t vdot_lane_f32_f16(float32x2_t r, float16x4_t a, float16x4_t b, const int lane);\nfloat32x4_t vdotq_laneq_f32_f16(float32x4_t r, float16x8_t a, float16x8_t b, const int lane);\nfloat32x2_t vdot_laneq_f32_f16(float32x2_t r, float16x4_t a, float16x8_t b, const int lane);\nfloat32x4_t vdotq_lane_f32_f16(float32x4_t r, float16x8_t a, float16x4_t b, const int lane);\n```\n\nGitOrigin-RevId: 7139d0b46f6d71334b74049bbd540ac4b18a1bc5\n"
    },
    {
      "commit": "6db82a2a25a7436cb2a1dc9acdfd18da2614ad26",
      "tree": "5f0cf66800ba426fcad48b910b9348345b7c0af5",
      "parents": [
        "b7f02f57f32cce81611b10f9e70fa5a288ce3984"
      ],
      "author": {
        "name": "Donát Nagy",
        "email": "donat.nagy@ericsson.com",
        "time": "Tue Jun 02 13:10:07 2026 +0200"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Tue Jun 02 04:16:13 2026 -0700"
      },
      "message": "[NFC][analyzer] Simplify ExprEngine::Visit, eliminate NodeBuilders (#200837)\n\nRemove NodeBuilder use from `ExprEngine::Visit` to reduce its size from\n700 lines to 600 lines of source code.\n\nThis commit also moves the \"for instance method operators, make sure the\n\u0027this\u0027 argument has a valid region\" logic (~10 lines) from the huge\n`ExprEngine::Visit` to the more specific `VisitCallExpr`; and applies a\nfew other very minor code quality improvements.\n\nGitOrigin-RevId: 78cf6cde9fbb6f08dacb3de5bc529243f9f85d9b\n"
    },
    {
      "commit": "b7f02f57f32cce81611b10f9e70fa5a288ce3984",
      "tree": "6b588b13f66db972d2d7311321dcb65b1a4f1df4",
      "parents": [
        "2d9f7fb29e6c31e088e59cca502f246911cfbd38"
      ],
      "author": {
        "name": "Ming-Yi Lai",
        "email": "ming-yi.lai@mediatek.com",
        "time": "Tue Jun 02 18:12:39 2026 +0800"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Tue Jun 02 03:15:47 2026 -0700"
      },
      "message": "[clang][RISCV][Zicfilp] Force user to use `-mcf-branch-label-scheme\u003dunlabeled` (#152122)\n\nExpected Behavior:\n\nWhen `-fcf-protection\u003dbranch|full` is specified, it\u0027s an error to omit\n`-mcf-branch-label-scheme\u003dunlabeled`.\n\nContext:\n\nWhen using forward-edge control flow integrity feature based on the\nRISC-V Zicfilp extension, the `-mcf-branch-label-scheme` option selects\nthe encoding scheme used in the landing pad labels. The spec defines 2\nschemes: `func-sig` and `unlabeled`, with the former specified as the\ndefault. However the `func-sig` backend is still under active\ndevelopment and won\u0027t land anytime soon; in the meanwhile, the\n`unlabeled` scheme almost has complete support in the toolchain now.\n\nGiven that Clang currently accepts and defaults to\n`-mcf-branch-label-scheme\u003dfunc-sig` but doesn\u0027t work correctly, we want\nto formally forbid the user from using\n`-mcf-branch-label-scheme\u003dfunc-sig` for now until the `func-sig` scheme\nis properly supported. We choose to do so by forcing the user to specify\n`-mcf-branch-label-scheme\u003dunlabeled` (instead of changing the default to\n`unlabeled`) so the default of `func-sig` can be retained according to\nthe spec and build scripts targeting the current Clang can be\nforward-compatible with the Clang that supports the complete `func-sig`\nscheme.\n\nGitOrigin-RevId: f341dab9edfa97d7a4a01f9bc0fe981055dc5cc9\n"
    },
    {
      "commit": "2d9f7fb29e6c31e088e59cca502f246911cfbd38",
      "tree": "947d8b2ddd3851efc6f714c9dcde05c23e7ce16e",
      "parents": [
        "2ff528fd7e1c475e4cfbf78b433bda2b248147bf"
      ],
      "author": {
        "name": "markyang92",
        "email": "80628866+markyang92@users.noreply.github.com",
        "time": "Tue Jun 02 17:59:53 2026 +0900"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Tue Jun 02 02:02:50 2026 -0700"
      },
