Nico Weber [Thu, 5 Sep 2019 18:03:18 +0000 (18:03 +0000)]
gn build: (manually) merge r358706
llvm-svn: 371102
Roman Lebedev [Thu, 5 Sep 2019 17:41:02 +0000 (17:41 +0000)]
[InstCombine] foldICmpBinOp(): consider inverted check in 'unsigned sub overflow' check
A follow-up for r329011.
This may be changed to produce @llvm.sub.with.overflow in a later patch,
but for now just make things more consistent overall.
A few observations stem from this:
* There does not seem to be a similar one-instruction fold for uadd-overflow
* I'm not sure we'll want to canonicalize `B u> A` as `usub.with.overflow`,
so since the `icmp` here no longer refers to `sub`,
reconstructing `usub.with.overflow` will be problematic,
and will likely require standalone pass (similar to DivRemPairs).
https://rise4fun.com/Alive/Zqs
Name: (A - B) u> A --> B u> A
%t0 = sub i8 %A, %B
%r = icmp ugt i8 %t0, %A
=>
%r = icmp ugt i8 %B, %A
Name: (A - B) u<= A --> B u<= A
%t0 = sub i8 %A, %B
%r = icmp ule i8 %t0, %A
=>
%r = icmp ule i8 %B, %A
Name: C u< (C - D) --> C u< D
%t0 = sub i8 %C, %D
%r = icmp ult i8 %C, %t0
=>
%r = icmp ult i8 %C, %D
Name: C u>= (C - D) --> C u>= D
%t0 = sub i8 %C, %D
%r = icmp uge i8 %C, %t0
=>
%r = icmp uge i8 %C, %D
llvm-svn: 371101
Roman Lebedev [Thu, 5 Sep 2019 17:40:49 +0000 (17:40 +0000)]
[InstCombine] foldICmpBinOp(): consider inverted check in 'unsigned add overflow' check
A follow-up for r342004.
This will be changed to produce @llvm.add.with.overflow in a later patch,
but for now just make things more consistent overall.
https://rise4fun.com/Alive/qxE
Name: (Op1 + X) u< Op1 --> ~Op1 u< X
%t0 = add i8 %Op1, %X
%r = icmp ult i8 %t0, %Op1
=>
%n = xor i8 %Op1, -1
%r = icmp ult i8 %n, %X
Name: (Op1 + X) u>= Op1 --> ~Op1 u>= X
%t0 = add i8 %Op1, %X
%r = icmp uge i8 %t0, %Op1
=>
%n = xor i8 %Op1, -1
%r = icmp uge i8 %n, %X
;-------------------------------------------------------------------------------
Name: Op0 u> (Op0 + X) --> X u> ~Op0
%t0 = add i8 %Op0, %X
%r = icmp ugt i8 %Op0, %t0
=>
%n = xor i8 %Op0, -1
%r = icmp ugt i8 %X, %n
Name: Op0 u<= (Op0 + X) --> X u<= ~Op0
%t0 = add i8 %Op0, %X
%r = icmp ule i8 %Op0, %t0
=>
%n = xor i8 %Op0, -1
%r = icmp ule i8 %X, %n
llvm-svn: 371100
Roman Lebedev [Thu, 5 Sep 2019 17:40:37 +0000 (17:40 +0000)]
[InstCombine][NFC] Tests for 'unsigned sub overflow' check
----------------------------------------
Name: unsigned sub, overflow, v0
%sub = sub i8 %x, %y
%ov = icmp ugt i8 %sub, %x
=>
%agg = usub_overflow i8 %x, %y
%sub = extractvalue {i8, i1} %agg, 0
%ov = extractvalue {i8, i1} %agg, 1
Done: 1
Optimization is correct!
----------------------------------------
Name: unsigned sub, no overflow, v0
%sub = sub i8 %x, %y
%ov = icmp ule i8 %sub, %x
=>
%agg = usub_overflow i8 %x, %y
%sub = extractvalue {i8, i1} %agg, 0
%not.ov = extractvalue {i8, i1} %agg, 1
%ov = xor %not.ov, -1
Done: 1
Optimization is correct!
llvm-svn: 371099
Roman Lebedev [Thu, 5 Sep 2019 17:40:28 +0000 (17:40 +0000)]
[InstCombine][NFC] Tests for 'unsigned add overflow' check
----------------------------------------
Name: unsigned add, overflow, v0
%add = add i8 %x, %y
%ov = icmp ult i8 %add, %x
=>
%agg = uadd_overflow i8 %x, %y
%add = extractvalue {i8, i1} %agg, 0
%ov = extractvalue {i8, i1} %agg, 1
Done: 1
Optimization is correct!
----------------------------------------
Name: unsigned add, overflow, v1
%add = add i8 %x, %y
%ov = icmp ult i8 %add, %y
=>
%agg = uadd_overflow i8 %x, %y
%add = extractvalue {i8, i1} %agg, 0
%ov = extractvalue {i8, i1} %agg, 1
Done: 1
Optimization is correct!
----------------------------------------
Name: unsigned add, no overflow, v0
%add = add i8 %x, %y
%ov = icmp uge i8 %add, %x
=>
%agg = uadd_overflow i8 %x, %y
%add = extractvalue {i8, i1} %agg, 0
%not.ov = extractvalue {i8, i1} %agg, 1
%ov = xor %not.ov, -1
Done: 1
Optimization is correct!
----------------------------------------
Name: unsigned add, no overflow, v1
%add = add i8 %x, %y
%ov = icmp uge i8 %add, %y
=>
%agg = uadd_overflow i8 %x, %y
%add = extractvalue {i8, i1} %agg, 0
%not.ov = extractvalue {i8, i1} %agg, 1
%ov = xor %not.ov, -1
Done: 1
Optimization is correct!
llvm-svn: 371098
Cyndy Ishida [Thu, 5 Sep 2019 17:33:44 +0000 (17:33 +0000)]
[TextAPI] update interface file for filtered iter
Summary:
This is a simple change that allows easy iterator semantics for symbols held in interface file.
Not being used, so harmless change right now, but will be once TBD-v4 is submitted.
Reviewers: ributzka, steven_wu
Reviewed By: ributzka
Subscribers: javed.absar, kristof.beyls, dexonsmith, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D67204
llvm-svn: 371097
DeForest Richards [Thu, 5 Sep 2019 17:30:52 +0000 (17:30 +0000)]
Docs: Move Documentation sections to separate pages.
