Simon Pilgrim [Mon, 2 Mar 2020 13:26:39 +0000 (13:26 +0000)]
EHScopeStack::Cleanup has virtual functions so the destructor should be too.
Fixes cppcheck warning.
David Green [Mon, 2 Mar 2020 14:26:32 +0000 (14:26 +0000)]
[LoopVectorizer] Change types of lists from pointers to references. NFC
getReductionVars, getInductionVars and getFirstOrderRecurrences were all
being returned from LoopVectorizationLegality as pointers to lists. This
just changes them to be references, cleaning up the interface slightly.
Differential Revision: https://reviews.llvm.org/D75448
Martijn Vels [Wed, 26 Feb 2020 20:55:49 +0000 (15:55 -0500)]
Partially inline basic_string::operator=(const basic_string&)
Summary:
This change partially inlines operator=(const basic_string&) where both the input and current instance are short strings, making the assignment a fixed length inlined memcpy.
Assignments where either of the strings are long are delegate to __assign_no_alias<__is_short>(), which is templated for the long / short branch already observed in the caller.
Stable:
```
--------------------------------------------------------------------------------
Benchmark Time CPU Iterations
--------------------------------------------------------------------------------
BM_StringAssignStr_Empty_Opaque 2.65 ns 2.66 ns
263745536
BM_StringAssignStr_Empty_Transparent 2.95 ns 2.96 ns
236494848
BM_StringAssignStr_Small_Opaque 2.93 ns 2.94 ns
237301760
BM_StringAssignStr_Small_Transparent 2.69 ns 2.69 ns
265809920
BM_StringAssignStr_Large_Opaque 19.6 ns 19.6 ns
35573760
BM_StringAssignStr_Large_Transparent 19.1 ns 19.1 ns
36716544
BM_StringAssignStr_Huge_Opaque 1901 ns 1901 ns 364544
BM_StringAssignStr_Huge_Transparent 1889 ns 1889 ns 360448
```
Unstable
```
--------------------------------------------------------------------------------
Benchmark Time CPU Iterations
--------------------------------------------------------------------------------
BM_StringAssignStr_Empty_Opaque 1.29 ns 1.29 ns
540454912
BM_StringAssignStr_Empty_Transparent 1.11 ns 1.12 ns
628482048
BM_StringAssignStr_Small_Opaque 1.29 ns 1.29 ns
541216768
BM_StringAssignStr_Small_Transparent 1.11 ns 1.11 ns
629469184
BM_StringAssignStr_Large_Opaque 15.6 ns 15.6 ns
44945408
BM_StringAssignStr_Large_Transparent 14.9 ns 14.9 ns
46764032
BM_StringAssignStr_Huge_Opaque 1713 ns 1713 ns 401408
BM_StringAssignStr_Huge_Transparent 1704 ns 1704 ns 397312
```
Subscribers: libcxx-commits
Tags: #libc
Differential Revision: https://reviews.llvm.org/D75211
Graham Hunter [Fri, 28 Feb 2020 13:53:40 +0000 (13:53 +0000)]
[OpenMP] Allow const parameters in declare simd linear clause
Reviewers: ABataev, kkwli0, jdoerfert, fpetrogalli
Reviewed By: ABataev, fpetrogalli
Differential Revision: https://reviews.llvm.org/D75350
Stephan Herhut [Mon, 2 Mar 2020 14:47:48 +0000 (15:47 +0100)]
[MLIR][GPU] fix loop trip count computation in LoopsToGPU
Summary: Added brackets to fix the loop trip count computation.
The brackets ensure the bounds are subtracted before we divide
the result by the step of the loop.
Differential Revision: https://reviews.llvm.org/D75449
Sanjay Patel [Mon, 2 Mar 2020 14:33:11 +0000 (09:33 -0500)]
[CodeGen] avoid running the entire optimizer pipeline in clang test file; NFC
I'm making the CHECK lines vague enough that they pass at -O0.
If that is too vague (we really want to check the data flow
to verify that the variables are not mismatched, etc), then
we can adjust those lines again to more closely match the output
at -O0 rather than -O1.
This change is based on the post-commit comments for:
https://github.com/llvm/llvm-project/commit/
83f4372f3a708ceaa800feff8b1bd92ae2c3be5f
http://lists.llvm.org/pipermail/cfe-commits/Week-of-Mon-
20200224/307888.html
Nicolas Vasilache [Sat, 29 Feb 2020 18:40:59 +0000 (13:40 -0500)]
[mlir] NFC - Move Vector structure from RunnerUtils.h to CRunnerUtils.h
Summary: The Vector struct does not require a C++ runtime.
Differential Revision: https://reviews.llvm.org/D75409
Haojian Wu [Mon, 2 Mar 2020 08:58:14 +0000 (09:58 +0100)]
[clangd] No need to query ctor refs in cross-file rename.
Summary:
This patch reverts https://github.com/llvm/llvm-project/commit/
2c5ee78de113484978450b834498e1b0e2aab5c4,
now kythe (https://github.com/kythe/kythe/issues/4381) supports returning ctors refs as part of class references, so
there is no need to query the ctor refs in the index (this would also
make the results worse, lots of duplications)
Reviewers: sammccall
Subscribers: ilya-biryukov, MaskRay, jkorous, arphaman, kadircet, usaxena95, cfe-commits
Tags: #clang
Differential Revision: https://reviews.llvm.org/D75439
Sanjay Patel [Mon, 2 Mar 2020 14:09:39 +0000 (09:09 -0500)]
[CodeGen] avoid running the entire optimizer pipeline in clang test file; NFC
There are no failures from the first set of RUN lines here,
so the CHECKs were already vague enough to not be affected
by optimizations. The final RUN line does induce some kind
of failure, so I'll try to fix that separately in a
follow-up.
Jonathan Coe [Mon, 2 Mar 2020 13:55:54 +0000 (13:55 +0000)]
[clang-format] Handle NullCoalescing and NullConditional operators in C#
Summary:
Disable merging of Type? into a single token.
