Preparatory patch for MachO feature.

Support for masked N-D vector transfer ops will be added in a subsequent commit.

Differential Revision: https://reviews.llvm.org/D101132

LLVM should be smarter about *known* malloc's alignment and this knowledge may enable other optimizations.

Originally started as LLVM patch - https://reviews.llvm.org/D100862 but this logic should be really in Clang.

Reviewed By: rjmccall

Differential Revision: https://reviews.llvm.org/D100879

This commit adds support for broadcast dimensions in permutation maps of vector transfer ops.

Also fixes a bug in VectorToSCF that generated incorrect in-bounds checks for broadcast dimensions.

Differential Revision: https://reviews.llvm.org/D101019

This reverts commit 4f5da356.

This causes some  buildbot failures, e.g.
https://lab.llvm.org/buildbot/#/builders/139/builds/3019

Use SCF for loops instead of Affine for loops.

Differential Revision: https://reviews.llvm.org/D101013

Updates the docs to reflect the changes introduced to the `add_mlir_doc`
CMake macro with https://reviews.llvm.org/D100517.

If we are using a simplified value, we need to add an extra
dependency this value , because changes to the class of the
simplified value may require us to invalidate any decision based on
that value.

This is done by adding such values as additional users, however the
current code does not excludes temporary instructions.

At the moment, this means that we miss those dependencies for
phi-of-ops, because they are temporary instructions at this point. We
instead need to add the extra dependencies to the root instruction of
the phi-of-ops.

This patch pushes the responsibility of adding extra users to the
callers of createExpression & performSymbolicEvaluation. At those
points, it is clearer which real instruction to pick.

Alternatively we could either pass the 'real' instruction as additional
argument or use another map, but I think the approach in the patch makes
things a bit easier to follow.

Fixes PR35074.

Reviewed By: asbirlea

Differential Revision: https://reviews.llvm.org/D99987

This commit adds support for dimension permutations in permutation maps of vector transfer ops.

Differential Revision: https://reviews.llvm.org/D101007

Fix style/clang-tidy warning, trim stale includes and forward
declarations, and cleanup/fix stale comments.

Differential Revision: https://reviews.llvm.org/D101021

Strided 1D vector transfer ops are 1D transfers operating on a memref dimension different from the last one. Such transfer ops do not accesses contiguous memory blocks (vectors), but access memory in a strided fashion. In the absence of a mask, strided 1D vector transfer ops can also be lowered using matrix.column.major.* LLVM instructions (in a later commit).

Subsequent commits will extend the pass to handle the remaining missing permutation maps (broadcasts, transposes, etc.).

Differential Revision: https://reviews.llvm.org/D100946

Make following function return void:

    addLabel()
    addSectionLabel()
    addSectionDelta()

This aligns with other attributes adding functions.

Reviewed By: dblaikie

Differential Revision: https://reviews.llvm.org/D101022

ConstantFoldingMIRBuilder was an experiment which is not used for
anything. The constant folding functionality is now part of
CSEMIRBuilder.

Differential Revision: https://reviews.llvm.org/D101050

EMITBKEY is emitted for PAC-RET+bkey, which is a non machine instructions.

PR: 49957

Reviewed By: eugenis

Differential Revision: https://reviews.llvm.org/D100996

When building preamble, clangd truncates file contents. This yielded
errnous warnings in some cases.

This patch fixes the issue by turning off no-newline-at-eof warnings whenever
the file has more contents than the preamble.

Fixes https://github.com/clangd/clangd/issues/744.

Differential Revision: https://reviews.llvm.org/D100501

Fixes PR47627

This fix suppresses rerolling a loop which has an unrerollable
instruction.

Sample IR for the explanation below:

```
define void @foo([2 x i32]* nocapture %a) {
entry:
  br label %loop

loop:
  ; base instruction
  %indvar = phi i64 [ 0, %entry ], [ %indvar.next, %loop ]

  ; unrerollable instructions
  %stptrx = getelementptr inbounds [2 x i32], [2 x i32]* %a, i64 %indvar, i64 0
  store i32 999, i32* %stptrx, align 4

  ; extra simple arithmetic operations, used by root instructions
  %plus20 = add nuw nsw i64 %indvar, 20
  %plus10 = add nuw nsw i64 %indvar, 10

  ; root instruction 0
  %ldptr0 = getelementptr inbounds [2 x i32], [2 x i32]* %a, i64 %plus20, i64 0
  %value0 = load i32, i32* %ldptr0, align 4
  %stptr0 = getelementptr inbounds [2 x i32], [2 x i32]* %a, i64 %plus10, i64 0
  store i32 %value0, i32* %stptr0, align 4

  ; root instruction 1
  %ldptr1 = getelementptr inbounds [2 x i32], [2 x i32]* %a, i64 %plus20, i64 1
  %value1 = load i32, i32* %ldptr1, align 4
  %stptr1 = getelementptr inbounds [2 x i32], [2 x i32]* %a, i64 %plus10, i64 1
  store i32 %value1, i32* %stptr1, align 4

  ; loop-increment and latch
  %indvar.next = add nuw nsw i64 %indvar, 1
  %exitcond = icmp eq i64 %indvar.next, 5
  br i1 %exitcond, label %exit, label %loop

exit:
  ret void
}
```

