Ties Stuij [Fri, 28 Aug 2020 14:08:02 +0000 (15:08 +0100)]
[ARM] Follow AACPS standard for volatile bit-fields access width
This patch resumes the work of D16586.
According to the AAPCS, volatile bit-fields should
be accessed using containers of the widht of their
declarative type. In such case:
```
struct S1 {
short a : 1;
}
```
should be accessed using load and stores of the width
(sizeof(short)), where now the compiler does only load
the minimum required width (char in this case).
However, as discussed in D16586,
that could overwrite non-volatile bit-fields, which
conflicted with C and C++ object models by creating
data race conditions that are not part of the bit-field,
e.g.
```
struct S2 {
short a;
int b : 16;
}
```
Accessing `S2.b` would also access `S2.a`.
The AAPCS Release 2020Q2
(https://documentation-service.arm.com/static/
5efb7fbedbdee951c1ccf186?token=)
section 8.1 Data Types, page 36, "Volatile bit-fields -
preserving number and width of container accesses" has been
updated to avoid conflict with the C++ Memory Model.
Now it reads in the note:
```
This ABI does not place any restrictions on the access widths of bit-fields where the container
overlaps with a non-bit-field member or where the container overlaps with any zero length bit-field
placed between two other bit-fields. This is because the C/C++ memory model defines these as being
separate memory locations, which can be accessed by two threads simultaneously. For this reason,
compilers must be permitted to use a narrower memory access width (including splitting the access into
multiple instructions) to avoid writing to a different memory location. For example, in
struct S { int a:24; char b; }; a write to a must not also write to the location occupied by b, this requires at least two
memory accesses in all current Arm architectures. In the same way, in struct S { int a:24; int:0; int b:8; };,
writes to a or b must not overwrite each other.
```
Patch D16586 was updated to follow such behavior by verifying that we
only change volatile bit-field access when:
- it won't overlap with any other non-bit-field member
- we only access memory inside the bounds of the record
- avoid overlapping zero-length bit-fields.
Regarding the number of memory accesses, that should be preserved, that will
be implemented by D67399.
Differential Revision: https://reviews.llvm.org/D72932
The following people contributed to this patch:
- Diogo Sampaio
- Ties Stuij
Volkan Keles [Tue, 8 Sep 2020 16:46:38 +0000 (09:46 -0700)]
GlobalISel: Combine `op undef, x` to 0
https://reviews.llvm.org/D86611
Heejin Ahn [Sun, 6 Sep 2020 17:36:07 +0000 (10:36 -0700)]
[WebAssembly] Fix fixEndsAtEndOfFunction for try-catch
When the function return type is non-void and `end` instructions are at
the very end of a function, CFGStackify's `fixEndsAtEndOfFunction`
function fixes the corresponding block/loop/try's type to match the
function's return type. This is applied to consecutive `end` markers at
the end of a function. For example, when the function return type is
`i32`,
```
block i32 ;; return type is fixed to i32
...
loop i32 ;; return type is fixed to i32
...
end_loop
end_block
end_function
```
But try-catch is a little different, because it consists of two parts:
a try part and a catch part, and both parts' return type should satisfy
the function's return type. Which means,
```
try i32 ;; return type is fixed to i32
...
block i32 ;; this should be changed i32 too!
...
end_block
catch
...
end_try
end_function
```
As you can see in this example, it is not sufficient to only `end`
instructions at the end of a function; in case of `try`, we should
check instructions before `catch`es, in case their corresponding `try`'s
type has been fixed.
This changes `fixEndsAtEndOfFunction`'s algorithm to use a worklist
that contains a reverse iterator, each of which is a starting point for
a new backward `end` instruction search.
Fixes https://bugs.llvm.org/show_bug.cgi?id=47413.
Reviewed By: dschuff, tlively
Differential Revision: https://reviews.llvm.org/D87207
Simon Pilgrim [Tue, 8 Sep 2020 16:21:28 +0000 (17:21 +0100)]
LiveRegUnits.h - reduce MachineRegisterInfo.h include. NFC.
We only need to include MachineInstrBundle.h, but exposes an implicit dependency in MachineOutliner.h.
Also, remove duplicate includes from LiveRegUnits.cpp + MachineOutliner.cpp.
Lubomir Litchev [Thu, 3 Sep 2020 20:15:39 +0000 (13:15 -0700)]
Add an option for unrolling loops up to a factor.
Currently, there is no option to allow for unrolling a loop up to a specific factor (specified by the user).
The code for doing that is there and there are benefits when unrolling is done to smaller loops (smaller than the factor specified).
Reviewed By: bondhugula
Differential Revision: https://reviews.llvm.org/D87111
Heejin Ahn [Tue, 8 Sep 2020 16:20:06 +0000 (09:20 -0700)]
[clang-tidy] Fix linking for FrontendOpenMP
Without this, builds with `-DBUILD_SHARED_LIBS=ON` fail.
Ronak Chauhan [Mon, 7 Sep 2020 09:10:00 +0000 (14:40 +0530)]
[AMDGPU] Support disassembly for AMDGPU kernel descriptors
Decode AMDGPU Kernel descriptors as assembler directives.