Merge ?? ?. and ?[ into a single token.
Reviewers: krasimir, MyDeveloperDay
Reviewed By: krasimir
Subscribers: cfe-commits
Tags: #clang-format, #clang
Differential Revision: https://reviews.llvm.org/D75368
Stephan Herhut [Mon, 2 Mar 2020 10:34:25 +0000 (11:34 +0100)]
[MLIR] Add includes to PointerLikeTypeTraits where needed.
Summary:
This is to ensure that the template declaration is seen before
any template specialization.
Reviewers: mravishankar, antiagainst, rriddle!
Differential Revision: https://reviews.llvm.org/D75442
Kadir Cetinkaya [Fri, 28 Feb 2020 08:25:40 +0000 (09:25 +0100)]
[clangd] Get rid of lexer usage in locateMacroAt
Reviewers: sammccall, hokein
Subscribers: ilya-biryukov, MaskRay, jkorous, arphaman, usaxena95, cfe-commits
Tags: #clang
Differential Revision: https://reviews.llvm.org/D75331
Kadir Cetinkaya [Thu, 27 Feb 2020 15:02:44 +0000 (16:02 +0100)]
[clangd] Get rid of unnecessary source transformations in locateMacroAt
Summary:
All callers are already passing spelling locations to locateMacroAt.
Also there's no point at looking at macro expansion for figuring out undefs as
it is forbidden to have PP directives inside macro bodies.
Also fixes a bug when the previous sourcelocation is unavailable.
Reviewers: sammccall, hokein
Subscribers: ilya-biryukov, MaskRay, jkorous, arphaman, usaxena95, cfe-commits
Tags: #clang
Differential Revision: https://reviews.llvm.org/D75259
Luke Geeson [Fri, 14 Feb 2020 13:33:32 +0000 (13:33 +0000)]
[ARM] Add Cortex-M55 Support for clang and llvm
This patch upstreams support for the ARM Armv8.1m cpu Cortex-M55.
In detail adding support for:
- mcpu option in clang
- Arm Target Features in clang
- llvm Arm TargetParser definitions
details of the CPU can be found here:
https://developer.arm.com/ip-products/processors/cortex-m/cortex-m55
Reviewers: chill
Reviewed By: chill
Subscribers: dmgreen, kristof.beyls, hiraditya, cfe-commits,
llvm-commits
Tags: #clang, #llvm
Differential Revision: https://reviews.llvm.org/D74966
Simon Pilgrim [Mon, 2 Mar 2020 11:30:42 +0000 (11:30 +0000)]
Fix shadow variable warning. NFC.
Simon Pilgrim [Mon, 2 Mar 2020 11:19:48 +0000 (11:19 +0000)]
Fix shadow variable warning. NFC.
Simon Pilgrim [Mon, 2 Mar 2020 11:18:41 +0000 (11:18 +0000)]
[CostModel][X86] Add vXi1 extract/insert cost tests
Joe Turner [Thu, 27 Feb 2020 19:18:38 +0000 (20:18 +0100)]
[clang-tidy] Copy the Ranges field from the Diagnostic when creating the ClangTidyError
Differential Revision: https://reviews.llvm.org/D68887
Balázs Kéri [Mon, 2 Mar 2020 10:48:19 +0000 (11:48 +0100)]
[analyzer][StreamChecker] Using function description objects - NFC.
Summary:
Have a description object for the stream functions
that can store different aspects of a single stream operation.
I plan to extend the structure with other members,
for example pre-callback and index of the stream argument.
Reviewers: Szelethus, baloghadamsoftware, NoQ, martong, Charusso, xazax.hun
Reviewed By: Szelethus
Subscribers: rnkovacs, xazax.hun, baloghadamsoftware, szepet, a.sidorin, mikhail.ramalho, Szelethus, donat.nagy, dkrupp, gamesh411, Charusso, martong, cfe-commits
Tags: #clang
Differential Revision: https://reviews.llvm.org/D75158
Awanish Pandey [Mon, 2 Mar 2020 05:22:12 +0000 (10:52 +0530)]
Reland "[DebugInfo][clang][DWARF5]: Added support for debuginfo generation for defaulted parameters
in C++ templates."
This was reverted in
802b22b5c8c30bebc1695a217478be02653c6b53 due to
missing .bc file and a chromium bot failure.
https://bugs.chromium.org/p/chromium/issues/detail?id=1057559#c1
This revision address both of them.
Summary:
This patch adds support for debuginfo generation for defaulted
parameters in clang and also extends corresponding DebugMetadata/IR to support this feature.
Reviewers: probinson, aprantl, dblaikie
Reviewed By: aprantl, dblaikie
Differential Revision: https://reviews.llvm.org/D73462
Simon Pilgrim [Mon, 2 Mar 2020 10:56:29 +0000 (10:56 +0000)]
Fix operator precedence warning. NFCI.
Alex Zinenko [Mon, 2 Mar 2020 10:40:50 +0000 (11:40 +0100)]
[mlir] mlir-opt: print a newline after the top-level module
A printer refactoring removed automatic newline printing in the printer
of a ModuleOp. As a consequence, mlir-opt no longer printed a newline
after the closing brace of a module, which made it hard to distinguish
when used from command line. Print the newline character explicitly in
mlir-opt.
