[qt/clang/llvm]: Summary of bulk changes made

Git repository change summary for qt/clang/llvm
Pushed by mirror-service into branch 'upstream/users/shiltian/fix-simplify-cfg-convergent-issue'.
Changed from e059f901a2b458f0eb8eb3d056d92909697d2093 to 71956dc466232dd8bd3bf81657a3360d4328f053
Acknowledgement was received that this change introduces only existing code that has been pushed to another public open source repository.

This change contains the following new commits:

Git commit 5a98e96135e223890cbf0640584e30c0db695ecf by GitHub (on behalf of Michael Kruse) on 16/07/2026 at 11:19..
[Flang][CMake] Use fakeflang as dispatcher (#209482)

With #203481 where `runtimes-configure` does not depend on `flang`
anymore, the following race could happen in incremental builds: While
ninja/make builds some of the libFortran*.so libraries in a
BUILD_SHARED_LIBS=ON build, CMake tries to execute flang which tries to
load the .so libraries while they are linked. This race condition can
only happen in incremental builds because before the real bin/flang is
built completely the first time, it points to fakeflang which would be
working. Only in an incremental build, would the real bin/flang already
been built created before, pointing to the currently re-built flang.
I did not think this would be possible because CMake only needs to probe
flang once and thereafter (in the incremental build) remembers the probe
result in `CMakeFiles/<version>/CMakeFortranCompiler.cmake`. CMake does
not support changing the compiler once configured and errors-out if you
try to change `CMAKE_Fortran_COMPILER`. I do not know why/when CMake is
re-probing the compiler in some incremental builds, but has been
reported to do so in #205360.

This patch avoids any replacement of build artifacts and instead decides
in fakeflang when to call real flang or to emulate the Flang
preprocessor. The decision is made based of on the FLANG_BOOTSTRAP_PROBE
environment variable that is set only in the configure step. The feature
that allows this is unfortunately only available starting in CMake 4.2.
All suppported CMake versions before CMake 4.2 use the filename
`CMakeFortranCompilerId.F`, so make the decision based on that when
using earlier versions of CMake.

Tested with CMake 3.20, 3.28, 3.31, 4.1, and 4.2.

Fixes #205360
https://invent.kde.org/qt/clang/llvm/-/commit/5a98e96135e223890cbf0640584e30c0db695ecf

Git commit 57ceb12780157d68f7c14d5ed8f18bc87dc768fd by GitHub (on behalf of Wenju He) on 16/07/2026 at 11:21..
[DebugInfo][NFC] Fix PDB/Native/LinePrinter.h build warning -Wmicrosoft-unqualified-friend (#209715)

Fix build warning on Windows, which is error under -Werror: unqualified
friend declaration referring to type outside of the nearest enclosing
namespace is a Microsoft extension; add a nested name specifier
[-Wmicrosoft-unqualified-friend]
https://invent.kde.org/qt/clang/llvm/-/commit/57ceb12780157d68f7c14d5ed8f18bc87dc768fd

Git commit 5e4f44041d32234ea8227cbb2473173bd5edfc1c by GitHub (on behalf of Mariusz Sikora) on 16/07/2026 at 11:22..
[AMDGPU] Clean up GFX13 feature list (#208416)

GFX13 features were previously mostly inherited from GFX12/GFX1250.
- define an explicit gfx13 target parser feature list, including
gfx13-insts and cvt-pknorm-vop3-insts.
- remove transpose load F4/F6 support from gfx13 (previously copied from
gfx12)
https://invent.kde.org/qt/clang/llvm/-/commit/5e4f44041d32234ea8227cbb2473173bd5edfc1c

Git commit 3faf46d218d3cb888c818b3e99d7e05b06ca68fa by GitHub (on behalf of Lang Hames) on 16/07/2026 at 11:32..
[orc-rt] Session::OnCallControllerReturnFn, comments. NFC. (#210035)

