This was removed in D39932 but turned out this is actually needed
because runtimes such as compiler-rt and libc++ rely on common options
processing for setting certain flags such as -ffunction-sections and
-fdata-sections.

Differential Revision: https://reviews.llvm.org/D45507

llvm-svn: 329778

[X86] Remove 128/256-bit masked pmaddubsw and pmaddwd intrinsics. Replace 512-bit masked intrinsic with unmasked intrinsic and a select.

The 128/256-bit versions were no longer used by clang. It uses the legacy SSE/AVX2 version and a select. The 512-bit was changed to the same for consistency.

llvm-svn: 329774

[X86] In X86FlagsCopyLowering, when rewriting a memory setcc we need to emit an explicit MOV8mr instruction.

Previously the code only knew how to handle setcc to a register.

This should fix a crash in the chromium build.

llvm-svn: 329771

llvm-svn: 329769

With -fno-plt, for example, calls to printf when getting converted to puts
still use the PLT. This patch checks for the metadata "RtLibUseGOT" and
annotates the declaration with the right attributes.

Differential Revision: https://reviews.llvm.org/D45180

llvm-svn: 329768

llvm-svn: 329767

With -fno-plt, global value references can use GOTPCREL and RIP must be used.

Differential Revision: https://reviews.llvm.org/D45460

llvm-svn: 329765

Author: Samuel Pitoiset

ds_read_b128 and ds_write_b128 have been recently enabled
under the amdgpu-ds128 option because the performance benefit
is unclear.

Though, using 128-bit loads/stores for the local address space
appears to introduce regressions in tessellation shaders. Not
sure what is broken, but as ds_read_b128/ds_write_b128 are not
enabled by default, just introduce a global option and enable
128-bit only if requested (until it's fixed/used correctly).

v2: - fix regressions in merge-stores.ll and multiple_tails.ll

Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=105464
llvm-svn: 329764

llvm-svn: 329763

Summary:
When inserting MOVs to avoid Falkor HWPF collisions, the non-base
register operand of load instructions (e.g. a register offset) was not
being considered live, so it could potentially have been used as a
scratch register, clobbering the actual offset value.

Reviewers: mcrosier

Subscribers: rengolin, javed.absar, kristof.beyls, llvm-commits

Differential Revision: https://reviews.llvm.org/D45502

llvm-svn: 329761

This is based on an example that was recently posted on llvm-dev:

void *propagate_null(void* b, int* g) {
  if (!b) {
    return 0;
  }
  (*g)++;
  return b;
}

https://godbolt.org/g/xYk3qG

The original code or constant propagation in other passes has obscured the fact 
that the phi can be removed completely.

Differential Revision: https://reviews.llvm.org/D45448

llvm-svn: 329755

Summary:
The verification rules for the intrinsics for atomic memcpy, atomic memmove,
and atomic memset are basically code clones. This change merges their verification
rules into a single block to remove duplication.

llvm-svn: 329753

Summary:
Darwin dynamic linker can handle weak symbols in ConstDataSection.
ReadonReadOnlyWithRel symbols should be emitted in ConstDataSection
instead of normal DataSection.

rdar://problem/39298457

Reviewers: dexonsmith, kledzik

Subscribers: llvm-commits

Differential Revision: https://reviews.llvm.org/D45472

llvm-svn: 329752

Summary:
A simple refactor to remove duplicate code in the definitions of
MemSetInst, AtomicMemSetInst, and AnyMemSetInst. Introduce a
templated base class that contains all of the methods unique to
a memset intrinsic, and derive these three classes from that.

llvm-svn: 329747

Recommit r329716 "Add missing nullptr check before getSection() to AArch64MachObjectWriter::recordRelocation"

This commit fixes the bot failures that were coming up before with r329716.

The fix was to move the check for "isInSection()" inside of the if condition
and emit the error there instead of waiting to get past the unreachable statement.

This should work in debug and release builds now.

llvm-svn: 329746

Summary:
A simple refactor to remove duplicate code in the definitions of
MemTransferInst, AtomicMemTransferInst, and AnyMemTransferInst.
Introduce a templated base class that contains all of the methods
unique to a memory transfer intrinsic, and derive these three
classes from that.

llvm-svn: 329744

rdar://39175175

llvm-svn: 329743

Reviewers: craig.topper, zvi, echristo

Reviewed By: craig.topper

Differential Revision: https://reviews.llvm.org/D45055

llvm-svn: 329742

llvm-svn: 329736

[X86] Change the name string for the newly add DF flag register to 'dirflag' to match the clobber name supported by clang for MS inline assembly.