Andrzej Warzynski [Wed, 19 Feb 2020 12:25:30 +0000 (12:25 +0000)]
[AArch64][SVE] Add intrinsics for non-temporal gather-loads/scatter-stores
Summary:
This patch adds the following LLVM IR intrinsics for SVE:
1. non-temporal gather loads
* @llvm.aarch64.sve.ldnt1.gather
* @llvm.aarch64.sve.ldnt1.gather.uxtw
* @llvm.aarch64.sve.ldnt1.gather.scalar.offset
2. non-temporal scatter stores
* @llvm.aarch64.sve.stnt1.scatter
* @llvm.aarch64.sve.ldnt1.gather.uxtw
* @llvm.aarch64.sve.ldnt1.gather.scalar.offset
These intrinsic are mapped to the corresponding SVE instructions
(example for half-words, zero-extending):
* ldnt1h { z0.s }, p0/z, [z0.s, x0]
* stnt1h { z0.s }, p0/z, [z0.s, x0]
Note that for non-temporal gathers/scatters, the SVE spec defines only
one instruction type: "vector + scalar". For this reason, we swap the
arguments when processing intrinsics that implement the "scalar +
vector" addressing mode:
* @llvm.aarch64.sve.ldnt1.gather
* @llvm.aarch64.sve.ldnt1.gather.uxtw
* @llvm.aarch64.sve.stnt1.scatter
* @llvm.aarch64.sve.ldnt1.gather.uxtw
In other words, all intrinsics for gather-loads and scatter-stores
implemented in this patch are mapped to the same load and store
instruction, respectively.
The sve2_mem_gldnt_vs multiclass (and it's counterpart for scatter
stores) from SVEInstrFormats.td was split into:
* sve2_mem_gldnt_vec_vs_32_ptrs (32bit wide base addresses)
* sve2_mem_gldnt_vec_vs_62_ptrs (64bit wide base addresses)
This is consistent with what we did for
@llvm.aarch64.sve.ld1.scalar.offset and highlights the actual split in
the spec and the implementation.
Reviewed by: sdesmalen
Differential Revision: https://reviews.llvm.org/D74858
Simon Tatham [Mon, 2 Mar 2020 09:06:09 +0000 (09:06 +0000)]
[ARM,MVE] Add ACLE intrinsics for VCVT[ANPM] family.
Summary:
These instructions convert a vector of floats to a vector of integers
of the same size, with assorted non-default rounding modes.
Implemented in IR as target-specific intrinsics, because as far as I
can see there are no matches for that functionality in the standard IR
intrinsics list.
Reviewers: MarkMurrayARM, dmgreen, miyuki, ostannard
Reviewed By: dmgreen
Subscribers: kristof.beyls, hiraditya, cfe-commits, llvm-commits
Tags: #clang, #llvm
Differential Revision: https://reviews.llvm.org/D75255
Simon Tatham [Mon, 2 Mar 2020 09:06:00 +0000 (09:06 +0000)]
[ARM,MVE] Add ACLE intrinsics for VCVT.F32.F16 family.
Summary:
These instructions make a vector of `<4 x float>` by widening every
other lane of a vector of `<8 x half>`.
I wondered about representing these using standard IR, along the lines
of a shufflevector to extract elements of the input into a `<4 x half>`
followed by an `fpext` to turn that into `<4 x float>`. But it looks as
if that would take a lot of work in isel lowering to make it match any
pattern I could sensibly write in Tablegen, and also I haven't been
able to think of any other case where that pattern might be generated
in IR, so there wouldn't be any extra code generation win from doing
it that way.
Therefore, I've just used another target-specific intrinsic. We can
always change it to the other way later if anyone thinks of a good
reason.
(In order to put the intrinsic definition near similar things in
`IntrinsicsARM.td`, I've also lifted the definition of the
`MVEMXPredicated` multiclass higher up the file, without changing it.)
Reviewers: MarkMurrayARM, dmgreen, miyuki, ostannard
Reviewed By: miyuki
Subscribers: kristof.beyls, hiraditya, cfe-commits, llvm-commits
Tags: #clang, #llvm
Differential Revision: https://reviews.llvm.org/D75254
Simon Tatham [Mon, 2 Mar 2020 09:05:48 +0000 (09:05 +0000)]
[ARM,MVE] Correct MC operands in VCVT.F32.F16. (NFC)
Summary:
The two MVE instructions that convert between v4f32 and v8f16 were
implemented as instances of the same class, with the same MC operand
list.
But that's not really appropriate, because the narrowing conversion
only partially overwrites its output register (it only has 4 f16
values to write into a vector of 8), so even when unpredicated, it
needs a $Qd_src input, a constraint tying that to the $Qd output, and
a vpred_n.
The widening conversion is better represented like any other
instruction that completely replaces its output when unpredicated: it
should have no $Qd_src operand, and instead, a vpred_r containing a
$inactive parameter. That's a better match to other similar
instructions, such as its integer analogue, the VMOVL instruction that
makes a v4i32 by sign- or zero-extending every other lane of a v8i16.
This commit brings the widening VCVT.F32.F16 into line with the other
instructions that behave like it. That means you can write isel
patterns that use it unpredicated, without having to add a pointless
undefined $QdSrc operand.
No existing code generation uses that instruction yet, so there should
be no functional change from this fix.
Reviewers: MarkMurrayARM, dmgreen, miyuki, ostannard
Reviewed By: dmgreen
Subscribers: kristof.beyls, hiraditya, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D75253
Simon Tatham [Mon, 2 Mar 2020 09:05:35 +0000 (09:05 +0000)]
[ARM,MVE] Add ACLE intrinsics for VQMOV[U]N family.
Summary:
These instructions work like VMOVN (narrowing a vector of wide values
to half size, and overwriting every other lane of an output register
with the result), except that the narrowing conversion is saturating.
They come in three signedness flavours: signed to signed, unsigned to
unsigned, and signed to unsigned. All are represented in IR by a
target-specific intrinsic that takes two separate 'unsigned' flags.
Reviewers: MarkMurrayARM, dmgreen, miyuki, ostannard
Reviewed By: dmgreen
Subscribers: kristof.beyls, hiraditya, cfe-commits, llvm-commits
Tags: #clang, #llvm
Differential Revision: https://reviews.llvm.org/D75252
Pavel Labath [Mon, 2 Mar 2020 10:28:48 +0000 (11:28 +0100)]
[lld] Fix test failure from
d978656fd06
Tweak the test to account for the slightly different wording of the
error message.
Pavel Labath [Tue, 25 Feb 2020 14:40:49 +0000 (15:40 +0100)]
[DWARF] Use DWARFDataExtractor::getInitialLength to parse debug_names
Summary:
In this patch I've done a slightly bigger rewrite to also remove the
hardcoded header lengths.
