RPCS3/llvm
1
0
Fork 0
mirror of https://github.com/RPCS3/llvm.git synced 2026-08-26 18:26:51 -04:00
Code Issues Packages Projects Releases 10 Wiki Activity
144,394 Commits 39 Branches 10 Tags
77d2e8630d0942fe827bc94271c06bed209a1929
Commit Graph

17108 Commits

Author SHA1 Message Date
Adam Nemet 49e45b454f [LV] Also port failure remarks to new OptimizationRemarkEmitter API 
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@293866 91177308-0d34-0410-b5e6-96231b3b80d8
 2017-02-02 05:41:51 +00:00
Matt Arsenault b324cd9254 InferAddressSpaces: Handle more cases with constant select operands 
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@293859 91177308-0d34-0410-b5e6-96231b3b80d8
 2017-02-02 03:37:22 +00:00
Davide Italiano 861b997813 [IPSCCP] Restore the old behaviour (pre r293799). 
It's not clear the change I made a good idea, and it definitely needs
further discussion. Thanks to Eli for pointing out.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@293846 91177308-0d34-0410-b5e6-96231b3b80d8
 2017-02-02 00:46:54 +00:00
Matt Arsenault 2eab96a48c InferAddressSpaces: clang-format some things 
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@293843 91177308-0d34-0410-b5e6-96231b3b80d8
 2017-02-02 00:28:25 +00:00
Sanjay Patel 06248e4909 [InstCombine] move folds for shift-shift pairs; NFCI 
Although this is 'no-functional-change-intended', I'm adding tests
for shl-shl and lshr-lshr pairs because there is no existing test 
coverage for those folds.

It seems like we should be able to remove some code from foldShiftedShift()
at this point because we're handling those patterns on the general path.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@293814 91177308-0d34-0410-b5e6-96231b3b80d8
 2017-02-01 21:31:34 +00:00
Michael Kuperstein e7c7da772e Shut up another GCC warning about operator precedence. NFC. 
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@293812 91177308-0d34-0410-b5e6-96231b3b80d8
 2017-02-01 21:06:33 +00:00
Jun Bum Lim d44f51cf08 [JumpThread] No need to erase BB from LoopHeaders. NFC. 
Summary: No need to try to ease BB from LoopHeaders as we already know that BB is not in LoopHeaders.

Reviewers: hsung, majnemer, mcrosier, haicheng, rengolin

Reviewed By: rengolin

Subscribers: llvm-commits

Differential Revision: https://reviews.llvm.org/D29232

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@293802 91177308-0d34-0410-b5e6-96231b3b80d8
 2017-02-01 19:06:55 +00:00
Davide Italiano 0eeb4b90fe [IPSCCP] Don't propagate return values of functions marked as noinline. 
This tries to address what Hal defined (in the post-commit review of
r293727) a long-standing problem with noinline, where we end up
de facto inlining trivial functions e.g.

__attribute__((noinline)) int patatino(void) { return 5; }

because of return value propagation.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@293799 91177308-0d34-0410-b5e6-96231b3b80d8
 2017-02-01 18:52:20 +00:00
Matthew Simpson aa96bc9c68 [LV] Move interleaved access helper functions to VectorUtils (NFC) 
This patch moves some helper functions related to interleaved access
vectorization out of LoopVectorize.cpp and into VectorUtils.cpp. We would like
to use these functions in a follow-on patch that improves interleaved load and
store lowering in (ARM/AArch64)ISelLowering.cpp. One of the functions was
already duplicated there and has been removed.

Differential Revision: https://reviews.llvm.org/D29398

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@293788 91177308-0d34-0410-b5e6-96231b3b80d8
 2017-02-01 17:45:46 +00:00
Sanjoy Das 387f151ad9 [InstCombine] Allow InstCombine to merge adjacent guards 
Summary:
If there are two adjacent guards with different conditions, we can
remove one of them and include its condition into the condition of
another one. This patch allows InstCombine to merge them by the
following pattern:

    guard(a); guard(b) -> guard(a & b).