Reviewed By: scott.linder, jhenderson, kzhuravl
Differential Revision: https://reviews.llvm.org/D80713
mydeveloperday [Tue, 8 Sep 2020 15:39:11 +0000 (16:39 +0100)]
[clang-format] Handle shifts within conditions
In some situation shifts can be treated as a template, and is thus formatted as one. So, by doing a couple extra checks to assure that the condition doesn't contain a template, and is in fact a bit shift should solve this problem.
This is a fix for [[ https://bugs.llvm.org/show_bug.cgi?id=46969 | bug 46969 ]]
Reviewed By: MyDeveloperDay
Patch By: Saldivarcher
Differential Revision: https://reviews.llvm.org/D86581
Louis Dionne [Tue, 8 Sep 2020 15:29:32 +0000 (11:29 -0400)]
[libc++] Allow overriding the cached value of LIBCXX_TEST_CONFIG
Louis Dionne [Tue, 8 Sep 2020 15:17:10 +0000 (11:17 -0400)]
[libc++] Make sure we always print all available features
Previously, we'd only print the features added through the new config,
however printing all the features is important for debugging purposes.
Jonas Paulsson [Wed, 19 Aug 2020 10:01:03 +0000 (12:01 +0200)]
[DAGTypeLegalizer] Handle ZERO_EXTEND of promoted type in WidenVecRes_Convert.
On SystemZ, a ZERO_EXTEND of an i1 vector handled by WidenVecRes_Convert()
always ended up being scalarized, because the type action of the input is
promotion which was previously an unhandled case in this method.
This fixes https://bugs.llvm.org/show_bug.cgi?id=47132.
Differential Revision: https://reviews.llvm.org/D86268
Patch by Eli Friedman.
Review: Ulrich Weigand
Haojian Wu [Tue, 8 Sep 2020 14:26:48 +0000 (16:26 +0200)]
[clang-tidy] Fix dynamic build failures after
156b127945a8c923d141e608b7380427da024376
Hans Wennborg [Tue, 8 Sep 2020 14:09:33 +0000 (16:09 +0200)]
Revert
3e782bf809 "[Sema][MSVC] warn at dynamic_cast when /GR- is given"
This caused more warnings than expected, see https://crbug.com/1126019
Also reverts the follow-up
7907e5516.
> Differential Revision: https://reviews.llvm.org/D86369
Nico Weber [Tue, 8 Sep 2020 14:02:00 +0000 (10:02 -0400)]
StructPackAlignCheck: Fix a -Winconsistent-missing-override warning
Nico Weber [Tue, 8 Sep 2020 14:00:24 +0000 (10:00 -0400)]
[gn build] (manually) port
156b127945a8
Florian Hahn [Fri, 4 Sep 2020 15:44:58 +0000 (16:44 +0100)]
[DSE,MemorySSA] Increase walker limit a bit.
This slightly bumps the walker limit so that it covers more cases while
not increasing compile-time too much:
http://llvm-compile-time-tracker.com/compare.php?from=
0fc1c2b51ba0cfb9145139af35be638333865251&to=
91144a50ea4fa82c0c877e77784f60371640b263&stat=instructions
Sam Parker [Tue, 8 Sep 2020 13:41:42 +0000 (14:41 +0100)]
[NFC][ARM] Precommit test
Benjamin Kramer [Tue, 8 Sep 2020 13:40:14 +0000 (15:40 +0200)]
[mlir][VectorOps] Fix more GCC5 weirdness
VectorToSCF.cpp:515:47: error: specialization of 'template<class TransferOpTy> mlir::LogicalResult mlir::VectorTransferRewriter<TransferOpTy>::matchAndRewrite(mlir::Operation*, mlir::PatternRewriter&) const' in different namespace [-fpermissive]
Haojian Wu [Tue, 8 Sep 2020 13:34:52 +0000 (15:34 +0200)]
[clang] Limit the maximum level of fold-expr expansion.
Introduce a new diagnostic, and respect the bracket-depth (256) by default.
Differential Revision: https://reviews.llvm.org/D86936
Frank Derry Wanye [Tue, 8 Sep 2020 13:35:14 +0000 (09:35 -0400)]
Add a new altera check for structure packing and alignment.
The altera struct pack align lint check finds structs that are inefficiently
packed or aligned and recommends packing/aligning of the structs using the
packed and aligned attributes as needed in a warning.
Sanjay Patel [Tue, 8 Sep 2020 13:17:01 +0000 (09:17 -0400)]
[InstCombine] add bitwise logic fold tests for D86395; NFC
Denys Petrov [Fri, 4 Sep 2020 12:03:09 +0000 (15:03 +0300)]
[analyzer] [NFC] Introduce refactoring of PthreadLockChecker
Change capitalization of some names due to LLVM naming rules.
Change names of some variables to make them more speaking.
Rework similar bug reports into one common function.
Prepare code for the next patches to reduce unrelated changes.
Differential Revision: https://reviews.llvm.org/D87138
Raul Tambre [Sat, 5 Sep 2020 14:52:23 +0000 (17:52 +0300)]
[CMake] Remove dead FindPythonInterp code
LLVM has bumped the minimum required CMake version to 3.13.4, so this has become dead code.