Reviewers: jhenderson, dblaikie, ikudrin
Subscribers: hiraditya, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D75119
Pavel Labath [Tue, 25 Feb 2020 14:17:57 +0000 (15:17 +0100)]
[DWARF] Use getInitialLength in range list parsing
Summary:
This could be considered obvious, but I am putting it up to illustrate
the usefulness/impact of the getInitialLength change.
Reviewers: dblaikie, jhenderson, ikudrin
Subscribers: hiraditya, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D75117
Pavel Labath [Fri, 14 Feb 2020 14:41:38 +0000 (15:41 +0100)]
[DWARFDebugLine] Use new DWARFDataExtractor::getInitialLength
Summary:
The error messages change somewhat, but I believe the overall
informational value remains unchanged.
Reviewers: jhenderson, dblaikie, ikudrin
Subscribers: hiraditya, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D75116
serge-sans-paille [Mon, 2 Mar 2020 10:06:07 +0000 (11:06 +0100)]
Fix Base64Test - for StringRef size
Original failures: http://lab.llvm.org:8011/builders/clang-with-lto-ubuntu/builds/15975/steps/test-stage1-compiler/logs/stdio
Jim Lin [Mon, 2 Mar 2020 09:11:08 +0000 (17:11 +0800)]
[git-clang-format] Fix typo in help message
Anna Welker [Mon, 2 Mar 2020 09:14:37 +0000 (09:14 +0000)]
[ARM][MVE] Restrict allowed types of gather/scatter offsets
The MVE gather instructions smaller than 32bits zext extend the values
in the offset register, as opposed to sign extending them. We need to
make sure that the code that we select from is suitably extended, which
this patch attempts to fix by tightening up the offset checks.
Differential Revision: https://reviews.llvm.org/D75361
Kang Zhang [Mon, 2 Mar 2020 09:50:01 +0000 (09:50 +0000)]
[NFC][PowerPC] Move some alias definition from PPCInstrInfo.td to PPCInstr64Bit.td
Summary:
Some 64-bit instructions alias definition is in PPCInstrInfo.td, it should be
moved to PPCInstr64Bit.td.
Sagar Jain [Mon, 2 Mar 2020 09:19:57 +0000 (10:19 +0100)]
[MLIR] Added llvm.freeze
This patch adds llvm.freeze & processes undef constants from LLVM IR.
Syntax:
LLVM IR
`<result> = freeze ty <val>`
MLIR LLVM Dialect:
`llvm.freeze val attr-dict : type`
Example:
LLVM IR: `%3 = freeze i32 5`
MLIR: `%6 = llvm.freeze %5 : !llvm.i32`
Differential Revision: https://reviews.llvm.org/D75329
LLVM GN Syncbot [Mon, 2 Mar 2020 09:02:51 +0000 (09:02 +0000)]
[gn build] Port
5a1958f2673
serge-sans-paille [Mon, 24 Feb 2020 16:21:32 +0000 (17:21 +0100)]
Syndicate, test and fix base64 implementation
Move Base64 implementation from clangd/SemanticHighlighting to
llvm/Support/Base64, fix its implementation and provide a decent test suite.
Previous implementation code was using + operator instead of | to combine some
results, which is a problem when shifting signed values. (0xFF << 16) is
implicitly converted to a (signed) int, and thus results in 0xffff0000, which is
negative. Combining negative numbers with a + in that context is not what we
want to do.
This fixes https://github.com/llvm/llvm-project/issues/149.
Differential Revision: https://reviews.llvm.org/D75057
Haojian Wu [Mon, 2 Mar 2020 08:46:47 +0000 (09:46 +0100)]
[clangd] Remove the deprecated clangdServer::rename API, NFC.
There is no actual user of it now.
Alex Brachet [Mon, 2 Mar 2020 08:47:21 +0000 (03:47 -0500)]
[libc] Add sigprocmask
Summary: This patch adds `sigprocmask`, `sigemptyset` and `sigaddset`
Reviewers: sivachandra, MaskRay, gchatelet
Reviewed By: sivachandra
Subscribers: mgorny, tschuett, libc-commits
Differential Revision: https://reviews.llvm.org/D75026
Hans Wennborg [Mon, 2 Mar 2020 08:26:00 +0000 (09:26 +0100)]
Revert "[DebugInfo][clang][DWARF5]: Added support for debuginfo generation for defaulted parameters"
The Bitcode/DITemplateParameter-5.0.ll test is failing:
FAIL: LLVM :: Bitcode/DITemplateParameter-5.0.ll (5894 of 36324)
******************** TEST 'LLVM :: Bitcode/DITemplateParameter-5.0.ll' FAILED ********************
Script:
--
: 'RUN: at line 1'; /usr/local/google/home/thakis/src/llvm-project/out/gn/bin/llvm-dis -o - /usr/local/google/home/thakis/src/llvm-project/llvm/test/Bitcode/DITemplateParameter-5.0.ll.bc | /usr/local/google/home/thakis/src/llvm-project/out/gn/bin/FileCheck /usr/local/google/home/thakis/src/llvm-project/llvm/test/Bitcode/DITemplateParameter-5.0.ll
--
Exit Code: 2
Command Output (stderr):
--
It looks like the Bitcode/DITemplateParameter-5.0.ll.bc file was never checked in.
This reverts commit
c2b437d53d40b6dc5603c97f527398f477d9c5f1.
Awanish Pandey [Mon, 2 Mar 2020 05:22:12 +0000 (10:52 +0530)]
[DebugInfo][clang][DWARF5]: Added support for debuginfo generation for defaulted parameters
in C++ templates.
Summary:
This patch adds support for debuginfo generation for defaulted
parameters in clang and also extends corresponding DebugMetadata/IR to support this feature.
