OSDN Git Service

android-x86/external-llvm.git
5 years ago[InstCombine] [NFC] add testcases for canonicalizing (-X srem Y) to -(X srem Y).
Chen Zheng [Sat, 13 Apr 2019 07:34:55 +0000 (07:34 +0000)]
[InstCombine] [NFC] add testcases for canonicalizing (-X srem Y) to -(X srem Y).

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@358327 91177308-0d34-0410-b5e6-96231b3b80d8

5 years ago[StackMaps] Update llvm-readobj to parse V3 Stackmaps
Philip Reames [Sat, 13 Apr 2019 03:55:13 +0000 (03:55 +0000)]
[StackMaps] Update llvm-readobj to parse V3 Stackmaps

This updates the StackMap parser in the llvm-readobj tool to parse version 3 StackMaps, which were bumped in https://reviews.llvm.org/D32629.

Version 3 StackMaps differ in that they have a uint16 sized "location size" field which was added to the Location block in a StackMap record. The record has additional padding for alignment. This was a backwards incompatible change resulting in a StackMap version bump.

Patch By: jacob.hughes@kcl.ac.uk (with a rewrite of tests by me)
Differential Revision: https://reviews.llvm.org/D59020

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@358325 91177308-0d34-0410-b5e6-96231b3b80d8

5 years ago[StackMaps] Add location size to llvm-readobj -stackmap output
Philip Reames [Sat, 13 Apr 2019 03:08:45 +0000 (03:08 +0000)]
[StackMaps] Add location size to llvm-readobj -stackmap output

The size field of a location can be different for each entry, so it is useful to have this displayed in the output of llvm-readobj -stackmap. Below is an example of how the output would look:

Record ID: 2882400000, instruction offset: 16
   3 locations:
     #1: Constant 1, size: 8
     #2: Constant 2, size: 8
     #3: Constant 3, size: 8
   0 live-outs: [ ]

Patch By: jacob.hughes@kcl.ac.uk (with heavy modification by me)
Differential Revision: https://reviews.llvm.org/D59169

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@358324 91177308-0d34-0410-b5e6-96231b3b80d8

5 years ago[llvm-readobj] Minor style tweak for consistency sake [NFC]
Philip Reames [Sat, 13 Apr 2019 02:23:08 +0000 (02:23 +0000)]
[llvm-readobj] Minor style tweak for consistency sake [NFC]

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@358323 91177308-0d34-0410-b5e6-96231b3b80d8

5 years ago[StackMaps] Remove format version from the class name [NFC]
Philip Reames [Sat, 13 Apr 2019 02:02:56 +0000 (02:02 +0000)]
[StackMaps] Remove format version from the class name [NFC]

Motivation is to reduce silly diffs when we change the format.  For instance, this causes most of D59020 to disappear.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@358322 91177308-0d34-0410-b5e6-96231b3b80d8

5 years ago[StackMaps] Add explicit location size accessor to the stackmap parser
Philip Reames [Sat, 13 Apr 2019 01:50:50 +0000 (01:50 +0000)]
[StackMaps] Add explicit location size accessor to the stackmap parser

The reserved uint8 field in the location block of the stackmap record is used to denote the size of the location.

Patch By: jacob.hughes@kcl.ac.uk
Differential Revision: https://reviews.llvm.org/D59167

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@358319 91177308-0d34-0410-b5e6-96231b3b80d8

5 years ago[AArch64][GlobalISel] Enable copy elision in the pre-legalizer combine and fix a...
Amara Emerson [Sat, 13 Apr 2019 00:33:25 +0000 (00:33 +0000)]
[AArch64][GlobalISel] Enable copy elision in the pre-legalizer combine and fix a crash.

This enables the simple copy combine that already exists in the CombinerHelper.
However, it exposed a bug in the GISelChangeObserver where it wouldn't clear a
set of MIs to process, and so would end up causing a crash when deleted MIs were
being added to the combiner worklist again.

Differential Revision: https://reviews.llvm.org/D60579

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@358318 91177308-0d34-0410-b5e6-96231b3b80d8

5 years ago[WebAssembly] Add DataCount section to object files
Thomas Lively [Fri, 12 Apr 2019 22:27:48 +0000 (22:27 +0000)]
[WebAssembly] Add DataCount section to object files

Summary:
This ensures that object files will continue to validate as
WebAssembly modules in the presence of bulk memory operations. Engines
that don't support bulk memory operations will not recognize the
DataCount section and will report validation errors, but that's ok
because object files aren't supposed to be run directly anyway.

Reviewers: aheejin, dschuff, sbc100

Subscribers: jgravelle-google, hiraditya, sunfish, rupprecht, llvm-commits

Tags: #llvm

Differential Revision: https://reviews.llvm.org/D60623

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@358315 91177308-0d34-0410-b5e6-96231b3b80d8

5 years ago[GlobalISel] Fix a crash when handling an invalid MVT during call lowering.
Amara Emerson [Fri, 12 Apr 2019 22:05:46 +0000 (22:05 +0000)]
[GlobalISel] Fix a crash when handling an invalid MVT during call lowering.

This crash was introduced in r358032 as we try to construct an EVT from an MVT
in order to find the register type for the calling conv. Fall back instead of
trying to do this with an invalid MVT coming from i256.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@358314 91177308-0d34-0410-b5e6-96231b3b80d8

5 years ago[MemorySSA] Add previous def to cache when found, even if trivial.
Alina Sbirlea [Fri, 12 Apr 2019 21:58:52 +0000 (21:58 +0000)]
[MemorySSA] Add previous def to cache when found, even if trivial.

Summary:
When inserting a new Def, MemorySSA may be have non-minimal number of Phis.
While inserting, the walk to find the previous definition may cleanup minimal Phis.
When the last definition is trivial to obtain, we do not cache it.

It is possible while getting the previous definition for a Def to get two different answers:
- one that was straight-forward to find when walking the first path (a trivial phi in this case), and
- another that follows a cleanup of the trivial phi, it determines it may need additional Phi nodes, it inserts them and returns a new phi in the same position as the former trivial one.
While the Phis added for the second path are all redundant, they are not complete (the walk is only done upwards), and they are not properly cleaned up afterwards.

A way to fix this problem is to cache the straight-forward answer we got on the first walk.
The caching is only kept for the duration of a getPreviousDef call, and for Phis we use TrackingVH, so removing the trivial phi will lead to replacing it with the next dominating phi in the cache.
Resolves PR40749.

Reviewers: george.burgess.iv

Subscribers: jlebar, Prazek, llvm-commits

Tags: #llvm

Differential Revision: https://reviews.llvm.org/D60634

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@358313 91177308-0d34-0410-b5e6-96231b3b80d8

5 years ago[AArch64][GlobalISel] Fix a crash when selecting shufflevectors with an undef mask...
Amara Emerson [Fri, 12 Apr 2019 21:31:21 +0000 (21:31 +0000)]
[AArch64][GlobalISel] Fix a crash when selecting shufflevectors with an undef mask element.