In the loop rerolling pass, `%indvar` and `%indvar.next` are appended
to the `LoopIncs` vector in the `LoopReroll::DAGRootTracker::findRoots`
function.

Before this fix, two instructions with `unrerollable instructions`
comment above are marked as `IL_All` at the end of the
`LoopReroll::DAGRootTracker::collectUsedInstructions` function,
as well as instructions with `extra simple arithmetic operations`
comment and `loop-increment and latch` comment. It is incorrect
because `IL_All` means that the instruction should be executed in all
iterations of the rerolled loop but the `store` instruction should
not.

This fix rejects instructions which may have side effects and don't
belong to def-use chains of any root instructions and reductions.

See https://bugs.llvm.org/show_bug.cgi?id=47627 for more information.

Differential Revision: https://reviews.llvm.org/D100820

The previous condition in the assert was over strict. We ought to allow
the same immidiate value being loaded more than once. The intention for
the assert is to check the same AMX register uses multiple different
immidiate shapes. So this fix supposes to be NFC.

Reviewed By: LuoYuanke

Differential Revision: https://reviews.llvm.org/D101124

We request no intersections between AMX instructions and their shapes'
def when we insert ldtilecfg. However, this is not always ture resulting
from not only users don't follow AMX API model, but also optimizations.

This patch adds a mechanism that tries to hoist AMX shapes' def as well.
It only hoists shapes inside a BB, we can improve it for cases across
BBs in future. Currently, it only hoists shapes of which all sources' def
above the first AMX instruction. We can improve for the case that only
source that moves an immediate value to a register below AMX instruction.

Differential Revision: https://reviews.llvm.org/D101067

Add __uintr_frame structure and use UIRET instruction for functions with
x86 interrupt calling convention when UINTR is present.

Reviewed By: LuoYuanke

Differential Revision: https://reviews.llvm.org/D99708

Differential Revision: https://reviews.llvm.org/D100819

This is mostly NFC except that for end of BB not previous slot is used.
Idx is used to find a def of sibling live interval in that slot.
The def on end of MBB and on previous slot of end MBB should be the same,
so it should be NFC.

Reviewers: reames, qcolombet, MatzeB, wmi, rnk
Reviewed By: rnk
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D100922

I first had a more invasive patch for this (D101069), but while trying
to get that polished for review I realized that lld's current symbol
merging semantics mean that only a very small code change is needed.
So this goes with the smaller patch for now.

This has no effect on projects that build with -fvisibility=hidden
(e.g.  chromium), since these see .private_extern symbols instead.

It does have an effect on projects that build with -fvisibility-inlines-hidden
(e.g. llvm) in -O2 builds, where LLVM's GlobalOpt pass will promote most inline
functions from .weak_definition to .weak_def_can_be_hidden.

Before this patch:

    % ls -l out/gn/bin/clang out/gn/lib/libclang.dylib
    -rwxr-xr-x  1 thakis  staff  113059936 Apr 22 11:51 out/gn/bin/clang
    -rwxr-xr-x  1 thakis  staff   86370064 Apr 22 11:51 out/gn/lib/libclang.dylib
    % out/gn/bin/llvm-objdump --macho --weak-bind out/gn/bin/clang | wc -l
        8291
    % out/gn/bin/llvm-objdump --macho --weak-bind out/gn/lib/libclang.dylib | wc -l
        5698

With this patch:

    % ls -l out/gn/bin/clang out/gn/lib/libclang.dylib
    -rwxr-xr-x  1 thakis  staff  111721096 Apr 22 11:55 out/gn/bin/clang
    -rwxr-xr-x  1 thakis  staff   85291208 Apr 22 11:55 out/gn/lib/libclang.dylib
    thakis@MBP llvm-project % out/gn/bin/llvm-objdump --macho --weak-bind out/gn/bin/clang | wc -l
         725
    thakis@MBP llvm-project % out/gn/bin/llvm-objdump --macho --weak-bind out/gn/lib/libclang.dylib | wc -l
         542