Reviewers: probinson, aprantl, dblaikie
Reviewed By: aprantl, dblaikie
Differential Revision: https://reviews.llvm.org/D73462
Fangrui Song [Mon, 2 Mar 2020 06:36:55 +0000 (22:36 -0800)]
[PowerPC][test] Improve .got2 and .toc tests
There is no .got2 test for powerpc32.
There is no comdat variable test for powerpc{32,64}.
Serguei Katkov [Fri, 28 Feb 2020 10:34:33 +0000 (17:34 +0700)]
[InlineSpiller] Relax re-materialization restriction for statepoint
We should be careful to allow count of re-materialization of operands to be less
then number of physical registers.
STATEPOINT instruction has a variable number of operands and potentially very big.
So re-materialization for all operands is disabled at the moment if restrict-statepoint-remat is true.
The patch relaxes the re-materialization restriction for STATEPOINT instruction allowing it for
fixed operands. Specifically it is about call target.
Reviewers: reames
Reviewed By: reames
Subscribers: llvm-commits, qcolombet, hiraditya
Differential Revision: https://reviews.llvm.org/D75335
Fangrui Song [Sun, 1 Mar 2020 02:40:58 +0000 (18:40 -0800)]
[ELF][PPC32] Don't report "relocation refers to a discarded section" for .got2
Similar to D63182 [ELF][PPC64] Don't report "relocation refers to a discarded section" for .toc
Reviewed By: Bdragon28
Differential Revision: https://reviews.llvm.org/D75419
Jim Lin [Mon, 2 Mar 2020 02:04:21 +0000 (10:04 +0800)]
[Sparc] Fix incorrect operand for matching CMPri pattern
Summary:
It should be normal constant instead of target constant.
Pattern CMPri can be matched if the constant can be fitted into immediate field.
Otherwise, pattern CMPrr will be matched.
This fixed bug https://bugs.llvm.org/show_bug.cgi?id=44091.
Reviewers: dcederman, jyknight
Reviewed By: jyknight
Subscribers: jonpa, hiraditya, fedor.sergeev, jrtc27, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D75227
Craig Topper [Mon, 2 Mar 2020 00:42:04 +0000 (16:42 -0800)]
[DAGCombiner][X86] Disable narrowExtractedVectorLoad if the element type size isn't byte sized
The address calculation for the offset assumes that you can calculate the offset by multiplying the index by the store size of the element. But that only works if the element's store size is exactly its real size since we store vectors tightly packed in memory. There are improvements we could make to this like special casing extracting element 0. I think we could also handle cases where the extracted VT is byte sized and the index is aligned with the extract element count.
Differential Revision: https://reviews.llvm.org/D75377
Shengchen Kan [Sun, 1 Mar 2020 07:43:53 +0000 (15:43 +0800)]
[X86] Not track size of the boudaryalign fragment during the layout
Summary:
Currently the boundaryalign fragment caches its size during the process
of layout and then it is relaxed and update the size in each iteration. This
behaviour is unnecessary and ugly.
Reviewers: annita.zhang, reames, MaskRay, craig.topper, LuoYuanke, jyknight
Reviewed By: MaskRay
Subscribers: hiraditya, dexonsmith, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D75404
Craig Topper [Fri, 28 Feb 2020 04:53:17 +0000 (20:53 -0800)]
[X86][TwoAddressInstructionPass] Teach tryInstructionCommute to continue checking for commutable FMA operands in more cases.
Previously we would only check for another commutable operand if the first commute was an aggressive commute.
But if we have two kill operands and neither is tied to the def at the start, we should consider both operands as the one to use as the new def.
This improves the loop in the fma-commute-loop.ll test. This test is derived from a post from discourse here https://llvm.discourse.group/t/unnecessary-vmovapd-instructions-generated-can-you-hint-in-favor-of-vfmadd231pd/582
Differential Revision: https://reviews.llvm.org/D75016
Craig Topper [Sun, 1 Mar 2020 18:41:20 +0000 (10:41 -0800)]
[DAGCombiner] Don't emit select_cc from visitSINT_TO_FP/visitUINT_TO_FP. Use plain select instead.
Select_cc isn't used by all targets. X86 doesn't have optimizations
for it.
Since we already know the input to the sint_to_fp/uint_to_fp is
a setcc we can just emit a plain select using that setcc as the
condition. Other DAG combines can turn that into a select_cc on
targets that support it.
Differential Revision: https://reviews.llvm.org/D75415
Lang Hames [Sun, 1 Mar 2020 17:34:31 +0000 (09:34 -0800)]
[JITLink] Update DEBUG_TYPE string for llvm-jitlink.
Apparently LLVM_DEBUG doesn't like dashes in strings.
Stefanos Baziotis [Sun, 1 Mar 2020 17:35:58 +0000 (19:35 +0200)]
Fix [ADT][NFC] SCCIterator: Change hasLoop() to hasCycle()
Stefanos Baziotis [Sun, 1 Mar 2020 17:17:21 +0000 (19:17 +0200)]
[ADT][NFC] SCCIterator: Change hasLoop() to hasCycle()
Reid Kleckner [Sun, 1 Mar 2020 16:45:22 +0000 (08:45 -0800)]
Attempt to fix ZLIB CMake logic on Windows
CMake doesn't seem to like it when you regex search for "^".
Reid Kleckner [Sun, 1 Mar 2020 15:47:55 +0000 (07:47 -0800)]
[WinEH] Fix inttoptr+phi optimization in presence of catchswitch
getFirstInsertionPt's return value must be checked for validity before
casting it to Instruction*. Don't attempt to insert casts after a phi in
a catchswitch block.
Fixes PR45033, introduced in D37832.
Reviewed By: davidxl, hfinkel
Differential Revision: https://reviews.llvm.org/D75381
Sanjay Patel [Sun, 1 Mar 2020 14:09:22 +0000 (09:09 -0500)]
[DAGCombiner] recognize shuffle (shuffle X, Mask0), Mask --> splat X
We get the simple cases of this via demanded elements and other folds,
but that doesn't work if the values have >1 use, so add a dedicated
match for the pattern.