Rename Session::OnCallHandlerCompleteFn to
Session::OnControllerCallReturnFn and add a comment to clarify its
purpose: This callback is used to handle retured values from calls to
the controller.
https://invent.kde.org/qt/clang/llvm/-/commit/3faf46d218d3cb888c818b3e99d7e05b06ca68fa

Git commit 8866215f30953c7df146276ac10753de941b1fcb by GitHub (on behalf of mkovacevic99) on 16/07/2026 at 11:35..
[Orc] Update the comment so it doesn't mention TargetTriple argument that no longer exists (#210023)

Completes the fix for the
https://github.com/llvm/llvm-project/issues/80097
https://invent.kde.org/qt/clang/llvm/-/commit/8866215f30953c7df146276ac10753de941b1fcb

Git commit 3d87de86a24d1bff2821e4eb2f2d28e58605a394 by GitHub (on behalf of Philip742) on 16/07/2026 at 11:53..
[mlir][tosa] Add support for MXFP in CONCAT (#209719)

Adds support to profile compliance for MXFP variants of CONCAT

Signed-off-by: Philip Wilkinson <[email protected]>
https://invent.kde.org/qt/clang/llvm/-/commit/3d87de86a24d1bff2821e4eb2f2d28e58605a394

Git commit 18fc5a8adbe5dba7dc281bacdb149e95dc7dc1be by GitHub (on behalf of Pengcheng Wang) on 16/07/2026 at 12:12..
[TableGen][RISCV] Generate pre/post-RA statistics for macro fusions (#209394)

Teach `MacroFusionPredicatorEmitter` to emit `Statistic` counters
recording how often each macro fusion is matched, and wire them
up for RISC-V.

This is based on the idea in #186499, generalized so counters are
generated automatically for any target using the emitter.

Because both the pre-RA and post-RA schedulers run MacroFusion,
a single counter would conflate the two. Counters are split into
pre-RA/post-RA variants, distinguished by the `NoVRegs` property,
addressing the double-counting concern from #186499:

```cpp
  STATISTIC(NumTuneLDADDFusionPreRA,  "Times TuneLDADDFusion Triggered (pre-ra)");
  STATISTIC(NumTuneLDADDFusionPostRA, "Times TuneLDADDFusion Triggered (post-ra)");
  ...
    if (SecondMI.getMF()->getProperties().hasNoVRegs())
      ++NumTuneLDADDFusionPostRA;
    else
      ++NumTuneLDADDFusionPreRA;
```

Besides, RISC-V counters use the `riscv-macro-fusion` debug type.

Co-authored-by: Sam Elliott <[email protected]>
Assisted-by: TRAE CLI (DeepSeek V4 Pro)
https://invent.kde.org/qt/clang/llvm/-/commit/18fc5a8adbe5dba7dc281bacdb149e95dc7dc1be

Git commit db7356dde8312a86e79ec0d7b7f844878a1a743c by GitHub (on behalf of Petr Hosek) on 16/07/2026 at 12:15..
[libc++] Partially revert #208330 (#209928)

We found out that the new mechanism for detecting overriden functions
does not support Arm Pointer Authentication (PAuth). Addressing this
limitation is going to require changes to Clang. In the meantime, we
switched back to the old mechanism when PAuth is enabled.
https://invent.kde.org/qt/clang/llvm/-/commit/db7356dde8312a86e79ec0d7b7f844878a1a743c

Git commit 3bf897eaac63ea4e6b8c4f77d84f833f63167e3d by GitHub (on behalf of Alexey Bataev) on 16/07/2026 at 12:25..
[SLP][NFC]Add a test with the non-profitable instcount check across the loops, NFC



Reviewers: 

Pull Request: https://github.com/llvm/llvm-project/pull/210045
https://invent.kde.org/qt/clang/llvm/-/commit/3bf897eaac63ea4e6b8c4f77d84f833f63167e3d

Git commit ee5cff53cf4f0c23254bda94c7249e4ea008d457 by GitHub (on behalf of Kseniya Tikhomirova) on 16/07/2026 at 12:27..
[libsycl] Fix mock error in unittest helper (#209726)