If a shufflevector's mask vector has an element with "undef" then the generic
instruction defining that element register is a G_IMPLICT_DEF instead of G_CONSTANT.
This fixes the selector to handle this case, and for now assumes that undef just means
zero. In future we'll optimize this case properly.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@358312 91177308-0d34-0410-b5e6-96231b3b80d8

5 years ago[WebAssembly] Add mutable-globals to bleeding-edge CPU
Thomas Lively [Fri, 12 Apr 2019 20:39:53 +0000 (20:39 +0000)]
[WebAssembly] Add mutable-globals to bleeding-edge CPU

Summary: This brings the backend in line with Clang.

Reviewers: aheejin, dschuff

Subscribers: sbc100, jgravelle-google, hiraditya, sunfish, llvm-commits

Tags: #llvm

Differential Revision: https://reviews.llvm.org/D60594

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@358310 91177308-0d34-0410-b5e6-96231b3b80d8

5 years ago[ConstantRange] Clarify makeGuaranteedNoWrapRegion() guarantees; NFC
Nikita Popov [Fri, 12 Apr 2019 19:36:47 +0000 (19:36 +0000)]
[ConstantRange] Clarify makeGuaranteedNoWrapRegion() guarantees; NFC

makeGuaranteedNoWrapRegion() is actually makeExactNoWrapRegion() as
long as only one of NUW or NSW is specified. This is not obvious from
the current documentation, and some code seems to think that it is
only exact for single-element ranges. Clarify docs and add tests to
be more confident this really holds.

There are currently no users of makeGuaranteedNoWrapRegion() that
pass both NUW and NSW. I think it would be best to drop support for
this entirely and then rename the function to makeExactNoWrapRegion().

Knowing that the no-wrap region is exact is useful, because we can
backwards-constrain values. What I have in mind in particular is
that LVI should be able to constrain values on edges where the
with.overflow overflow flag is false.

Differential Revision: https://reviews.llvm.org/D60598

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@358305 91177308-0d34-0410-b5e6-96231b3b80d8

5 years ago[SCEV] Add option to forget everything in SCEV.
Alina Sbirlea [Fri, 12 Apr 2019 19:16:07 +0000 (19:16 +0000)]
[SCEV] Add option to forget everything in SCEV.

Summary:
Create a method to forget everything in SCEV.
Add a cl::opt and PassManagerBuilder option to use this in LoopUnroll.

Motivation: Certain Halide applications spend a very long time compiling in forgetLoop, and prefer to forget everything and rebuild SCEV from scratch.
Sample difference in compile time reduction: 21.04 to 14.78 using current ToT release build.
Testcase showcasing this cannot be opensourced and is fairly large.

The option disabled by default, but it may be desirable to enable by
default. Evidence in favor (two difference runs on different days/ToT state):

File Before (s) After (s)
clang-9.bc 7267.91 6639.14
llvm-as.bc 194.12 194.12
llvm-dis.bc 62.50 62.50
opt.bc 1855.85 1857.53

File Before (s) After (s)
clang-9.bc 8588.70 7812.83
llvm-as.bc 196.20 194.78
llvm-dis.bc 61.55 61.97
opt.bc 1739.78 1886.26

Reviewers: sanjoy

Subscribers: mehdi_amini, jlebar, zzheng, javed.absar, dmgreen, jdoerfert, llvm-commits

Tags: #llvm

Differential Revision: https://reviews.llvm.org/D60144

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@358304 91177308-0d34-0410-b5e6-96231b3b80d8

5 years ago[MemorySSA] Small fix for the clobber limit.
Alina Sbirlea [Fri, 12 Apr 2019 18:48:46 +0000 (18:48 +0000)]
[MemorySSA] Small fix for the clobber limit.

Summary:
After introducing the limit for clobber walking, `walkToPhiOrClobber` would assert that the limit is at least 1 on entry.
The test included triggered that assert.

The callsite in `tryOptimizePhi` making the calls to `walkToPhiOrClobber` is structured like this:
```
while (true) {
   if (getBlockingAccess()) { // calls walkToPhiOrClobber
   }
   for (...) {
     walkToPhiOrClobber();
   }
}
```

The cleanest fix is to check if the limit was reached inside `walkToPhiOrClobber`, and give an allowence of 1.
This approach not make any alias() calls (no calls to instructionClobbersQuery), so the performance condition is enforced.
The limit is set back to 0 if not used, as this provides info on the fact that we stopped before reaching a true clobber.

Reviewers: george.burgess.iv

Subscribers: jlebar, Prazek, llvm-commits

Tags: #llvm

Differential Revision: https://reviews.llvm.org/D60479

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@358303 91177308-0d34-0410-b5e6-96231b3b80d8

5 years ago[InstCombine] Fix a nasty miscompile introduced w/masked.gather demanded elts
Philip Reames [Fri, 12 Apr 2019 18:26:56 +0000 (18:26 +0000)]
[InstCombine] Fix a nasty miscompile introduced w/masked.gather demanded elts

This fixes a miscompile which was introduced in r356510 (https://reviews.llvm.org/D57372).

The problem is that the original patch removed pointer operands where the load results we're demanded, but without considering the legality of the load itself.  If the masked.gather had active, but undemanded, lanes, then we could end up creating a load which loaded from an undef address.  The result could be a segfault, or, in theory, an arbitrary read from a random memory location into an used register.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@358299 91177308-0d34-0410-b5e6-96231b3b80d8

5 years ago[CVP] Set NSW/NUW flags when simplifying with.overflow
Nikita Popov [Fri, 12 Apr 2019 18:18:17 +0000 (18:18 +0000)]
[CVP] Set NSW/NUW flags when simplifying with.overflow

When CVP determines that a with.overflow intrinsic cannot overflow,
it currently inserts a simple add/sub. As we already determined that
there can be no overflow, we should add the appropriate NUW/NSW flag.

Differential Revision: https://reviews.llvm.org/D60585

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@358298 91177308-0d34-0410-b5e6-96231b3b80d8

5 years ago[KnownBits] Add computeForAddCarry()
Nikita Popov [Fri, 12 Apr 2019 18:18:08 +0000 (18:18 +0000)]
[KnownBits] Add computeForAddCarry()

This is for D60460. computeForAddSub() essentially already supports
carries because it has to deal with subtractions. This revision
extracts a lower-level computeForAddCarry() function, which allows
computing the known bits for add (carry known zero), sub (carry known
one) and addcarry (carry unknown).

As we don't seem to have any yet, I've added a unit test file for
KnownBits and exhaustive tests for the new computeForAddCarry()
functionality, as well the existing computeForAddSub() function.

Differential Revision: https://reviews.llvm.org/D60522

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@358297 91177308-0d34-0410-b5e6-96231b3b80d8

5 years ago[Tests] Checkin a test demonstrating a miscompile so that patch which fixes it shows...
Philip Reames [Fri, 12 Apr 2019 18:11:58 +0000 (18:11 +0000)]
[Tests] Checkin a test demonstrating a miscompile so that patch which fixes it shows a clear diff

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@358296 91177308-0d34-0410-b5e6-96231b3b80d8

5 years agoSimplify decoupling between RuntimeDyld/RuntimeDyldChecker, add 'got_addr' util.
Lang Hames [Fri, 12 Apr 2019 18:07:28 +0000 (18:07 +0000)]
Simplify decoupling between RuntimeDyld/RuntimeDyldChecker, add 'got_addr' util.