Reviewed By: #libc, ldionne
Differential Revision: https://reviews.llvm.org/D87189
Ehsan Toosi [Mon, 24 Aug 2020 11:19:50 +0000 (13:19 +0200)]
[mlir] Fix bug in copy removal
A crash could happen due to copy removal. The bug is fixed and two more
test cases are added.
Differential Revision: https://reviews.llvm.org/D87128
Benjamin Kramer [Tue, 8 Sep 2020 12:02:46 +0000 (14:02 +0200)]
[mlir][VectorOps]
Put back anonymous namespace to work around GCC5 bug.
VectorToSCF.cpp:241:61: error: specialization of 'template<class ConcreteOp> mlir::LogicalResult {anonymous}::NDTransferOpHelper<ConcreteOp>::doReplace()' in different namespace [-fpermissive]
Simon Pilgrim [Tue, 8 Sep 2020 12:01:09 +0000 (13:01 +0100)]
[Codegen][X86] Begin moving X86 specific codegen tests into X86 subfolder.
Discussed with @craig.topper and @spatel - this is to try and tidyup the codegen folder and move the x86 specific tests (as opposed to general tests that just happen to use x86 triples) into subfolders. Its up to other targets if they follow suit.
It also helps speed up test iterations as using wildcards on lit commands often misses some filenames.
Simon Pilgrim [Tue, 8 Sep 2020 11:46:00 +0000 (12:46 +0100)]
X86CallLowering.cpp - improve auto const/pointer/reference qualifiers. NFCI.
Fix clang-tidy warnings by ensuring auto variables are more cleanly qualified, or just avoid auto entirely.
Simon Pilgrim [Tue, 8 Sep 2020 11:45:08 +0000 (12:45 +0100)]
X86DomainReassignment.cpp - improve auto const/pointer/reference qualifiers. NFCI.
Fix clang-tidy warnings by ensuring auto variables are more cleanly qualified, or just avoid auto entirely.
Xing GUO [Tue, 8 Sep 2020 11:55:14 +0000 (19:55 +0800)]
[DWARFYAML] Make the debug_ranges section optional.
This patch makes the debug_ranges section optional. When we specify an
empty debug_ranges section, yaml2obj only emits the section header.
Reviewed By: jhenderson
Differential Revision: https://reviews.llvm.org/D87263
Ehsan Toosi [Thu, 20 Aug 2020 10:56:19 +0000 (12:56 +0200)]
[mlir] remove BufferAssignmentPlacer from BufferAssignmentOpConversionPattern
BufferPlacement has been removed, as allocations are no longer placed during the conversion.
Differential Revision: https://reviews.llvm.org/D87079
Simon Pilgrim [Tue, 8 Sep 2020 10:59:38 +0000 (11:59 +0100)]
[OpenMP] Fix typo in CodeGenFunction::EmitOMPWorksharingLoop (PR46412)
Fixes issue noticed by static analysis where we have a copy+paste typo, testing ScheduleKind.M1 twice instead of ScheduleKind.M2.
Differential Revision: https://reviews.llvm.org/D87250
Benjamin Kramer [Tue, 8 Sep 2020 09:57:50 +0000 (11:57 +0200)]
[mlir][VectorOps] Clean up outdated comments. NFCI.
While there
- De-templatify code that can use function_ref
- Make BoundCaptures usable when they're const
- Address post-submit review comment (static function into global namespace)
Eduardo Caldas [Mon, 7 Sep 2020 17:47:09 +0000 (17:47 +0000)]
[SyntaxTree] Fix crash on functions with default arguments.
* Do not visit `CXXDefaultArgExpr`
* To build `CallArguments` nodes, just go through non-default arguments
Differential Revision: https://reviews.llvm.org/D87249
Eduardo Caldas [Mon, 7 Sep 2020 08:40:49 +0000 (08:40 +0000)]
[SyntaxTree] Ignore implicit `CXXFunctionalCastExpr` wrapping constructor
Differential Revision: https://reviews.llvm.org/D87229
Eduardo Caldas [Thu, 27 Aug 2020 09:44:09 +0000 (09:44 +0000)]
[SyntaxTree] Ignore implicit leaf `CXXConstructExpr`
Differential Revision: https://reviews.llvm.org/D86700
Eduardo Caldas [Thu, 27 Aug 2020 09:43:14 +0000 (09:43 +0000)]
[SyntaxTree] Ignore implicit non-leaf `CXXConstructExpr`
Differential Revision: https://reviews.llvm.org/D86699
Jakub Lichman [Tue, 8 Sep 2020 09:42:25 +0000 (09:42 +0000)]
[mlir] Fix of broken build on windows caused by using uint
Sam Tebbs [Mon, 7 Sep 2020 09:39:14 +0000 (10:39 +0100)]
[ARM][LowOverheadLoops] Remove modifications to the correct element
count register
After my patch at D86087, code that now uses the mov operand rather than
the vctp operand will no longer remove modifications to the vctp operand
as they should. This patch fixes that by explicitly removing
modifications to the vctp operand rather than the register used as the
element count.
Qiu Chaofan [Tue, 8 Sep 2020 09:20:00 +0000 (17:20 +0800)]
Revert "[PowerPC] Implement instruction clustering for stores"
This reverts commit
3c0b3250230b3847a2a47dfeacfdb794c2285f02, (along
with
ea795304 and
bb39eb9e) since it breaks test with UB sanitizer.