Signed-off-by: Tikhomirova, Kseniya <[email protected]>
https://invent.kde.org/qt/clang/llvm/-/commit/ee5cff53cf4f0c23254bda94c7249e4ea008d457

Git commit 66e8cc5ec83d7ec5e506c53b11c1d22887c0c418 by GitHub (on behalf of Florian Hahn) on 16/07/2026 at 12:51..
[ConstraintElim] Handle (X | Y) >s -1 as X >s -1 && Y >s -1. (#209743)

InstCombine canonicalizes X >=s 0 && Y >=s 0 as (X | Y) >=s 0. Teach
ConstraintElimination to recover the signed-positive information by
looking through compares of binary ORs.

Alive2 Proof: https://alive2.llvm.org/ce/z/sqNF8M

Compile-time impact is neutral

https://llvm-compile-time-tracker.com/compare.php?from=88acd428fd72f44312408c3fb6165992fb3b043a&to=a2aa4cacb953b32d6319939b9aee3226a7294334&stat=instructions%3Au

PR: https://github.com/llvm/llvm-project/pull/209743
https://invent.kde.org/qt/clang/llvm/-/commit/66e8cc5ec83d7ec5e506c53b11c1d22887c0c418

Git commit 6388cfc492e82702dfc1aa965adab7e85c7ad3d8 by GitHub (on behalf of Nikita Popov) on 16/07/2026 at 13:23..
[AMDGPU] Add missing FMF flags in tests (NFC) (#210055)

The flags should be on the select, not the fcmp for nsz to have
any effect.
https://invent.kde.org/qt/clang/llvm/-/commit/6388cfc492e82702dfc1aa965adab7e85c7ad3d8

Git commit 424fec2368d615dff78b4167e78e9bbb06fc4214 by GitHub (on behalf of Jessica Clarke) on 16/07/2026 at 13:36..
[NFC][test][AliasSet] Canonicalise memset.ll's overload suffices (#210059)

The second overload suffix here corresponds to the type of the length,
the third argument, which is an i64 not an i32. Canonicalise this to
avoid confusion from inconsistency.

Whilst here, fix the whitespace used to align the arguments in the
intrinsic declarations, a remnant of losing the two characters in "i8"
when migrating to opaque pointers.

Fixes: 9f0916129024 ("[AST] Add test coverage of memsets")
Fixes: ac924305793c ("[AliasSet] Convert tests to opaque pointers
(NFC)")
https://invent.kde.org/qt/clang/llvm/-/commit/424fec2368d615dff78b4167e78e9bbb06fc4214

Git commit 006d13e206401545c011028d85749071acdc5326 by GitHub (on behalf of Alex Duran) on 16/07/2026 at 13:45..
[OFFLOAD][L0][NFC] Cleanup dead code (#210060)

As we have done several refactorings lately, some dead code/data was
left behind inadvertedly. This is just cleaning it up.
https://invent.kde.org/qt/clang/llvm/-/commit/006d13e206401545c011028d85749071acdc5326

Git commit 6ba12f1878384fa0d5be092b511f67e7e808772b by GitHub (on behalf of Simon Pilgrim) on 16/07/2026 at 13:47..
[X86] fold-signbit-reduction-cmp.ll - test for each x86-64 cpu level (#210051)
https://invent.kde.org/qt/clang/llvm/-/commit/6ba12f1878384fa0d5be092b511f67e7e808772b

Git commit 26c200d88e720fa88e8692ace13a493696ac591f by GitHub (on behalf of Ayokunle Amodu) on 16/07/2026 at 13:53..
[CIR][AMDGPU] Add support for AMDGCN sin builtins (#197784)

Adds codegen for the following AMDGCN sin builtins:

- __builtin_amdgcn_sinf (float)
- __builtin_amdgcn_sinh (half)
- __builtin_amdgcn_sin_bf16 (bfloat16)

These are lowered to the corresponding `llvm.amdgcn.sin` intrinsic.
https://invent.kde.org/qt/clang/llvm/-/commit/26c200d88e720fa88e8692ace13a493696ac591f

Git commit 8739d89c4c2475eefc4638882311515fc67f9f9b by GitHub (on behalf of Yuan Suo) on 16/07/2026 at 14:01..
[LifetimeSafety] Introduce buildOriginFlowChain to use-after-invalidation (#208891)

After completing the implementation of `buildOriginFlowChain`, we should
integrate it into the remaining diagnostics. I started with
`use-after-invalidation` because there is already an overload of
`buildOriginFlowChain` for `UseFact`.