Updates the links on the homepage by moving the User Guides, Programming Documentation, and Subsystem Documentation sections to separate pages. Also changes "Overview" to "About" at the top of the LLVM Docs homepage. This work is part of the Google Season of Docs project.
llvm-svn: 371096
Sanjay Patel [Thu, 5 Sep 2019 17:28:17 +0000 (17:28 +0000)]
[x86] fix horizontal math bug exposed by improved demanded elements analysis (PR43225)
https://bugs.llvm.org/show_bug.cgi?id=43225
llvm-svn: 371095
Craig Topper [Thu, 5 Sep 2019 17:24:34 +0000 (17:24 +0000)]
[X86] Add a FIXME about why the CWD/CDQ/CQO have a bogus implicit def of the A register. NFC
The instructions copy the sign bit of the A register to every bit
of the D register. But they don't write to the A register.
llvm-svn: 371094
Craig Topper [Thu, 5 Sep 2019 17:24:25 +0000 (17:24 +0000)]
[X86] Remove unneeded CHECK lines from a test. NFC
llvm-svn: 371093
Craig Topper [Thu, 5 Sep 2019 17:24:15 +0000 (17:24 +0000)]
[X86] Fix stale comment. NFC
We aren't checking for a concat here. We're just always splitting
256-bit stores.
llvm-svn: 371092
Nico Weber [Thu, 5 Sep 2019 17:22:55 +0000 (17:22 +0000)]
gn build: (manually) merge r371003
llvm-svn: 371091
Adrian McCarthy [Thu, 5 Sep 2019 17:22:30 +0000 (17:22 +0000)]
Fix windows-x86-debug compilation with python enabled using multi-target generator
[Patch by Leonid Mashinskiy]
Visual Studio CMake generator is multi-target and does not define
CMAKE_BUILD_TYPE, so Debug build on VS was failing due selection of release
python library. This patch reverts back some of latest changes and fixes
building by raw VS using CMake expression generators.
Differential Revision: https://reviews.llvm.org/D66994
llvm-svn: 371090
Denis Bakhvalov [Thu, 5 Sep 2019 17:00:32 +0000 (17:00 +0000)]
[MergedLoadStoreMotion] Sink stores to BB with more than 2 predecessors
If we have:
bb5:
br i1 %arg3, label %bb6, label %bb7
bb6:
%tmp = getelementptr inbounds i32, i32* %arg1, i64 2
store i32 3, i32* %tmp, align 4
br label %bb9
bb7:
%tmp8 = getelementptr inbounds i32, i32* %arg1, i64 2
store i32 3, i32* %tmp8, align 4
br label %bb9
bb9: ; preds = %bb4, %bb6, %bb7
...
We can't sink stores directly into bb9.
This patch creates new BB that is successor of %bb6 and %bb7
and sinks stores into that block.
SplitFooterBB is the parameter to the pass that controls
that behavior.
Change-Id: I7fdf50a772b84633e4b1b860e905bf7e3e29940f
Differential: https://reviews.llvm.org/D66234
llvm-svn: 371089
Sanjay Patel [Thu, 5 Sep 2019 16:58:18 +0000 (16:58 +0000)]
[x86] add test for horizontal math bug (PR43225); NFC
llvm-svn: 371088
Alina Sbirlea [Thu, 5 Sep 2019 16:58:15 +0000 (16:58 +0000)]
[MemorySSA] Verify MSSAUpdater exists.
llvm-svn: 371087
Hiroshi Yamauchi [Thu, 5 Sep 2019 16:56:55 +0000 (16:56 +0000)]
[PGO][CHR] Speed up following long, interlinked use-def chains.
Summary:
Avoid visiting an instruction more than once by using a map.
This is similar to https://reviews.llvm.org/rL361416.
Reviewers: davidxl
Reviewed By: davidxl
Subscribers: llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D67198
llvm-svn: 371086
Fangrui Song [Thu, 5 Sep 2019 16:32:31 +0000 (16:32 +0000)]
[ELF] Initialize PhdrEntry::p_align to maxPageSize for PT_LOAD
```
Writer<ELFT>::run
assignFileOffsets
setFileOffset
computeFileOffset
os->ptLoad->p_align may be smaller than config->maxPageSize
setPhdrs
p_align = max(p_align, config->maxPageSize)
```
If we move the config->maxPageSize logic to the constructor of
PhdrEntry, computeFileOffset can be simplified.
Reviewed By: ruiu
Differential Revision: https://reviews.llvm.org/D67211
llvm-svn: 371085
Alina Sbirlea [Thu, 5 Sep 2019 16:25:24 +0000 (16:25 +0000)]
[MemorySSA] Update MemorySSA when removing debug.value calls.
llvm-svn: 371084
Krzysztof Parzyszek [Thu, 5 Sep 2019 16:19:47 +0000 (16:19 +0000)]
[Hexagon] Fix type in HexagonTargetLowering::ReplaceNodeResults
llvm-svn: 371083
Guillaume Chatelet [Thu, 5 Sep 2019 15:44:33 +0000 (15:44 +0000)]
[Alignment][NFC] Change internal representation of TargetLowering.h
Summary:
This is patch is part of a series to introduce an Alignment type.
See this thread for context: http://lists.llvm.org/pipermail/llvm-dev/2019-July/133851.html
See this patch for the introduction of the type: https://reviews.llvm.org/D64790
Reviewers: courbet
Subscribers: hiraditya, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D67226
llvm-svn: 371082
Ilya Biryukov [Thu, 5 Sep 2019 15:30:05 +0000 (15:30 +0000)]
[clangd] Initialize int field to zero. NFC
To make sure we do not have uninitialized values and undefined behavior.
llvm-svn: 371081
Alexandre Ganea [Thu, 5 Sep 2019 15:24:49 +0000 (15:24 +0000)]
[DebugInfo] Add debug location to stubs generated by CGDeclCXX and mark them as artificial
Differential Revision: https://reviews.llvm.org/D66328
llvm-svn: 371080
David Candler [Thu, 5 Sep 2019 15:17:25 +0000 (15:17 +0000)]
[ARM] Add support for the s,j,x,N,O inline asm constraints
A number of inline assembly constraints are currently supported by LLVM, but rejected as invalid by Clang:
Target independent constraints:
s: An integer constant, but allowing only relocatable values
ARM specific constraints:
j: An immediate integer between 0 and 65535 (valid for MOVW)
x: A 32, 64, or 128-bit floating-point/SIMD register: s0-s15, d0-d7, or q0-q3
N: An immediate integer between 0 and 31 (Thumb1 only)
O: An immediate integer which is a multiple of 4 between -508 and 508. (Thumb1 only)
This patch adds support to Clang for the missing constraints along with some checks to ensure that the constraints are used with the correct target and Thumb mode, and that immediates are within valid ranges (at least where possible). The constraints are already implemented in LLVM, but just a couple of minor corrections to checks (V8M Baseline includes MOVW so should work with 'j', 'N' and 'O' shouldn't be valid in Thumb2) so that Clang and LLVM are in line with each other and the documentation.