This patch reduces the number of functions in the interface between RuntimeDyld
and RuntimeDyldChecker by combining "GetXAddress" and "GetXContent" functions
into "GetXInfo" functions that return a struct describing both the address and
content. The GetStubOffset function is also replaced with a pair of utilities,
GetStubInfo and GetGOTInfo, that fit the new scheme. For RuntimeDyld both of
these functions will return the same result, but for the new JITLink linker
(https://reviews.llvm.org/D58704) these will provide the addresses of PLT stubs
and GOT entries respectively.

For JITLink's use, a 'got_addr' utility has been added to the rtdyld-check
language, and the syntax of 'got_addr' and 'stub_addr' has been changed: both
functions now take two arguments, a 'stub container name' and a target symbol
name. For llvm-rtdyld/RuntimeDyld the stub container name is the object file
name and section name, separated by a slash. E.g.:

rtdyld-check: *{8}(stub_addr(foo.o/__text, y)) = y

For the upcoming llvm-jitlink utility, which creates stubs on a per-file basis
rather than a per-section basis, the container name is just the file name. E.g.:

jitlink-check: *{8}(got_addr(foo.o, y)) = y

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@358295 91177308-0d34-0410-b5e6-96231b3b80d8

5 years ago[Hexagon] Fix reuse bug in Vector Loop Carried Reuse pass
Brendon Cahoon [Fri, 12 Apr 2019 16:37:12 +0000 (16:37 +0000)]
[Hexagon] Fix reuse bug in Vector Loop Carried Reuse pass

The Hexagon Vector Loop Carried Reuse pass was allowing reuse between
two shufflevectors with different masks. The reason is that the masks
are not instruction objects, so the code that checks each operand
just skipped over the operands.

This patch fixes the bug by checking if the operands are the same
when they are not instruction objects. If the objects are not the
same, then the code assumes that reuse cannot occur.

Differential Revision: https://reviews.llvm.org/D60019

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@358292 91177308-0d34-0410-b5e6-96231b3b80d8

5 years ago[DAGCombiner] narrow shuffle of concatenated vectors
Sanjay Patel [Fri, 12 Apr 2019 16:31:56 +0000 (16:31 +0000)]
[DAGCombiner] narrow shuffle of concatenated vectors

// shuffle (concat X, undef), (concat Y, undef), Mask -->
// concat (shuffle X, Y, Mask0), (shuffle X, Y, Mask1)

The ARM changes with 'vtrn' and narrowed 'vuzp' are improvements.

The x86 changes look neutral or better. There's one test with an
extra instruction, but that could be reversed for a subtarget with
the right attributes. But by default, we want to avoid the 256-bit
op when possible (in my motivating benchmark, a handful of ymm ops
sprinkled into a sequence of xmm ops are triggering frequency
throttling on Haswell resulting in significantly worse perf).

Differential Revision: https://reviews.llvm.org/D60545

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@358291 91177308-0d34-0410-b5e6-96231b3b80d8

5 years ago[PDB Docs] Add some prose describing public and global symbols.
Zachary Turner [Fri, 12 Apr 2019 15:51:40 +0000 (15:51 +0000)]
[PDB Docs] Add some prose describing public and global symbols.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@358289 91177308-0d34-0410-b5e6-96231b3b80d8

5 years agoAdd options for MaxLoadsPerMemcmp(OptSize).
Hiroshi Yamauchi [Fri, 12 Apr 2019 15:05:46 +0000 (15:05 +0000)]
Add options for MaxLoadsPerMemcmp(OptSize).

Reviewers: davidxl

Reviewed By: davidxl

Subscribers: llvm-commits

Tags: #llvm

Differential Revision: https://reviews.llvm.org/D60587

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@358287 91177308-0d34-0410-b5e6-96231b3b80d8

5 years ago[X86][SSE] Recognise vXi1 boolean anyof/allof reduction patterns
Simon Pilgrim [Fri, 12 Apr 2019 14:22:57 +0000 (14:22 +0000)]
[X86][SSE] Recognise vXi1 boolean anyof/allof reduction patterns

Currently combineHorizontalPredicateResult only handles anyof/allof reduction patterns of legal types, which can be tricky to match as type legalization of bools can introduce bitcasts/truncs/extensions.

This patch extends combineHorizontalPredicateResult to recognise vXi1 bool reductions as well and uses the existing combineBitcastvxi1 helper to create the MOVMSK necessary to then compare the signmask result.

This ensures the accuracy of the reduction costs added in D60403 which assume the MOVMSK generation.

Differential Revision: https://reviews.llvm.org/D60610

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@358286 91177308-0d34-0410-b5e6-96231b3b80d8

5 years agoRevert r358268 "[DebugInfo] DW_OP_deref_size in PrologEpilogInserter."
Hans Wennborg [Fri, 12 Apr 2019 12:54:52 +0000 (12:54 +0000)]
Revert r358268 "[DebugInfo] DW_OP_deref_size in PrologEpilogInserter."

It causes clang to crash while building Chromium. See https://crbug.com/952230
for reproducer.

> The PrologEpilogInserter need to insert a DW_OP_deref_size before
> prepending a memory location expression to an already implicit
> expression to avoid having the existing expression act on the memory
> address instead of the value behind it.
>
> The reason for using DW_OP_deref_size and not plain DW_OP_deref is that
> big-endian targets need to read the right size as simply truncating a
> larger read would yield the wrong result (LSB bytes are not at the lower
> address).
>
> Differential Revision: https://reviews.llvm.org/D59687

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@358281 91177308-0d34-0410-b5e6-96231b3b80d8

5 years ago[llvm-objcopy] Fill .symtab_shndx section correctly
Eugene Leviant [Fri, 12 Apr 2019 11:59:30 +0000 (11:59 +0000)]
[llvm-objcopy] Fill .symtab_shndx section correctly

Differential revision: https://reviews.llvm.org/D60555

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@358278 91177308-0d34-0410-b5e6-96231b3b80d8

5 years agoUse llvm::upper_bound. NFC
Fangrui Song [Fri, 12 Apr 2019 11:31:16 +0000 (11:31 +0000)]
Use llvm::upper_bound. NFC

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@358277 91177308-0d34-0410-b5e6-96231b3b80d8

5 years ago[PowerPC] Add initialization for some ppc passes
Kang Zhang [Fri, 12 Apr 2019 09:59:40 +0000 (09:59 +0000)]
[PowerPC] Add initialization for some ppc passes

Summary:

Some llc debug options need pass-name as the parameters.
But if we use the pass-name ppc-early-ret, we will get below error:
llc test.ll -stop-after ppc-early-ret
LLVM ERROR: "ppc-early-ret" pass is not registered.
Below pass-names have the pass is not registered error:
ppc-ctr-loops
ppc-ctr-loops-verify
ppc-loop-preinc-prep
ppc-toc-reg-deps
ppc-vsx-copy
ppc-early-ret
ppc-vsx-fma-mutate
ppc-vsx-swaps
ppc-reduce-cr-ops
ppc-qpx-load-splat
ppc-branch-coalescing
ppc-branch-select

Reviewed By: jsji

Differential Revision: https://reviews.llvm.org/D60248

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@358271 91177308-0d34-0410-b5e6-96231b3b80d8

5 years ago[DebugInfo] Fix pr41175 Dead Store Elimination missing debug loc
Jeremy Morse [Fri, 12 Apr 2019 09:47:35 +0000 (09:47 +0000)]
[DebugInfo] Fix pr41175 Dead Store Elimination missing debug loc

Bug: https://bugs.llvm.org/show_bug.cgi?id=41175

In the bug test case the DSE pass is shortening the range of memory that a
memset is working on. A getelementptr is generated so that the new
starting address can be passed to memset. This instruction was not given
a DebugLoc.

