OSDN Git Service

android-x86/external-llvm.git
5 years ago[llvm-objcopy][NFCI] Fix build failure with GCC
Seiya Nuta [Tue, 25 Jun 2019 01:08:21 +0000 (01:08 +0000)]
[llvm-objcopy][NFCI] Fix build failure with GCC

Here is unreachable since the switch statement above is exhaustive.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@364263 91177308-0d34-0410-b5e6-96231b3b80d8

5 years agoAMDGPU/GlobalISel: Fix regbankselect for amdgcn.class
Matt Arsenault [Tue, 25 Jun 2019 01:07:22 +0000 (01:07 +0000)]
AMDGPU/GlobalISel: Fix regbankselect for amdgcn.class

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@364262 91177308-0d34-0410-b5e6-96231b3b80d8

5 years ago[InstCombine][NFC] Add test to show missing fold for icmp ult/uge (shl %x, C2), C1.
Huihui Zhang [Tue, 25 Jun 2019 00:14:02 +0000 (00:14 +0000)]
[InstCombine][NFC] Add test to show missing fold for icmp ult/uge (shl %x, C2), C1.

Summary:
'shl' inequality test

```
  icmp ult/uge (shl %x, C2), C1 iff C1 is power of two
```

can be simplified as 'and' equality test

```
  icmp eq/ne (and %x, (lshr -C1, C2)), 0.
```

Reviewers: lebedev.ri, efriedma

Reviewed By: lebedev.ri

Subscribers: llvm-commits

Tags: #llvm

Differential Revision: https://reviews.llvm.org/D63670

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@364256 91177308-0d34-0410-b5e6-96231b3b80d8

5 years ago[InstCombine] Fold icmp eq/ne (and %x, C), 0 iff (-C) is power of two -> %x u</u...
Huihui Zhang [Tue, 25 Jun 2019 00:09:10 +0000 (00:09 +0000)]
[InstCombine] Fold  icmp eq/ne (and %x, C), 0 iff (-C) is power of two -> %x u</u>= (-C)  earlier.

Summary:
To generate simplified IR, make sure fold
  (X & ~C) ==/!= 0 --> X u</u>= C+1

is scheduled before fold
  ((X << Y) & C) == 0 -> (X & (C >> Y)) == 0.

https://rise4fun.com/Alive/7ZN

Reviewers: lebedev.ri, efriedma, spatel, craig.topper

Reviewed By: lebedev.ri

Subscribers: hiraditya, llvm-commits

Tags: #llvm

Differential Revision: https://reviews.llvm.org/D63505

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@364255 91177308-0d34-0410-b5e6-96231b3b80d8

5 years ago[llvm-objcopy][NFC] Refactor output target parsing
Seiya Nuta [Tue, 25 Jun 2019 00:02:04 +0000 (00:02 +0000)]
[llvm-objcopy][NFC] Refactor output target parsing

Summary:
Use an enum instead of string to hold the output file format in Config.InputFormat and Config.OutputFormat. It's essential to support other output file formats other than ELF.

Reviewers: espindola, alexshap, rupprecht, jhenderson

Reviewed By: rupprecht, jhenderson

Subscribers: jyknight, compnerd, emaste, arichardson, fedor.sergeev, jakehehrlich, MaskRay, llvm-commits

Tags: #llvm

Differential Revision: https://reviews.llvm.org/D63239

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@364254 91177308-0d34-0410-b5e6-96231b3b80d8

5 years agoDataExtractor: use decodeSLEB128 to implement getSLEB128
David Blaikie [Mon, 24 Jun 2019 23:45:18 +0000 (23:45 +0000)]
DataExtractor: use decodeSLEB128 to implement getSLEB128

Should've been NFC, but turns out DataExtractor had better test coverage
for decoding SLEB128 than the decodeSLEB128 did - revealing a couple of
bugs (one in the error handling, another in sign extension). So fixed
those to get the DataExtractor tests passing again.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@364253 91177308-0d34-0410-b5e6-96231b3b80d8

5 years ago[llvm-objcopy][MachO] Fix strict-aliasing warning. NFCI
Seiya Nuta [Mon, 24 Jun 2019 23:39:01 +0000 (23:39 +0000)]
[llvm-objcopy][MachO] Fix strict-aliasing warning. NFCI

Summary:
Use MachOObjectFile::isRelocationScattered instead of reinterpret_cast.

Fixes https://bugs.llvm.org/show_bug.cgi?id=42360

Reviewers: alexshap, rupprecht, jhenderson

Reviewed By: alexshap

Subscribers: dendibakh, bjope, uabelho, jakehehrlich, llvm-commits

Tags: #llvm

Differential Revision: https://reviews.llvm.org/D63699

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@364252 91177308-0d34-0410-b5e6-96231b3b80d8

5 years agoRevert "[NVPTX][NFC] Fix documentation for shfl instructions." The
Tim Shen [Mon, 24 Jun 2019 23:29:20 +0000 (23:29 +0000)]
Revert "[NVPTX][NFC] Fix documentation for shfl instructions." The
original documentation is correct as it matches the C++ builtins.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@364250 91177308-0d34-0410-b5e6-96231b3b80d8

5 years ago[NFC] Fix tests added in r364225 which failed on Windows due to incorrect path separa...
Douglas Yung [Mon, 24 Jun 2019 23:16:32 +0000 (23:16 +0000)]
[NFC] Fix tests added in r364225 which failed on Windows due to incorrect path separators.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@364249 91177308-0d34-0410-b5e6-96231b3b80d8

5 years ago[NVPTX][NFC] Fix documentation for shfl instructions.
Tim Shen [Mon, 24 Jun 2019 23:16:32 +0000 (23:16 +0000)]
[NVPTX][NFC] Fix documentation for shfl instructions.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@364248 91177308-0d34-0410-b5e6-96231b3b80d8

5 years ago[NFC] Add missing consts into memoryaccess_def_iterator
Vitaly Buka [Mon, 24 Jun 2019 22:42:53 +0000 (22:42 +0000)]
[NFC] Add missing consts into memoryaccess_def_iterator

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@364247 91177308-0d34-0410-b5e6-96231b3b80d8

5 years ago[InstCombine] squash is-not-power-of-2 using ctpop
Sanjay Patel [Mon, 24 Jun 2019 22:35:26 +0000 (22:35 +0000)]
[InstCombine] squash is-not-power-of-2 using ctpop

This is the Demorgan'd 'not' of the pattern handled in:
D63660 / rL364153

This is another intermediate IR step towards solving PR42314:
https://bugs.llvm.org/show_bug.cgi?id=42314

We can test if a value is not a power-of-2 using ctpop(X) > 1,
so combining that with an is-zero check of the input is the
same as testing if not exactly 1 bit is set:

(X == 0) || (ctpop(X) u> 1) --> ctpop(X) != 1

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@364246 91177308-0d34-0410-b5e6-96231b3b80d8