Differential Revision: https://reviews.llvm.org/D65863
Change-Id: I18076619e319bac35fbb60f590c069145c9d9a0a
llvm-svn: 371079
Simon Pilgrim [Thu, 5 Sep 2019 15:07:07 +0000 (15:07 +0000)]
[X86][SSE] EltsFromConsecutiveLoads - ignore non-zero offset base loads (PR43227)
As discussed on D64551 and PR43227, we don't correctly handle cases where the base load has a non-zero byte offset.
Until we can properly handle this, we must bail from EltsFromConsecutiveLoads.
llvm-svn: 371078
Raphael Isemann [Thu, 5 Sep 2019 14:51:23 +0000 (14:51 +0000)]
[lldb][NFC] Add dimensions to pexpect tests failing on sanitizer bot
I'm still unable to reproduce this locally, so let's see if this
is enough to make the bots happy. The missing dimension parameter
seems to be the biggest difference between these tests and
TestIOHandlerCompletion.py which runs just fine on the same bot,
so that's the motivation for trying this out.
llvm-svn: 371077
Alexander Kornienko [Thu, 5 Sep 2019 14:48:31 +0000 (14:48 +0000)]
[clang-tidy] Make most ArgumentCommentCheck options local, as they should be
llvm-svn: 371076
Alexander Kornienko [Thu, 5 Sep 2019 14:48:23 +0000 (14:48 +0000)]
Add a bugprone-argument-comment option: IgnoreSingleArgument.
Summary:
Add bugprone-argument-comment option: IgnoreSingleArgument.
When true, the check will ignore the single argument.
Sometimes, it's not necessary to add comment to single argument.
For example:
> std::string name("Yubo Xie");
> pScreen->SetWidth(1920);
> pScreen->SetHeight(1080);
This option can ignore such single argument in bugprone-argument-comment check.
Reviewers: alexfh
Reviewed By: alexfh
Subscribers: cfe-commits
Tags: #clang
Patch by Yubo Xie.
Differential Revision: https://reviews.llvm.org/D67056
llvm-svn: 371075
Fangrui Song [Thu, 5 Sep 2019 14:25:57 +0000 (14:25 +0000)]
[yaml2obj] Write the section header table after section contents
Linkers (ld.bfd/gold/lld) place the section header table at the very
end. This allows tools to strip it, which is optional in executable/shared objects.
In addition, if we add or section, the size of the section header table
will change. Placing the section header table in the end keeps section
offsets unchanged.
yaml2obj currently places the section header table immediately after the
program header. Follow what linkers do to make offset updating easier.
Reviewed By: grimar
Differential Revision: https://reviews.llvm.org/D67221
llvm-svn: 371074
Guillaume Chatelet [Thu, 5 Sep 2019 14:17:08 +0000 (14:17 +0000)]
[LLVM][Alignment][NFC] Fixing wrong documentation
llvm-svn: 371073
Alexander Kornienko [Thu, 5 Sep 2019 14:13:57 +0000 (14:13 +0000)]
[clang-tidy] Fix bugprone-argument-comment bug: negative literal number is not checked.
Summary:
For example:
```
void foo(int a);
foo(-2);
```
should be fixed as:
```
foo(/*a=*/-2);
```
This change tries to fix this issue.
Reviewers: alexfh, hokein, aaron.ballman
Reviewed By: alexfh, aaron.ballman
Subscribers: xazax.hun, cfe-commits
Tags: #clang, #clang-tools-extra
Patch by Yubo Xie.
Differential Revision: https://reviews.llvm.org/D67084
llvm-svn: 371072
George Rimar [Thu, 5 Sep 2019 14:02:58 +0000 (14:02 +0000)]
[llvm-readelf] - Allow dumping dynamic symbols when there is no program headers.
D62179 introduced a regression. llvm-readelf lose the ability to dump the dynamic symbols
when there is .dynamic section with a DT_SYMTAB, but there are no program headers:
https://reviews.llvm.org/D62179#1652778
Below is a program flow before the D62179 change:
1) Find SHT_DYNSYM.
2) Find there is no PT_DYNAMIC => don't try to parse it.
3) Print dynamic symbols using information about them found on step (1).
And after the change it became:
1) Find SHT_DYNSYM.
2) Find there is no PT_DYNAMIC => find SHT_DYNAMIC.
3) Parse dynamic table, but fail to handle the DT_SYMTAB because of the absence of the PT_LOAD. Report the "Virtual address is not in any segment" error.
This patch fixes the issue. For doing this it checks that the value of DT_SYMTAB was
mapped to a segment. If not - it ignores it.
Differential revision: https://reviews.llvm.org/D67078
llvm-svn: 371071
Louis Dionne [Thu, 5 Sep 2019 13:50:28 +0000 (13:50 +0000)]
[libc++] Revert "Make `vector` unconditionally move elements when exceptions are disabled."
This reverts r370502, which broke the use case of a copy-only T (with a
deleted move constructor) when exceptions are disabled. Until we figure
out the right behavior, I'm reverting the commit.
llvm-svn: 371068
Louis Dionne [Thu, 5 Sep 2019 13:50:18 +0000 (13:50 +0000)]
[libc++] Add a test for resizing of a vector with copy-only elements
See https://reviews.llvm.org/D62228#1658620
llvm-svn: 371067
Justin Hibbits [Thu, 5 Sep 2019 13:38:46 +0000 (13:38 +0000)]
Add -m(no)-spe to clang
Summary:
r337347 added support for the Signal Processing Engine (SPE) to LLVM.
This follows that up with the clang side.
This adds -mspe and -mno-spe, to match GCC.