To fix the bug, copy the DebugLoc from the memset instruction.

Patch by Orlando Cazalet-Hyams!

Differential Revision: https://reviews.llvm.org/D60556

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@358270 91177308-0d34-0410-b5e6-96231b3b80d8

5 years ago[DebugInfo] DW_OP_deref_size in PrologEpilogInserter.
Markus Lavin [Fri, 12 Apr 2019 08:23:55 +0000 (08:23 +0000)]
[DebugInfo] DW_OP_deref_size in PrologEpilogInserter.

The PrologEpilogInserter need to insert a DW_OP_deref_size before
prepending a memory location expression to an already implicit
expression to avoid having the existing expression act on the memory
address instead of the value behind it.

The reason for using DW_OP_deref_size and not plain DW_OP_deref is that
big-endian targets need to read the right size as simply truncating a
larger read would yield the wrong result (LSB bytes are not at the lower
address).

Differential Revision: https://reviews.llvm.org/D59687

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@358268 91177308-0d34-0410-b5e6-96231b3b80d8

5 years agoFix missing arguments in tutorial
Hans Wennborg [Fri, 12 Apr 2019 08:23:28 +0000 (08:23 +0000)]
Fix missing arguments in tutorial

In tutorial "8. Kaleidoscope: Compiling to Object Code" a call to
TargetMachine->addPassesToEmitFile(pass, dest, FileType) is missing
nullptr as its 3rd value.

Patch by Sajjad Heydari!

Differential revision: https://reviews.llvm.org/D60369

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@358267 91177308-0d34-0410-b5e6-96231b3b80d8

5 years agoMove getNumFrameInfos and getDwarfFrameInfos out of line and remove
Eric Christopher [Fri, 12 Apr 2019 07:42:35 +0000 (07:42 +0000)]
Move getNumFrameInfos and getDwarfFrameInfos out of line and remove
the MCDwarf.h include.

This removes 50 transitive dependencies for a modification of
MCDwarf.h in a build of llc for a pair of out of line functions
and reduces the build overhead of 'touch MCDwarf.h" by 15% without
impacting test time of check-llvm.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@358264 91177308-0d34-0410-b5e6-96231b3b80d8

5 years agoAdd explicit dependencies on MCSection.h and MCDwarf.h to the .cpp
Eric Christopher [Fri, 12 Apr 2019 07:40:01 +0000 (07:40 +0000)]
Add explicit dependencies on MCSection.h and MCDwarf.h to the .cpp
files rather than rely on transitive includes from MCStreamer.h.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@358263 91177308-0d34-0410-b5e6-96231b3b80d8

5 years ago[ConstantFold] Don't evaluate FP or FP vector casts or truncations when simplifying...
Fangrui Song [Fri, 12 Apr 2019 07:34:30 +0000 (07:34 +0000)]
[ConstantFold] Don't evaluate FP or FP vector casts or truncations when simplifying icmp

Fix PR41476

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@358262 91177308-0d34-0410-b5e6-96231b3b80d8

5 years agoRevert "[PowerPC] Add initialization for some ppc passes"
Eric Christopher [Fri, 12 Apr 2019 07:16:58 +0000 (07:16 +0000)]
Revert "[PowerPC] Add initialization for some ppc passes"

This reverts commit 6f8f98ce8de7c0e4ebd7fa2e1fd9507fe8d1c317 as it
is breaking nearly every bot.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@358260 91177308-0d34-0410-b5e6-96231b3b80d8

5 years ago[llvm-readobj] Change variables' name to match LLVM-style. NFC.
Xing GUO [Fri, 12 Apr 2019 07:09:41 +0000 (07:09 +0000)]
[llvm-readobj] Change variables' name to match LLVM-style. NFC.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@358259 91177308-0d34-0410-b5e6-96231b3b80d8

5 years agoMove addInitialFrameState out of line and remove the MCDwarf.h include.
Eric Christopher [Fri, 12 Apr 2019 06:57:45 +0000 (06:57 +0000)]
Move addInitialFrameState out of line and remove the MCDwarf.h include.

This removes 50 transitive dependencies for a modification of
MCDwarf.h in a build of llc for a single out of line function
and reduces the build overhead by 20% without impacting test
time of check-llvm.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@358258 91177308-0d34-0410-b5e6-96231b3b80d8

5 years ago[TargetLowering][X86] Teach SimplifyDemandedBits to use ShrinkDemandedOp on ISD:...
Craig Topper [Fri, 12 Apr 2019 06:49:28 +0000 (06:49 +0000)]
[TargetLowering][X86] Teach SimplifyDemandedBits to use ShrinkDemandedOp on ISD::SHL nodes.

If the upper bits of the SHL result aren't used, we might be able to use a narrower shift. For example, on X86 this can turn a 64-bit into 32-bit enabling a smaller encoding.

Differential Revision: https://reviews.llvm.org/D60358

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@358257 91177308-0d34-0410-b5e6-96231b3b80d8

5 years ago[PowerPC] Add initialization for some ppc passes
Kang Zhang [Fri, 12 Apr 2019 06:35:15 +0000 (06:35 +0000)]
[PowerPC] Add initialization for some ppc passes

Summary:

Some llc debug options need pass-name as the parameters.
But if we use the pass-name ppc-early-ret, we will get below error:
llc test.ll -stop-after ppc-early-ret
LLVM ERROR: "ppc-early-ret" pass is not registered.
Below pass-names have the pass is not registered error:
ppc-ctr-loops
ppc-ctr-loops-verify
ppc-loop-preinc-prep
ppc-toc-reg-deps
ppc-vsx-copy
ppc-early-ret
ppc-vsx-fma-mutate
ppc-vsx-swaps
ppc-reduce-cr-ops
ppc-qpx-load-splat
ppc-branch-coalescing
ppc-branch-select

Reviewed By: jsji

Differential Revision: https://reviews.llvm.org/D60248

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@358256 91177308-0d34-0410-b5e6-96231b3b80d8

5 years agoMove addFrameInst out of line and remove the MCDwarf.h include.
Eric Christopher [Fri, 12 Apr 2019 06:31:59 +0000 (06:31 +0000)]
Move addFrameInst out of line and remove the MCDwarf.h include.

This removes 500 transitive dependencies for a modification of
MCDwarf.h in a build of llc for a single out of line function
and reduces the build overhead by more than half without impacting
test time of check-llvm.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@358255 91177308-0d34-0410-b5e6-96231b3b80d8

5 years agoInclude what's used in a few cpp files - these were getting transitive
Eric Christopher [Fri, 12 Apr 2019 06:16:33 +0000 (06:16 +0000)]
Include what's used in a few cpp files - these were getting transitive
includes from MCDwarf.h.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@358254 91177308-0d34-0410-b5e6-96231b3b80d8

5 years ago[PowerPC] More precise exploitation of P9 maddld instruction when operands are constant
Zi Xuan Wu [Fri, 12 Apr 2019 05:21:31 +0000 (05:21 +0000)]
[PowerPC] More precise exploitation of P9 maddld instruction when operands are constant

There are 3 operands of maddld, (add (mul %1, %2), %3) and sometimes
they are constant. If there is constant operand, it takes extra li to
materialize the operand, and one more extra register too. So it's not
profitable to use maddld to optimize mul-add pattern.

