I don't think this test case can be handled correctly on AAPCS64.
The ABI says that the caller passes the address of the return object
in x8.  x8 is a caller-spilled (aka "volatile") register, and the
function is not required to preserve x8 or to copy the address back
into x8 on function exit like the SysV x86_64 ABI does with rax.
(from aapcs64: "there is no requirement for the callee to preserve the
value stored in x8")

From my quick reading of ABISysV_arm64, I worry that it may actually be
using the value in x8 at function exit, assuming it still has the
address of the return object -

    if (is_return_value) {
      // We are assuming we are decoding this immediately after returning from
      // a function call and that the address of the structure is in x8
      reg_info = reg_ctx->GetRegisterInfoByName("x8", 0);

This will work on trivial test programs / examples, but if the function
does another function call, or overwrites x8 as a scratch register, lldb
will provide incorrect values to the user.

ABIMacOSX_arm64 doesn't do this, but it also doesn't flag the value
as unavailable so we're providing incorrect values to the user all
the time.  I expect my fix will be to make ABIMacOSX_arm64 flag
the return value as unretrievable, unless I've misread the ABI.

This is equivalent to the existing `import_name` and `import_module`
attributes which control the import names in the final wasm binary
produced by lld.

This maps the existing

This attribute currently requires a string rather than using the
symbol name for a couple of reasons:

1. Avoid confusion with static and dynamic linking which is
   based on symbol name.  Exporting a function from a wasm module using
   this directive is orthogonal to both static and dynamic linking.
2. Avoids name mangling.

Differential Revision: https://reviews.llvm.org/D70520

Fix for https://bugs.llvm.org/show_bug.cgi?id=40846.

This adds a combine for cases where a (a + b) < a style overflow
check is performed, but with a + b being the result of
uadd.with.overflow, so the overflow result is also already available
and we can just use it. Subsequently GVN/CSE will deduplicate the extracts.

We can run into this situation if you have both a uadd.with.overflow
and a manual add + overflow check in the same function (on the same
operands), in which case GVN will rewrite the add to the with.overflow
result and leave you with this pattern.

The implementation is a bit ugly because I'm handling the various
canonicalization edge cases.

This does not yet handle the negated version of this pattern.

Differential Revision: https://reviews.llvm.org/D58644

This patch introduced additional PointerEscape callbacks after conservative
calls for output parameters. This should not really affect the current
checkers but the upcoming FuchsiaHandleChecker relies on this heavily.

Differential Revision: https://reviews.llvm.org/D71224

If the pointer was loaded/stored before the null check, the check
is redundant and can be removed. For now the optimizers do not
remove the nullptr check, see https://gcc.godbolt.org/z/H2r5GG.
The patch allows to use more nonnull constraints. Also, it found
one more optimization in some PowerPC test. This is my first llvm
review, I am free to any comments.

Differential Revision: https://reviews.llvm.org/D71177

Tests for D71177.

The checker was trying to analyze the body of every method in Objective-C
@implementation clause but the sythesized accessor stubs that were introduced
into it by 2073dd2d have no bodies.

While analyzing code `memcmp(a, NULL, n);', where `a' has an unconstrained
symbolic value, the analyzer was emitting a warning about the *first* argument
being a null pointer, even though we'd rather have it warn about the *second*
argument.

This happens because CStringChecker first checks whether the two argument
buffers are in fact the same buffer, in order to take the fast path.
This boils down to assuming `a == NULL' to true. Then the subsequent check
for null pointer argument "discovers" that `a' is null.

Don't take the fast path unless we are *sure* that the buffers are the same.
Otherwise proceed as normal.

Differential Revision: https://reviews.llvm.org/D71322

Differential Revision: https://reviews.llvm.org/D71321

Fix for https://bugs.llvm.org/show_bug.cgi?id=44236. This code was
originally introduced in rG36512330.
However, the attribute copying was done in the wrong place (in general
call replacement, not thunk generation) and a proper fix was
implemented in D12581.