While testing, I noticed the following diagnostics:

```txt
test.cpp:9:23: note: result of call to 'begin<std::vector<int>>' aliases the storage of parameter 'v'
    9 |   auto it = std::find(std::begin(v), std::end(v), 3); // expected-warning {{parameter 'v' is later invalidated}} \
      |                       ^~~~~~~~~~~~~
test.cpp:9:13: note: result of call to 'find<__gnu_cxx::__normal_iterator<int *, std::vector<int>>, int>' aliases the storage of parameter 'v'
    9 |   auto it = std::find(std::begin(v), std::end(v), 3); // expected-warning {{parameter 'v' is later invalidated}} \
      |             ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
```

I think these diagnostics could be made clearer, but I'd prefer to
address that in a separate PR.

---------

Signed-off-by: Yuan Suo <[email protected]>
https://invent.kde.org/qt/clang/llvm/-/commit/8739d89c4c2475eefc4638882311515fc67f9f9b

Git commit 19595357e32b36383a025fddaaf8c38f90658b7b by GitHub (on behalf of Simon Pilgrim) on 16/07/2026 at 14:05..
[X86] combineAdd - remove unnecessary add(a,concat(vpmaddwd(x,y),vpmaddwd(z,w))) -> vpdpwssd(a,concat(x,z),concat(y,w)) fold (#209838)

We already handle vpmaddwd concatenation in combineConcatVectorOps - but
512-bit vpmaddwd is only available on BWI targets, which was missing
from the tests

Technically if there's a target that has AVX512VNNI but not BWI, then
this fold would be useful, but there's no such target and there's a
limit to "what if" optimisations we really need.

A small attempt at cleaning out some not very useful code from
X86ISelLowering.cpp - I'm intending to do a lot more of this for 24.x / #143088
https://invent.kde.org/qt/clang/llvm/-/commit/19595357e32b36383a025fddaaf8c38f90658b7b

Git commit e3d3b3b0cf3672480d6f781a9726cec1e7983d14 by GitHub (on behalf of DylanFleming-arm) on 16/07/2026 at 14:20..
[libc][mathvec] Add loop over scalar for unary FP32 (#199273)

Adds all FP32 unary functions currently supported in scalar LLVM LIBM to
LIBMVEC as loop over scalars.

Includes some cleanup of existing mathvec testing infrastructure
https://invent.kde.org/qt/clang/llvm/-/commit/e3d3b3b0cf3672480d6f781a9726cec1e7983d14

Git commit 1cb7e838cd47ecad4050948c0c907ecb1f466ac3 by GitHub (on behalf of Jay Foad) on 16/07/2026 at 14:32..
[AMDGPU] Simplify some isVALU calls. NFC. (#210057)
https://invent.kde.org/qt/clang/llvm/-/commit/1cb7e838cd47ecad4050948c0c907ecb1f466ac3

Git commit 122efb26893f7e214d0223615c3a4abcc0ac7796 by GitHub (on behalf of Nikolas Klauser) on 16/07/2026 at 14:46..
[libc++] Optimize std::is_heap_until (#209141)

Fixes #208299
https://invent.kde.org/qt/clang/llvm/-/commit/122efb26893f7e214d0223615c3a4abcc0ac7796

Git commit 6e7ff09291652cc784a095707029852e249902f9 by GitHub (on behalf of Jay Foad) on 16/07/2026 at 14:51..
[AMDGPU] Improve debug printing of regunits in SIInsertWaitcnts (#210066)

Use the standard printRegUnit helper.
https://invent.kde.org/qt/clang/llvm/-/commit/6e7ff09291652cc784a095707029852e249902f9