We already have this transform in IR, but it doesn't help the
motivating x86 tests (based on PR42024) because the shuffles don't
exist until after legalization and other combines have happened.
The AArch64 test shows a minimal IR example of the problem.
Differential Revision: https://reviews.llvm.org/D75348
Jun Ma [Sun, 1 Mar 2020 12:55:06 +0000 (20:55 +0800)]
[Coroutines][New pass manager] Move CoroElide pass to right position
Differential Revision: https://reviews.llvm.org/D75345
Jun Ma [Sun, 1 Mar 2020 13:37:41 +0000 (21:37 +0800)]
Revert "[Coroutines][new pass manager] Move CoroElide pass to right position"
This reverts commit
4c0a133a412cd85381469e20f88ee7bf5d2ded8e.
Jun Ma [Sun, 1 Mar 2020 12:55:06 +0000 (20:55 +0800)]
[Coroutines][new pass manager] Move CoroElide pass to right position
Differential Revision: https://reviews.llvm.org/D75345
Craig Topper [Sun, 1 Mar 2020 08:01:38 +0000 (00:01 -0800)]
[X86] Don't add DELETED_NODES to DAG combine worklist after calling SimplifyDemandedBits/SimplifyDemandedVectorElts.
These AddToWorklist calls were added in
84cd968f75bbd6e0fbabecc29d2c1090263adec7.
It's possible the SimplifyDemandedBits/SimplifyDemandedVectorElts
triggered CSE that deleted N. Detect that and avoid adding N
to the worklist.
Fixes PR45067.
Fangrui Song [Sat, 29 Feb 2020 20:08:08 +0000 (12:08 -0800)]
[PowerPC] Move .got2/.toc logic from PPCLinuxAsmPrinter::doFinalization() to emitEndOfAsmFile()
Delete redundant .p2align 2 and improve tests.
Mehdi Amini [Sun, 1 Mar 2020 01:11:12 +0000 (01:11 +0000)]
Fix MLIR build by adding missing header after cleanup in
af450eab
Juneyoung Lee [Sat, 29 Feb 2020 22:22:03 +0000 (07:22 +0900)]
[ValueTracking] Let getGuaranteedNonFullPoisonOp consider assume, remove mentioning about br
Summary:
This patch helps getGuaranteedNonFullPoisonOp handle llvm.assume call.
Also, a comment about the semantics of branch is removed to prevent confusion.
As llvm.assume does, branching on poison directly raises UB (as LangRef says), and this allows transformations such as introduction of llvm.assume on branch condition at each successor, or freely replacing values after conditional branch (such as at loop exit).
Handling br is not addressed in this patch. It makes SCEV more accurate, causing existing LoopVectorize/IndVar/etc tests to fail.
Reviewers: spatel, lebedev.ri, nlopes
Reviewed By: nlopes
Subscribers: hiraditya, javed.absar, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D75397
Juneyoung Lee [Sat, 29 Feb 2020 04:10:31 +0000 (13:10 +0900)]
[ValueTracking] A value is never undef or poison if it must raise UB
Summary:
This patch helps isGuaranteedNotToBeUndefOrPoison return true if the value
makes the program always undefined.
According to value tracking functions' comments, it is not still in consensus
whether a poison value can be bitwise or not, so conservatively only the case with
i1 is considered.
Reviewers: spatel, lebedev.ri, reames, nlopes, regehr
Reviewed By: nlopes
Subscribers: uenoku, hiraditya, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D75396
Juneyoung Lee [Sat, 29 Feb 2020 05:58:07 +0000 (14:58 +0900)]
[GVN] Fold equivalent freeze instructions
Summary:
This patch defines two freeze instructions to have the same value number if they are equivalent.
This is allowed because GVN replaces all uses of a duplicated instruction with another.
If it partially rewrites use, it is not allowed. e.g)
```
a = freeze(x)
b = freeze(x)
use(a)
use(a)
use(b)
=>
use(a)
use(b) // This is not allowed!
use(b)
```
Reviewers: fhahn, reames, spatel, efriedma
Reviewed By: fhahn
Subscribers: lebedev.ri, hiraditya, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D75398
Craig Topper [Sat, 29 Feb 2020 06:45:05 +0000 (22:45 -0800)]
[X86] Tighten up the SDTypeProfile for X86ISD::CVTNE2PS2BF16. NFCI
Reid Kleckner [Sat, 29 Feb 2020 18:23:54 +0000 (10:23 -0800)]
Avoid including FileSystem.h from MemoryBuffer.h
Lots of headers pass around MemoryBuffer objects, but very few open
them. Let those that do include FileSystem.h.
Saves ~250 includes of Chrono.h & FileSystem.h:
$ diff -u thedeps-before.txt thedeps-after.txt | grep '^[-+] ' | sort | uniq -c | sort -nr
254 - ../llvm/include/llvm/Support/FileSystem.h
253 - ../llvm/include/llvm/Support/Chrono.h
237 - ../llvm/include/llvm/Support/NativeFormatting.h
237 - ../llvm/include/llvm/Support/FormatProviders.h
192 - ../llvm/include/llvm/ADT/StringSwitch.h
190 - ../llvm/include/llvm/Support/FormatVariadicDetails.h
...
This requires duplicating the file_t typedef, which is unfortunate. I
sunk the choice of mapping mode down into the cpp file using variable
template specializations instead of class members in headers.
Stephen Neuendorffer [Sat, 29 Feb 2020 19:51:04 +0000 (11:51 -0800)]
Revert "[MLIR] Move from using target_link_libraries to LINK_LIBS for llvm libraries."
This reverts commit
7a6c68977114b91097d693e9cfcb636631f61f91.
This breaks the build with cmake 3.13.4, but succeeds with cmake 3.15.3
Stephen Neuendorffer [Sat, 29 Feb 2020 19:50:45 +0000 (11:50 -0800)]
Revert "[MLIR] Remove redundant library dependencies"
This reverts commit
c4c8fbde649af8359cbfcbb5cef90fbe195c16fc.
