Conflicting debug info for function arguments causes hard-to-debug
assertions in the DWARF backend, so the Verifier should reject it.
For performance reasons this only checks function arguments from
non-inlined debug intrinsics for now.

rdar://problem/30520286

llvm-svn: 295749

Summary:
Rework the code that was sinking/duplicating (icmp and, 0) sequences
into blocks where they were being used by conditional branches to form
more tbz instructions on AArch64.  The new code is more general in that
it just looks for 'and's that have all icmp 0's as users, with a target
hook used to select which subset of 'and' instructions to consider.
This change also enables 'and' sinking for X86, where it is more widely
beneficial than on AArch64.

The 'and' sinking/duplicating code is moved into the optimizeInst phase
of CodeGenPrepare, where it can take advantage of the fact the
OptimizeCmpExpression has already sunk/duplicated any icmps into the
blocks where they are used.  One minor complication from this change is
that optimizeLoadExt needed to be updated to always mark 'and's it has
determined should be in the same block as their feeding load in the
InsertedInsts set to avoid an infinite loop of hoisting and sinking the
same 'and'.

This change fixes a regression on X86 in the tsan runtime caused by
moving GVNHoist to a later place in the optimization pipeline (see
PR31382).

Reviewers: t.p.northover, qcolombet, MatzeB

Subscribers: aemerson, mcrosier, sebpop, llvm-commits

Differential Revision: https://reviews.llvm.org/D28813

llvm-svn: 295746

Differential Revision: http://reviews.llvm.org/D29913

llvm-svn: 295745

llvm-svn: 295740

There appears never to have been a time that the reference was updated.

llvm-svn: 295739

To avoid depending on kernel headers, we just repeat the single define
we need, which is likely never going to change.

Patch by Joakim Sindholt <opensource@zhasha.com>

llvm-svn: 295738

Reviewed by Vedant Kumar (D30178)

llvm-svn: 295737

VPERMI2Q doesn't have shuffle decoding from re-materializable constants.

llvm-svn: 295736

llvm-svn: 295735

Add a missing test that I forgot to svn add in my previous commit

llvm-svn: 295734

As i64 isn't a value type on 32-bit targets, we need to fold the VZEXT_LOAD into VPBROADCASTQ.

llvm-svn: 295733

PC isn't allowed in the source operand of t2MOVr, so change the register class
to one without PC. SP handling is slightly trickier and changes depending on if
we're in ARMv8, so do that in checkTargetMatchPredicate.

Differential Revision: https://reviews.llvm.org/D30199

llvm-svn: 295732

llvm-svn: 295731

Summary: On windows 10, the ucrt DLL is performing allocations before the function hooking and there are multiple allocations not handled by Asan. When a free occur at the end of the process, asan is reporting desallocations not malloc-ed.

Reviewers: rnk, kcc

Reviewed By: rnk, kcc

Subscribers: kcc, llvm-commits, kubamracek, chrisha, dberris

Differential Revision: https://reviews.llvm.org/D25946

llvm-svn: 295730

As i64 isn't a value type on 32-bit targets, we fail to fold the VZEXT_LOAD into VPBROADCASTQ.

Also shows that we're not decoding VPERMIV3 shuffles very well....

llvm-svn: 295729

Summary:
Historically, NetBSD, FreeBSD and OpenBSD have defined the macro ABICALLS in
the preprocessor when -mabicalls is in effect.

Mainline GCC later defined __mips_abicalls when -mabicalls is in effect.

This patch teaches the preprocessor to define these macros when appropriate.

NetBSD does not require the ABICALLS macro.

This resolves PR/31694.

Thanks to Sean Bruno for highlighting this issue!

Reviewers: slthakur, seanbruno

Reviewed By: seanbruno

Subscribers: joerg, brad, emaste, seanbruno, cfe-commits

Differential Revision: https://reviews.llvm.org/D29032

llvm-svn: 295728

Patch fixes PR32024.

Sections that were not marked as Live has null output section.
Previously we tried to access that field and segfaulted.

Differential revision: https://reviews.llvm.org/D30188

llvm-svn: 295727

We only ever use the wrapped domain of AccessSchedule, so stop
creating an entire union_map and then pulling the domain out.

Reviewers: grosser
Tags: #polly

Contributed-by: Siddharth Bhat <siddu.druid@gmail.com>

Differential Revision: https://reviews.llvm.org/D30179

llvm-svn: 295726

r295240 tweaked LLD to generate a symbol table when passing in
--export-dynamic, even when creating static executables. Add a test to
make sure this never regresses.

Reviewed by:	ruiu, rafael
Differential Revision:	https://reviews.llvm.org/D30175

llvm-svn: 295725

llvm-svn: 295724

This matches what is already done during shuffle lowering and helps prevent the need for a zero-vector in cases where shuffles match both patterns.

llvm-svn: 295723

Previously we evaluated the values of LMA incorrectly for next cases:

.text : AT(ADDR(.text) - 0xffffffff80000000) { ... }
.data : AT(ADDR(.data) - 0xffffffff80000000) { ... }
.init.begin : AT(ADDR(.init.begin) - 0xffffffff80000000) { ... }

Reason was that we evaluated offset when VA was not assigned. For case above
we ended up with 3 loads that has similar LMA and it was incorrect.
That is critical for linux kernel.