This should fix the failure found by Chromium reported here https://bugs.chromium.org/p/chromium/issues/detail?id=831158

The test case will be added in clang.

llvm-svn: 329734

llvm-svn: 329733

On 2nd reading, putting the C example after the bit about
multiple regions makes this flow better.

llvm-svn: 329732

Summary: Noticed by Andrea Di Biagio while reviewing r329369

Reviewers: whitequark, harlanhaskins

Reviewed By: harlanhaskins

Subscribers: llvm-commits, abergmeier-dsfishlabs

Differential Revision: https://reviews.llvm.org/D45496

llvm-svn: 329731

llvm-svn: 329729

Revert 329716 "Add missing nullptr check before getSection() to AArch64MachObjectWriter::recordRelocation"

This broke a bunch of bots so I'm reverting while I figure it out.

llvm-svn: 329728

  
This is copied from Andrea's text in PR36875:
https://bugs.llvm.org/show_bug.cgi?id=36875

As noted there, this is a hack...but it's a good one!
It's important to show potential workflows up-front
with examples, so customers can copy and experiment
with them.

llvm-svn: 329726

llvm-svn: 329724

There was a missing not line. Also, tail call before ret -> call before ret.

llvm-svn: 329723

Delay printing the newline until after the opening bracket was
printed, e.g.
  BUNDLE implicit-def $r1, implicit-def $r21, implicit $r1 {
    renamable $r1 = S2_asr_i_r renamable $r1, 1
    renamable $r21 = A2_tfrsi 0
  }
instead of
  BUNDLE implicit-def $r1, implicit-def $r21, implicit $r1
 {    renamable $r1 = S2_asr_i_r renamable $r1, 1
    renamable $r21 = A2_tfrsi 0
  }

llvm-svn: 329719

Caused a build failure in check-tsan.

llvm-svn: 329718

There was missing nullptr check before a call to getSection() in
recordRelocation. This would result in a segfault in code like the attached
test.

This adds the missing check and a test which makes sure we get the expected 
error output.

llvm-svn: 329716

Summary:
We would like the UMR debugging tool[0] to be able to provide
disassembly for currently live waves based on plain memory
dumps, and we want to leverage the LLVM disassembler for this.

This mostly works, except that UMR clearly can't provide real
symbol info, so it wants to set DisInfo == nullptr.

[0] https://cgit.freedesktop.org/amd/umr/

Reviewers: arsenm, rampitec, artem.tamazov, dp

Subscribers: kzhuravl, wdng, yaxunl, dstuttard, tpr, t-tye, llvm-commits

Differential Revision: https://reviews.llvm.org/D45477

Change-Id: Ibb2c5af2e66f2e100b4702fd81308e1932bc4ee6
llvm-svn: 329715

llvm-svn: 329712

llvm-svn: 329711

llvm-svn: 329709

[llvm-mca] Move the logic that prints dispatch unit statistics from BackendStatistics to its own view.

This patch moves the logic that collects and analyzes dispatch events to the
DispatchStatistics view.

Added flag -dispatch-stats to print statistics related to the dispatch logic.

llvm-svn: 329708

llvm-svn: 329707

Summary:
This type is created on-demand and used as the base type for array
ranges. Since it is "special", its construction did not go through the
createTypeDIE function and so it was never inserted into the accelerator
table, although it clearly belongs there.

I add an explicit addAccelType call to insert it into the table.

During review, we also decided to rename the type to something more
unique to avoid confusion in case the user has own "sizetype" type. The
new name for the type size __ARRAY_SIZE_TYPE__.

Reviewers: JDevlieghere, aprantl, dblaikie

Subscribers: llvm-commits

Differential Revision: https://reviews.llvm.org/D45445

llvm-svn: 329705

llvm-svn: 329704

Failed<ErrorInfoBase>() did not compile, because it was attempting to
create a copy of the Error object when passing it to the nested matcher,
which was not possible because ErrorInfoBase is abstract.

This commit fixes the problem by making sure we pass the ErrorInfo
object by reference, which also improves the handling of non-abstract
objects, as we avoid potentially slicing an object during the copy.

llvm-svn: 329703