Stephen Neuendorffer [Sat, 29 Feb 2020 19:50:31 +0000 (11:50 -0800)]
Revert "[MLIR] Move from add_dependencies() to DEPENDS"
This reverts commit
31e07d716a083b85df226c31f6ba72bc477f58fc.
Simon Pilgrim [Sat, 29 Feb 2020 19:24:21 +0000 (19:24 +0000)]
Fix Wdocumentation warning - use tparam for template parameters. NFC.
Simon Pilgrim [Sat, 29 Feb 2020 19:22:40 +0000 (19:22 +0000)]
[MachineInst] Remove dead code. NFCI.
The MachineFunction MF value is not used any more and is always null.
Simon Pilgrim [Sat, 29 Feb 2020 19:11:00 +0000 (19:11 +0000)]
Make argument const to silence cppcheck warning. NFCI.
Stephen Neuendorffer [Sat, 29 Feb 2020 19:08:19 +0000 (11:08 -0800)]
Revert "[MLIR] Add support for libMLIR.so"
This reverts commit
e17d9c11d49ab42ec19620679e981bb6147a2856.
It breaks the build.
Stephen Neuendorffer [Sat, 29 Feb 2020 19:06:50 +0000 (11:06 -0800)]
Revert "[MLIR] Fixes for BUILD_SHARED_LIBS=on"
This reverts commit
777e97cc1a8c1fe44245bdcb4fd9386ae2b3cc9d.
Petr Hosek [Thu, 6 Feb 2020 23:24:07 +0000 (15:24 -0800)]
[CMake] Link against ZLIB::ZLIB
This is the imported target that find_package(ZLIB) defines.
Differential Revision: https://reviews.llvm.org/D74176
Petr Hosek [Thu, 6 Feb 2020 03:07:43 +0000 (19:07 -0800)]
[CMake] Use PUBLIC link mode for static libraries
Using INTERFACE prevents the use of imported libraries as we've done
in 00b3d49 because these aren't linked against the target, they're
only made part of the interface. This doesn't affect the output since
static libraries aren't being linked into, but it enables the use of
imported libraries.
Differential Revision: https://reviews.llvm.org/D74106
Simon Pilgrim [Sat, 29 Feb 2020 18:56:53 +0000 (18:56 +0000)]
[X86][F16C] Remove cvtph2ps intrinsics and use generic half2float conversion (PR37554)
This removes everything but int_x86_avx512_mask_vcvtph2ps_512 which provides the SAE variant, but even this can use the fpext generic if the rounding control is the default.
Differential Revision: https://reviews.llvm.org/D75162
Stephen Neuendorffer [Thu, 27 Feb 2020 22:53:12 +0000 (14:53 -0800)]
[MLIR] Fixes for BUILD_SHARED_LIBS=on
Differential Revision: https://reviews.llvm.org/D75308
Valentin Churavy [Sun, 9 Feb 2020 03:27:54 +0000 (19:27 -0800)]
[MLIR] Add support for libMLIR.so
Putting this up mainly for discussion on
how this should be done. I am interested in MLIR from
the Julia side and we currently have a strong preference
to dynamically linking against the LLVM shared library,
and would like to have a MLIR shared library.
This patch adds a new cmake function add_mlir_library()
which accumulates a list of targets to be compiled into
libMLIR.so. Note that not all libraries make sense to
be compiled into libMLIR.so. In particular, we want
to avoid libraries which primarily exist to support
certain tools (such as mlir-opt and mlir-cpu-runner).
Note that the resulting libMLIR.so depends on LLVM, but
does not contain any LLVM components. As a result, it
is necessary to link with libLLVM.so to avoid linkage
errors. So, libMLIR.so requires LLVM_BUILD_LLVM_DYLIB=on
FYI, Currently it appears that LLVM_LINK_LLVM_DYLIB is broken
because mlir-tblgen is linked against libLLVM.so and
and independent LLVM components.
Previous version of this patch broke depencies on TableGen
targets. This appears to be because it compiled all
libraries to OBJECT libraries (probably because cmake
is generating different target names). Avoiding object
libraries results in correct dependencies.
(updated by Stephen Neuendorffer)
Differential Revision: https://reviews.llvm.org/D73130
Stephen Neuendorffer [Thu, 20 Feb 2020 18:48:51 +0000 (10:48 -0800)]
[MLIR] Move from add_dependencies() to DEPENDS
add_llvm_library and add_llvm_executable may need to create new targets with
appropriate dependencies. As a result, it is not sufficient in some
configurations (namely LLVM_BUILD_LLVM_DYLIB=on) to only call
add_dependencies(). Instead, the explicit TableGen dependencies must
be passed to add_llvm_library() or add_llvm_executable() using the DEPENDS
keyword.
Differential Revision: https://reviews.llvm.org/D74930
Stephen Neuendorffer [Thu, 20 Feb 2020 06:56:38 +0000 (22:56 -0800)]
[MLIR] Remove redundant library dependencies
In cmake, it is redundant to have a target list under target_link_libraries()
and add_dependency(). This patch removes the redundant dependency from
add_dependency().
Differential Revision: https://reviews.llvm.org/D74929
Stephen Neuendorffer [Tue, 18 Feb 2020 23:07:35 +0000 (15:07 -0800)]
[MLIR] Move from using target_link_libraries to LINK_LIBS for llvm libraries.
When compiling libLLVM.so, add_llvm_library() manipulates the link libraries
being used. This means that when using add_llvm_library(), we need to pass
the list of libraries to be linked (using the LINK_LIBS keyword) instead of
using the standard target_link_libraries call. This is preparation for
properly dealing with creating libMLIR.so as well.
Differential Revision: https://reviews.llvm.org/D74864
Mehdi Amini [Sat, 29 Feb 2020 06:04:38 +0000 (06:04 +0000)]
Ensure that multi-threading is disabled when enabling IRPrinting with module scope
This is avoid the user to shoot themselves in the foot and encounter
strange crashes that are confusing until one run with TSAN.