5 years agoAMDGPU/GlobalISel: Add tests for regbankselect of v2s16 and/or/xor
Matt Arsenault [Mon, 24 Jun 2019 22:21:02 +0000 (22:21 +0000)]
AMDGPU/GlobalISel: Add tests for regbankselect of v2s16 and/or/xor

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@364244 91177308-0d34-0410-b5e6-96231b3b80d8

5 years ago[SLP] NFC: Fixed typo in comment
Vasileios Porpodas [Mon, 24 Jun 2019 21:40:48 +0000 (21:40 +0000)]
[SLP] NFC: Fixed typo in comment

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@364237 91177308-0d34-0410-b5e6-96231b3b80d8

5 years agoInstCombine: Preserve nuw when reassociating nuw ops [3/3]
Matt Arsenault [Mon, 24 Jun 2019 21:37:03 +0000 (21:37 +0000)]
InstCombine: Preserve nuw when reassociating nuw ops [3/3]

Alive says this is OK.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@364235 91177308-0d34-0410-b5e6-96231b3b80d8

5 years agoInstCombine: Preserve nuw when reassociating nuw ops [2/3]
Matt Arsenault [Mon, 24 Jun 2019 21:37:02 +0000 (21:37 +0000)]
InstCombine: Preserve nuw when reassociating nuw ops [2/3]

Alive says this is OK.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@364234 91177308-0d34-0410-b5e6-96231b3b80d8

5 years agoInstCombine: Preserve nuw when reassociating nuw ops [1/3]
Matt Arsenault [Mon, 24 Jun 2019 21:36:59 +0000 (21:36 +0000)]
InstCombine: Preserve nuw when reassociating nuw ops [1/3]

Alive says this is OK.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@364233 91177308-0d34-0410-b5e6-96231b3b80d8

5 years ago[NFC][Reassociate] Add unary FNeg tests to fast-ReassociateVector.ll
Cameron McInally [Mon, 24 Jun 2019 21:36:09 +0000 (21:36 +0000)]
[NFC][Reassociate] Add unary FNeg tests to fast-ReassociateVector.ll

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@364232 91177308-0d34-0410-b5e6-96231b3b80d8

5 years agoNFC: DataExtractor: use decodeULEB128 to implement getULEB128
David Blaikie [Mon, 24 Jun 2019 20:43:36 +0000 (20:43 +0000)]
NFC: DataExtractor: use decodeULEB128 to implement getULEB128

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@364230 91177308-0d34-0410-b5e6-96231b3b80d8

5 years ago[CVP] Reenable nowrap flag inference
Nikita Popov [Mon, 24 Jun 2019 20:13:13 +0000 (20:13 +0000)]
[CVP] Reenable nowrap flag inference

Inference of nowrap flags in CVP has been disabled, because it
triggered a bug in LFTR (https://bugs.llvm.org/show_bug.cgi?id=31181).
This issue has been fixed in D60935, so we should be able to reenable
nowrap flag inference now.

Differential Revision: https://reviews.llvm.org/D62776

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@364228 91177308-0d34-0410-b5e6-96231b3b80d8

5 years ago[InstCombine] add tests for more variants of isPowerOf2; NFC
Sanjay Patel [Mon, 24 Jun 2019 20:11:40 +0000 (20:11 +0000)]
[InstCombine] add tests for more variants of isPowerOf2; NFC

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@364227 91177308-0d34-0410-b5e6-96231b3b80d8

5 years agollvm-symbolizer: Add a FRAME command.
Peter Collingbourne [Mon, 24 Jun 2019 20:03:23 +0000 (20:03 +0000)]
llvm-symbolizer: Add a FRAME command.

This command prints a description of the referenced function's stack frame.
For each formal parameter and local variable, the tool prints:

- function name
- variable name
- file/line of declaration
- FP-relative variable location (if available)
- size in bytes
- HWASAN tag offset

This information will be used by the HWASAN runtime to identify local
variables in UAR reports.

Differential Revision: https://reviews.llvm.org/D63468

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@364225 91177308-0d34-0410-b5e6-96231b3b80d8

5 years ago[InstCombine] Regenerate test pr17827. NFCI.
Huihui Zhang [Mon, 24 Jun 2019 19:49:42 +0000 (19:49 +0000)]
[InstCombine] Regenerate test pr17827. NFCI.

Prep work for upcoming patch D63505.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@364224 91177308-0d34-0410-b5e6-96231b3b80d8

5 years ago[CodeGen] Add missing vector type legalization for ctlz_zero_undef
Roland Froese [Mon, 24 Jun 2019 19:27:07 +0000 (19:27 +0000)]
[CodeGen] Add missing vector type legalization for ctlz_zero_undef

Widen vector result type for ctlz_zero_undef and cttz_zero_undef the same as
ctlz and cttz.

Differential Revision: https://reviews.llvm.org/D63463

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@364221 91177308-0d34-0410-b5e6-96231b3b80d8

5 years ago[Tests] Add cases where we're failing to discharge provably loop exits (tests for...
Philip Reames [Mon, 24 Jun 2019 19:26:17 +0000 (19:26 +0000)]
[Tests] Add cases where we're failing to discharge provably loop exits (tests for D63733)

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@364220 91177308-0d34-0410-b5e6-96231b3b80d8

5 years ago[SLP] Support unary FNeg vectorization
Cameron McInally [Mon, 24 Jun 2019 19:24:23 +0000 (19:24 +0000)]
[SLP] Support unary FNeg vectorization

Differential Revision: https://reviews.llvm.org/D63609

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@364219 91177308-0d34-0410-b5e6-96231b3b80d8

5 years agoRemove flag for no longer supported MSVC version
Nico Weber [Mon, 24 Jun 2019 18:47:53 +0000 (18:47 +0000)]
Remove flag for no longer supported MSVC version

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@364218 91177308-0d34-0410-b5e6-96231b3b80d8

5 years agoAMDGPU/GlobalISel: Select G_TRUNC
Matt Arsenault [Mon, 24 Jun 2019 18:02:18 +0000 (18:02 +0000)]
AMDGPU/GlobalISel: Select G_TRUNC

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@364215 91177308-0d34-0410-b5e6-96231b3b80d8

5 years agoAMDGPU/GlobalISel: RegBankSelect for amdgcn.class
Matt Arsenault [Mon, 24 Jun 2019 18:00:47 +0000 (18:00 +0000)]
AMDGPU/GlobalISel: RegBankSelect for amdgcn.class

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@364214 91177308-0d34-0410-b5e6-96231b3b80d8

5 years ago[PowerPC][UpdateTestChecks] powerpc- triple support
Jinsong Ji [Mon, 24 Jun 2019 18:00:34 +0000 (18:00 +0000)]
[PowerPC][UpdateTestChecks] powerpc- triple support

There are quite some old testcases with powerpc- triple,
we should add this triple support so that we can update them with script.