Patch updates the offset after VA calculation. That fixes the issue.

Differential revision: https://reviews.llvm.org/D30163

llvm-svn: 295722

Currently just contains one case where we combine to VZEXT_MOVL instead of VZEXT which would avoid the need for a zero vector to be generated

llvm-svn: 295721

Previously LLD would error out just "ld.lld: error: unable to move location counter backward"
What does not really reveal the place of issue,
Patch adds location to the output.

Differential revision: https://reviews.llvm.org/D30187

llvm-svn: 295720

Summary:
This is a fix for assertion failure in
`getInverseMinMaxSelectPattern` when ABS is passed in as a select pattern.

We should not be invoking the simplification rule for
ABS(MIN(~ x,y))) or ABS(MAX(~x,y)) combinations.

Added a test case which would cause an assertion failure without the patch.

Reviewers: sanjoy, majnemer

Subscribers: llvm-commits

Differential Revision: https://reviews.llvm.org/D30051

llvm-svn: 295719

Differential Revision: https://reviews.llvm.org/D30189

llvm-svn: 295718

llvm-svn: 295717

For now, we hardcode a BLX instruction, and generate an ADJCALLSTACKDOWN/UP pair
with amount 0.

llvm-svn: 295716

The check doesn't really know if the code it is warning about came before
or after the header guard, so phrase it more neutral instead of complaining
about code before the header guard. The location for the warning is still
not optimal, but I don't think fixing that is worth the effort, the
preprocessor doesn't give us a better location.

Differential Revision: https://reviews.llvm.org/D30191

llvm-svn: 295715

llvm-svn: 295714

Implement the -polly-delicm pass. The pass intends to undo the
effects of LoopInvariantCodeMotion (LICM) which adds additional scalar
dependencies into SCoPs. DeLICM will try to map those scalars back to
the array elements they were promoted from, as long as the array
element is unused.

The is the main patch from the DeLICM/DePRE patch series. It does not
yet undo GVN PRE for which additional information about known values
is needed and does not handle PHI write accesses that have have no
target. As such its usefulness is limited. Patches for these issues
including regression tests for error situatons will follow.

Reviewers: grosser

Differential Revision: https://reviews.llvm.org/D24716

llvm-svn: 295713

Summary:
The code was attempting to copy the shared pointer member in order to
guarantee atomicity, but this is not enough. Instead, protect the
pointer with a proper read-write mutex.

This bug was present here for a long time, but my recent refactors must
have altered the timings slightly, such that now this fails fairly often
when running the tests: the test runner runs the "log disable" command
just as the thread monitoring the lldb-server child is about to report
that the server has exited.

I add a test case for this. It's not possible to reproduce the race
deterministically in normal circumstances, but I have verified that
before the fix, the test failed when run under tsan, and was running
fine afterwards.

Reviewers: clayborg, zturner

Subscribers: lldb-commits

Differential Revision: https://reviews.llvm.org/D30168

llvm-svn: 295712

use llvm::to_string instead of std:: version.

llvm-svn: 295711

case where the class template has a parameter pack.

Checking of the template arguments expects an "as-written" template argument
list, which in particular does not have any parameter packs. So flatten the
packs into separate arguments before passing them in.

llvm-svn: 295710

They are all covered by the SSSE3 intrinsics test with SSSE3, AVX, and AVX512 command lines.

llvm-svn: 295708

[X86] Remove sse4.2 intrinsic tests from the avx intrinsics test file. Fix some other consistency issues.

They are all covered by the SSE4.2 intrinsics test with SSE4.2, AVX, and AVX512 command lines.

Merge sse42.ll into the other intrinsics test. Rename sse42_64.ll to be named like other intrinsic tests.

llvm-svn: 295707

They are all covered by the SSE4.1 intrinsics test with SSE4.1, AVX, and AVX512 command lines.

llvm-svn: 295706

They are all covered by the SSE3 intrinsics test with SSE2, AVX, and AVX512 command lines.

llvm-svn: 295705

The new method introduced under "-lsr-exp-narrow" option (currenlty set to true).

Summary:

The method is based on registers number mathematical expectation and should be
 generally closer to optimal solution.
Please see details in comments to
 "LSRInstance::NarrowSearchSpaceByDeletingCostlyFormulas()" function
 (in lib/Transforms/Scalar/LoopStrengthReduce.cpp).

Reviewers: qcolombet

Differential Revision: http://reviews.llvm.org/D29862

From: Evgeny Stupachenko <evstupac@gmail.com>
llvm-svn: 295704

Previously ASSERT we implemented returned expression value.
Ex:
. = ASSERT(0x100);
would set Dot value to 0x100

Form of assert when it is assigned to Dot was implemented for 
compatibility with very old GNU ld which required it.
Some scripts in the wild, including linux kernel scripts 
use such ASSERTs at the end for doing different checks.

Currently we fail with "unable to move location counter backward"
for such scripts. Patch changes ASSERT to return location counter 
value to fix that.

Differential revision: https://reviews.llvm.org/D30171

llvm-svn: 295703