Previously this code was just unnecessary but harmless (because
FunctionComparator ensured that the attributes of the two functions
are exactly the same), but since byval was changed to accept a type
this copying is actively wrong and may result in malformed IR.

Differential Revision: https://reviews.llvm.org/D71173

Summary:
The following intrinsics for unary narrowing operations are added:
 * @llvm.aarch64.sve.sqxtnb
 * @llvm.aarch64.sve.uqxtnb
 * @llvm.aarch64.sve.sqxtunb
 * @llvm.aarch64.sve.sqxtnt
 * @llvm.aarch64.sve.uqxtnt
 * @llvm.aarch64.sve.sqxtunt

Reviewers: sdesmalen, rengolin, efriedma

Reviewed By: efriedma

Subscribers: tschuett, kristof.beyls, hiraditya, llvm-commits

Tags: #llvm

Differential Revision: https://reviews.llvm.org/D71270

I'm not sure how .. is resolved on Windows. Disable it for now to make
the bots happy again.

This fixes crashes with $noreg operands.

The default characters for trim, ltrim and rtrim are " \t\n\v\f\r" but
only spaces were tested. Test that the others are trimmed as well.

This is an alternate fix for the bug discussed in D70595.
This also includes minimal tests for other in-tree targets
to show the problem more generally.

We check the number of uses as a predicate for whether some
value is free to negate, but that use count can change as we
rewrite the expression in getNegatedExpression(). So something
that was marked free to negate during the cost evaluation
phase becomes not free to negate during the rewrite phase (or
the inverse - something that was not free becomes free).
This can lead to a crash/assert because we expect that
everything in an expression that is negatible to be handled
in the corresponding code within getNegatedExpression().

This patch skips the use check during the rewrite phase.
So we determine that some expression isNegatibleForFree
(identically to without this patch), but during the rewrite,
don't rely on use counts to decide how to create the optimal
expression.

Differential Revision: https://reviews.llvm.org/D70975

directive.

Fixed capturing of the if condition if no modifer was specified in this
condition. Previously could capture it only in outer region and it could
lead to a compiler crash.

When writing out a profile, avoid allocating a page on the stack for the
purpose of writing out zeroes, as some embedded environments do not have
enough stack space to accomodate this.

Instead, use a small, fixed-size zero buffer that can be written
repeatedly.

For a synthetic file with >100,000 functions, I did not measure a
significant difference in profile write times. We are removing a
page-length zero-fill `memset()` in favor of several smaller buffered
`fwrite()` calls: in practice, I am not sure there is much of a
difference. The performance impact is only expected to affect the
continuous sync mode (%c) -- zero padding is less than 8 bytes in all
other cases.

rdar://57810014

Differential Revision: https://reviews.llvm.org/D71323

Extend the virtual working directory test with a few edge cases that are
not currently tested.

According to OpenMP 5.0, if clause can be used in for simd directive. If
condition in the if clause if false, the non-vectorized version of the
loop must be executed.

This is a half-implemented feature that as far as we can tell was
never used by anything since its original inclusion in 2014. This
patch removes it to make remaining the code easier to understand.

Differential Revision: https://reviews.llvm.org/D71310

Summary:
The original design used --foo but the upstream complained that ELF only
options should be -z foo. See https://sourceware.org/ml/binutils/2019-04/msg00151.html
https://sourceware.org/git/?p=binutils-gdb.git;a=commitdiff;h=8bf6d176b0a442a8091d338d4af971591d19922c
made the rename.

Our --force-bti and --pac-plt implement the same functionality, so it
seems wise to be consistent with GNU ld.

Reviewed By: peter.smith

Subscribers: emaste, arichardson, kristof.beyls, llvm-commits

Tags: #llvm

Differential Revision: https://reviews.llvm.org/D71327

Summary:
The current da printer shows the dependence without indicating
which instructions are being considered as the src vs dst. It
also silently ignores call instructions, despite the fact that
they create confused dependence edges to other memory
instructions. This patch addresses these two issues plus a
couple of minor non-functional improvements.