Differential Revision: https://reviews.llvm.org/D63723

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@364213 91177308-0d34-0410-b5e6-96231b3b80d8

5 years agoAMDGPU/GlobalISel: Split VALU s64 G_ZEXT/G_SEXT in RegBankSelect
Matt Arsenault [Mon, 24 Jun 2019 17:54:12 +0000 (17:54 +0000)]
AMDGPU/GlobalISel: Split VALU s64 G_ZEXT/G_SEXT in RegBankSelect

Scalar extends to s64 can use S_BFE_{I64|U64}, but vector extends need
to extend to the 32-bit half, and then to 64.

I'm not sure what the line should be between what RegBankSelect
handles, and what instruction select does, but for now I'm erring on
the side of RegBankSelect for future post-RBS combines.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@364212 91177308-0d34-0410-b5e6-96231b3b80d8

5 years ago[llvm-objdump] Match GNU objdump on symbol types shown in disassembly
Yuanfang Chen [Mon, 24 Jun 2019 17:47:56 +0000 (17:47 +0000)]
[llvm-objdump] Match GNU objdump on symbol types shown in disassembly
output.

STT_OBJECT and STT_COMMON are dumped as data, not disassembled.

https://bugs.llvm.org/show_bug.cgi?id=41947

Differential Revision: https://reviews.llvm.org/D62964

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@364211 91177308-0d34-0410-b5e6-96231b3b80d8

5 years ago[AMDGPU] Allow any value in unused src0 field in v_nop
Tim Renouf [Mon, 24 Jun 2019 17:35:20 +0000 (17:35 +0000)]
[AMDGPU] Allow any value in unused src0 field in v_nop

Summary:
The LLVM disassembler assumes that the unused src0 operand of v_nop is
zero. Other tools can put another value in that field, which is still
valid. This commit fixes the LLVM disassembler to recognize such an
encoding as v_nop, in the same way as we already do for s_getpc.

Differential Revision: https://reviews.llvm.org/D63724

Change-Id: Iaf0363eae26ff92fc4ebc716216476adbff37a6f

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@364208 91177308-0d34-0410-b5e6-96231b3b80d8

5 years ago[X86] Don't a vzext_movl in LowerBuildVectorv16i8/LowerBuildVectorv8i16 if there...
Craig Topper [Mon, 24 Jun 2019 17:28:41 +0000 (17:28 +0000)]
[X86] Don't a vzext_movl in LowerBuildVectorv16i8/LowerBuildVectorv8i16 if there are no zeroes in the vector we're building.

In LowerBuildVectorv16i8 we took care to use an any_extend if the first pair is in the lower 16-bits of the vector and no elements are 0. So bits [31:16] will be undefined. But we still emitted a vzext_movl to ensure that bits [127:32] are 0. If we don't need any zeroes we should be consistent and make all of 127:16 undefined.

In LowerBuildVectorv8i16 we can just delete the vzext_movl code because we only use the scalar_to_vector when there are no zeroes. So the vzext_movl is always unnecessary.

Found while investigating whether (vzext_movl (scalar_to_vector (loadi32)) patterns are necessary. At least one of the cases where they were necessary was where the loadi32 matched 32-bit aligned 16-bit extload. Seemed weird that we required vzext_movl for that case.

Differential Revision: https://reviews.llvm.org/D63700

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@364207 91177308-0d34-0410-b5e6-96231b3b80d8

5 years ago[X86] Cleanups and safety checks around the isFNEG
Craig Topper [Mon, 24 Jun 2019 17:28:26 +0000 (17:28 +0000)]
[X86] Cleanups and safety checks around the isFNEG

This patch does a few things to start cleaning up the isFNEG function.

-Remove the Op0/Op1 peekThroughBitcast calls that seem unnecessary. getTargetConstantBitsFromNode has its own peekThroughBitcast inside. And we have a separate peekThroughBitcast on the return value.
-Add a check of the scalar size after the first peekThroughBitcast to ensure we haven't changed the element size and just did something like f32->i32 or f64->i64.
-Remove an unnecessary check that Op1's type is floating point after the peekThroughBitcast. We're just going to look for a bit pattern from a constant. We don't care about its type.
-Add VT checks on several places that consume the return value of isFNEG. Due to the peekThroughBitcasts inside, the type of the return value isn't guaranteed. So its not safe to use it to build other nodes without ensuring the type matches the type being used to build the node. We might be able to replace these checks with bitcasts instead, but I don't have a test case so a bail out check seemed better for now.

Differential Revision: https://reviews.llvm.org/D63683

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@364206 91177308-0d34-0410-b5e6-96231b3b80d8

5 years ago[AArch64] Regenerate vcvt tests. NFCI.
Simon Pilgrim [Mon, 24 Jun 2019 17:18:20 +0000 (17:18 +0000)]
[AArch64] Regenerate vcvt tests. NFCI.

Prep work for an upcoming patch

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@364205 91177308-0d34-0410-b5e6-96231b3b80d8

5 years ago[AArch64] Regenerate 2velem tests. NFCI.
Simon Pilgrim [Mon, 24 Jun 2019 16:58:19 +0000 (16:58 +0000)]
[AArch64] Regenerate 2velem tests. NFCI.

Prep work for an upcoming patch

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@364204 91177308-0d34-0410-b5e6-96231b3b80d8

5 years ago[AArch64] Regenerate merge-store tests. NFCI.
Simon Pilgrim [Mon, 24 Jun 2019 16:57:12 +0000 (16:57 +0000)]
[AArch64] Regenerate merge-store tests. NFCI.

Prep work for an upcoming patch

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@364203 91177308-0d34-0410-b5e6-96231b3b80d8

5 years ago[X86] Regenerate fast fadd reduction tests. NFCI
Simon Pilgrim [Mon, 24 Jun 2019 16:25:30 +0000 (16:25 +0000)]
[X86] Regenerate fast fadd reduction tests. NFCI

Fix whitespace.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@364200 91177308-0d34-0410-b5e6-96231b3b80d8

5 years agoAMDGPU/GlobalISel: Fix selecting G_IMPLICIT_DEF for s1
Matt Arsenault [Mon, 24 Jun 2019 16:24:03 +0000 (16:24 +0000)]
AMDGPU/GlobalISel: Fix selecting G_IMPLICIT_DEF for s1

Try to fail for scc, since I don't think that should ever be produced.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@364199 91177308-0d34-0410-b5e6-96231b3b80d8

5 years ago[bindings/go] Add debug information accessors
Ayke van Laethem [Mon, 24 Jun 2019 16:23:17 +0000 (16:23 +0000)]
[bindings/go] Add debug information accessors

Add debug information accessors, as provided in the following patches:

https://reviews.llvm.org/D46627 (DILocation)
https://reviews.llvm.org/D52693 metadata kind
https://reviews.llvm.org/D60481 get/set debug location on a Value
https://reviews.llvm.org/D60489 (DIScope)

The API as proposed in this patch is similar to the current Value API,
with a single root type and methods that are only valid for certain
subclasses. I have considered just implementing generic Line() calls
(that are valid on all DINodes that have a line) but the implementation
of that got a bit awkward without support from the C API. I've also
considered creating generic getters like a Metadata.DebugLoc() that
returns a DebugLoc, but there is a mismatch between the Go DI nodes in
the LLVM API and the actual DINode class hierarchy, so that's also hard
to get right (without being confusing or breaking the API).