Differential Revision: https://reviews.llvm.org/D60181

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@358253 91177308-0d34-0410-b5e6-96231b3b80d8

5 years agoMCDwarfLineTableheader::tryGetFile : replace a loop with llvm::find
Fangrui Song [Fri, 12 Apr 2019 04:55:10 +0000 (04:55 +0000)]
MCDwarfLineTableheader::tryGetFile : replace a loop with llvm::find

Note, `DirIndex++` below is incorrect for DWARF 5, but it can be fixed
later after the file index is fixed.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@358251 91177308-0d34-0410-b5e6-96231b3b80d8

5 years agoMove a couple of optional references to just optional to make the
Eric Christopher [Fri, 12 Apr 2019 03:49:13 +0000 (03:49 +0000)]
Move a couple of optional references to just optional to make the
forwarding APIs look similar.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@358250 91177308-0d34-0410-b5e6-96231b3b80d8

5 years ago[MC] Fix typo: .symtab_shndxr -> .symtab_shndx
Fangrui Song [Fri, 12 Apr 2019 02:16:15 +0000 (02:16 +0000)]
[MC] Fix typo: .symtab_shndxr -> .symtab_shndx

This special section is named .symtab_shndx, according to gABI Chapter 4
Sections, and the name is used by some other tools. Though the section
type SHT_SYMTAB_SHNDX is what really matters, let's fix the typo
introduced in rL204769 :)

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@358247 91177308-0d34-0410-b5e6-96231b3b80d8

5 years agoUse llvm::lower_bound. NFC
Fangrui Song [Fri, 12 Apr 2019 02:02:06 +0000 (02:02 +0000)]
Use llvm::lower_bound. NFC

This reapplies rL358161. That commit inadvertently reverted an exegesis file to an old version.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@358246 91177308-0d34-0410-b5e6-96231b3b80d8

5 years agoRemove a parameter that was being passed around that we had at the
Eric Christopher [Fri, 12 Apr 2019 01:02:02 +0000 (01:02 +0000)]
Remove a parameter that was being passed around that we had at the
local callsite.

NFC.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@358244 91177308-0d34-0410-b5e6-96231b3b80d8

5 years agollvm-undname: Use UNREACHABLE after exhaustive switch returning everywhere
Nico Weber [Thu, 11 Apr 2019 23:23:00 +0000 (23:23 +0000)]
llvm-undname: Use UNREACHABLE after exhaustive switch returning everywhere

No behavior change.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@358241 91177308-0d34-0410-b5e6-96231b3b80d8

5 years agollvm-undname: Name a bool param, no behavior change
Nico Weber [Thu, 11 Apr 2019 23:20:18 +0000 (23:20 +0000)]
llvm-undname: Name a bool param, no behavior change

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@358240 91177308-0d34-0410-b5e6-96231b3b80d8

5 years agollvm-undname: Fix out-of-bounds read on invalid intrinsic function code
Nico Weber [Thu, 11 Apr 2019 23:11:33 +0000 (23:11 +0000)]
llvm-undname: Fix out-of-bounds read on invalid intrinsic function code

Found by inspection.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@358239 91177308-0d34-0410-b5e6-96231b3b80d8

5 years agollvm-undname: Don't crash on incomplete enum tag manglings
Nico Weber [Thu, 11 Apr 2019 22:59:25 +0000 (22:59 +0000)]
llvm-undname: Don't crash on incomplete enum tag manglings

Found by inspection.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@358238 91177308-0d34-0410-b5e6-96231b3b80d8

5 years agollvm-undname: Fix crash on incomplete virtual this adjusts
Nico Weber [Thu, 11 Apr 2019 22:47:18 +0000 (22:47 +0000)]
llvm-undname: Fix crash on incomplete virtual this adjusts

Found by oss-fuzz.

Also remove an else-after-return, this part has no behavior change.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@358237 91177308-0d34-0410-b5e6-96231b3b80d8

5 years ago[X86AsmPrinter] refactor static functions into private methods. NFC
Nick Desaulniers [Thu, 11 Apr 2019 22:47:13 +0000 (22:47 +0000)]
[X86AsmPrinter] refactor static functions into private methods. NFC

Summary:
A lot of the code for printing special cases of operands in this
translation unit are static functions. While I too have suffered many
years of abuse at the hands of C, we should prefer private methods,
particularly when you start passing around *this as your first argument,
which is a code smell.

This will help make generic vs arch specific asm printing easier, as it
brings X86AsmPrinter more in line with other arch's derived AsmPrinters.
We will then be able to more easily move architecture generic code to
the base class, and architecture specific code to the derived classes.

Some other small refactorings while we're here:
- the parameter Op is now consistently OpNo
- add spaces around binary expressions. I know we're not millionaires
  but c'mon.

Reviewers: echristo

Reviewed By: echristo

Subscribers: smeenai, hiraditya, llvm-commits, srhines, craig.topper

Tags: #llvm

Differential Revision: https://reviews.llvm.org/D60577

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@358236 91177308-0d34-0410-b5e6-96231b3b80d8

5 years agollvm-undname: Fix crash on invalid name in a template parameter pointer to member arg
Nico Weber [Thu, 11 Apr 2019 22:23:35 +0000 (22:23 +0000)]
llvm-undname: Fix crash on invalid name in a template parameter pointer to member arg

Found by oss-fuzz.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@358234 91177308-0d34-0410-b5e6-96231b3b80d8

5 years ago[Pipeliner] Fix incorrect loop carried dependence calculation
Brendon Cahoon [Thu, 11 Apr 2019 21:57:51 +0000 (21:57 +0000)]
[Pipeliner] Fix incorrect loop carried dependence calculation

The isLoopCarriedDep function does not correctly compute loop
carried dependences when the array index offset is negative
or the stride is smallar than the access size.

Patch by Denis Antrushin.

Differential Revision: https://reviews.llvm.org/D60135

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@358233 91177308-0d34-0410-b5e6-96231b3b80d8

5 years ago[CVP] Generate full test checks for overflows.ll; NFC
Nikita Popov [Thu, 11 Apr 2019 21:10:39 +0000 (21:10 +0000)]
[CVP] Generate full test checks for overflows.ll; NFC

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@358229 91177308-0d34-0410-b5e6-96231b3b80d8

5 years ago[ConstantRange] Add unsignedMulMayOverflow()
Nikita Popov [Thu, 11 Apr 2019 21:10:33 +0000 (21:10 +0000)]
[ConstantRange] Add unsignedMulMayOverflow()

Same as the other ConstantRange overflow checking methods, but for
unsigned mul. In this case there is no cheap overflow criterion, so
using umul_ov for the implementation.