Benjamin Kramer [Mon, 7 Sep 2020 15:39:16 +0000 (17:39 +0200)]
[mlir][VectorOps] Redo the scalar loop emission in VectoToSCF to pad instead of clipping
This replaces the select chain for edge-padding with an scf.if that
performs the memory operation when the index is in bounds and uses the
pad value when it's not. For transfer_write the same mechanism is used,
skipping the store when the index is out of bounds.
The integration test has a bunch of cases of how I believe this should
work.
Differential Revision: https://reviews.llvm.org/D87241
Jakub Lichman [Tue, 8 Sep 2020 08:31:52 +0000 (08:31 +0000)]
[mlir] Conv ops vectorization pass
In this commit a new way of convolution ops lowering is introduced.
The conv op vectorization pass lowers linalg convolution ops
into vector contractions. This lowering is possible when conv op
is first tiled by 1 along specific dimensions which transforms
it into dot product between input and kernel subview memory buffers.
This pass converts such conv op into vector contraction and does
all necessary vector transfers that make it work.
Differential Revision: https://reviews.llvm.org/D86619
Serge Guelton [Thu, 25 Jun 2020 09:57:01 +0000 (05:57 -0400)]
Provide anchor for compiler extensions
This patch is cherry-picked from
04b0a4e22e3b4549f9d241f8a9f37eebecb62a31, and
amended to prevent an undefined reference to `llvm::EnableABIBreakingChecks'
Shivanshu Goyal [Tue, 8 Sep 2020 08:17:05 +0000 (10:17 +0200)]
getClangStripDependencyFileAdjuster(): Do not remove -M args when using MSVC cl driver
MSVC's cl.exe has a few command line arguments which start with -M such
as "-MD", "-MDd", "-MT", "-MTd", "-MP".
These arguments are not dependency file generation related, and these
arguments were being removed by getClangStripDependencyFileAdjuster()
which was wrong.
Differential revision: https://reviews.llvm.org/D86999
Nicolas Vasilache [Tue, 8 Sep 2020 07:39:23 +0000 (03:39 -0400)]
[mlir][Vector] Make VectorToSCF deterministic
Differential Revision: https://reviews.llvm.org/D87273
Xing GUO [Tue, 8 Sep 2020 08:08:42 +0000 (16:08 +0800)]
[obj2yaml] Stop parsing the debug_str section when it encounters a string without the null terminator.
When obj2yaml encounters a string without the null terminator, it should
stop parsing the debug_str section. This patch addresses comments in
[D86867](https://reviews.llvm.org/D86867#inline-803291).
Reviewed By: jhenderson
Differential Revision: https://reviews.llvm.org/D87261
Richard Barton [Thu, 3 Sep 2020 10:44:03 +0000 (11:44 +0100)]
[flang] Convert release notes to markdown
Switch ReleaseNotes from .rst to .md to match the other docs.
At the same time, fix the version number for master.
Max Kazantsev [Tue, 8 Sep 2020 07:33:47 +0000 (14:33 +0700)]
[Test] More tests where IndVars fails to eliminate a range check
Qiu Chaofan [Tue, 8 Sep 2020 07:30:16 +0000 (15:30 +0800)]
[PowerPC] Fix getMemOperandWithOffsetWidth
Commit
3c0b3250 introduced memory cluster under pwr10 target, but a
check for operands was unexpectedly removed. This adds it back to avoid
regression.
Simon Wallis [Tue, 8 Sep 2020 07:04:52 +0000 (08:04 +0100)]
[AARCH64][RegisterCoalescer] clang miscompiles zero-extension to long long
Implement AArch64 variant of shouldCoalesce() to detect a known failing case
and prevent the coalescing of a 32-bit copy into a 64-bit sign-extending load.
Do not coalesce in the following case:
COPY where source is bottom 32 bits of a 64-register,
and destination is a 32-bit subregister of a 64-bit register,
ie it causes the rest of the register to be implicitly set to zero.
A mir test has been added.
In the test case, the 32-bit copy implements a 32 to 64 bit zero extension
and relies on the upper 32 bits being zeroed.
Coalescing to the result of the 64-bit load meant overwriting
the upper 32 bits incorrectly when the loaded byte was negative.
Reviewed By: john.brawn
Differential Revision: https://reviews.llvm.org/D85956
Mikael Holmen [Tue, 8 Sep 2020 06:05:47 +0000 (08:05 +0200)]
[PowerPC] Add parentheses to silence gcc warning
Without gcc 7.4 warns with
../lib/Target/PowerPC/PPCInstrInfo.cpp:2284:25: warning: suggest parentheses around '&&' within '||' [-Wparentheses]
BaseOp1.isFI() &&
~~~~~~~~~~~~~~~^~
"Only base registers and frame indices are supported.");
~
Martin Storsjö [Tue, 8 Sep 2020 06:26:39 +0000 (09:26 +0300)]
[clang] Remove a stray semicolon, fixing pedantic GCC warnings. NFC.