Differential Revision: https://reviews.llvm.org/D63056

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@364198 91177308-0d34-0410-b5e6-96231b3b80d8

5 years agoHexagon: Rename another copy of Register class
Matt Arsenault [Mon, 24 Jun 2019 16:16:19 +0000 (16:16 +0000)]
Hexagon: Rename another copy of Register class

For some reason clang is happy with the conflict, but MSVC is not.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@364196 91177308-0d34-0410-b5e6-96231b3b80d8

5 years agoARC: Fix -Wimplicit-fallthrough
Matt Arsenault [Mon, 24 Jun 2019 16:16:16 +0000 (16:16 +0000)]
ARC: Fix -Wimplicit-fallthrough

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@364195 91177308-0d34-0410-b5e6-96231b3b80d8

5 years agoGlobalISel: Remove unsigned variant of SrcOp
Matt Arsenault [Mon, 24 Jun 2019 16:16:12 +0000 (16:16 +0000)]
GlobalISel: Remove unsigned variant of SrcOp

Force using Register.

One downside is the generated register enums require explicit
conversion.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@364194 91177308-0d34-0410-b5e6-96231b3b80d8

5 years agoCodeGen: Introduce a class for registers
Matt Arsenault [Mon, 24 Jun 2019 15:50:29 +0000 (15:50 +0000)]
CodeGen: Introduce a class for registers

Avoids using a plain unsigned for registers throughoug codegen.
Doesn't attempt to change every register use, just something a little
more than the set needed to build after changing the return type of
MachineOperand::getReg().

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@364191 91177308-0d34-0410-b5e6-96231b3b80d8

5 years ago[AMDGPU] Remove unused variable AllSGPRSpilledToVGPRs. NFC
Bjorn Pettersson [Mon, 24 Jun 2019 15:50:18 +0000 (15:50 +0000)]
[AMDGPU] Remove unused variable AllSGPRSpilledToVGPRs. NFC

Summary:
Removing the unused variable AllSGPRSpilledToVGPRs in
SIFrameLowering::processFunctionBeforeFrameFinalized
to avoid
  error: variable 'AllSGPRSpilledToVGPRs' set but not used
  [-Werror=unused-but-set-variable]

Reviewers: arsenm, nhaehnle

Reviewed By: nhaehnle

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

Tags: #llvm

Differential Revision: https://reviews.llvm.org/D63721

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@364190 91177308-0d34-0410-b5e6-96231b3b80d8

5 years agoHexagon: Rename Register class
Matt Arsenault [Mon, 24 Jun 2019 15:27:29 +0000 (15:27 +0000)]
Hexagon: Rename Register class

This avoids a naming conflict in a future patch.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@364188 91177308-0d34-0410-b5e6-96231b3b80d8

5 years ago[InstCombine] reduce funnel-shift i16 X, X, 8 to bswap X
Sanjay Patel [Mon, 24 Jun 2019 15:20:49 +0000 (15:20 +0000)]
[InstCombine] reduce funnel-shift i16 X, X, 8 to bswap X

Prefer the more exact intrinsic to remove a use of the input value
and possibly make further transforms easier (we will still need
to match patterns with funnel-shift of wider types as pieces of
bswap, especially if we want to canonicalize to funnel-shift with
constant shift amount). Discussed in D46760.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@364187 91177308-0d34-0410-b5e6-96231b3b80d8

5 years agoAMDGPU/GlobalISel: Fix RegBankSelect for s1 sext/zext/anyext
Matt Arsenault [Mon, 24 Jun 2019 14:53:58 +0000 (14:53 +0000)]
AMDGPU/GlobalISel: Fix RegBankSelect for s1 sext/zext/anyext

This needs different handling if the source is known to be a valid
condition or not. Handle turning it into shifts or a select during
regbankselect.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@364186 91177308-0d34-0410-b5e6-96231b3b80d8

5 years agoAMDGPU: Fold frame index into MUBUF
Matt Arsenault [Mon, 24 Jun 2019 14:53:56 +0000 (14:53 +0000)]
AMDGPU: Fold frame index into MUBUF

This matters for byval uses outside of the entry block, which appear
as copies.

Previously, the only folding done was during selection, which could
not see the underlying frame index. For any uses outside the entry
block, the frame index was materialized in the entry block relative to
the global scratch wave offset.

This may produce worse code in cases where the offset ends up not
fitting in the MUBUF offset field. A better heuristic would be helpfu
for extreme frames.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@364185 91177308-0d34-0410-b5e6-96231b3b80d8

5 years ago[InstCombine] add tests for funnel-shift to bswap; NFC
Sanjay Patel [Mon, 24 Jun 2019 14:47:02 +0000 (14:47 +0000)]
[InstCombine] add tests for funnel-shift to bswap; NFC

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@364184 91177308-0d34-0410-b5e6-96231b3b80d8

5 years agoAMDGPU: Cleanup checking when spills need emergency slots
Matt Arsenault [Mon, 24 Jun 2019 14:34:40 +0000 (14:34 +0000)]
AMDGPU: Cleanup checking when spills need emergency slots

Address fixme, which should no longer be a problem since r363757.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@364182 91177308-0d34-0410-b5e6-96231b3b80d8

5 years ago[InstCombine] SliceUpIllegalIntegerPHI - bail on out of range shifts
Simon Pilgrim [Mon, 24 Jun 2019 13:13:36 +0000 (13:13 +0000)]
[InstCombine] SliceUpIllegalIntegerPHI - bail on out of range shifts

trunc(lshr) handling - if the shift is out of range (undefined) then bail like we do for non-constant shifts.

Fixes OSS Fuzz #15217

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@364181 91177308-0d34-0410-b5e6-96231b3b80d8

5 years ago[DAGCombine] visitMUL - allow shift by zero in MulByConstant.
Simon Pilgrim [Mon, 24 Jun 2019 12:47:17 +0000 (12:47 +0000)]
[DAGCombine] visitMUL - allow shift by zero in MulByConstant.

This can occur under certain circumstances when undefs are created later on in the constant multipliers (e.g. in this case due to SimplifyDemandedVectorElts). Its better to let the shift by zero to occur and perform any cleanup afterward.