Differential Revision: https://reviews.llvm.org/D60574

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@358228 91177308-0d34-0410-b5e6-96231b3b80d8

5 years ago[ConstantRangeTest] Fix typos in test names; NFC
Nikita Popov [Thu, 11 Apr 2019 21:10:19 +0000 (21:10 +0000)]
[ConstantRangeTest] Fix typos in test names; NFC

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@358227 91177308-0d34-0410-b5e6-96231b3b80d8

5 years ago[cmake] Fix dependency issue in TableGen
Pengxuan Zheng [Thu, 11 Apr 2019 21:05:15 +0000 (21:05 +0000)]
[cmake] Fix dependency issue in TableGen

Summary:
There is a bug in add_tablegen which causes cmake to fail with the following
error message if LLVM_TABLEGEN is set.

CMake Error at cmake/modules/TableGen.cmake:147 (add_dependencies):
  The dependency target "LLVM-tablegen-host" of target "CLANG-tablegen-host"
  does not exist.
Call Stack (most recent call first):
  tools/clang/utils/TableGen/CMakeLists.txt:3 (add_tablegen)

The issue happens because setting LLVM_TABLEGEN causes cmake to skip generating
the LLVM-tablegen-host target. As a result, a non-existent target was added for
CLANG-tablegen-host causing cmake to fail.

In order to fix this issue, this patch adds a guard to check the validity of the
dependency target before adding it as a dependency.

Reviewers: aganea, smeenai

Reviewed By: aganea

Subscribers: mgorny, llvm-commits

Tags: #llvm

Differential Revision: https://reviews.llvm.org/D60576

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@358226 91177308-0d34-0410-b5e6-96231b3b80d8

5 years ago[PGO] Better handling of profile hash mismatch
Rong Xu [Thu, 11 Apr 2019 20:54:17 +0000 (20:54 +0000)]
[PGO] Better handling of profile hash mismatch

We currently assume profile hash conflicts will be caught by an upfront
check and we assert for the cases that escape the check. The assumption
is not always true as there are chances of conflict. This patch prints
a warning and skips annotating the function for the escaped cases,.

Differential Revision: https://reviews.llvm.org/D60154

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@358225 91177308-0d34-0410-b5e6-96231b3b80d8

5 years ago[AArch64][GlobalISel] Flesh out vector load/store support for more types.
Amara Emerson [Thu, 11 Apr 2019 20:40:01 +0000 (20:40 +0000)]
[AArch64][GlobalISel] Flesh out vector load/store support for more types.

Some of these were legalizing into smaller vector types unnecessarily,
others were simply not supported yet.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@358223 91177308-0d34-0410-b5e6-96231b3b80d8

5 years ago[AArch64][GlobalISel] Legalization and ISel support for load/stores of vectors of...
Amara Emerson [Thu, 11 Apr 2019 20:32:24 +0000 (20:32 +0000)]
[AArch64][GlobalISel] Legalization and ISel support for load/stores of vectors of pointers.

Loads and store of values with type like <2 x p0> currently don't get imported
because SelectionDAG has no knowledge of pointer types. To leverage the existing
support for vector load/stores, we can bitcast the value to have s64 element
types instead. We do this as a custom legalization.

This patch also adds support for general loads of <2 x s64>, and relaxes some
type conditions on selecting G_BITCAST.

Differential Revision: https://reviews.llvm.org/D60534

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@358221 91177308-0d34-0410-b5e6-96231b3b80d8

5 years ago[DebugInfo] Combine Trivial and NonTrivial flags
Aaron Smith [Thu, 11 Apr 2019 20:25:10 +0000 (20:25 +0000)]
[DebugInfo] Combine Trivial and NonTrivial flags

Summary:
Companion to https://reviews.llvm.org/D59347

Reviewers: rnk, zturner, probinson, dblaikie, deadalnix

Subscribers: aprantl, jdoerfert, llvm-commits

Tags: #llvm

Differential Revision: https://reviews.llvm.org/D59348

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@358220 91177308-0d34-0410-b5e6-96231b3b80d8

5 years ago[X86] Restrict vselect handling in scalarizeExtEltFP to only case to pre type legaliz...
Craig Topper [Thu, 11 Apr 2019 19:57:44 +0000 (19:57 +0000)]
[X86] Restrict vselect handling in scalarizeExtEltFP to only case to pre type legalization where the setcc result type is vXi1.

If the vector setcc has been legalized then we will need to convert a vector boolean of 0 or -1 to a scalar boolean of 0 or 1.

The added test case previously crashed in 32-bit mode by creating a setcc with an i64 condition that type legalization couldn't expand.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@358218 91177308-0d34-0410-b5e6-96231b3b80d8

5 years ago[X86] Add 32-bit command line to extractelement-fp.ll so I can add a test case for...
Craig Topper [Thu, 11 Apr 2019 19:57:24 +0000 (19:57 +0000)]
[X86] Add 32-bit command line to extractelement-fp.ll so I can add a test case for a 32-bit only crasher. NFC

This is a bit ugly for ABI reasons about how floats/doubles are returned.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@358217 91177308-0d34-0410-b5e6-96231b3b80d8

5 years ago[X86] Add patterns for using movss/movsd for atomic load/store of f32/64. Remove...
Craig Topper [Thu, 11 Apr 2019 19:19:52 +0000 (19:19 +0000)]
[X86] Add patterns for using movss/movsd for atomic load/store of f32/64. Remove atomic fadd pseudos use isel patterns instead.

This patch adds patterns for turning bitcasted atomic load/store into movss/sd.

It also removes the pseudo instructions for atomic RMW fadd. Instead just adding isel patterns for folding an atomic load into addss/sd. And relying on the new movss/sd store pattern to handle the write part.

This also makes the fadd patterns use VEX and EVEX instructions when AVX or AVX512F are enabled.

Differential Revision: https://reviews.llvm.org/D60394

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@358215 91177308-0d34-0410-b5e6-96231b3b80d8

5 years agoRecommit r358211 "[X86] Use FILD/FIST to implement i64 atomic load on 32-bit targets...
Craig Topper [Thu, 11 Apr 2019 19:19:42 +0000 (19:19 +0000)]
Recommit r358211 "[X86] Use FILD/FIST to implement i64 atomic load on 32-bit targets with X87, but no SSE2"

With correct test checks this time.

If we have X87, but not SSE2 we can atomicaly load an i64 value into the significand of an 80-bit extended precision x87 register using fild. We can then use a fist instruction to convert it back to an i64 integ

This matches what gcc and icc do for this case and removes an existing FIXME.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@358214 91177308-0d34-0410-b5e6-96231b3b80d8

5 years agoRevert r358211 "[X86] Use FILD/FIST to implement i64 atomic load on 32-bit targets...
Craig Topper [Thu, 11 Apr 2019 19:04:38 +0000 (19:04 +0000)]
Revert r358211 "[X86] Use FILD/FIST to implement i64 atomic load on 32-bit targets with X87, but no SSE2"

I seem to have messed up the test checks.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@358212 91177308-0d34-0410-b5e6-96231b3b80d8

5 years ago[X86] Use FILD/FIST to implement i64 atomic load on 32-bit targets with X87, but...
Craig Topper [Thu, 11 Apr 2019 18:40:21 +0000 (18:40 +0000)]
[X86] Use FILD/FIST to implement i64 atomic load on 32-bit targets with X87, but no SSE2

If we have X87, but not SSE2 we can atomicaly load an i64 value into the significand of an 80-bit extended precision x87 register using fild. We can then use a fist instruction to convert it back to an i64 integer and store it to a stack temporary. From there we can do two 32-bit loads to get the value into integer registers without worrying about atomicness.