Andrew Wei [Tue, 8 Sep 2020 05:14:53 +0000 (13:14 +0800)]
[LSR] Canonicalize a formula before insert it into the list
In GenerateConstantOffsetsImpl, we may generate non canonical Formula
if BaseRegs of that Formula is updated and includes a recurrent expr reg
related with current loop while its ScaledReg is not.
Patched by: mdchen
Reviewed By: qcolombet
Differential Revision: https://reviews.llvm.org/D86939
Fangrui Song [Tue, 8 Sep 2020 04:44:26 +0000 (21:44 -0700)]
[ELF] --symbol-ordering-file: optimize a loop
Johannes Doerfert [Tue, 18 Aug 2020 20:32:21 +0000 (15:32 -0500)]
[Attributor][FIX] Don't crash on internalizing linkonce_odr hidden functions
The CloneFunctionInto has implicit requirements with regards to the
linkage and visibility of the function. We now update these after we did
the CloneFunctionInto on the copy with the same linkage and visibility
as the original.
Johannes Doerfert [Fri, 4 Sep 2020 04:42:33 +0000 (23:42 -0500)]
[Attributor][NFC] Cleanup internalize test case
One run line was different and probably introduced for the manually
added function attribute & name checks. We can do this with the script
and a check prefix used for the other run lines as well.
Johannes Doerfert [Thu, 3 Sep 2020 21:13:28 +0000 (16:13 -0500)]
[Attributor][NFC] Change variable spelling
Johannes Doerfert [Sat, 5 Sep 2020 18:26:20 +0000 (13:26 -0500)]
[Attributor][NFC] Clang tidy: no else after continue
Johannes Doerfert [Thu, 3 Sep 2020 16:08:39 +0000 (11:08 -0500)]
[Attributor][NFC] Expand `auto` types (clang-fix-it)
Johannes Doerfert [Fri, 28 Aug 2020 01:34:07 +0000 (20:34 -0500)]
[Attributor][FIX] Properly return changed if the IR was modified
Deleting or replacing anything is certainly a modification. This caused
a later assertion in IPSCCP when compiling 400.perlbench with the new PM.
I'm not sure how to test this.
Max Kazantsev [Tue, 8 Sep 2020 04:14:36 +0000 (11:14 +0700)]
[Test] Auto-generated checks for some IndVarSimplify tests
Zequan Wu [Tue, 8 Sep 2020 03:55:05 +0000 (20:55 -0700)]
[Sema] fix /gr warning test case
Qiu Chaofan [Tue, 8 Sep 2020 03:03:09 +0000 (11:03 +0800)]
[PowerPC] Implement instruction clustering for stores
On Power10, it's profitable to schedule some stores with adjacent target
address together. This patch implements this feature.
Reviewed By: steven.zhang
Differential Revision: https://reviews.llvm.org/D86754
Alexander Shaposhnikov [Tue, 8 Sep 2020 01:29:48 +0000 (18:29 -0700)]
[llvm-objcopy] Consolidate and unify version tests
In this diff the tests which verify version printing functionality are refactored.
Since they are not specific to a particular format we move them into tool-version.test
and slightly unify (similarly to tool-name.test and tool-help-message.test).
Test plan: make check-all
Differential revision: https://reviews.llvm.org/D87211
Kiran Kumar T P [Tue, 8 Sep 2020 01:22:07 +0000 (06:52 +0530)]
[flang][OpenMP] Enhance parser support for atomic construct to OpenMP 5.0
Summary:
This patch enhances parser support for atomic construct to OpenMP 5.0.
2.17.7 atomic -> ATOMIC [clause [,]] atomic-clause [[,] clause] |
ATOMIC [clause]
clause -> memory-order-clause | HINT(hint-expression)
memory-order-clause -> SEQ_CST | ACQ_REL | RELEASE | ACQUIRE | RELAXED
atomic-clause -> READ | WRITE | UPDATE | CAPTURE
The patch includes code changes and testcase modifications.
Reviewed By: DavidTruby, kiranchandramohan, sameeranjoshi
Differential Revision: https://reviews.llvm.org/D82931
Craig Topper [Tue, 8 Sep 2020 00:57:39 +0000 (17:57 -0700)]
[builtins] Inline __paritysi2 into __paritydi2 and inline __paritydi2 into __parityti2.
No point in making __parityti2 go through 2 calls to get to
__paritysi2.
Reviewed By: MaskRay, efriedma
Differential Revision: https://reviews.llvm.org/D87218
Mehdi Amini [Tue, 8 Sep 2020 00:56:10 +0000 (00:56 +0000)]
Update SVG images to be properly cropped (NFC)
Mehdi Amini [Tue, 8 Sep 2020 00:06:37 +0000 (00:06 +0000)]
Add a doc/tutorial on traversing the IR
Reviewed By: stephenneuendorffer
Differential Revision: https://reviews.llvm.org/D87221
Mehdi Amini [Mon, 7 Sep 2020 23:58:54 +0000 (23:58 +0000)]
Add documentation for getDependentDialects() in the PassManagement infra docs
Reviewed By: rriddle
Differential Revision: https://reviews.llvm.org/D87181
Zequan Wu [Fri, 21 Aug 2020 20:42:20 +0000 (13:42 -0700)]
[Sema][MSVC] warn at dynamic_cast when /GR- is given
Differential Revision: https://reviews.llvm.org/D86369
Florian Hahn [Mon, 7 Sep 2020 21:52:10 +0000 (22:52 +0100)]
[DSE,MemorySSA] Add an early check for read clobbers to traversal.