Fixes OSS Fuzz #15429

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@364179 91177308-0d34-0410-b5e6-96231b3b80d8

5 years ago[ConstantFolding] Use hasVectorInstrinsicScalarOpd. NFC
Bjorn Pettersson [Mon, 24 Jun 2019 12:07:17 +0000 (12:07 +0000)]
[ConstantFolding] Use hasVectorInstrinsicScalarOpd. NFC

Summary:
Use the hasVectorInstrinsicScalarOpd helper function
in ConstantFoldVectorCall.

Reviewers: rengolin, RKSimon, dblaikie

Reviewed By: rengolin, RKSimon

Subscribers: tschuett, hiraditya, llvm-commits

Tags: #llvm

Differential Revision: https://reviews.llvm.org/D63705

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@364178 91177308-0d34-0410-b5e6-96231b3b80d8

5 years ago[Scalarizer] Add scalarizer support for smul.fix.sat
Bjorn Pettersson [Mon, 24 Jun 2019 12:07:11 +0000 (12:07 +0000)]
[Scalarizer] Add scalarizer support for smul.fix.sat

Summary:
Handle smul.fix.sat in the scalarizer. This is done by
adding smul.fix.sat to the set of "isTriviallyVectorizable"
intrinsics.

The addition of smul.fix.sat in isTriviallyVectorizable and
hasVectorInstrinsicScalarOpd can also be seen as a preparation
to be able to use hasVectorInstrinsicScalarOpd in ConstantFolding.

Reviewers: rengolin, RKSimon, dblaikie

Reviewed By: rengolin

Subscribers: hiraditya, llvm-commits

Tags: #llvm

Differential Revision: https://reviews.llvm.org/D63704

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@364177 91177308-0d34-0410-b5e6-96231b3b80d8

5 years ago[docs][llvm-nm] Add missing options to documentation
James Henderson [Mon, 24 Jun 2019 10:50:49 +0000 (10:50 +0000)]
[docs][llvm-nm] Add missing options to documentation

There were several options missing from the documentation. This patch
adds them as well as improving some wording and separating the Mach-O
only options into a separate section.

Fixes https://bugs.llvm.org/show_bug.cgi?id=42234.

Reviewed by: MaskRay

Differential Revision: https://reviews.llvm.org/D63655

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@364176 91177308-0d34-0410-b5e6-96231b3b80d8

5 years ago[sancov] Avoid unnecessary unique_ptr
Fangrui Song [Mon, 24 Jun 2019 10:23:47 +0000 (10:23 +0000)]
[sancov] Avoid unnecessary unique_ptr

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@364175 91177308-0d34-0410-b5e6-96231b3b80d8

5 years ago[ARM] Add MVE interleaving load/store family.
Simon Tatham [Mon, 24 Jun 2019 10:00:39 +0000 (10:00 +0000)]
[ARM] Add MVE interleaving load/store family.

This adds the family of loads and stores with names like VLD20.8 and
VST42.32, which load and store parts of multiple q-registers in such a
way that executing both VLD20 and VLD21, or all four of VLD40..VLD43,
will distribute 2 or 4 vectors' worth of memory data across the lanes
of the same number of registers but in a transposed order.

In addition to the Tablegen descriptions of the instructions
themselves, this patch also adds encode and decode support for the
QQPR and QQQQPR register classes (representing the range of loaded or
stored vector registers), and tweaks to the parsing system for lists
of vector registers to make it return the right format in this case
(since, unlike NEON, MVE regards q-registers as primitive, and not
just an alias for two d-registers).

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@364172 91177308-0d34-0410-b5e6-96231b3b80d8

5 years ago[docs][llvm-nm] Improve symbol code documentation
James Henderson [Mon, 24 Jun 2019 09:53:02 +0000 (09:53 +0000)]
[docs][llvm-nm] Improve symbol code documentation

The existing symbol code documentation was very incomplete. This patch
adds the missing codes, and defines them based on the current code
behaviour.

Fixes https://bugs.llvm.org/show_bug.cgi?id=42231.

Reviewed by: rupprecht, mtrent, MaskRay

Differential Revision: https://reviews.llvm.org/D63327

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@364171 91177308-0d34-0410-b5e6-96231b3b80d8

5 years ago[Support] Fix error handling in DataExtractor::get[US]LEB128
Pavel Labath [Mon, 24 Jun 2019 09:11:24 +0000 (09:11 +0000)]
[Support] Fix error handling in DataExtractor::get[US]LEB128

Summary:
These functions are documented as not modifying the offset argument if
the extraction fails (just like other DataExtractor functions). However,
while reviewing D63591 we discovered that this is not the case -- if the
function reaches the end of the data buffer, it will just return the
value parsed until that point and set offset to point to the end of the
buffer.

This fixes the functions to act as advertised, and adds a regression
test.

Reviewers: dblaikie, probinson, bkramer

Subscribers: kristina, llvm-commits

Tags: #llvm

Differential Revision: https://reviews.llvm.org/D63645

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@364169 91177308-0d34-0410-b5e6-96231b3b80d8

5 years agoFollow up of rL363913. NFC.
Sjoerd Meijer [Mon, 24 Jun 2019 08:44:29 +0000 (08:44 +0000)]
Follow up of rL363913. NFC.

Minor reshuffle in AArch64 targetparser unittest, solving a potential problem
with querying iterators too early.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@364168 91177308-0d34-0410-b5e6-96231b3b80d8

5 years ago[llvm-readobj/llvm-readelf] - Eliminate the elf-groups.x86_64 precompiled binary...
George Rimar [Mon, 24 Jun 2019 08:29:54 +0000 (08:29 +0000)]
[llvm-readobj/llvm-readelf] - Eliminate the elf-groups.x86_64 precompiled binary from the inputs.

We do not need the elf-groups.x86_64. In one of the tests, it was
used for no solid reason, and for the second test case we can use
YAML input with SHT_GROUP sections.

The patch performs a cleanup of one of the test cases, removes another
one completely (since during the review was found out it actually
duplicates one of the existent tests) and removes the precompiled binary.

Differential revision: https://reviews.llvm.org/D63647

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@364167 91177308-0d34-0410-b5e6-96231b3b80d8

5 years ago[X86] Turn v16i16->v16i8 truncate+store into a any_extend+truncstore if we avx512f...
Craig Topper [Sun, 23 Jun 2019 23:51:21 +0000 (23:51 +0000)]
[X86] Turn v16i16->v16i8 truncate+store into a any_extend+truncstore if we avx512f, but not avx512bw.

Ideally we'd be able to represent this truncate as a any_extend to
v16i32 and a truncate, but SelectionDAG doens't know how to not
fold those together.