This matches what gcc and icc do for this case and removes an existing FIXME.

Differential Revision: https://reviews.llvm.org/D60156

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@358211 91177308-0d34-0410-b5e6-96231b3b80d8

5 years ago[X86] Pre-commit i64 volatile test case for D60156. NFC
Craig Topper [Thu, 11 Apr 2019 18:40:08 +0000 (18:40 +0000)]
[X86] Pre-commit i64 volatile test case for D60156. NFC

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@358210 91177308-0d34-0410-b5e6-96231b3b80d8

5 years agoRevert "Use llvm::lower_bound. NFC"
Ali Tamur [Thu, 11 Apr 2019 17:35:20 +0000 (17:35 +0000)]
Revert "Use llvm::lower_bound. NFC"

This reverts commit rL358161.

This patch have broken the test:
llvm/test/tools/llvm-exegesis/X86/uops-CMOV16rm-noreg.s

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@358199 91177308-0d34-0410-b5e6-96231b3b80d8

5 years agoFix sphinx documentation warning.
Zachary Turner [Thu, 11 Apr 2019 17:30:03 +0000 (17:30 +0000)]
Fix sphinx documentation warning.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@358198 91177308-0d34-0410-b5e6-96231b3b80d8

5 years ago[PDB Docs] Add skeleton of documentation for CodeView symbols.
Zachary Turner [Thu, 11 Apr 2019 17:29:48 +0000 (17:29 +0000)]
[PDB Docs] Add skeleton of documentation for CodeView symbols.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@358197 91177308-0d34-0410-b5e6-96231b3b80d8

5 years agoNew document skeleton describing how to add a constrained floating-point
Kevin P. Neal [Thu, 11 Apr 2019 17:16:03 +0000 (17:16 +0000)]
New document skeleton describing how to add a constrained floating-point
intrinsic.

Reviewed by: andrew.w.kaylor, cameron.mcinally
Differential Revision: https://reviews.llvm.org/D59833

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@358194 91177308-0d34-0410-b5e6-96231b3b80d8

5 years ago[ConstantFold] ExtractConstantBytes - handle shifts on large integer types
Simon Pilgrim [Thu, 11 Apr 2019 16:39:31 +0000 (16:39 +0000)]
[ConstantFold] ExtractConstantBytes - handle shifts on large integer types

Use APInt instead of getZExtValue from the ConstantInt until we can confirm that the shift amount is in range.

Reduced from OSS-Fuzz #14169 - https://bugs.chromium.org/p/oss-fuzz/issues/detail?id=14169

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@358192 91177308-0d34-0410-b5e6-96231b3b80d8

5 years ago[DAGCombiner] refactor narrowing of extracted vector binop; NFC
Sanjay Patel [Thu, 11 Apr 2019 15:59:47 +0000 (15:59 +0000)]
[DAGCombiner] refactor narrowing of extracted vector binop; NFC

There's a TODO comment about handling patterns with insert_subvector,
and we do want to match that.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@358187 91177308-0d34-0410-b5e6-96231b3b80d8

5 years ago[X86] SimplifyDemandedVectorElts - add X86ISD::VPERMV3 mask support
Simon Pilgrim [Thu, 11 Apr 2019 15:29:15 +0000 (15:29 +0000)]
[X86] SimplifyDemandedVectorElts - add X86ISD::VPERMV3 mask support

Completes SimplifyDemandedVectorElts's basic variable shuffle mask support which should help D60512 + D60562

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@358186 91177308-0d34-0410-b5e6-96231b3b80d8

5 years agoMake llvm-nm -help great again
Serge Guelton [Thu, 11 Apr 2019 15:22:48 +0000 (15:22 +0000)]
Make llvm-nm -help great again

Only display help from the llvm-nm category instead of all llvm options, which make it much more usable.
There's still an issue with -s, which is probably a bug in llvm::cl and worth another commit.

Differential Revision: https://reviews.llvm.org/D60411

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@358185 91177308-0d34-0410-b5e6-96231b3b80d8

5 years ago[RISCV] Diagnose invalid second input register operand when using %tprel_add
Roger Ferrer Ibanez [Thu, 11 Apr 2019 15:13:12 +0000 (15:13 +0000)]
[RISCV] Diagnose invalid second input register operand when using %tprel_add

RISCVMCCodeEmitter::expandAddTPRel asserts that the second operand must be
x4/tp. As we are not currently checking this in the RISCVAsmParser, the assert
is easy to trigger due to wrong assembly input.

This patch does a late check of this constraint.

An alternative could be using a singleton register class for x4/tp similar to
the current one for sp. Unfortunately it does not result in a good diagnostic.
Because add is an overloaded mnemonic, if no matching is possible, the
diagnostic of the first failing alternative seems to be used as the diagnostic
itself. This means that this case the %tprel_add is diagnosed as an invalid
operand (because the real add instruction only has 3 operands).

Differential Revision: https://reviews.llvm.org/D60528

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@358183 91177308-0d34-0410-b5e6-96231b3b80d8

5 years ago[X86][AVX] Tweak X86ISD::VPERMV3 demandedelts test
Simon Pilgrim [Thu, 11 Apr 2019 15:09:03 +0000 (15:09 +0000)]
[X86][AVX] Tweak X86ISD::VPERMV3 demandedelts test

Original test was too dependent on the order of the combines that could cause the inserted element being demanded after all

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@358182 91177308-0d34-0410-b5e6-96231b3b80d8

5 years ago[X86] Add MM register mapping from CodeView to MC register id
Luo, Yuanke [Thu, 11 Apr 2019 15:01:03 +0000 (15:01 +0000)]
[X86] Add MM register mapping from CodeView to MC register id

Differential Revision: https://reviews.llvm.org/D60437

Change-Id: I2183a6d825d0284b22705d423b88882992b236c5

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@358179 91177308-0d34-0410-b5e6-96231b3b80d8

5 years ago[llvm] [lit] Add target-x86* features
Michal Gorny [Thu, 11 Apr 2019 14:58:39 +0000 (14:58 +0000)]
[llvm] [lit] Add target-x86* features

Add a 'target-x86' and 'target-x86_64' feature sthat indicates that
the default target is 32-bit or 64-bit x86, appropriately.  Combined
with 'native' feature, we're going to use this to control x86-specific
LLDB native process tests.

Differential Revision: https://reviews.llvm.org/D60474

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@358177 91177308-0d34-0410-b5e6-96231b3b80d8

5 years agoYAMLIO: Fix serialization of strings with embedded nuls
Pavel Labath [Thu, 11 Apr 2019 14:57:34 +0000 (14:57 +0000)]
YAMLIO: Fix serialization of strings with embedded nuls

Summary:
A bug/typo in Output::scalarString caused us to round-trip a StringRef
through a const char *. This meant that any strings with embedded nuls
were unintentionally cut short at the first such character. (It also
could have caused accidental buffer overruns, but it seems that all
StringRefs coming into this functions were formed from null-terminated
strings.)

This patch fixes the bug and adds an appropriate test.