Reviewers: reames, apilipenko, igor-laevsky, anna, sanjoy

Reviewed By: sanjoy

Subscribers: llvm-commits

Differential Revision: https://reviews.llvm.org/D29378

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@293778 91177308-0d34-0410-b5e6-96231b3b80d8
 2017-02-01 16:34:55 +00:00
Artur Pilipenko 38847f25fe [LoopPredication] Add a new line to debug output in LoopPredication pass 
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@293762 91177308-0d34-0410-b5e6-96231b3b80d8
 2017-02-01 12:25:38 +00:00
Florian Hahn 7b41d2df25 [LoopUnroll] Use addClonedBlockToLoopInfo to add loop header to LI (NFC). 
Summary:
I have a similar patch up for review already (D29173). If you prefer I
can squash them both together.

Also I think there more potential for code sharing between
LoopUnroll.cpp and LoopUnrollRuntime.cpp. Do you think patches for
that would be worthwhile? 

Reviewers: mkuper, mzolotukhin

Reviewed By: mkuper, mzolotukhin

Subscribers: llvm-commits

Differential Revision: https://reviews.llvm.org/D29311

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@293758 91177308-0d34-0410-b5e6-96231b3b80d8
 2017-02-01 10:39:35 +00:00
Justin Bogner 770cb8b1b5 SanitizerCoverage: Support sanitizer guard section on darwin 
MachO's sections need a segment as well as a section name, and the
section start and end symbols are spelled differently than on ELF.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@293733 91177308-0d34-0410-b5e6-96231b3b80d8
 2017-02-01 02:38:39 +00:00
Davide Italiano fb36e6d33b [IPSCCP] Teach how to not propagate return values of naked functions. 
Differential Revision:  https://reviews.llvm.org/D29360

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@293727 91177308-0d34-0410-b5e6-96231b3b80d8
 2017-02-01 01:01:22 +00:00
Matt Arsenault 40fe3a93e9 InferAddressSpaces: Handle select 
This fails to handle some cases where one of the inputs is
a constant to be fixed in a later commit.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@293723 91177308-0d34-0410-b5e6-96231b3b80d8
 2017-02-01 00:08:53 +00:00
Matt Arsenault adb43778a7 InferAddressSpaces: Remove dead declaration 
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@293720 91177308-0d34-0410-b5e6-96231b3b80d8
 2017-01-31 23:57:20 +00:00
Matt Arsenault 08e88e4167 InferAddressSpaces: Avoid double map lookup 
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@293719 91177308-0d34-0410-b5e6-96231b3b80d8
 2017-01-31 23:48:44 +00:00
Matt Arsenault 34a0d5fe90 InferAddressSpaces: Fix broken casting of constants 
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@293718 91177308-0d34-0410-b5e6-96231b3b80d8
 2017-01-31 23:48:40 +00:00
Daniel Berlin 07b87bc280 NewGVN: Dead argument cleanup 
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@293708 91177308-0d34-0410-b5e6-96231b3b80d8
 2017-01-31 22:32:03 +00:00
Daniel Berlin bb5c6551c0 NewGVN: Cleanup conditions to match reality 
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@293707 91177308-0d34-0410-b5e6-96231b3b80d8
 2017-01-31 22:32:01 +00:00
Daniel Berlin 05b6b111d2 NewGVN: Add basic support for symbolic comparison evaluation 
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@293706 91177308-0d34-0410-b5e6-96231b3b80d8
 2017-01-31 22:31:58 +00:00
Daniel Berlin 740306c544 NewGVN: Formatting cleanup after lookupOperandLeader change 
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@293705 91177308-0d34-0410-b5e6-96231b3b80d8
 2017-01-31 22:31:56 +00:00
Daniel Berlin dc4b13e22c NewGVN: Remove the unsued two arguments from lookupOperandLeader. 
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@293704 91177308-0d34-0410-b5e6-96231b3b80d8
 2017-01-31 22:31:53 +00:00
Daniel Berlin e41f4f18c1 NewGVN: Cleanup header files we are using. 
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@293703 91177308-0d34-0410-b5e6-96231b3b80d8
 2017-01-31 22:31:50 +00:00
Davide Italiano 3ba2159d59 [NewGVN] Preserve TargetLibraryInfo analysis. 
We can maybe preserve more but this is a first step.
Ack'ed by Danny on IRC.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@293694 91177308-0d34-0410-b5e6-96231b3b80d8
 2017-01-31 21:53:18 +00:00
Taewook Oh 73b0db7269 Do not propagate DebugLoc across basic blocks 
Summary:
DebugLoc shouldn't be propagated across basic blocks to prevent incorrect stepping and imprecise sample profile result. rL288903 addressed the wrong DebugLoc propagation issue by limiting the copy of DebugLoc when GVN removes a fully redundant load that is dominated by some other load. However, DebugLoc is still incorrectly propagated in the following example:


```
1:  extern int g;
2: 
3:  void foo(int x, int y, int z) {
4:    if (x)
5:      g = 0;
6:    else
7:      g = 1;
8:
9:    int i = 0;
10:   for ( ; i < y ; i++)
11:     if (i > z)
12:       g++;
13: }