Depending on the benchmark, this early exit can save a substantial
amount of compile-time:
http://llvm-compile-time-tracker.com/compare.php?from=
505f2d817aa8e07ba98e5fd4a8f6ff0666f89df1&to=
eb4e441147f9b4b7a5fcbbc57428cadbe9e01f10&stat=instructions
Fangrui Song [Mon, 7 Sep 2020 21:44:53 +0000 (14:44 -0700)]
[asan][test] Use --image-base for Linux/asan_prelink_test.cpp if ld is LLD
LLD supports -Ttext but with the option there is still a PT_LOAD at address zero
and thus the Linux kernel will map it to a different address and the test will fail.
Use --image-base instead.
Roman Lebedev [Mon, 7 Sep 2020 20:54:06 +0000 (23:54 +0300)]
Reland [SimplifyCFG][LoopRotate] SimplifyCFG: disable common instruction hoisting by default, enable late in pipeline
This was reverted in
503deec2183d466dad64b763bab4e15fd8804239
because it caused gigantic increase (3x) in branch mispredictions
in certain benchmarks on certain CPU's,
see https://reviews.llvm.org/D84108#2227365.
It has since been investigated and here are the results:
https://lists.llvm.org/pipermail/llvm-commits/Week-of-Mon-
20200907/827578.html
> It's an amazingly severe regression, but it's also all due to branch
> mispredicts (about 3x without this). The code layout looks ok so there's
> probably something else to deal with. I'm not sure there's anything we can
> reasonably do so we'll just have to take the hit for now and wait for
> another code reorganization to make the branch predictor a bit more happy :)
>
> Thanks for giving us some time to investigate and feel free to recommit
> whenever you'd like.
>
> -eric
So let's just reland this.
Original commit message:
I've been looking at missed vectorizations in one codebase.
One particular thing that stands out is that some of the loops
reach vectorizer in a rather mangled form, with weird PHI's,
and some of the loops aren't even in a rotated form.
After taking a more detailed look, that happened because
the loop's headers were too big by then. It is evident that
SimplifyCFG's common code hoisting transform is at fault there,
because the pattern it handles is precisely the unrotated
loop basic block structure.
Surprizingly, `SimplifyCFGOpt::HoistThenElseCodeToIf()` is enabled
by default, and is always run, unlike it's friend, common code sinking
transform, `SinkCommonCodeFromPredecessors()`, which is not enabled
by default and is only run once very late in the pipeline.
I'm proposing to harmonize this, and disable common code hoisting
until //late// in pipeline. Definition of //late// may vary,
here currently i've picked the same one as for code sinking,
but i suppose we could enable it as soon as right after
loop rotation happens.
Experimentation shows that this does indeed unsurprizingly help,
more loops got rotated, although other issues remain elsewhere.
Now, this undoubtedly seriously shakes phase ordering.
This will undoubtedly be a mixed bag in terms of both compile- and
run- time performance, codesize. Since we no longer aggressively
hoist+deduplicate common code, we don't pay the price of said hoisting
(which wasn't big). That may allow more loops to be rotated,
so we pay that price. That, in turn, that may enable all the transforms
that require canonical (rotated) loop form, including but not limited to
vectorization, so we pay that too. And in general, no deduplication means
more [duplicate] instructions going through the optimizations. But there's still
late hoisting, some of them will be caught late.
As per benchmarks i've run {
F12360204}, this is mostly within the noise,
there are some small improvements, some small regressions.
One big regression i saw i fixed in rG8d487668d09fb0e4e54f36207f07c1480ffabbfd, but i'm sure
this will expose many more pre-existing missed optimizations, as usual :S
llvm-compile-time-tracker.com thoughts on this:
http://llvm-compile-time-tracker.com/compare.php?from=
e40315d2b4ed1e38962a8f33ff151693ed4ada63&to=
c8289c0ecbf235da9fb0e3bc052e3c0d6bff5cf9&stat=instructions
* this does regress compile-time by +0.5% geomean (unsurprizingly)
* size impact varies; for ThinLTO it's actually an improvement
The largest fallout appears to be in GVN's load partial redundancy
elimination, it spends *much* more time in
`MemoryDependenceResults::getNonLocalPointerDependency()`.
Non-local `MemoryDependenceResults` is widely-known to be, uh, costly.
There does not appear to be a proper solution to this issue,
other than silencing the compile-time performance regression
by tuning cut-off thresholds in `MemoryDependenceResults`,
at the cost of potentially regressing run-time performance.
D84609 attempts to move in that direction, but the path is unclear
and is going to take some time.