Reviewers: sammccall, jhenderson

Subscribers: kristina, llvm-commits

Tags: #llvm

Differential Revision: https://reviews.llvm.org/D60505

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@358176 91177308-0d34-0410-b5e6-96231b3b80d8

5 years ago[X86][AVX] Add X86ISD::VPERMV3 demandedelts test
Simon Pilgrim [Thu, 11 Apr 2019 14:48:46 +0000 (14:48 +0000)]
[X86][AVX] Add X86ISD::VPERMV3 demandedelts test

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@358175 91177308-0d34-0410-b5e6-96231b3b80d8

5 years ago[X86] SimplifyDemandedVectorElts - add X86ISD::VPERMV mask support
Simon Pilgrim [Thu, 11 Apr 2019 14:35:45 +0000 (14:35 +0000)]
[X86] SimplifyDemandedVectorElts - add X86ISD::VPERMV mask support

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@358174 91177308-0d34-0410-b5e6-96231b3b80d8

5 years ago[X86][AVX] Add X86ISD::VPERMV demandedelts test
Simon Pilgrim [Thu, 11 Apr 2019 14:26:32 +0000 (14:26 +0000)]
[X86][AVX] Add X86ISD::VPERMV demandedelts test

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@358173 91177308-0d34-0410-b5e6-96231b3b80d8

5 years ago[DAGCombiner][x86] scalarize inserted vector FP ops
Sanjay Patel [Thu, 11 Apr 2019 14:21:57 +0000 (14:21 +0000)]
[DAGCombiner][x86] scalarize inserted vector FP ops

// bo (build_vec ...undef, x, undef...), (build_vec ...undef, y, undef...) -->
// build_vec ...undef, (bo x, y), undef...

The lifetime of the nodes in these examples is different for variables versus constants,
but they are all build vectors briefly, so I'm proposing to catch them in this form to
handle all of the leading examples in the motivating test file.

Before we have build vectors, we might have insert_vector_element. After that, we might
have scalar_to_vector and constant pool loads.

It's going to take more work to ensure that FP vector operands are getting simplified
with undef elements, so this transform can apply more widely. In a non-loose FP environment,
we are likely simplifying FP elements to NaN values rather than undefs.

We also need to allow more opcodes down this path. Eg, we don't handle FP min/max flavors
yet.

Differential Revision: https://reviews.llvm.org/D60514

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@358172 91177308-0d34-0410-b5e6-96231b3b80d8

5 years ago[AArch64] Add lowering pattern for llvm.aarch64.neon.vcvtfxs2fp.f16.i64
Diogo N. Sampaio [Thu, 11 Apr 2019 14:19:43 +0000 (14:19 +0000)]
[AArch64] Add lowering pattern for llvm.aarch64.neon.vcvtfxs2fp.f16.i64

Summary:  Add lowering pattern for llvm.aarch64.neon.vcvtfxs2fp.f16.i64

Reviewers: pbarrio, DavidSpickett, LukeGeeson

Reviewed By: LukeGeeson

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

Tags: #llvm

Differential Revision: https://reviews.llvm.org/D60259

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@358171 91177308-0d34-0410-b5e6-96231b3b80d8

5 years ago[X86] SimplifyDemandedVectorElts - add X86ISD::VPERMILPV mask support
Simon Pilgrim [Thu, 11 Apr 2019 14:15:01 +0000 (14:15 +0000)]
[X86] SimplifyDemandedVectorElts - add X86ISD::VPERMILPV mask support

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@358170 91177308-0d34-0410-b5e6-96231b3b80d8

5 years ago[X86][AVX] Add X86ISD::VPERMILPV demandedelts tests
Simon Pilgrim [Thu, 11 Apr 2019 14:09:35 +0000 (14:09 +0000)]
[X86][AVX] Add X86ISD::VPERMILPV demandedelts tests

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@358168 91177308-0d34-0410-b5e6-96231b3b80d8

5 years ago[X86] SimplifyDemandedVectorElts - add X86ISD::VPERMIL2 mask support
Simon Pilgrim [Thu, 11 Apr 2019 14:04:19 +0000 (14:04 +0000)]
[X86] SimplifyDemandedVectorElts - add X86ISD::VPERMIL2 mask support

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@358167 91177308-0d34-0410-b5e6-96231b3b80d8

5 years ago[X86][XOP] Add X86ISD::VPERMIL2 demandedelts test
Simon Pilgrim [Thu, 11 Apr 2019 13:52:43 +0000 (13:52 +0000)]
[X86][XOP] Add X86ISD::VPERMIL2 demandedelts test

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@358166 91177308-0d34-0410-b5e6-96231b3b80d8

5 years ago[X86] SimplifyDemandedVectorElts - add VPPERM support
Simon Pilgrim [Thu, 11 Apr 2019 13:30:38 +0000 (13:30 +0000)]
[X86] SimplifyDemandedVectorElts - add VPPERM support

We need to add support for all variable shuffle mask ops, but VPPERM is the only one that already has test coverage.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@358165 91177308-0d34-0410-b5e6-96231b3b80d8

5 years ago[ValueTracking] Change if-else chain into switch in computeKnownBitsFromAssume
Sander de Smalen [Thu, 11 Apr 2019 13:02:19 +0000 (13:02 +0000)]
[ValueTracking] Change if-else chain into switch in computeKnownBitsFromAssume

This is a follow-up patch to D60504 to further improve
performance issues in computeKnownBitsFromAssume.

The patch is NFC, but may improve compile-time performance
if the compiler isn't clever enough to do the optimization
itself.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@358163 91177308-0d34-0410-b5e6-96231b3b80d8

5 years agoTest commit access
Oliver Stannard [Thu, 11 Apr 2019 12:53:33 +0000 (12:53 +0000)]
Test commit access

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@358162 91177308-0d34-0410-b5e6-96231b3b80d8

5 years agoUse llvm::lower_bound. NFC
Fangrui Song [Thu, 11 Apr 2019 10:25:41 +0000 (10:25 +0000)]
Use llvm::lower_bound. NFC

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@358161 91177308-0d34-0410-b5e6-96231b3b80d8

5 years ago[MCA] Remove wrong comments from a test. NFC
Andrea Di Biagio [Thu, 11 Apr 2019 10:15:04 +0000 (10:15 +0000)]
[MCA] Remove wrong comments from a test. NFC

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@358160 91177308-0d34-0410-b5e6-96231b3b80d8

5 years ago[ADT] Fix template parameter names of llvm::{upper|lower}_bound
Ilya Biryukov [Thu, 11 Apr 2019 09:00:36 +0000 (09:00 +0000)]
[ADT] Fix template parameter names of llvm::{upper|lower}_bound

Summary:
Rename template parameter for a search value from 'ForwardIt' to 'T'.
While here, also use perfect forwarding to pass the value to STL algos.

Reviewers: sammccall

Reviewed By: sammccall

Subscribers: dexonsmith, llvm-commits

Tags: #llvm

Differential Revision: https://reviews.llvm.org/D60510

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@358158 91177308-0d34-0410-b5e6-96231b3b80d8

5 years agotry to fix the sphinx build some more
Hans Wennborg [Thu, 11 Apr 2019 07:46:25 +0000 (07:46 +0000)]
try to fix the sphinx build some more

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@358156 91177308-0d34-0410-b5e6-96231b3b80d8