```
Below is LLVM IR representation of the program before GVN:


```
@g = external local_unnamed_addr global i32, align 4

; Function Attrs: nounwind uwtable
define void @foo(i32 %x, i32 %y, i32 %z) local_unnamed_addr #0 !dbg !4 {
entry:
  %not.tobool = icmp eq i32 %x, 0, !dbg !8
  %.sink = zext i1 %not.tobool to i32, !dbg !8
  store i32 %.sink, i32* @g, align 4, !tbaa !9
  %cmp8 = icmp sgt i32 %y, 0, !dbg !13
  br i1 %cmp8, label %for.body.preheader, label %for.end, !dbg !17

for.body.preheader:                               ; preds = %entry
  br label %for.body, !dbg !19

for.body:                                         ; preds = %for.body.preheader, %for.inc
  %i.09 = phi i32 [ %inc4, %for.inc ], [ 0, %for.body.preheader ]
  %cmp1 = icmp sgt i32 %i.09, %z, !dbg !19
  br i1 %cmp1, label %if.then2, label %for.inc, !dbg !21

if.then2:                                         ; preds = %for.body
  %0 = load i32, i32* @g, align 4, !dbg !22, !tbaa !9
  %inc = add nsw i32 %0, 1, !dbg !22
  store i32 %inc, i32* @g, align 4, !dbg !22, !tbaa !9
  br label %for.inc, !dbg !23

for.inc:                                          ; preds = %for.body, %if.then2
  %inc4 = add nuw nsw i32 %i.09, 1, !dbg !24
  %exitcond = icmp ne i32 %inc4, %y, !dbg !13
  br i1 %exitcond, label %for.body, label %for.end.loopexit, !dbg !17

for.end.loopexit:                                 ; preds = %for.inc
  br label %for.end, !dbg !26

for.end:                                          ; preds = %for.end.loopexit, %entry
  ret void, !dbg !26
}

```
where 


```
!21 = !DILocation(line: 11, column: 9, scope: !15)
!22 = !DILocation(line: 12, column: 8, scope: !20)
!23 = !DILocation(line: 12, column: 7, scope: !20)
!24 = !DILocation(line: 10, column: 20, scope: !25)
```

And below is after GVN:


```
@g = external local_unnamed_addr global i32, align 4

define void @foo(i32 %x, i32 %y, i32 %z) local_unnamed_addr !dbg !4 {
entry:
  %not.tobool = icmp eq i32 %x, 0, !dbg !8
  %.sink = zext i1 %not.tobool to i32, !dbg !8
  store i32 %.sink, i32* @g, align 4, !tbaa !9
  %cmp8 = icmp sgt i32 %y, 0, !dbg !13
  br i1 %cmp8, label %for.body.preheader, label %for.end, !dbg !17

for.body.preheader:                               ; preds = %entry
  br label %for.body, !dbg !19

for.body:                                         ; preds = %for.inc, %for.body.preheader
  %0 = phi i32 [ %1, %for.inc ], [ %.sink, %for.body.preheader ], !dbg !21
  %i.09 = phi i32 [ %inc4, %for.inc ], [ 0, %for.body.preheader ]
  %cmp1 = icmp sgt i32 %i.09, %z, !dbg !19
  br i1 %cmp1, label %if.then2, label %for.inc, !dbg !22

if.then2:                                         ; preds = %for.body
  %inc = add nsw i32 %0, 1, !dbg !21
  store i32 %inc, i32* @g, align 4, !dbg !21, !tbaa !9
  br label %for.inc, !dbg !23

for.inc:                                          ; preds = %if.then2, %for.body
  %1 = phi i32 [ %inc, %if.then2 ], [ %0, %for.body ]
  %inc4 = add nuw nsw i32 %i.09, 1, !dbg !24
  %exitcond = icmp ne i32 %inc4, %y, !dbg !13
  br i1 %exitcond, label %for.body, label %for.end.loopexit, !dbg !17

for.end.loopexit:                                 ; preds = %for.inc
  br label %for.end, !dbg !26

for.end:                                          ; preds = %for.end.loopexit, %entry
  ret void, !dbg !26
}

```
As you see, GVN removes the load in if.then2 block and creates a phi instruction in for.body for it. The problem is that DebugLoc of remove load instruction is propagated to the newly created phi instruction, which is wrong. rL288903 cannot handle this case because ValuesPerBlock.size() is not 1 in this example when the load is removed.