If we look at stats before/after diffs, some excerpts:
* RawSpeed (the target) {
F12360200}
* -14 (-73.68%) loops not rotated due to the header size (yay)
* -272 (-0.67%) `"Number of live out of a loop variables"` - good for vectorizer
* -3937 (-64.19%) common instructions hoisted
* +561 (+0.06%) x86 asm instructions
* -2 basic blocks
* +2418 (+0.11%) IR instructions
* vanilla test-suite + RawSpeed + darktable {
F12360201}
* -36396 (-65.29%) common instructions hoisted
* +1676 (+0.02%) x86 asm instructions
* +662 (+0.06%) basic blocks
* +4395 (+0.04%) IR instructions
It is likely to be sub-optimal for when optimizing for code size,
so one might want to change tune pipeline by enabling sinking/hoisting
when optimizing for size.
Reviewed By: mkazantsev
Differential Revision: https://reviews.llvm.org/D84108
This reverts commit
503deec2183d466dad64b763bab4e15fd8804239.
Nikita Popov [Mon, 7 Sep 2020 19:07:02 +0000 (21:07 +0200)]
[KnownBits] Avoid some copies (NFC)
These lambdas don't need copies, use const reference.
Nikita Popov [Sat, 5 Sep 2020 08:27:23 +0000 (10:27 +0200)]
[SCCP] Compute ranges for supported intrinsics
For intrinsics supported by ConstantRange, compute the result range
based on the argument ranges. We do this independently of whether
some or all of the input ranges are full, as we can often still
constrain the result in some way.
Differential Revision: https://reviews.llvm.org/D87183
Craig Topper [Mon, 7 Sep 2020 19:23:15 +0000 (12:23 -0700)]
[SelectionDAG][X86][ARM] Teach ExpandIntRes_ABS to use sra+add+xor expansion when ADDCARRY is supported.
Rather than using SELECT instructions, use SRA, UADDO/ADDCARRY and
XORs to expand ABS. This is the multi-part version of the sequence
we use in LegalizeDAG.
It's also the same as the Custom sequence uses for i64 on 32-bit
and i128 on 64-bit. So we can remove the X86 customization.
Reviewed By: RKSimon
Differential Revision: https://reviews.llvm.org/D87215
Sanjay Patel [Mon, 7 Sep 2020 19:47:57 +0000 (15:47 -0400)]
[InstCombine] improve fold of pointer differences
This was supposed to be an NFC cleanup, but there's
a real logic difference (did not drop 'nsw') visible
in some tests in addition to an efficiency improvement.
This is because in the case where we have 2 GEPs,
the code was *always* swapping the operands and
negating the result. But if we have 2 GEPs, we
should *never* need swapping/negation AFAICT.
This is part of improving flags propagation noticed
with PR47430.
Sanjay Patel [Mon, 7 Sep 2020 19:26:43 +0000 (15:26 -0400)]
[InstCombine] add ptr difference tests; NFC
Craig Topper [Mon, 7 Sep 2020 17:59:57 +0000 (10:59 -0700)]
[X86] Use the same sequence for i128 ISD::ABS on 64-bit targets as we use for i64 on 32-bit targets.
Differential Revision: https://reviews.llvm.org/D87214
Craig Topper [Mon, 7 Sep 2020 17:41:05 +0000 (10:41 -0700)]
[X86] Pre-commit new test case for D87214. NFC
Sanjay Patel [Mon, 7 Sep 2020 18:11:06 +0000 (14:11 -0400)]
[DAGCombiner] allow more store merging for non-i8 truncated ops
This is a follow-up suggested in D86420 - if we have a pair of stores
in inverted order for the target endian, we can rotate the source
bits into place.
The "be_i64_to_i16_order" test shows a limitation of the current
function (which might be avoided if we integrate this function with
the other cases in mergeConsecutiveStores). In the earlier
"be_i64_to_i16" test, we skip the first 2 stores because we do not
match the full set as consecutive or rotate-able, but then we reach
the last 2 stores and see that they are an inverted pair of 16-bit
stores. The "be_i64_to_i16_order" test alters the program order of
the stores, so we miss matching the sub-pattern.
Differential Revision: https://reviews.llvm.org/D87112
Eric Astor [Mon, 7 Sep 2020 18:00:05 +0000 (14:00 -0400)]
[ms] [llvm-ml] Allow use of locally-defined variables in expressions
MASM allows variables defined by equate statements to be used in expressions.
Reviewed By: thakis
Differential Revision: https://reviews.llvm.org/D86946
Eric Astor [Mon, 7 Sep 2020 17:58:55 +0000 (13:58 -0400)]
[ms] [llvm-ml] Fix STRUCT field alignment
MASM aligns fields to the _minimum_ of the STRUCT alignment value and the size of the next field.
Reviewed By: thakis
Differential Revision: https://reviews.llvm.org/D86945
Eric Astor [Mon, 7 Sep 2020 17:57:06 +0000 (13:57 -0400)]
[ms] [llvm-ml] Add support for bitwise named operators (AND, NOT, OR) in MASM
Add support for expressions of the form '1 or 2', etc.
Reviewed By: thakis
Differential Revision: https://reviews.llvm.org/D86944
Simon Pilgrim [Mon, 7 Sep 2020 17:35:06 +0000 (18:35 +0100)]
VPlan.h - remove unnecessary forward declarations. NFCI.
Already defined in includes.
Simon Pilgrim [Mon, 7 Sep 2020 17:15:26 +0000 (18:15 +0100)]
MipsISelLowering.h - remove CCState/CCValAssign forward declarations. NFCI.