Git commit c7a5b9c0ddd1a5e89a3e8b5a79a916f03e0689b7 by GitHub (on behalf of Sean Clarke) on 16/07/2026 at 14:53..
[LoopIdiomRecognize] Enable clmul optimization for CRC loops (#203405)

The current `optimizeCRCLoop` function always optimizes CRC loops to use
a 256-entry Sarwate lookup table to process a byte at a time (except on
Hexagon and with optsize). However, some targets are able to process
larger chunks of CRC data with carry-less/polynomial multiplication
instructions. Implement this approach in IR for such targets using the
`llvm.clmul` intrinsic. (Only with optsize for the moment, though.)

Like the current CRC loop optimization, this does not apply to the
Hexagon target, despite it having a `pmpyw` instruction.

Assisted-by: Claude Opus 4.8
https://invent.kde.org/qt/clang/llvm/-/commit/c7a5b9c0ddd1a5e89a3e8b5a79a916f03e0689b7

Git commit 73aa9102362f07e2f8fc537b3d2f72062d4a1e68 by GitHub (on behalf of Kseniya Tikhomirova) on 16/07/2026 at 14:54..
[libsycl] Implement nd_range kernel submissions (#206505)

This PR adds core SYCL 2020 nd_range kernel submissions (including
nd_item, group, and sub_group).

Assisted with GH Copilot.

---------

Signed-off-by: Tikhomirova, Kseniya <[email protected]>
https://invent.kde.org/qt/clang/llvm/-/commit/73aa9102362f07e2f8fc537b3d2f72062d4a1e68

Git commit 0852ac5649a0ac6959af6fbc2e4cf6cf4bad7238 by GitHub (on behalf of Kai Nacke) on 16/07/2026 at 14:55..
[PowerPC] Remove itinerary IIC_LdStDCBA (#207227)

The itinerary IIC_LdStDCBA is defined and used in scheduling models, but
no instruction is assigned to it. However, since the instruction is
basically retired (not part of the portable Power ISA), and all
scheduling models model it similar to IIC_LdStDCBF, the itinerary can be
removed.
https://invent.kde.org/qt/clang/llvm/-/commit/0852ac5649a0ac6959af6fbc2e4cf6cf4bad7238

Git commit 96cab305c0402e051d16e748911bd469b4ad57fe by GitHub (on behalf of Robert Imschweiler) on 16/07/2026 at 15:01..
[AMDGPU] Improve amdgcn_kill verification (#208220)
https://invent.kde.org/qt/clang/llvm/-/commit/96cab305c0402e051d16e748911bd469b4ad57fe

Git commit 8c22bbde747ce8c577318c3fb6cafaae6f1b8aec by GitHub (on behalf of Craig Topper) on 16/07/2026 at 15:04..
[RISCV][P-ext] Replace lowerBuildVectorAsRV32PNarrowingShift with shuffle lowering. (#209088)

Add an implementation of isExtractSubvectorCheap for extracting
v2i16/v4i8 from a v4i16/v8i8 vector on RV32. This allows the
build_vector+extract_elt pattern to be turned into
shuffle_vector+extract_subvectors.

Assisted-by: Claude
https://invent.kde.org/qt/clang/llvm/-/commit/8c22bbde747ce8c577318c3fb6cafaae6f1b8aec

Git commit 00674044ab7111ce25ac9a613fefeace7dff1125 by GitHub (on behalf of Benjamin Jurk) on 16/07/2026 at 15:15..
[PatternMatch] Add m_SpecificType(Type*) matcher (#208427)

Fixes #208401.
https://invent.kde.org/qt/clang/llvm/-/commit/00674044ab7111ce25ac9a613fefeace7dff1125

Git commit d333df5aeb31ecf867e1a7908a7814485f4aa7f0 by GitHub (on behalf of aokblast) on 16/07/2026 at 15:17..
[LLDB] Print stderr in HandleStop (#208875)

There are some testcases silently fail in stopping hook without
returning anything on FreeBSD. We dump the error from the handler to
make it easier to debug.
https://invent.kde.org/qt/clang/llvm/-/commit/d333df5aeb31ecf867e1a7908a7814485f4aa7f0

Git commit 7fc7d8d5d2c10db1cff7f53175ad4cb158f0890d by GitHub (on behalf of forking-google-bazel-bot[bot]) on 16/07/2026 at 15:20..
[Bazel] Fixes e3d3b3b (#210068)

This fixes e3d3b3b0cf3672480d6f781a9726cec1e7983d14.