We have isel patterns to use a vpmovzxwd+vpdmovdb for the truncate,
but we aren't able to simultaneously fold the load and the store
from the isel pattern. By pulling the truncate into the store we
can successfully hide it from the DAG combiner. Then we can isel
pattern match the truncstore and load+any_extend separately.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@364163 91177308-0d34-0410-b5e6-96231b3b80d8

5 years ago[GN] Generation failure caused by trailing space in file name
Petr Hosek [Sun, 23 Jun 2019 23:12:10 +0000 (23:12 +0000)]
[GN] Generation failure caused by trailing space in file name

When I executed gn.py gen out/gn I got the following error:

ERROR at //compiler-rt/lib/builtins/BUILD.gn:162:7: Only source, header, and object files belong in the sources of a static_library. //compiler-rt/lib/builtins/emutls.c  is not one of the valid types.
      "emutls.c ",
      ^----------
See //compiler-rt/lib/BUILD.gn:3:5: which caused the file to be included.
    "//compiler-rt/lib/builtins",
    ^---------------------------
It turns out to be that the latest gn doesn't accept ill-format file name. And the emutls.c above has a trailing space.
Remove the trailing space should work.

Patch By: myhsu
Differential Revision: https://reviews.llvm.org/D63449

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@364162 91177308-0d34-0410-b5e6-96231b3b80d8

5 years agoFix typo in comment; NFC
Sanjoy Das [Sun, 23 Jun 2019 19:22:13 +0000 (19:22 +0000)]
Fix typo in comment; NFC

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@364159 91177308-0d34-0410-b5e6-96231b3b80d8

5 years ago[X86] Fix isel pattern that was looking for a bitcasted load. Remove what appears...
Craig Topper [Sun, 23 Jun 2019 19:17:50 +0000 (19:17 +0000)]
[X86] Fix isel pattern that was looking for a bitcasted load. Remove what appears to be a copy/paste mistake.

DAG combine should ensure bitcasts of loads don't exist.

Also remove 3 patterns that are identical to the block above them.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@364158 91177308-0d34-0410-b5e6-96231b3b80d8

5 years ago[Tests] Autogen and improve test readability
Philip Reames [Sun, 23 Jun 2019 17:13:53 +0000 (17:13 +0000)]
[Tests] Autogen and improve test readability

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@364156 91177308-0d34-0410-b5e6-96231b3b80d8

5 years ago[IndVars] Remove dead instructions after folding trivial loop exit
Philip Reames [Sun, 23 Jun 2019 17:06:57 +0000 (17:06 +0000)]
[IndVars] Remove dead instructions after folding trivial loop exit

In rL364135, I taught IndVars to fold exiting branches in loops with a zero backedge taken count (i.e. loops that only run one iteration).  This extends that to eliminate the dead comparison left around.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@364155 91177308-0d34-0410-b5e6-96231b3b80d8

5 years agoSlotIndexes: delete unused functions
Fangrui Song [Sun, 23 Jun 2019 16:05:29 +0000 (16:05 +0000)]
SlotIndexes: delete unused functions

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@364154 91177308-0d34-0410-b5e6-96231b3b80d8

5 years ago[InstCombine] squash is-power-of-2 that uses ctpop
Sanjay Patel [Sun, 23 Jun 2019 14:22:37 +0000 (14:22 +0000)]
[InstCombine] squash is-power-of-2 that uses ctpop

This is another intermediate IR step towards solving PR42314:
https://bugs.llvm.org/show_bug.cgi?id=42314

We can test if a value is power-of-2-or-0 using ctpop(X) < 2,
so combining that with a non-zero check of the input is the
same as testing if exactly 1 bit is set:

(X != 0) && (ctpop(X) u< 2) --> ctpop(X) == 1

Differential Revision: https://reviews.llvm.org/D63660

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@364153 91177308-0d34-0410-b5e6-96231b3b80d8

5 years agoSlotIndexes: simplify IdxMBBPair operators
Fangrui Song [Sun, 23 Jun 2019 13:16:03 +0000 (13:16 +0000)]
SlotIndexes: simplify IdxMBBPair operators

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@364152 91177308-0d34-0410-b5e6-96231b3b80d8

5 years ago[SelectionDAG] Remove the code that attempts to calculate the alignment for the secon...
Craig Topper [Sun, 23 Jun 2019 07:00:46 +0000 (07:00 +0000)]
[SelectionDAG] Remove the code that attempts to calculate the alignment for the second half of a split masked load/store.

The code divides the alignment by 2 if the original alignment is
equal to the original VT size. But this wouldn't be correct
if the alignment was larger than the VT size.

The memory operand object already takes care of calling MinAlign
on the base alignment and the memory pointer offset. So we don't
need any special code at all.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@364151 91177308-0d34-0410-b5e6-96231b3b80d8

5 years ago[X86][SelectionDAG] Cleanup and simplify masked_load/masked_store in tablegen. Use...
Craig Topper [Sun, 23 Jun 2019 06:06:04 +0000 (06:06 +0000)]
[X86][SelectionDAG] Cleanup and simplify masked_load/masked_store in tablegen. Use more precise PatFrags for scalar masked load/store.

Rename masked_load/masked_store to masked_ld/masked_st to discourage
their direct use. We need to check truncating/extending and
compressing/expanding before using them. This revealed that
our scalar masked load/store patterns were misusing these.

With those out of the way, renamed masked_load_unaligned and
masked_store_unaligned to remove the "_unaligned". We didn't
check the alignment anyway so the name was somewhat misleading.

Make the aligned versions inherit from masked_load/store instead
from a separate identical version. Merge the 3 different alignments
PatFrags into a single version that uses the VT from the SDNode to
determine the size that the alignment needs to match.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@364150 91177308-0d34-0410-b5e6-96231b3b80d8

5 years ago[Support] Fix build under Emscripten
Keno Fischer [Sun, 23 Jun 2019 00:29:59 +0000 (00:29 +0000)]
[Support] Fix build under Emscripten

Summary:
Emscripten's libc doesn't define MNT_LOCAL, thus causing a build
failure in the fallback path. However, to the best of my knowledge,
it also doesn't support remote file system mounts, so we may simply
return `true` here (as we do for e.g. Fuchsia). With this fix, the
core LLVM libraries build correctly under emscripten (though some
of the tools and utils do not).

Reviewers: kripken
Differential Revision: https://reviews.llvm.org/D63688

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@364143 91177308-0d34-0410-b5e6-96231b3b80d8

5 years agoRevert [CommandLine] Remove OptionCategory and SubCommand caches from the Option...
Don Hinton [Sat, 22 Jun 2019 23:32:36 +0000 (23:32 +0000)]
Revert [CommandLine] Remove OptionCategory and SubCommand caches from the Option class.

This reverts r364134 (git commit a5b83bc9e3b8e8945b55068c762bd6c73621a4b0)

Caused errors in the asan bot, so the GeneralCategory global needs to
be changed to ManagedStatic.