Differential Revision: https://reviews.llvm.org/D75399
Fangrui Song [Sat, 29 Feb 2020 16:25:22 +0000 (08:25 -0800)]
[MC] Add MCStreamer::emitInt{8,16,32,64}
Similar to AsmPrinter::emitInt{8,16,32,64}.
Aaron Puchert [Sat, 29 Feb 2020 13:22:41 +0000 (14:22 +0100)]
Remove unused parameter from CXXRecordDecl::forallBases [NFC]
Summary:
Apparently all users of the function were fine with short-circuiting
and none cared to override the default argument.
Reviewers: aaron.ballman, rsmith
Reviewed By: aaron.ballman
Subscribers: cfe-commits
Tags: #clang
Differential Revision: https://reviews.llvm.org/D75319
Sanjay Patel [Fri, 28 Feb 2020 22:49:22 +0000 (17:49 -0500)]
[PassManager] add tests for vector pass enabling; NFC
Stefan Gränitz [Sat, 29 Feb 2020 11:52:19 +0000 (11:52 +0000)]
[ExecutionEngine] Add JITSymbolFlags::fromSummary(GlobalValueSummary*)
Summary: A function that creates JITSymbolFlags from a GlobalValueSummary. Similar functions exist: fromGlobalValue(), fromObjectSymbol()
Reviewers: lhames
Reviewed By: lhames
Subscribers: hiraditya, steven_wu, dexonsmith, arphaman, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D75082
Stefanos Baziotis [Sat, 29 Feb 2020 09:53:11 +0000 (11:53 +0200)]
First test commit - empty
Georgii Rymar [Fri, 14 Feb 2020 09:47:52 +0000 (12:47 +0300)]
[llvm-readobj] - Report warnings instead of errors for broken relocations.
This is a follow-up for https://reviews.llvm.org/D74545.
It adds test cases for each incorrect case returned in `getRelocationTarget`.
Differential revision: https://reviews.llvm.org/D74595
Benjamin Kramer [Sat, 29 Feb 2020 08:50:23 +0000 (09:50 +0100)]
ArrayRef'ize restoreCalleeSavedRegisters. NFCI.
restoreCalleeSavedRegisters can mutate the contents of the
CalleeSavedInfos, so use a MutableArrayRef.
Siva Chandra Reddy [Thu, 27 Feb 2020 22:30:24 +0000 (14:30 -0800)]
[libc] Add ability to generate enum types/values to HdrGen.
A target to generate the std C threads.h file has been added. This
utilizes the new feature added in this change.
Reviewers: phosek
Differential Revision: https://reviews.llvm.org/D75379
Shengchen Kan [Fri, 28 Feb 2020 14:27:53 +0000 (22:27 +0800)]
[X86] Move the function getOrCreateBoundaryAlignFragment
MCObjectStreamer is more suitable to create fragments than
X86AsmBackend, for example, the function getOrCreateDataFragment is
defined in MCObjectStreamer.
Differential Revision: https://reviews.llvm.org/D75351
Shengchen Kan [Fri, 28 Feb 2020 13:09:30 +0000 (21:09 +0800)]
[X86] Disable the NOP padding for branches when bundle is enabled
When bundle is enabled, data fragment itself has a space to emit NOP
to bundle-align instructions. The behaviour makes it impossible for
us to determine whether the macro fusion really happen when emitting
instructions. In addition, boundary-align fragment is also used to
emit NOPs to align instructions, currently using them together sometimes
makes code crazy.
Differential Revision: https://reviews.llvm.org/D75346
Fangrui Song [Sat, 29 Feb 2020 01:22:29 +0000 (17:22 -0800)]
[ELF][PPC32] Fix canonical PLTs when the order does not match the PLT order
Reviewed By: Bdragon28
Differential Revision: https://reviews.llvm.org/D75394
Fangrui Song [Sat, 29 Feb 2020 05:26:36 +0000 (21:26 -0800)]
[ELF] Delete two unneeded `referenced = true` after D65584
Greg Clayton [Sat, 29 Feb 2020 05:19:05 +0000 (21:19 -0800)]
Revert "Fix GSYM tests to run the yaml files and fix test failures on some machines."
This reverts commit
d334ce0b5acb945d6202d0ab6a17bdca530f50c1.
Michael Liao [Fri, 28 Feb 2020 05:01:08 +0000 (00:01 -0500)]
[cmake] Fix LLVM_USE_SPLIT_DWARF
Summary:
- Add `-gsplit-dwarf` as an option instead of a definition.
- Only add that option on compilers supporting dwarf splitting, such as clang
and gcc.
Reviewers: echristo, pcc
Subscribers: mgorny, aprantl, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D75328
Vedant Kumar [Sat, 29 Feb 2020 02:59:56 +0000 (18:59 -0800)]
Disable instrprof-merging.cpp on powerpc64 (D69471 follow up)
An execution count goes missing for a constructor, this needs
investigation:
http://lab.llvm.org:8011/builders/clang-ppc64be-linux/builds/45132/
```
/home/buildbots/ppc64be-clang-test/clang-ppc64be/llvm/compiler-rt/test/profile/instrprof-merging.cpp:28:16:
error: V1: expected string not found in input
A() {} // V1: [[@LINE]]{{ *}}|{{ *}}1
<stdin>:28:32: note: possible intended match here
28| | A() {} // V1: [[@LINE]]{{ *}}|{{ *}}1
```
Vedant Kumar [Sat, 29 Feb 2020 02:27:04 +0000 (18:27 -0800)]
Add cast to appease clang-armv7-linux-build-cache (D69471 followup)
http://lab.llvm.org:8011/builders/clang-armv7-linux-build-cache/builds/27075
error: non-constant-expression cannot be narrowed from type 'uint64_t'
(aka 'unsigned long long') to 'size_t' (aka 'unsigned int') in
initializer list [-Wc++11-narrowing]
return {MappingBuf, getDataSize<FuncRecordTy, Endian>(Record)};