Buildkite error link:
https://buildkite.com/llvm-project/upstream-bazel/builds?commit=e3d3b3b0cf3672480d6f781a9726cec1e7983d14

Co-authored-by: Google Bazel Bot <[email protected]>
https://invent.kde.org/qt/clang/llvm/-/commit/7fc7d8d5d2c10db1cff7f53175ad4cb158f0890d

Git commit 638d6736400891b03f211487ccb643a49f14f743 by GitHub (on behalf of Zeyi Xu) on 16/07/2026 at 15:28..
[LifetimeSafety] Add better lifetimebound fix-it spelling for C (#209967)

Use `__attribute__((lifetimebound))` as the default lifetimebound fix-it
spelling in pre-C23 C.
C++ and C23 continue to use `[[clang::lifetimebound]]` as the default
spelling.

When no explicit lifetimebound macro is configured, pre-C23 C only looks
for `__attribute__((lifetimebound))`, C++ and C23 keep looking for
`[[clang::lifetimebound]]` macros first, then
`__attribute__((lifetimebound))`.

Follow up of https://github.com/llvm/llvm-project/pull/204045
https://invent.kde.org/qt/clang/llvm/-/commit/638d6736400891b03f211487ccb643a49f14f743

Git commit 9ec7974432975674b0128ea85a54cb3e5b0ffa75 by GitHub (on behalf of LU-JOHN) on 16/07/2026 at 15:32..
[AMDGPU] Update libc/runtime target to use amdgpu (#210032)

Update libc/runtime targets to use "amdgpu" instead of "amdgcn". This
change is necessary to do libc testing with a
check-libc-amdgpu-amd-amdhsa target. The previous
check-libc-amdgcn-amd-amdhsa target already stopped working after
https://github.com/llvm/llvm-project/pull/209247.

Signed-off-by: John Lu <[email protected]>
https://invent.kde.org/qt/clang/llvm/-/commit/9ec7974432975674b0128ea85a54cb3e5b0ffa75

Git commit dfa1360735f5a834ce96c8c63594e2cdc3bfe0ea by GitHub (on behalf of Ayokunle Amodu) on 16/07/2026 at 15:42..
[CIR][AMDGPU] Add support for AMDGCN cos builtins (#197806)

Adds codegen for the following AMDGCN cos builtins:

- __builtin_amdgcn_cosf (float)
- __builtin_amdgcn_cosh (half)
- __builtin_amdgcn_cos_bf16 (bfloat16)

These are lowered to the corresponding `llvm.amdgcn.cos` intrinsic.
https://invent.kde.org/qt/clang/llvm/-/commit/dfa1360735f5a834ce96c8c63594e2cdc3bfe0ea

Git commit 8f0a479f46c98e1a098319786d0848e4b9e0a98d by GitHub (on behalf of jandrovins) on 16/07/2026 at 15:59..
[Instrumentor] Fix indirect arguments on gpu targets (#209835)

When an argument is passed indirectly,
`IRTCallDescription::createLLVMCall` spills it into a temporary alloca
and passes that alloca address to the runtime call. The alloca is
created in the target's alloca address space. On GPUs it is not address
space 0, which is what the runtime expects for indirect parameters. The
alloca pointer was passed as the argument, so its type didn't match the
call's declared parameter type, causing a "bad signature" assertion in
`CallInst::init` when instrumenting GPU code that requires indirection.
The solution is therefore to cast this pointer to address space 0 before
passing it the the instrumentor call.
https://invent.kde.org/qt/clang/llvm/-/commit/8f0a479f46c98e1a098319786d0848e4b9e0a98d