Differential Revision: https://reviews.llvm.org/D62105

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@364141 91177308-0d34-0410-b5e6-96231b3b80d8

5 years ago[X86][SSE] Fold extract_subvector(vselect(x,y,z),0) -> vselect(extract_subvector...
Simon Pilgrim [Sat, 22 Jun 2019 17:57:01 +0000 (17:57 +0000)]
[X86][SSE] Fold extract_subvector(vselect(x,y,z),0) -> vselect(extract_subvector(x,0),extract_subvector(y,0),extract_subvector(z,0))

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@364136 91177308-0d34-0410-b5e6-96231b3b80d8

5 years agoExploit a zero LoopExit count to eliminate loop exits
Philip Reames [Sat, 22 Jun 2019 17:54:25 +0000 (17:54 +0000)]
Exploit a zero LoopExit count to eliminate loop exits

This turned out to be surprisingly effective. I was originally doing this just for completeness sake, but it seems like there are a lot of cases where SCEV's exit count reasoning is stronger than it's isKnownPredicate reasoning.

Once this is in, I'm thinking about trying to build on the same infrastructure to eliminate provably untaken checks. There may be something generally interesting here.

Differential Revision: https://reviews.llvm.org/D63618

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@364135 91177308-0d34-0410-b5e6-96231b3b80d8

5 years ago[CommandLine] Remove OptionCategory and SubCommand caches from the Option class.
Don Hinton [Sat, 22 Jun 2019 17:22:50 +0000 (17:22 +0000)]
[CommandLine] Remove OptionCategory and SubCommand caches from the Option class.

Summary:
This change processes `OptionCategory`s and `SubCommand`s as they
 are seen instead of caching them in the Option class and processing
them later.  Doing so simplifies the work needed to be done by the Global
parser and significantly reduces the size of the Option class to a mere 64
bytes.

Removing  the `OptionCategory` cache saved 24 bytes, and removing
the `SubCommand` cache saved an additional 48 bytes, for a total of a
72 byte reduction.

Reviewers: beanz, zturner, MaskRay, serge-sans-paille

Reviewed By: serge-sans-paille

Subscribers: serge-sans-paille, tstellar, zturner, hiraditya, llvm-commits

Tags: #llvm

Differential Revision: https://reviews.llvm.org/D62105

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@364134 91177308-0d34-0410-b5e6-96231b3b80d8

5 years ago[NFC] Fix indentation in PPCAsmPrinter.cpp
Hubert Tong [Sat, 22 Jun 2019 16:03:29 +0000 (16:03 +0000)]
[NFC] Fix indentation in PPCAsmPrinter.cpp

After r248261, the indentation switches, inside a namespace definition,
between indenting and not indenting one level in for that namespace; the
abomination occurs in the middle of a class definition. Fix that.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@364133 91177308-0d34-0410-b5e6-96231b3b80d8

5 years ago[PowerPC][NFC] Move comment to the relevant function
Hubert Tong [Sat, 22 Jun 2019 16:02:02 +0000 (16:02 +0000)]
[PowerPC][NFC] Move comment to the relevant function

A comment that applies to a virtual destructor was placed on a class
constructor. Move the comment to where it belongs.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@364132 91177308-0d34-0410-b5e6-96231b3b80d8

5 years agoPDB docs: Delete trailing whitespace, wrap to 80 cols
Nico Weber [Sat, 22 Jun 2019 11:23:01 +0000 (11:23 +0000)]
PDB docs: Delete trailing whitespace, wrap to 80 cols

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@364131 91177308-0d34-0410-b5e6-96231b3b80d8

5 years ago[NewGVN] Fix copy/paste mistake in cast
Nikita Popov [Sat, 22 Jun 2019 10:20:13 +0000 (10:20 +0000)]
[NewGVN] Fix copy/paste mistake in cast

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@364130 91177308-0d34-0410-b5e6-96231b3b80d8

5 years ago[NewGVN] Remove dead SwitchEdges variable; NFC
Nikita Popov [Sat, 22 Jun 2019 10:20:07 +0000 (10:20 +0000)]
[NewGVN] Remove dead SwitchEdges variable; NFC

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@364129 91177308-0d34-0410-b5e6-96231b3b80d8

5 years ago[LFTR] Add tests for PR41998; NFC
Nikita Popov [Sat, 22 Jun 2019 09:57:59 +0000 (09:57 +0000)]
[LFTR] Add tests for PR41998; NFC

The limit for the pointer case is incorrect.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@364128 91177308-0d34-0410-b5e6-96231b3b80d8

5 years agoAArch64: Add support for reading pc using llvm.read_register.
Peter Collingbourne [Sat, 22 Jun 2019 03:03:25 +0000 (03:03 +0000)]
AArch64: Add support for reading pc using llvm.read_register.

This is useful for allowing code to efficiently take an address
that can be later mapped onto debug info. Currently the hwasan
pass achieves this by taking the address of the current function:
http://llvm-cs.pcc.me.uk/lib/Transforms/Instrumentation/HWAddressSanitizer.cpp#921

but this costs two instructions (plus a GOT entry in PIC code) per function
with stack variables. This will allow the cost to be reduced to a single
instruction.

Differential Revision: https://reviews.llvm.org/D63471

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@364126 91177308-0d34-0410-b5e6-96231b3b80d8

5 years ago[CMake] Delete redundant DEPENDS/LINK_LIBS from LineEditor/XRay
Fangrui Song [Sat, 22 Jun 2019 01:50:21 +0000 (01:50 +0000)]
[CMake] Delete redundant DEPENDS/LINK_LIBS from LineEditor/XRay

The link dependencies are already specified in LLVMBuild.txt

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@364125 91177308-0d34-0410-b5e6-96231b3b80d8

5 years agoMake GlobalISel depend on SelectionDAG after D63169
Fangrui Song [Sat, 22 Jun 2019 01:30:17 +0000 (01:30 +0000)]
Make GlobalISel depend on SelectionDAG after D63169

GlobalISel/IRTranslator.cpp now references SelectionDAG/FunctionLoweringInfo.cpp.
This fixes a link error in -DBUILD_SHARED_LIBS=on builds:

    ld.lld: error: undefined symbol: llvm::FunctionLoweringInfo::clear()
    >>> referenced by IRTranslator.cpp:2198 (../lib/CodeGen/GlobalISel/IRTranslator.cpp:2198)
    >>>               lib/CodeGen/GlobalISel/CMakeFiles/LLVMGlobalISel.dir/IRTranslator.cpp.o:(llvm::IRTranslator::finalizeFunction())

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@364124 91177308-0d34-0410-b5e6-96231b3b80d8

5 years agoFix UNSUPPORTED attribute from windows to system-windows.
Douglas Yung [Sat, 22 Jun 2019 01:14:29 +0000 (01:14 +0000)]
Fix UNSUPPORTED attribute from windows to system-windows.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@364122 91177308-0d34-0410-b5e6-96231b3b80d8

5 years ago[llvm-objdump] Allow --disassemble-functions to take demangled names
Yuanfang Chen [Sat, 22 Jun 2019 01:13:04 +0000 (01:13 +0000)]
[llvm-objdump] Allow --disassemble-functions to take demangled names

The --disassemble-functions switch takes demangled names when
--demangle is specified, otherwise the switch takes mangled names.

https://bugs.llvm.org/show_bug.cgi?id=41908

Reviewers: jhenderson, grimar, MaskRay, rupprecht

Differential Revision: https://reviews.llvm.org/D63524

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@364121 91177308-0d34-0410-b5e6-96231b3b80d8

5 years ago[llvm-objdump] Move --start-address >= --stop-address check out of the
Yuanfang Chen [Sat, 22 Jun 2019 00:22:57 +0000 (00:22 +0000)]
[llvm-objdump] Move --start-address >= --stop-address check out of the
-d code.