Linking clang becomes a tiny bit faster with this patch:

    x 100    0.67263818    0.77847815    0.69430709    0.69877208   0.017715892
    + 100    0.67209601    0.73323393    0.68600798    0.68917346   0.012824377
    Difference at 95.0% confidence
            -0.00959861 +/- 0.00428661
            -1.37364% +/- 0.613449%
            (Student's t, pooled s = 0.0154648)

This only happens if lld with the patch and lld without the patch are both
linked with an lld with the patch or both linked with an lld without the patch
(...or with ld64). I accidentally linked the lld with the patch with an lld
without the patch and the other way round at first. In that setup, no
difference is found. That makese sense, since having fewer weak imports will
make the linked output a bit faster too. So not only does this make linking
binaries such as clang a bit faster (since fewer exports need to be written to
the export trie by lld), the linked output binary binary is also a bit faster
(since dyld needs to process fewer dynamic imports).

This also happens to fix the one `check-clang` failure when using lld as host
linker, but mostly for silly reasons: See crbug.com/1183336, mostly comment 26.
The real bug here is that c-index-test links all of LLVM both statically and
dynamically, which is an ODR violation. Things just happen to work with this
patch.

So after this patch, check-clang, check-lld, check-llvm all pass with lld as
host linker :)

Differential Revision: https://reviews.llvm.org/D101080

- __got is in --bind output, so print that too (makes the test
  a bit stronger)
- WEAK_DEFINES, BINDS_TO_WEAK are in the mach-o header, so
  --private-header is enough, no need for --all-headers
  (makes the test a bit easier to work with when it fails)

Differential Revision: https://reviews.llvm.org/D101065

9931b1f7 switched this to checking for
the two specific subtargets, instead of the dedicated feature. This
broke supporting functions which force added the feature when emitting
targets that do not actually support them. This stil does not work for
the targets that use the gfx6/7 or gfx10 encodings.

The Linux kernel objtool diagnostic `call without frame pointer save/setup`
arise in multiple instrumentation passes (asan/tsan/gcov). With the mechanism
introduced in D100251, it's trivial to respect the command line
-m[no-]omit-leaf-frame-pointer/-f[no-]omit-frame-pointer, so let's do it.

Fix: https://github.com/ClangBuiltLinux/linux/issues/1236 (tsan)
Fix: https://github.com/ClangBuiltLinux/linux/issues/1238 (asan)

Also document the function attribute "frame-pointer" which is long overdue.

Differential Revision: https://reviews.llvm.org/D101016

Adds the six new concept constrained comparisons.

Differential Revision: https://reviews.llvm.org/D100429

This reverts commit 24c1ed3b.

during overload resolution, even when calling a copy constructor.

 - The register encoding state in the JSON crashlog format changes.
   Update the parser accordingly.
 - Print the register state when printing the symbolicated thread.

We had got it backwards... the minimum version of the target
should be higher than the min version of the object files, presumably
since new platforms are backwards-compatible with older formats.

Fixes PR50078.

The original commit (aa05439c) broke many tests that had inputs too
new for our target platform (10.0). This commit changes the inputs to
target 10.0, which was the simpler thing to do, but we should really
just have our lit.local.cfg default to targeting 10.15... we're not
likely to ever have proper support for the older versions anyway. I will
follow up with a change to that effect.

Differential Revision: https://reviews.llvm.org/D101114

RV32/64:
    grev
    grevi
    gorc
    gorci
    shfl
    shfli
    unshfl
    unshfli

RV64 ONLY:
    grevw
    greviw
    gorcw
    gorciw
    shflw
    shfli     (For non-existing shfliw)
    unshfli   (For non-existing unshfliw)

Reviewed By: craig.topper

Differential Revision: https://reviews.llvm.org/D100830

[clang][nfc] Split getOrCheckAMDGPUCodeObjectVersion

Separates detection of deprecated or invalid code object version from
returning the version. Written to avoid any behaviour change.

Precursor to a revision of D98746.

Reviewed By: yaxunl

Differential Revision: https://reviews.llvm.org/D101077

It looks like there's some old version of gcc that doesn't like this
static_assert (I couldn't reproduce the issue with gcc 8, 9 or 10).
Work around the issue by only checking the static_assert under clang,
which should provide sufficient coverage.

Should hopefully fix this buildbot:
https://lab.llvm.org/buildbot/#/builders/112/builds/5356

This reverts commit aa05439c.

Remove the dependence on standard C++ header
for overloaded math functions in HIP header
since standard C++ header is not available for hipRTC.

Reviewed by: Artem Belevich, Justin Lebar

Differential Revision: https://reviews.llvm.org/D100794