Subscribers: nemanjai, kbarton, cfe-commits
Differential Revision: https://reviews.llvm.org/D49754
llvm-svn: 371066
Simon Atanasyan [Thu, 5 Sep 2019 13:37:16 +0000 (13:37 +0000)]
[mips] Make a few test cases more tolerant to exact symbol addresses. NFC
llvm-svn: 371065
David Green [Thu, 5 Sep 2019 13:37:04 +0000 (13:37 +0000)]
[ARM] Fixup the creation of VPT blocks
This attempts to just fix the creation of VPT blocks, fixing up the iterating,
which instructions are considered in the bundle, and making sure that we do not
overrun the end of the block.
Differential Revision: https://reviews.llvm.org/D67219
llvm-svn: 371064
Guillaume Chatelet [Thu, 5 Sep 2019 13:09:42 +0000 (13:09 +0000)]
[LLVM][Alignment] Convert isLegalNTStore/isLegalNTLoad to llvm::Align
Summary:
This is patch is part of a serie to introduce an Alignment type.
See this thread for context: http://lists.llvm.org/pipermail/llvm-dev/2019-July/133851.html
See this patch for the introduction of the type: https://reviews.llvm.org/D64790
Reviewers: courbet
Subscribers: hiraditya, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D67223
llvm-svn: 371063
Simon Pilgrim [Thu, 5 Sep 2019 12:36:11 +0000 (12:36 +0000)]
[X86][SSE] Add (failing) test case for PR43227
llvm-svn: 371061
Petar Avramovic [Thu, 5 Sep 2019 11:20:32 +0000 (11:20 +0000)]
[MIPS GlobalISel] Select G_FENCE
G_FENCE comes form fence instruction. For MIPS fence is generated in
AtomicExpandPass when atomic instruction gets surrounded with fence
instruction when needed.
G_FENCE arguments don't have LLT, because of that there is no job for
legalizer and regbankselect. Instruction select G_FENCE for MIPS32.
Differential Revision: https://reviews.llvm.org/D67181
llvm-svn: 371056
Petar Avramovic [Thu, 5 Sep 2019 11:16:37 +0000 (11:16 +0000)]
[MIPS GlobalISel] Select llvm.trap intrinsic
Select G_INTRINSIC_W_SIDE_EFFECTS for Intrinsic::trap for MIPS32
via legalizeIntrinsic.
Differential Revision: https://reviews.llvm.org/D67180
llvm-svn: 371055
Petar Avramovic [Thu, 5 Sep 2019 11:12:01 +0000 (11:12 +0000)]
[MIPS GlobalISel] Lower SRet pointer arguments
Instead of returning structure by value clang usually adds pointer
to that structure as an argument. Pointers don't require special
handling no matter the SRet flag. Remove unsuccessful exit from
lowerCall for arguments with SRet flag if they are pointers.
Differential Revision: https://reviews.llvm.org/D67179
llvm-svn: 371054
Simon Pilgrim [Thu, 5 Sep 2019 10:38:39 +0000 (10:38 +0000)]
Revert rL370996 from llvm/trunk: [AArch64][GlobalISel] Teach AArch64CallLowering to handle basic sibling calls
This adds support for basic sibling call lowering in AArch64. The intent here is
to only handle tail calls which do not change the ABI (hence, sibling calls.)
At this point, it is very restricted. It does not handle
- Vararg calls.
- Calls with outgoing arguments.
- Calls whose calling conventions differ from the caller's calling convention.
- Tail/sibling calls with BTI enabled.
This patch adds
- `AArch64CallLowering::isEligibleForTailCallOptimization`, which is equivalent
to the same function in AArch64ISelLowering.cpp (albeit with the restrictions
above.)
- `mayTailCallThisCC` and `canGuaranteeTCO`, which are identical to those in
AArch64ISelLowering.cpp.
- `getCallOpcode`, which is exactly what it sounds like.
Tail/sibling calls are lowered by checking if they pass target-independent tail
call positioning checks, and checking if they satisfy
`isEligibleForTailCallOptimization`. If they do, then a tail call instruction is
emitted instead of a normal call. If we have a sibling call (which is always the
case in this patch), then we do not emit any stack adjustment operations. When
we go to lower a return, we check if we've already emitted a tail call. If so,
then we skip the return lowering.
For testing, this patch
- Adds call-translator-tail-call.ll to test which tail calls we currently lower,
which ones we don't, and which ones we shouldn't.
- Updates branch-target-enforcement-indirect-calls.ll to show that we fall back
as expected.
Differential Revision: https://reviews.llvm.org/D67189
........
This fails on EXPENSIVE_CHECKS builds due to a -verify-machineinstrs test failure in CodeGen/AArch64/dllimport.ll
llvm-svn: 371051
Simon Pilgrim [Thu, 5 Sep 2019 10:26:38 +0000 (10:26 +0000)]
[X86] X86SpeculativeLoadHardeningPass::canHardenRegister - fix out of bounds warning.
Fixes clang static-analyzer warning.
llvm-svn: 371050
Jonas Paulsson [Thu, 5 Sep 2019 10:20:05 +0000 (10:20 +0000)]
[SystemZ] Recognize INLINEASM_BR in backend
Handle the remaining cases also by handling asm goto in
SystemZInstrInfo::getBranchInfo().
Review: Ulrich Weigand
https://reviews.llvm.org/D67151
llvm-svn: 371048
Simon Pilgrim [Thu, 5 Sep 2019 10:18:24 +0000 (10:18 +0000)]
[X86] X86InstrInfo::optimizeCompareInstr - fix potential null dereference.
Fixes clang static-analyzer warning.
Technically the MachineInstr *Sub might still be null if we're comparing zero (IsCmpZero == true), although this probably won't happen as SrcReg2 is probably == 0.
llvm-svn: 371047
Sven van Haastregt [Thu, 5 Sep 2019 10:01:24 +0000 (10:01 +0000)]
[OpenCL] Add image type handling for builtins
Image types were previously available, but not working. This patch
adds image type handling.
Rename the image type definitions in the .td file to make them
consistent with other type names. Use abstract types to represent the
unqualified types. Instantiate access-qualified image types at the
point of use using, e.g. `ImageType<Image2d, "RO">`.
Add/update TableGen definitions for the read_image/write_image
builtin functions.
Patch by Pierre Gondois and Sven van Haastregt.
Differential Revision: https://reviews.llvm.org/D63480
llvm-svn: 371046
Guillaume Chatelet [Thu, 5 Sep 2019 10:00:22 +0000 (10:00 +0000)]
[LLVM][Alignment] Make functions using log of alignment explicit
Summary:
This patch renames functions that takes or returns alignment as log2, this patch will help with the transition to llvm::Align.