Reviewers: aprantl, andreadb, wolfgangp

Reviewed By: andreadb

Subscribers: davide, llvm-commits

Differential Revision: https://reviews.llvm.org/D29254

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@293688 91177308-0d34-0410-b5e6-96231b3b80d8
 2017-01-31 20:57:13 +00:00
Davide Italiano ee994236c6 [Instcombine] Combine consecutive identical fences 
Differential Revision:  https://reviews.llvm.org/D29314

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@293661 91177308-0d34-0410-b5e6-96231b3b80d8
 2017-01-31 18:09:05 +00:00
Arnold Schwaighofer b77943578a Don't combine stores to a swifterror pointer operand to a different type 
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@293658 91177308-0d34-0410-b5e6-96231b3b80d8
 2017-01-31 17:53:49 +00:00
Dehao Chen 500e57803c Explicitly promote indirect calls before sample profile annotation. 
Summary: In iterative sample pgo where profile is collected from PGOed binary, we may see indirect call targets promoted and inlined in the profile. Before profile annotation, we need to make this happen in order to annotate correctly on IR. This patch explicitly promotes these indirect calls and inlines them before profile annotation.

Reviewers: xur, davidxl

Reviewed By: davidxl

Subscribers: llvm-commits

Differential Revision: https://reviews.llvm.org/D29040

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@293657 91177308-0d34-0410-b5e6-96231b3b80d8
 2017-01-31 17:49:37 +00:00
Sanjay Patel 963ec7630c fix formatting; NFC 
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@293652 91177308-0d34-0410-b5e6-96231b3b80d8
 2017-01-31 17:25:42 +00:00
Silviu Baranga 150da15398 [InstCombine] Make sure that LHS and RHS have the same type in 
transformToIndexedCompare

If they don't have the same type, the size of the constant
index would need to be adjusted (and this wouldn't be always
possible).

Alternatively we could try the analysis with the initial
RHS value, which would guarantee that the two sides have
the same type. However it is unlikely that in practice this
would pass our transformation requirements.

Fixes PR31808 (https://llvm.org/bugs/show_bug.cgi?id=31808).



git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@293629 91177308-0d34-0410-b5e6-96231b3b80d8
 2017-01-31 14:04:15 +00:00
Florian Hahn 2938de42db [LoopUnroll] Use addClonedBlockToLoopInfo to clone the top level loop (NFC) 
Summary:
rL293124 added the necessary infrastructure to properly add the cloned
top level loop to LoopInfo, which means we do not have to do it manually
in CloneLoopBlocks.

@mkuper sorry for not pointing this out during my review of D29156, I just
realized that today.