Authored By: bmahjour

Reviewer: dmgreen, fhahn, philip.pfaffe, chandlerc

Reviewed By: dmgreen, fhahn

Tags: #llvm

Differential Revision: https://reviews.llvm.org/D71088

This fixes a crash when debug instructions are in between 2 stores.

The backend already supports the @llvm.minimum and @llvm.maximum
intrinsics, but we had no test cases for those.  Add tests.

[X86] Erase dead LEA instruction after converting it to MOV in FixupLEAPass::processInstrForSlow3OpLEA.

AMDGPU was the last in tree target to use this tablegen mode. I plan to
split up the global intrinsic enum similar to the way that clang
diagnostics are split up today. I don't plan to build on this mode.

Reviewers: arsenm, echristo, efriedma

Reviewed By: echristo

Differential Revision: https://reviews.llvm.org/D71318

Before z14, we did not have any FMA instruction for 128-bit
floating-point, so the @llvm.fma.f128 intrinsic needs to be
expanded to a libcall on those platforms.

This worked correctly for regular FMA, but was implemented
incorrectly for the strict version.  This was not noticed
because we did not have test coverage for this case.

This patch fixes that incorrect expansion and adds the
missing test cases.

Summary:
This patch adds a method to determine if a loop is in rotated form (the latch is
an exiting block). It also modifies the getLoopGuardBranch method to use this
new method. This method can also be used in Loopfusion. Once this patch lands I
will make the corresponding changes there.

Reviewers: jdoerfert, Meinersbur, dmgreen, etiotto, Whitney, fhahn, hfinkel

Reviewed By: Meinersbur

Subscribers: hiraditya, llvm-commits

Tags: #llvm

Differential Revision: https://reviews.llvm.org/D65958

This simplifies code where no extra details are required
Also don't write out detail when it is empty.

Differential Revision: https://reviews.llvm.org/D71347

There are a few places that check specific string attributes have
particular values, and assert if they are something else. The verifier
should catch these kinds of cases.

Sometimes the return value of a comparison operator call is
`UnkownVal`. Since no assumptions can be made on `UnknownVal`,
this leeds to keeping impossible execution paths in the
exploded graph resulting in poor performance and false
positives. To overcome this we replace unknown results of
iterator comparisons by conjured symbols.

Differential Revision: https://reviews.llvm.org/D70244

On some edge cases such as Chromium compiled with full instrumentation we
have a .text section over twice the size of the maximum branch range and
the instrumented code generation containing many examples of the erratum
sequence. The combination of Thunks and many erratum sequences causes
finalizeAddressDependentContent() to not converge. We end up with:
start
- Thunk Creation (disturbs addresses after thunks, creating more patches)
- Patch Creation (disturbs addresses after patches, creating more thunks)
- goto start

In most images with few thunks and patches the mutual disturbance does not
cause convergence problems. As the .text size and number of patches go up
the risk increases.

A way to prevent the thunk creation from interfering with patch creation is
to round up the size of the thunks to a 4KiB boundary when the
erratum patch is enabled. As the erratum sequence only triggers when an
instruction sequence starts at 0xff8 or 0xffc modulo (4 KiB) by making the
thunks not affect addresses modulo (4 KiB) we prevent thunks from
interfering with the patch.

The patches themselves could be aggregated in the same way that Thunks are
within ThunkSections and we could round up the size in the same way. This
would reduce the number of patches created in a .text section size >
128 MiB but would not likely help convergence problems.

Differential Revision: https://reviews.llvm.org/D71281

fixes (remaining part of) pr44071, other part in D71242

The code to insert patch section merges them with a comparison function that
uses logic of the form:
return (isa<PatchSection>(a) && !isa<PatchSection>(b));
If the PatchSections don't implement classof this check fails if b is also
a SyntheticSection. This can result in the patches being out of range if
the SyntheticSection is big, for example a ThunkSection with lots of thunks.

Differential Revision: https://reviews.llvm.org/D71242

fixes (part of) pr44071