The renaming makes it explicit that we deal with log(alignment) instead of a power of two alignment.
A few renames uncovered dubious assignments:
- `MirParser`/`MirPrinter` was expecting powers of two but `MachineFunction` and `MachineBasicBlock` were using deal with log2(align). This patch fixes it and updates the documentation.
- `MachineBlockPlacement` exposes two flags (`align-all-blocks` and `align-all-nofallthru-blocks`) supposedly interpreted as power of two alignments, internally these values are interpreted as log2(align). This patch updates the documentation,
- `MachineFunctionexposes` exposes `align-all-functions` also interpreted as power of two alignment, internally this value is interpreted as log2(align). This patch updates the documentation,
Reviewers: lattner, thegameg, courbet
Subscribers: dschuff, arsenm, jyknight, dylanmckay, sdardis, nemanjai, jvesely, nhaehnle, javed.absar, hiraditya, kbarton, fedor.sergeev, asb, rbar, johnrusso, simoncook, apazos, sabuasal, niosHD, jrtc27, MaskRay, zzheng, edward-jones, atanasyan, rogfer01, MartinMosbeck, brucehoult, the_o, dexonsmith, PkmX, jocewei, jsji, Jim, s.egerton, llvm-commits, courbet
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D65945
llvm-svn: 371045
Dmitri Gribenko [Thu, 5 Sep 2019 09:48:39 +0000 (09:48 +0000)]
[libclang] Refactored SharedParsedRegionsStorage
Summary:
Removed the `PPRegionSetTy` typedef because it is only used 3 times, and
obscures code more than it helps.
Renamed SharedParsedRegionsStorage to ThreadSafeParsedRegions, because
that better reflects the reason for this type to exist.
Replaced the `copyTo()` method that had an out parameter with a getter.
Renamed the `merge()` method to `addParsedRegions()`.
Renamed `ParsedSrcLocationsTracker::ParsedRegions` to
`ParsedRegionsSnapshot`, which better reflects its role.
Subscribers: arphaman, cfe-commits
Tags: #clang
Differential Revision: https://reviews.llvm.org/D67077
llvm-svn: 371041
Russell Gallop [Thu, 5 Sep 2019 09:26:04 +0000 (09:26 +0000)]
Fix time-trace breaking flame graph assumptions
-ftime-trace could break flame-graph assumptions on Windows, with an
inner scope overrunning outer scopes. This was due to the way that times
were truncated. Changed this so time_points for the flame-graph are
truncated instead of durations, preserving the relative order of event
starts and ends.
I have tried to retain the extra precision for the totals, which count
thousands or millions of events.
Added assert to check this property holds in future.
Fixes PR43043
Differential Revision: https://reviews.llvm.org/D66411
llvm-svn: 371039
Haojian Wu [Thu, 5 Sep 2019 09:26:03 +0000 (09:26 +0000)]
[clangd][vscode] Add a flag to enable semantic highlighting in clangd
Reviewers: ilya-biryukov
Subscribers: MaskRay, jkorous, arphaman, kadircet, cfe-commits
Tags: #clang
Differential Revision: https://reviews.llvm.org/D67096
llvm-svn: 371038
Haojian Wu [Thu, 5 Sep 2019 09:14:04 +0000 (09:14 +0000)]
[clangd][vscode] Make SemanticHighlightingFeature more self-contained.
Summary:
so that we don't have too many usage from the client side (just a single
occurrance for register), this also aligns with how other builtin feature
being implemented in vscode.
Reviewers: ilya-biryukov
Subscribers: MaskRay, jkorous, arphaman, kadircet, cfe-commits
Tags: #clang
Differential Revision: https://reviews.llvm.org/D67165
llvm-svn: 371036
Hans Wennborg [Thu, 5 Sep 2019 09:07:05 +0000 (09:07 +0000)]
Win: handle \\?\UNC\ prefix in realPathFromHandle (PR43204)
After r361885, realPathFromHandle() ends up getting called on the working
directory on each Clang invocation. This unveiled that the code didn't work for
paths on network shares.
For example, if one maps the local dir c:\src\tmp to x:
net use x: \\localhost\c$\tmp
and run e.g. "clang -c foo.cc" in x:\, realPathFromHandle will get
\\?\UNC\localhost\c$\src\tmp\ back from GetFinalPathNameByHandleW, and would
strip off the initial \\?\ prefix, ending up with a path that doesn't work.
This patch makes the prefix stripping a little smarter to handle this case.
Differential revision: https://reviews.llvm.org/D67166
llvm-svn: 371035
George Rimar [Thu, 5 Sep 2019 08:59:28 +0000 (08:59 +0000)]
[lib/ObjectYAML] - Cleanup the private interface of ELFState<ELFT>. NFCI.
In D62809 I accidentally added "ELFState<ELFT> &State" as the
first parameter to two methods. There is no reason for having that.
I removed this argument and also moved finalizeStrings declaration to
remove an excessive 'private:' tag.
Differential revision: https://reviews.llvm.org/D67157
llvm-svn: 371033
Ilya Biryukov [Thu, 5 Sep 2019 08:59:06 +0000 (08:59 +0000)]
[Sema] Refactor LookupVisibleDecls. NFC
Summary:
We accumulated some configuration parameters for LookupVisibleDecls that
are being passed unchanged to recursive calls, e.g. LoadExternal and
IncludeDependentBases.
At the same time, there is a bunch of parameters that can change in the
recursive invocations.
It is hard to tell the difference between those groups, making the code
hard to follow.
This change introduces a helper struct and factors out the non-changing
bits into fields, making recursive calls in the implementation code easier
to read.
Reviewers: sammccall
Reviewed By: sammccall
Subscribers: riccibruno, doug.gregor, jkorous, arphaman, kadircet, cfe-commits
Tags: #clang
Differential Revision: https://reviews.llvm.org/D65752
llvm-svn: 371032
George Rimar [Thu, 5 Sep 2019 08:52:26 +0000 (08:52 +0000)]
Recommit r371023 "[lib/ObjectYAML] - Stop calling error(1) when mapping the st_other field of a symbol."
Fix: added missing return "return 0;"
Original commit message:
This eliminates one of the error(1) call in this lib.
It is different from the others because happens on a fields mapping stage
and can be easily fixed.