Reviewers: mzolotukhin, chandlerc, mkuper

Reviewed By: mkuper

Subscribers: llvm-commits, mkuper

Differential Revision: https://reviews.llvm.org/D29173

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@293615 91177308-0d34-0410-b5e6-96231b3b80d8
 2017-01-31 11:13:44 +00:00
Matt Arsenault 6a569f5700 InferAddressSpaces: Rename constant 
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@293594 91177308-0d34-0410-b5e6-96231b3b80d8
 2017-01-31 02:17:41 +00:00
Matt Arsenault 1394a16b28 InferAddressSpaces: Handle icmp 
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@293593 91177308-0d34-0410-b5e6-96231b3b80d8
 2017-01-31 02:17:32 +00:00
Matt Arsenault 35b092a75f InferAddressSpaces: Support memory intrinsics 
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@293587 91177308-0d34-0410-b5e6-96231b3b80d8
 2017-01-31 01:56:57 +00:00
Matt Arsenault 6be67912a8 InferAddressSpaces: Support atomics 
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@293584 91177308-0d34-0410-b5e6-96231b3b80d8
 2017-01-31 01:40:38 +00:00
Matt Arsenault de6cb7e695 InferAddressSpaces: Don't replace volatile users 
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@293582 91177308-0d34-0410-b5e6-96231b3b80d8
 2017-01-31 01:30:16 +00:00
Matt Arsenault 9be098398c NVPTX: Move InferAddressSpaces to generic code 
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@293579 91177308-0d34-0410-b5e6-96231b3b80d8
 2017-01-31 01:10:58 +00:00
Sanjay Patel 126da6d52b [InstCombine] enable (X <<nsw C1) >>s C2 --> X <<nsw (C1 - C2) for vectors with splat constants 
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@293570 91177308-0d34-0410-b5e6-96231b3b80d8
 2017-01-30 23:35:52 +00:00
Benjamin Kramer cc71b6de7d [ICP] Fix bool conversion warning and actually write out the reason instead of dropping it. 
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@293564 91177308-0d34-0410-b5e6-96231b3b80d8
 2017-01-30 23:11:29 +00:00
Sanjay Patel 222619278d [InstCombine] enable more lshr(shl X, C1), C2 folds for vectors with splat constants 
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@293562 91177308-0d34-0410-b5e6-96231b3b80d8
 2017-01-30 23:01:05 +00:00
Dehao Chen 7bac4378dd Expose isLegalToPromot as a global helper function so that SamplePGO pass can call it for legality check. 
Summary: SamplePGO needs to check if it is legal to promote a target before it actually promotes it.

Reviewers: davidxl

Reviewed By: davidxl

Subscribers: llvm-commits

Differential Revision: https://reviews.llvm.org/D29306

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@293559 91177308-0d34-0410-b5e6-96231b3b80d8
 2017-01-30 22:46:37 +00:00
Dehao Chen ff08fc9e81 Revert r292979 which causes compile time failure. 
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@293557 91177308-0d34-0410-b5e6-96231b3b80d8
 2017-01-30 22:26:05 +00:00
Matt Arsenault fed7884d87 LSR: Don't drop address space when type doesn't match 
For targets with different addressing modes in each address space,
if this is dropped querying isLegalAddressingMode later with this
will give a nonsense result, breaking the isLegalUse assertions.

This is a candidate for the 4.0 release branch.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@293542 91177308-0d34-0410-b5e6-96231b3b80d8
 2017-01-30 19:50:17 +00:00
Sanjay Patel 1790b5fe1c [InstCombine] enable (X >>?exact C1) << C2 --> X >>?exact (C1-C2) for vectors with splat constants 
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@293524 91177308-0d34-0410-b5e6-96231b3b80d8
 2017-01-30 18:40:23 +00:00
Daniel Berlin 6d874b003f NewGVN: Instead of changeToUnreachable, insert an instruction SimplifyCFG will turn into unreachable when it runs 
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@293515 91177308-0d34-0410-b5e6-96231b3b80d8
 2017-01-30 18:12:56 +00:00
Sanjay Patel 49e7dfece8 [InstCombine] use auto with obvious type; NFC 
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@293508 91177308-0d34-0410-b5e6-96231b3b80d8
 2017-01-30 17:38:55 +00:00
Sanjay Patel 093eb4f6ff [InstCombine] enable (X <<nsw C1) >>s C2 --> X <<nsw (C1-C2) for vectors with splat constants 
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@293507 91177308-0d34-0410-b5e6-96231b3b80d8
 2017-01-30 17:19:32 +00:00
Daniel Berlin 77a24c81b1 Revert "NewGVN: Make unreachable blocks be marked with unreachable" 
This reverts commit r293196

Besides making things look nicer, ATM, we'd like to preserve analysis
more than we'd like to destroy the CFG.  We'll probably revisit in the future

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@293501 91177308-0d34-0410-b5e6-96231b3b80d8
 2017-01-30 17:06:55 +00:00
Sanjay Patel f406e8165b [InstCombine] fixed to propagate 'exact' on lshr 
The original shift is bigger, so this may qualify as 'obvious', 
but here's an attempt at an Alive-based proof:

Name: exact
Pre: (C1 u< C2)
%a = shl i8 %x, C1
%b = lshr exact i8 %a, C2 
  =>
%c = lshr exact i8 %x, C2 - C1
%b = and i8 %c, ((1 << width(C1)) - 1) u>> C2

Optimization is correct!



git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@293498 91177308-0d34-0410-b5e6-96231b3b80d8
 2017-01-30 16:53:03 +00:00