Git commit 689fa1d6d74a465130ad0b1705813e14d1cd5b1a by GitHub (on behalf of Hardik Kumar) on 16/07/2026 at 16:00..
Fix Unintialised char *CommentLexer resulting in potential UB. (#210067)

The patch intialised the  pointer to a .
Previously the pointer would have some garbage value which might have
resulted in UB.

closes #210034
https://invent.kde.org/qt/clang/llvm/-/commit/689fa1d6d74a465130ad0b1705813e14d1cd5b1a

Git commit f57bdc62619417a2ee9076ac05db9d115c3ba464 by GitHub (on behalf of Florian Hahn) on 16/07/2026 at 16:03..
[ConstraintElim] Handle (X & Y) <s 0 as X <s 0 && Y <s 0. (#210033)

InstCombine canonicalizes X <s 0 && Y <s 0 as (X & Y) <s 0. Teach
ConstraintElimination to recover the signed-negative information by
looking through compares of binary ANDs.

Alive2 Proof: https://alive2.llvm.org/ce/z/R6DWqD

Depends on https://github.com/llvm/llvm-project/pull/209743 

PR: https://github.com/llvm/llvm-project/pull/210033
https://invent.kde.org/qt/clang/llvm/-/commit/f57bdc62619417a2ee9076ac05db9d115c3ba464

Git commit c9ec54a8a781166546c1001a8bfb2448fbacda44 by GitHub (on behalf of Craig Topper) on 16/07/2026 at 16:04..
[RISCV] Add DUMMY_REG_PAIR_WITH_X0->GPR copy to copyPhysReg. (#209964)

The register coalescer can convert (extract_subreg x0_pair, sub_gpr_odd)
to a read of DUMMY_REG_PAIR_WITH_X0. This patch treats it the same as a
copy from X0.

I don't know if this is a good long term fix, but I think it's ok for
now. I will continue to look for other options.
https://invent.kde.org/qt/clang/llvm/-/commit/c9ec54a8a781166546c1001a8bfb2448fbacda44

Git commit a5b70908d9cacd0aa1c918c920e8a6118b4340ce by GitHub (on behalf of Shilei Tian) on 16/07/2026 at 16:13..
[AMDGPU] Guard more intrinsics with target features (#203956)
https://invent.kde.org/qt/clang/llvm/-/commit/a5b70908d9cacd0aa1c918c920e8a6118b4340ce

Git commit 53dae32b8a3e973a1677bd2745b9128741d69f59 by GitHub (on behalf of Kseniya Tikhomirova) on 16/07/2026 at 16:21..
[libsycl] fix nd_range template resolution (#210081)

Signed-off-by: Tikhomirova, Kseniya <[email protected]>
https://invent.kde.org/qt/clang/llvm/-/commit/53dae32b8a3e973a1677bd2745b9128741d69f59

Git commit c78de0baf7000e535fee63aca59ab4894ecde7f7 by Shilei Tian on 16/07/2026 at 16:33..
[NFC] Pre-commit a test case for a SimplifyCFG issue
https://invent.kde.org/qt/clang/llvm/-/commit/c78de0baf7000e535fee63aca59ab4894ecde7f7

Git commit 71956dc466232dd8bd3bf81657a3360d4328f053 by Shilei Tian on 16/07/2026 at 16:33..
[SimplifyCFG] Do not thread branches into uncontrolled convergent regions

SimplifyCFG's foldCondBranchOnValueKnownInPredecessor can thread an edge past
a block that acts as a reconvergence point. If the threaded destination reaches
an uncontrolled convergent operation before returning to the threaded-through
block, the transform can change which dynamic instance of the convergent
operation is executed.

Add a conservative destination scan for this fold and skip the threading
candidate when it can reach an uncontrolled convergent call before returning
to the original block. Controlled convergent operations using convergence
control tokens are left alone.

Fixes ROCM-26496.
https://invent.kde.org/qt/clang/llvm/-/commit/71956dc466232dd8bd3bf81657a3360d4328f053