Differential revision: https://reviews.llvm.org/D67150
llvm-svn: 371030
Raphael Isemann [Thu, 5 Sep 2019 08:45:20 +0000 (08:45 +0000)]
[lldb][NFC] Stabilize gui/basic test by waiting on prompt
We should wait on the GUI to exit before tearing down the test.
llvm-svn: 371028
Hans Wennborg [Thu, 5 Sep 2019 08:43:00 +0000 (08:43 +0000)]
Revert r361885 "[Driver] Fix -working-directory issues"
This made clang unable to open files using relative paths on network shares on
Windows (PR43204). On the bug it was pointed out that createPhysicalFileSystem()
is not terribly mature, and using it is risky. Reverting for now until there's
a clear way forward.
> Currently the `-working-directory` option does not actually impact the working
> directory for all of the clang driver, it only impacts how files are looked up
> to make sure they exist. This means that that clang passes the wrong paths
> to -fdebug-compilation-dir and -coverage-notes-file.
>
> This patch fixes that by changing all the places in the driver where we convert
> to absolute paths to use the VFS, and then calling setCurrentWorkingDirectory on
> the VFS. This also changes the default VFS for `Driver` to use a virtualized
> working directory, instead of changing the process's working directory.
>
> Differential Revision: https://reviews.llvm.org/D62271
This also revertes the part of r369938 which checked that -working-directory works.
llvm-svn: 371027
George Rimar [Thu, 5 Sep 2019 08:39:44 +0000 (08:39 +0000)]
Revert r371023 "[lib/ObjectYAML] - Stop calling error(1) when mapping the st_other field of a symbol." (2)
Forgot to revert the cpp file.
llvm-svn: 371025
George Rimar [Thu, 5 Sep 2019 08:38:29 +0000 (08:38 +0000)]
Revert r371023 "[lib/ObjectYAML] - Stop calling error(1) when mapping the st_other field of a symbol."
It broke BBots:
http://lab.llvm.org:8011/builders/lld-x86_64-darwin13/builds/36387/steps/build_Lld/logs/stdio
http://lab.llvm.org:8011/builders/clang-x86_64-debian-fast/builds/17117/steps/test/logs/stdio
llvm-svn: 371024
George Rimar [Thu, 5 Sep 2019 08:28:43 +0000 (08:28 +0000)]
[lib/ObjectYAML] - Stop calling error(1) when mapping the st_other field of a symbol.
This eliminates one of the error(1) call in this lib.
It is different from the others because happens on a fields mapping stage
and can be easily fixed.
Differential revision: https://reviews.llvm.org/D67150
llvm-svn: 371023
Kadir Cetinkaya [Thu, 5 Sep 2019 08:11:21 +0000 (08:11 +0000)]
[clang-tidy] Fix definitions in headers check to respect qualifiers
Summary:
The check was generating a fix without taking qualifiers in return type
into account. This patch changes the insertion location to be before qualifers.
Reviewers: gribozavr
Subscribers: xazax.hun, cfe-commits
Tags: #clang
Differential Revision: https://reviews.llvm.org/D67213
llvm-svn: 371022
Balazs Keri [Thu, 5 Sep 2019 07:59:45 +0000 (07:59 +0000)]
Rename of constants in ASTImporterVisibilityTest. NFC.
Reviewers: martong, a.sidorin, shafik
Reviewed By: shafik
Subscribers: shafik, rnkovacs, dkrupp, Szelethus, gamesh411, cfe-commits
Tags: #clang
Differential Revision: https://reviews.llvm.org/D67174
llvm-svn: 371021
Pavel Labath [Thu, 5 Sep 2019 07:38:55 +0000 (07:38 +0000)]
[dotest] Remove top-level Makefile
This appears to be a really ancient way of running the test suite. It
is also completely disfunctional now, and probably has been for a long
time.
llvm-svn: 371020
Pavel Labath [Thu, 5 Sep 2019 07:35:45 +0000 (07:35 +0000)]
[dotest] Centralize initialization commands even more
Summary:
In r367234 we introduced a central place to hold the set up commands for
the various ways we have of launching lldb. However, a number of
commands still remained outside of that.
This patch moves the remaining set up commands into this function, which
allows us to remove manual clang module path setting code in
TestBatchMode.
One unfortunate victim of this approach is TestSTTYBeforeAndAfter which,
due to how it launches lldb (pexpect->expect->lldb), fails get the
quoting right. It would be possible to fix the quoting there, it would be a bit
icky, and none of the commands in this list are really relevant for what this
test is doing, so I just remove the commands outright.
Reviewers: JDevlieghere, jankratochvil
Subscribers: lldb-commits
Differential Revision: https://reviews.llvm.org/D67173
llvm-svn: 371019
Pavel Labath [Thu, 5 Sep 2019 07:09:31 +0000 (07:09 +0000)]
Obliterate LLDB_CONFIGURATION_BUILDANDINTEGRATION
Summary:
With the XCode project gone, there doesn't seem to be anything setting
this macro anymore -- and the macro wasn't doing much anyway.
Reviewers: jingham, sgraenitz
Subscribers: emaste, lldb-commits
Differential Revision: https://reviews.llvm.org/D66742
llvm-svn: 371018
Pavel Labath [Thu, 5 Sep 2019 07:05:15 +0000 (07:05 +0000)]
Breakpad: Basic support for STACK WIN unwinding
Summary:
This patch makes it possible to unwind via breakpad STACK WIN records.
It is "basic" because two important features are missing:
- support for the .raSearch keyword
- support for multiple STACK WIN records within a single function
Right now, we just reject the .raSearch records, and always pick the
first record for the whole function
SymbolFileBreakpad, and so I think it can serve as a good example of
what is needed of the symbol file and unwinding machinery to make this
work.
However, it is already useful for unwinding in some situations, and it
sets up the general framework for the parsing of these kinds of records,
which reduces the size of the followup patches implementing the two
other components.
Reviewers: amccarth, rnk, markmentovai
Subscribers: lldb-commits
Differential Revision: https://reviews.llvm.org/D67067
llvm-svn: 371017
Igor Kudrin [Thu, 5 Sep 2019 07:02:28 +0000 (07:02 +0000)]
[DWARF] Fix referencing Range List Tables from CUs for DWARF64.
As DW_AT_rnglists_base points after the header and headers have
different sizes for DWARF32 and DWARF64, we have to use the format
of the CU to adjust the offset correctly in order to extract
the referenced range list table.
The patch also changes the type of RangeSectionBase because in DWARF64
it is 8-bytes long.