Summary:
Move it into `main` function so the checking is effective for all actions
user may do with llvm-objdump; notably, -r and -s in addition to existing -d.

Match GNU behavior.

Reviewers: jhenderson, grimar, MaskRay, rupprecht

Subscribers: llvm-commits

Tags: #llvm

Differential Revision: https://reviews.llvm.org/D63631

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@364118 91177308-0d34-0410-b5e6-96231b3b80d8

5 years agoAArch64: Prefer FP-relative debug locations in HWASANified functions.
Peter Collingbourne [Sat, 22 Jun 2019 00:06:51 +0000 (00:06 +0000)]
AArch64: Prefer FP-relative debug locations in HWASANified functions.

To help produce better diagnostics for stack use-after-return, we'd like
to be able to determine the addresses of each HWASANified function's local
variables given a small amount of information recorded on entry to the
function. Currently we require all HWASANified functions to use frame pointers
and record (PC, FP) on function entry. This works better than recording SP
because FP cannot change during the function, unlike SP which can change
e.g. due to dynamic alloca.

However, most variables currently end up using SP-relative locations in their
debug info. This prevents us from recomputing the address of most variables
because the distance between SP and FP isn't recorded in the debug info. To
address this, make the AArch64 backend prefer FP-relative debug locations
when producing debug info for HWASANified functions.

Differential Revision: https://reviews.llvm.org/D63300

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@364117 91177308-0d34-0410-b5e6-96231b3b80d8

5 years agogn build: Merge r364046.
Peter Collingbourne [Sat, 22 Jun 2019 00:03:53 +0000 (00:03 +0000)]
gn build: Merge r364046.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@364116 91177308-0d34-0410-b5e6-96231b3b80d8

5 years ago[COFF, ARM64] Fix encoding of debugtrap for Windows
Tom Tan [Fri, 21 Jun 2019 23:38:05 +0000 (23:38 +0000)]
[COFF, ARM64] Fix encoding of debugtrap for Windows

On Windows ARM64, intrinsic __debugbreak is compiled into brk #0xF000 which is
mapped to llvm.debugtrap in Clang. Instruction brk #F000 is the defined break
point instruction on ARM64 which is recognized by Windows debugger and
exception handling code, so llvm.debugtrap should map to it instead of
redirecting to llvm.trap (brk #1) as the default implementation.

Differential Revision: https://reviews.llvm.org/D63635

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@364115 91177308-0d34-0410-b5e6-96231b3b80d8

5 years agoRevert [SLP] Look-ahead operand reordering heuristic.
Reid Kleckner [Fri, 21 Jun 2019 23:10:25 +0000 (23:10 +0000)]
Revert [SLP] Look-ahead operand reordering heuristic.

This reverts r364084 (git commit 5698921be2d567f6abf925479ac9f5a376d6d74f)

It caused crashes while compiling a file in Chrome. Reduction
forthcoming.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@364111 91177308-0d34-0410-b5e6-96231b3b80d8

5 years ago[llvm-lipo] Implement -thin
Shoaib Meenai [Fri, 21 Jun 2019 21:59:01 +0000 (21:59 +0000)]
[llvm-lipo] Implement -thin

Creates thin output file of specified arch_type from the fat input file.

Patch by Anusha Basana <anushabasana@fb.com>

Differential Revision: https://reviews.llvm.org/D63341

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@364107 91177308-0d34-0410-b5e6-96231b3b80d8

5 years ago[ASan] Use dynamic shadow on 32-bit iOS and simulators
Julian Lettner [Fri, 21 Jun 2019 21:01:39 +0000 (21:01 +0000)]
[ASan] Use dynamic shadow on 32-bit iOS and simulators

The VM layout on iOS is not stable between releases. On 64-bit iOS and
its derivatives we use a dynamic shadow offset that enables ASan to
search for a valid location for the shadow heap on process launch rather
than hardcode it.

This commit extends that approach for 32-bit iOS plus derivatives and
their simulators.

rdar://50645192
rdar://51200372
rdar://51767702

Reviewed By: delcypher

Differential Revision: https://reviews.llvm.org/D63586

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@364105 91177308-0d34-0410-b5e6-96231b3b80d8

5 years ago[X86] Add test cases for incorrect shrinking of volatile vector loads from 128-bits...
Craig Topper [Fri, 21 Jun 2019 20:16:26 +0000 (20:16 +0000)]
[X86] Add test cases for incorrect shrinking of volatile vector loads from 128-bits to 32 or 64 bits. NFC

This is caused by isel patterns that look for vzmovl+load and
treat it the same as vzload.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@364101 91177308-0d34-0410-b5e6-96231b3b80d8

5 years agoAMDGPU: Fix not using s33 for scratch wave offset in kernels
Matt Arsenault [Fri, 21 Jun 2019 20:04:02 +0000 (20:04 +0000)]
AMDGPU: Fix not using s33 for scratch wave offset in kernels

Fixes missing piece from r363990.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@364099 91177308-0d34-0410-b5e6-96231b3b80d8

5 years ago[X86] Add DAG combine to turn (vzmovl (insert_subvector undef, X, 0)) into (insert_su...
Craig Topper [Fri, 21 Jun 2019 19:10:21 +0000 (19:10 +0000)]
[X86] Add DAG combine to turn (vzmovl (insert_subvector undef, X, 0)) into (insert_subvector allzeros, (vzmovl X), 0)

128/256 bit scalar_to_vectors are canonicalized to (insert_subvector undef, (scalar_to_vector), 0). We have isel patterns that try to match this pattern being used by a vzmovl to use a 128-bit instruction and a subreg_to_reg.

This patch detects the insert_subvector undef portion of this and pulls it through the vzmovl, creating a narrower vzmovl and an insert_subvector allzeroes. We can then match the insertsubvector into a subreg_to_reg operation by itself. Then we can fall back on existing (vzmovl (scalar_to_vector)) patterns.

Note, while the scalar_to_vector case is the motivating case I didn't restrict to just that case. I'm also wondering about shrinking any 256/512 vzmovl to an extract_subvector+vzmovl+insert_subvector(allzeros) but I fear that would have bad implications to shuffle combining.

I also think there is more canonicalization we can do with vzmovl with loads or scalar_to_vector with loads to create vzload.

Differential Revision: https://reviews.llvm.org/D63512

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@364095 91177308-0d34-0410-b5e6-96231b3b80d8