These are already defined in the CallingConvLower.h include.
Simon Pilgrim [Mon, 7 Sep 2020 16:50:58 +0000 (17:50 +0100)]
BTFDebug.h - reduce MachineInstr.h include to forward declaration. NFCI.
Simon Pilgrim [Mon, 7 Sep 2020 16:09:42 +0000 (17:09 +0100)]
LeonPasses.h - remove unnecessary includes. NFCI.
Reduce to forward declarations and move includes to LeonPasses.cpp where necessary.
Simon Pilgrim [Mon, 7 Sep 2020 15:56:57 +0000 (16:56 +0100)]
LeonPasses.h - remove orphan function declarations. NFCI.
The implementations no longer exist.
Sanjay Patel [Mon, 7 Sep 2020 16:37:59 +0000 (12:37 -0400)]
[InstCombine] improve folds for icmp with multiply operands (PR47432)
Check for no overflow along with an odd constant before
we lose information by converting to bitwise logic.
https://rise4fun.com/Alive/2Xl
Pre: C1 != 0
%mx = mul nsw i8 %x, C1
%my = mul nsw i8 %y, C1
%r = icmp eq i8 %mx, %my
=>
%r = icmp eq i8 %x, %y
Name: nuw ne
Pre: C1 != 0
%mx = mul nuw i8 %x, C1
%my = mul nuw i8 %y, C1
%r = icmp ne i8 %mx, %my
=>
%r = icmp ne i8 %x, %y
Name: odd ne
Pre: C1 % 2 != 0
%mx = mul i8 %x, C1
%my = mul i8 %y, C1
%r = icmp ne i8 %mx, %my
=>
%r = icmp ne i8 %x, %y
Sanjay Patel [Mon, 7 Sep 2020 15:40:59 +0000 (11:40 -0400)]
[InstCombine] move/add tests for icmp with mul operands; NFC
alex-t [Mon, 7 Sep 2020 15:57:27 +0000 (18:57 +0300)]
[AMDGPU] SILowerControlFlow::optimizeEndCF should remove empty basic block
optimizeEndCF removes EXEC restoring instruction case this instruction is the only one except the branch to the single successor and that successor contains EXEC mask restoring instruction that was lowered from END_CF belonging to IF_ELSE.
As a result of such optimization we get the basic block with the only one instruction that is a branch to the single successor.
In case the control flow can reach such an empty block from S_CBRANCH_EXEZ/EXECNZ it might happen that spill/reload instructions that were inserted later by register allocator are placed under exec == 0 condition and never execute.
Removing empty block solves the problem.
This change require further work to re-implement LIS updates. Recently, LIS is always nullptr in this pass. To enable it we need another patch to fix many places across the codegen.
Reviewed By: rampitec
Differential Revision: https://reviews.llvm.org/D86634
Momchil Velikov [Mon, 7 Sep 2020 15:16:52 +0000 (16:16 +0100)]
Reduce the number of memory allocations when displaying
a warning about clobbering reserved registers (NFC).
Also address some minor inefficiencies and style issues.
Differential Revision: https://reviews.llvm.org/D86088
Gabor Marton [Mon, 7 Sep 2020 15:15:15 +0000 (17:15 +0200)]
[analyzer][StdLibraryFunctionsChecker] Have proper weak dependencies
We want the generice StdLibraryFunctionsChecker to report only if there
are no specific checkers that would handle the argument constraint for a
function.
Note, the assumptions are still evaluated, even if the arguement
constraint checker is set to not report. This means that the assumptions
made in the generic StdLibraryFunctionsChecker should be an
over-approximation of the assumptions made in the specific checkers. But
most importantly, the assumptions should not contradict.
Differential Revision: https://reviews.llvm.org/D87240
Richard Barton [Mon, 7 Sep 2020 15:33:55 +0000 (16:33 +0100)]
[flang] Spelling and format edits to README.txt. NFC.
Gabor Marton [Thu, 23 Jul 2020 14:57:16 +0000 (16:57 +0200)]
[analyzer][StdLibraryFunctionsChecker] Add POSIX pthread handling functions
Differential Revision: https://reviews.llvm.org/D84415
Richard Barton [Mon, 7 Sep 2020 15:31:12 +0000 (16:31 +0100)]
[flang] Fix link to old repo location in doxygen mainpage. NFC.
Simon Pilgrim [Mon, 7 Sep 2020 15:39:42 +0000 (16:39 +0100)]
AntiDepBreaker.h - remove unnecessary ScheduleDAG.h include. NFCI.
Simon Pilgrim [Mon, 7 Sep 2020 15:17:31 +0000 (16:17 +0100)]
[Sparc] Add reduced funnel shift test case for PR47303
Simon Pilgrim [Mon, 7 Sep 2020 15:11:40 +0000 (16:11 +0100)]
[X86][SSE] Don't use LowerVSETCCWithSUBUS for unsigned compare with +ve operands (PR47448)
We already simplify the unsigned comparisons if we've found the operands are non-negative, but we were still calling LowerVSETCCWithSUBUS which resulted in the PR47448 regressions.
Simon Pilgrim [Mon, 7 Sep 2020 14:53:15 +0000 (15:53 +0100)]
[X86][SSE] Add test cases for PR47448