Differential Revision: https://reviews.llvm.org/D67098
llvm-svn: 371016
Pavel Labath [Thu, 5 Sep 2019 06:55:23 +0000 (06:55 +0000)]
[dotest] Delete trivial inline test makefiles
inline tests are able to generate these automatically
llvm-svn: 371015
Igor Kudrin [Thu, 5 Sep 2019 06:49:05 +0000 (06:49 +0000)]
[DWARF] Support DWARF64 in DWARFListTableHeader.
This enables 64-bit DWARF support for parsing range and location list tables.
Differential Revision: https://reviews.llvm.org/D66643
llvm-svn: 371014
Rui Ueyama [Thu, 5 Sep 2019 05:30:24 +0000 (05:30 +0000)]
Align output segments correctly
Previously, segments were aligned according to their first section's
alignment requirements. That was not correct, but segments are also
aligned to a page boundary, and a page boundary is usually much larger
than a section alignment requirement, so no one noticed this bug before.
Now, lld has --nmagic option which sets maxPageSize to 1 to effectively
disable page alignment, which reveals the issue.
Fixes https://bugs.llvm.org/show_bug.cgi?id=43212
Differential Revision: https://reviews.llvm.org/D67152
llvm-svn: 371013
Matt Arsenault [Thu, 5 Sep 2019 03:00:43 +0000 (03:00 +0000)]
AMDGPU: Add builtins for is_shared/is_private
llvm-svn: 371010
Matt Arsenault [Thu, 5 Sep 2019 02:20:39 +0000 (02:20 +0000)]
AMDGPU: Add intrinsics for address space identification
The library currently uses ptrtoint and directly checks the queue ptr
for this, which counts as a pointer capture.
llvm-svn: 371009
Matt Arsenault [Thu, 5 Sep 2019 02:20:32 +0000 (02:20 +0000)]
AMDGPU/GlobalISel: Restore insert point when getting aperture
Avoids SSA violations in a future patch.
llvm-svn: 371008
Matt Arsenault [Thu, 5 Sep 2019 02:20:29 +0000 (02:20 +0000)]
AMDGPU/GlobalISel: Fix placeholder value used for addrspacecast
llvm-svn: 371007
Matt Arsenault [Thu, 5 Sep 2019 02:20:25 +0000 (02:20 +0000)]
AMDGPU/GlobalISel: Fix assert on load from constant address
llvm-svn: 371006
Puyan Lotfi [Thu, 5 Sep 2019 02:10:41 +0000 (02:10 +0000)]
[mir-canon][NFC] Adding -verify-machineinstrs to mir-canon tests.
In the review process for some of the refactoring of MIRCanonicalizationPass it
was noted that some of the tests didn't have verifier enabled. Enabling here.
llvm-svn: 371005
Richard Smith [Thu, 5 Sep 2019 01:23:47 +0000 (01:23 +0000)]
[c++20] Fix some ambiguities in our mangling of lambdas with explicit
template parameters.
This finishes the implementation of the proposal described in
https://github.com/itanium-cxx-abi/cxx-abi/issues/31. (We already
implemented the <lambda-sig> extensions, but didn't take them into
account when computing mangling numbers, and didn't deal properly with
expanded parameter packs, and didn't disambiguate between different
levels of template parameters in manglings.)
llvm-svn: 371004
Yi Kong [Thu, 5 Sep 2019 01:05:05 +0000 (01:05 +0000)]
Revert "Revert "[builtins] Rounding mode support for addxf3/subxf3""
Test failure fixed.
This reverts commit
e204d244badb2e9765a1020f41c773f63da208f4.
llvm-svn: 371003
Artem Dergachev [Thu, 5 Sep 2019 00:44:56 +0000 (00:44 +0000)]
[analyzer] scan-build: handle --sysroot=/path in addition to --sysroot /path.
Current code assumes flags in CompilerLinkerOptionMap don't use =,
which isn't always true.
Patch by Chris Laplante!
Differential Revision: https://reviews.llvm.org/D66569
llvm-svn: 371002
Marshall Clow [Thu, 5 Sep 2019 00:38:36 +0000 (00:38 +0000)]
Reformat the beginning of the testing doc to make clear how to run all the tests.
llvm-svn: 371001
Reid Kleckner [Thu, 5 Sep 2019 00:34:01 +0000 (00:34 +0000)]
Use -mtriple to fix AMDGPU test sensitive to object file format
GOTPCREL32 doesn't exist on COFF, so it isn't used when this test runs
on Windows.
llvm-svn: 371000
Jonas Devlieghere [Wed, 4 Sep 2019 23:05:32 +0000 (23:05 +0000)]
[Disassembler] Simplify a few methods (2/2) (NFC)
Use early returns to highlight preconditions and make the code easier to
follow.
llvm-svn: 370998
Jessica Paquette [Wed, 4 Sep 2019 22:54:52 +0000 (22:54 +0000)]
[AArch64][GlobalISel] Teach AArch64CallLowering to handle basic sibling calls
This adds support for basic sibling call lowering in AArch64. The intent here is
to only handle tail calls which do not change the ABI (hence, sibling calls.)
At this point, it is very restricted. It does not handle
- Vararg calls.
- Calls with outgoing arguments.
- Calls whose calling conventions differ from the caller's calling convention.
- Tail/sibling calls with BTI enabled.
This patch adds
- `AArch64CallLowering::isEligibleForTailCallOptimization`, which is equivalent
to the same function in AArch64ISelLowering.cpp (albeit with the restrictions
above.)
- `mayTailCallThisCC` and `canGuaranteeTCO`, which are identical to those in
AArch64ISelLowering.cpp.
- `getCallOpcode`, which is exactly what it sounds like.
Tail/sibling calls are lowered by checking if they pass target-independent tail
call positioning checks, and checking if they satisfy
`isEligibleForTailCallOptimization`. If they do, then a tail call instruction is
emitted instead of a normal call. If we have a sibling call (which is always the
case in this patch), then we do not emit any stack adjustment operations. When
we go to lower a return, we check if we've already emitted a tail call. If so,
then we skip the return lowering.
For testing, this patch
- Adds call-translator-tail-call.ll to test which tail calls we currently lower,
which ones we don't, and which ones we shouldn't.
- Updates branch-target-enforcement-indirect-calls.ll to show that we fall back
as expected.