      "message": "[clang][cmake] Move perf-training out of CLANG_INCLUDE_TESTS (#192163)\n\nperf-training defines the generate-profdata target used by the PGO\nbootstrap build.\nHowever, it is currently enabled only when CLANG_INCLUDE_TESTS\u003dON.\nFor distribution builds such as Yocto/OE, tests are usually disabled by\nsetting this to OFF.\n\nBut perf-training is a PGO utility, not a test target, and it is\ncurrently gated by that block.\nAs a result, generate-profdata is unavailable to the PGO bootstrap build\nwhen\nCLANG_INCLUDE_TESTS\u003dOFF.\n\nMove perf-training out of the CLANG_INCLUDE_TESTS block.\n\nThis is safe because utils/perf-training/CMakeLists.txt adds targets\nonly when\nLLVM_BUILD_INSTRUMENTED or CLANG_BOLT is enabled, so moving it out does\nnot add\nany targets unless PGO or BOLT is actually in use.\n\nAlso fix the path in lit.site.cfg.in for standalone builds.\nWhen llvm_src_dir is set from CMAKE_SOURCE_DIR, standalone clang builds\nend up\nusing the clang source directory instead of the LLVM source directory.\nSet it to the proper LLVM source location instead.\n\nAssisted-by: claude\nGitOrigin-RevId: b0e72b7c3f240cdc7e20eb6692d4662e7315bbc6\n"
    },
    {
      "commit": "2ff528fd7e1c475e4cfbf78b433bda2b248147bf",
      "tree": "a10d1c442d127923ac5f9b6d30c4774ddb9f875f",
      "parents": [
        "d02f6c436205f7cfd09c7692f4246a182322f30c"
      ],
      "author": {
        "name": "Timm Baeder",
        "email": "tbaeder@redhat.com",
        "time": "Tue Jun 02 10:04:58 2026 +0200"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Tue Jun 02 01:11:12 2026 -0700"
      },
      "message": "[clang][bytecode] Reject invalid UETT_OpenMPRequiredSimdAlign nodes (#200997)\n\nGitOrigin-RevId: e174a520b21103890ebc7367952202914b2baf88\n"
    },
    {
      "commit": "d02f6c436205f7cfd09c7692f4246a182322f30c",
      "tree": "98ba6c773cd37376a20cd0ffbc55766e5d58472c",
      "parents": [
        "d821f3d48f268203e0b6db1bf53c2e3ee4991a99"
      ],
      "author": {
        "name": "Antonio Frighetto",
        "email": "me@antoniofrighetto.com",
        "time": "Tue Jun 02 09:22:12 2026 +0200"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Tue Jun 02 00:26:09 2026 -0700"
      },
      "message": "[clang][CodeGen] Drop TBAA metadata emission on FP libcalls (#200752)\n\nTBAA annotation on FP libcalls has been superseded by\nrecently-introduced `llvm.errno.tbaa` module-level metadata.\n\nGitOrigin-RevId: 6f44c750a1987b050490f9655dc88c0145a44813\n"
    },
    {
      "commit": "d821f3d48f268203e0b6db1bf53c2e3ee4991a99",
      "tree": "bb86ed6aa4971d68bcd4b0ea2c604a3a6f04bf19",
      "parents": [
        "4b284cc8d9cb97596151cd9295995d06a316dc6a"
      ],
      "author": {
        "name": "Abhinav Pradeep",
        "email": "abhinav.pradeep@oracle.com",
        "time": "Tue Jun 02 17:08:18 2026 +1000"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Tue Jun 02 00:11:08 2026 -0700"
      },
      "message": "[LifetimeSafety] Prevent false-negative lifetimebound verification when origin escapes in an unrelated manner (#200786)\n\nThis PR removes a false-negative `[[lifetimebound]]` verification result\nthat occurs when the annotated attribute escapes via unrelated origin\nescape kinds (escape through global, escape through field).\n\nChange summary:\n- `Checker.cpp` has function `checkAnnotations` which checks if an\nescaping `Origin` was an annotated parameter. Modified the logic there\nto only verify `[[lifetimebound]]` in the case that the escape was\nthrough a return statement.\n\nFixes #200412\n\n---------\n\nSigned-off-by: Abhinav Pradeep \u003cabhinav.pradeep@oracle.com\u003e\nGitOrigin-RevId: c22474fc02c1c26c81577d4cc147ed3b7c04a079\n"
    },
    {
      "commit": "4b284cc8d9cb97596151cd9295995d06a316dc6a",
      "tree": "7aae4ed24498e1975dcdaeed6182c4d5935048a7",
      "parents": [
        "901ce0b2c3de37cc18460ddd961ecaf982878cfb"
      ],
      "author": {
        "name": "Jiahao Guo",
        "email": "eoonguo@gmail.com",