Differential Revision: https://reviews.llvm.org/D67189
llvm-svn: 370996
Jonas Devlieghere [Wed, 4 Sep 2019 22:38:20 +0000 (22:38 +0000)]
[Disassembler] Simplify a few methods (NFC)
Use early returns to highlight preconditions and make the code easier to
follow.
llvm-svn: 370994
Richard Smith [Wed, 4 Sep 2019 22:14:50 +0000 (22:14 +0000)]
For PR43213, track whether template parameters are implicit through
template instantiation so we know whether to mangle them in
lambda-expressions.
llvm-svn: 370991
Nico Weber [Wed, 4 Sep 2019 21:34:21 +0000 (21:34 +0000)]
gn build: Merge r370985
llvm-svn: 370988
Richard Smith [Wed, 4 Sep 2019 21:31:22 +0000 (21:31 +0000)]
[www] Fix hyperlink syntax in attribute reference.
llvm-svn: 370987
Richard Smith [Wed, 4 Sep 2019 21:31:21 +0000 (21:31 +0000)]
[www] Update attribute reference for 'constinit'.
llvm-svn: 370986
Puyan Lotfi [Wed, 4 Sep 2019 21:29:10 +0000 (21:29 +0000)]
[mir-canon][NFC] Move MIR vreg renaming code to separate file for better reuse.
Moving MIRCanonicalizerPass vreg renaming code to MIRVRegNamerUtils so that it
can be reused in another pass (ie planing to write a standalone mir-namer pass).
I'm going to write a mir-namer pass so that next time someone has to author a
test in MIR, they can use it to cleanup the naming and make it more readable by
having the numbered vregs swapped out with named vregs.
Differential Revision: https://reviews.llvm.org/D67114
llvm-svn: 370985
Jonas Devlieghere [Wed, 4 Sep 2019 21:18:46 +0000 (21:18 +0000)]
[test] Escape path to match the literal string
This test was failing when you had things like `+` in your build path.
llvm-svn: 370983
Aaron Ballman [Wed, 4 Sep 2019 21:01:57 +0000 (21:01 +0000)]
Diagnose _Atomic as a C11 extension.
llvm-svn: 370982
Yuanfang Chen [Wed, 4 Sep 2019 20:58:15 +0000 (20:58 +0000)]
[DebugInfo] Emit DW_TAG_enumeration_type for referenced global enumerator.
This essentially reverts changes from r361400 while keeping behavior for
CodeView.
Reviewers: akhuang, rnk, probinson
Reviewed by: rnk
Subscribers: cfe-commits
Tags: #clang
Differential Revision: https://reviews.llvm.org/D67141
llvm-svn: 370981
Matt Arsenault [Wed, 4 Sep 2019 20:46:31 +0000 (20:46 +0000)]
AMDGPU/GlobalISel: Select G_BITREVERSE
llvm-svn: 370980
Matt Arsenault [Wed, 4 Sep 2019 20:46:15 +0000 (20:46 +0000)]
GlobalISel: Add basic legalization for G_BITREVERSE
llvm-svn: 370979
Johannes Doerfert [Wed, 4 Sep 2019 20:34:57 +0000 (20:34 +0000)]
[Attributor][Stats] Use the right statistics macro
llvm-svn: 370976
Johannes Doerfert [Wed, 4 Sep 2019 20:34:52 +0000 (20:34 +0000)]
[Attributor][Fix] Make sure we do not delete live code
Summary: Liveness needs to mark edges, not blocks as dead.
Reviewers: sstefan1, uenoku
Subscribers: hiraditya, bollu, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D67191
llvm-svn: 370975
Martin Storsjo [Wed, 4 Sep 2019 20:34:00 +0000 (20:34 +0000)]
[LLD] [COFF] Implement MinGW default manifest handling
In mingw environments, resources are normally compiled to resource
object files directly, instead of letting the linker convert them to
COFF format.
Since some time, GCC supports the notion of a default manifest object.
When invoking the linker, GCC looks for the default manifest object
file, and if found in the expected path, it is added to linker commands.
The default manifest is one that indicates support for the latest known
versions of windows, to implicitly unlock the modern behaviours of certain
APIs.
Not all mingw/gcc distributions include this file, but e.g. in msys2,
the default manifest object is distributed in a separate package (which
can be but might not always be installed).
This means that even if user projects only use one single resource
object file, the linker can end up with two resource object files,
and thus needs to support merging them.
The default manifest has a language id of zero, and GNU ld has got
logic for dropping a manifest with a zero language id, if there's
another manifest present with a nonzero language id. If there are
multiple manifests with a nonzero language id, the merging process
errors out.
Differential Revision: https://reviews.llvm.org/D66825
llvm-svn: 370974
Richard Smith [Wed, 4 Sep 2019 20:30:37 +0000 (20:30 +0000)]
[c++20] P1143R2: Add support for the C++20 'constinit' keyword.
This is mostly the same as the
[[clang::require_constant_initialization]] attribute, but has a couple
of additional syntactic and semantic restrictions.
In passing, I added a warning for the attribute form being added after
we have already seen the initialization of the variable (but before we
see the definition); that case previously slipped between the cracks and
the attribute was silently ignored.
llvm-svn: 370972
Leonard Chan [Wed, 4 Sep 2019 20:30:29 +0000 (20:30 +0000)]
[NewPM][Sancov] Make Sancov a Module Pass instead of 2 Passes
This patch merges the sancov module and funciton passes into one module pass.
The reason for this is because we ran into an out of memory error when
attempting to run asan fuzzer on some protobufs (pc.cc files). I traced the OOM
error to the destructor of SanitizerCoverage where we only call
appendTo[Compiler]Used which calls appendToUsedList. I'm not sure where precisely
in appendToUsedList causes the OOM, but I am able to confirm that it's calling
this function *repeatedly* that causes the OOM. (I hacked sancov a bit such that
I can still create and destroy a new sancov on every function run, but only call
appendToUsedList after all functions in the module have finished. This passes, but
when I make it such that appendToUsedList is called on every sancov destruction,
we hit OOM.)
I don't think the OOM is from just adding to the SmallSet and SmallVector inside
appendToUsedList since in either case for a given module, they'll have the same
max size. I suspect that when the existing llvm.compiler.used global is erased,
the memory behind it isn't freed. I could be wrong on this though.
This patch works around the OOM issue by just calling appendToUsedList at the
end of every module run instead of function run. The same amount of constants
still get added to llvm.compiler.used, abd we make the pass usage and logic
simpler by not having any inter-pass dependencies.
Differential Revision: https://reviews.llvm.org/D66988
llvm-svn: 370971