        "time": "Tue Jun 02 14:58:17 2026 +0800"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Tue Jun 02 00:01:09 2026 -0700"
      },
      "message": "[CIR] Add cir.lifetime.start and cir.lifetime.end Op (#199599)\n\n### summary\n\nAdds the `cir.lifetime.start` and `cir.lifetime.end` ops, which mark the\nbeginning and end of an alloca\u0027s live range. They mirror the LLVM\nintrinsics `llvm.lifetime.start` / `llvm.lifetime.end`\n\nGitOrigin-RevId: b3438c9af04b2ccf13b0adb6faf5e1fa0e3fb585\n"
    },
    {
      "commit": "901ce0b2c3de37cc18460ddd961ecaf982878cfb",
      "tree": "43afdac8e7092807930d1cfdfd271094c49d61c6",
      "parents": [
        "1b997818cc6bbf7bd004cf3625a150f9690735e9"
      ],
      "author": {
        "name": "Cullen Rhodes",
        "email": "cullen.rhodes@arm.com",
        "time": "Tue Jun 02 06:45:33 2026 +0100"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Mon Jun 01 22:51:10 2026 -0700"
      },
      "message": "[CMake][Release] Use llvm-bitcode-strip on Darwin for build with -ffat-lto-objects (#200764)\n\nBuilding with --fat-lto-objects was added in #140381 (cff9ae7a15a5). On\nmacOS the tests are failing when building release binaries with many\n\"The file was not recognized as a valid object file\" errors, e.g.:\n\n  2026-01-16T12:54:26.0928880Z\n\n/Users/runner/work/llvm-project/llvm-project/build/tools/clang/stage2-instrumented-bins/tools/clang/stage2-bins/_CPack_Packages/Darwin/TXZ/LLVM-22.1.0-rc1-macOS-ARM64/bin/llvm-strip:\n  error:\n\n\u0027/Users/runner/work/llvm-project/llvm-project/build/tools/clang/stage2-instrumented-bins/tools/clang/stage2-bins/_CPack_Packages/Darwin/TXZ/LLVM-22.1.0-rc1-macOS-ARM64/lib/libLLVMAArch64AsmParser.a(AArch64AsmParser.cpp.o)\u0027:\n  The file was not recognized as a valid object file\n\nIt\u0027s assuming bitcode is embedded in section .llvm.lto\n(llvm/lib/Object/ObjectFile.cpp:80) but on MachO it\u0027s in\n__LLVM,__bitcode (llvm/lib/Object/MachOObjectFile.cpp:2184)\n\nllvm-bitcode-strip is a driver for the MachO bitcode_strip utility which\nhandles this. Use this instead.\n\nFixes #176398.\n\nAssisted-by: codex\nGitOrigin-RevId: b8e768f6ae95343ac542b195265572ecd063fc2b\n"
    },
    {
      "commit": "1b997818cc6bbf7bd004cf3625a150f9690735e9",
      "tree": "6978c50c4c53e282fb80a2b5f1109fb93f9546e0",
      "parents": [
        "49c869982512d7b41ed9d230b462ba6bedbd02f7"
      ],
      "author": {
        "name": "Chaitanya",
        "email": "Krishna.Sankisa@amd.com",
        "time": "Tue Jun 02 09:15:15 2026 +0530"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Mon Jun 01 20:51:22 2026 -0700"
      },
      "message": "[CIR][AMDGPU] Implement lowering for __builtin_amdgcn_dispatch_ptr (#199880)\n\nPort `emitAMDGPUDispatchPtr` from OGCG. Emits the `amdgcn.dispatch.ptr`\nintrinsic and inserts an address-space cast when the builtin\u0027s expected\nreturn type differs.\n\nGitOrigin-RevId: 7290d876ad409596fe232e20bc4e83706fc5c05b\n"
    },
    {
      "commit": "49c869982512d7b41ed9d230b462ba6bedbd02f7",
      "tree": "43247bbd88e97cac210a8155c3f318649b7699c2",
      "parents": [
        "061169f6e5af1a12ba6fe571a1577f964cc5c9bf"
      ],
      "author": {
        "name": "Chaitanya",
        "email": "Krishna.Sankisa@amd.com",
        "time": "Tue Jun 02 09:14:46 2026 +0530"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Mon Jun 01 20:51:06 2026 -0700"
      },
      "message": "[CIR] Fix cir.call_llvm_intrinsic lowering for 0-result ops (#199516)\n\n`cir.call_llvm_intrinsic` declares `Optional\u003cCIR_AnyType\u003e:$result`, but\nthe lowering indexed `op-\u003egetResultTypes()[0]` unconditionally and OOBed\non void calls.\nGuard with `getNumResults()` and pick the void overload of\n`LLVM::CallIntrinsicOp::create` in `createCallLLVMIntrinsicOp`.\n\nGitOrigin-RevId: fba43a5b8923ea4f1d526ff0180a33543b65bc59\n"
    }
  ],
  "next": "061169f6e5af1a12ba6fe571a1577f964cc5c9bf"
}