Bill Wendling [Wed, 23 Feb 2022 09:20:48 +0000 (01:20 -0800)]
[NFC] Remove unnecessary "#include"s from header files
David Green [Wed, 23 Feb 2022 08:55:54 +0000 (08:55 +0000)]
[ARM] Recognize SSAT and USAT from SMIN/SMAX
We have some recognition of SSAT and USAT from SELECT_CC at the moment.
This extends the matching to SMIN/SMAX which can help catch more cases,
either from min/max being the canonical form in instcombine or from some
expanded nodes like fp_to_si_sat.
Differential Revision: https://reviews.llvm.org/D119819
Nikita Popov [Tue, 22 Feb 2022 16:53:14 +0000 (17:53 +0100)]
[InstCombine] Remove one-use limitation from X-Y==0 fold
This one-use limitation is artificial, we do not increase
instruction count if we perform the fold with multiple uses. The
motivating case is shown in @sub_eq_zero_select, where the one-use
limitation causes us to miss a subsequent select fold.
I believe the backend is pretty good about reusing flag-producing
subs for cmps with same operands, so I think doing this is fine.
Differential Revision: https://reviews.llvm.org/D120337
Nathan James [Wed, 23 Feb 2022 08:35:30 +0000 (08:35 +0000)]
[clang-tidy][NFC] Remove Tristate from CachedGlobList
The tristate is a little redundant as we can determine if the item was already in the cache based on the return from try_emplace.
Reviewed By: salman-javed-nz
Differential Revision: https://reviews.llvm.org/D120196
Nathan James [Wed, 23 Feb 2022 08:33:58 +0000 (08:33 +0000)]
[ASTMatchers] Expand isInline matcher to VarDecl
Add support to the `isInline` matcher for C++17's inline variables.
Reviewed By: aaron.ballman
Differential Revision: https://reviews.llvm.org/D118900
Chuanqi Xu [Wed, 23 Feb 2022 08:22:55 +0000 (16:22 +0800)]
[NFC] Add unittest for Decl::isInExportDeclContext
Stanislav Gatev [Fri, 18 Feb 2022 18:51:42 +0000 (18:51 +0000)]
[clang][dataflow] Add support for global storage values
This is part of the implementation of the dataflow analysis framework.
See "[RFC] A dataflow analysis framework for Clang AST" on cfe-dev.
Reviewed-by: ymandel, xazax.hun
Differential Revision: https://reviews.llvm.org/D120149
LLVM GN Syncbot [Wed, 23 Feb 2022 08:21:48 +0000 (08:21 +0000)]
[gn build] Port
a2fab82f33bb
Haojian Wu [Fri, 11 Feb 2022 13:09:15 +0000 (14:09 +0100)]
[pseudo] Implement LRTable.
This patch introduces a dense implementation of the LR parsing table, which is
used by LR parsers.
We build a SLR(1) parsing table from the LR(0) graph.
Statistics of the LR parsing table on the C++ spec grammar:
- number of states: 1449
- number of actions: 83069
- size of the table (bytes): 334928
Differential Revision: https://reviews.llvm.org/D118196
Lian Wang [Wed, 23 Feb 2022 08:01:51 +0000 (08:01 +0000)]
[RISCV] Add more test for GORCI aliaes in Zbp extension
Supplement tests for some aliaes of gorci.
RV32:
add orc4.h/orc2.h in rv32zbp.ll
add orc.h/orc16/orc8/orc4/orc2/orc in rv32zbp-intrinsic.ll
RV64:
add orc4.h/orc2.h in rv64zbp.ll
add orc.h/orc32/orc16/orc8/orc4/orc2/orc/orc16.w/orc8.w/
orc4.w/orc2.w/orc.w in rv64zbp-intrinsic.ll
Reviewed By: craig.topper
Differential Revision: https://reviews.llvm.org/D120388
Valentin Clement [Wed, 23 Feb 2022 07:27:10 +0000 (08:27 +0100)]
[flang] Lower real constant
This patch handles lowering of real constant.
This patch is part of the upstreaming effort from fir-dev branch.
Reviewed By: PeteSteinfeld
Differential Revision: https://reviews.llvm.org/D120354
Co-authored-by: Eric Schweitz <eschweitz@nvidia.com>
Co-authored-by: Jean Perier <jperier@nvidia.com>
Lian Wang [Wed, 23 Feb 2022 06:14:58 +0000 (06:14 +0000)]
[RISCV] Supplement more tests for GREVI aliaes in Zbp extension
Supplement tests for some aliaes of grevi.
RV32:
add rev4.h/rev2.h in rv32zbp.ll
add rev/rev2/rev4/rev8/rev16 in rv32zbp-intrinsic.ll
RV64:
add rev4.h/rev2.h in rv64zbp.ll
add rev.h/rev/rev2/rev4/rev8/rev16/rev32/rev.w/rev2.w/
rev4.w/rev8.w/rev16.w in rv64zbp-intrinsic.ll
Reviewed By: craig.topper
Differential Revision: https://reviews.llvm.org/D120304
minglotus-6 [Tue, 22 Feb 2022 03:17:06 +0000 (19:17 -0800)]
[SampleProfile] Handle the case when the option `MaxNumPromotions` is zero.
In places where `MaxNumPromotions` is used to allocated an array, bail out early to prevent allocating an array of length 0.
Differential Revision: https://reviews.llvm.org/D120295
Fangrui Song [Wed, 23 Feb 2022 05:42:15 +0000 (21:42 -0800)]
[ELF][test] Fix CU address_size in some gdb-index tests
Revert
251640ab575634256de6d6fde5c5359fe21efe91 which fixed the wrong thing.
While here, add `2>&1 | count 0` to assert no warning from lib/DebugInfo/DWARF.
Fangrui Song [Wed, 23 Feb 2022 04:29:08 +0000 (20:29 -0800)]
[ProfileData] Remove unused and racy FunctionSamples::Format after D51643
The write may be racy if ThinLTO creates multiple `InProcessThinBackend` instances.
Alex Lorenz [Mon, 14 Feb 2022 22:20:28 +0000 (14:20 -0800)]
[compiler-rt][builtins] build the macOS compiler-rt built-ins with Mac Catalyst support
This patch extends compiler-rt's cmake config to build macOS builtins with both macOS and Mac Catalyst support.
This is done by telling the compiler to emit macho files with two build version load commands.
Differential Revision: https://reviews.llvm.org/D118875
Fangrui Song [Wed, 23 Feb 2022 03:20:55 +0000 (19:20 -0800)]
[ELF][test] Terminate .debug_info with a null entry to fix warnings
Arthur Eubanks [Wed, 23 Feb 2022 02:28:32 +0000 (18:28 -0800)]
[clang] Remove Address::deprecated() calls in CodeGenFunction.cpp
Joseph Huber [Wed, 23 Feb 2022 02:21:33 +0000 (21:21 -0500)]
[OpenMP] Remove static allocator in linker wrapper
Summary:
We don't need this static allocator to survive the entire file, the
strings stored have a defined lifetime.
Joseph Huber [Wed, 23 Feb 2022 02:13:03 +0000 (21:13 -0500)]
[OpenMP] Unrecognized objects should not be considered failure
Summary:
This patch removes the error we recieve when attempting to extract
offloading sections. We shouldn't consider this a failure because
extracting bitcode isn't necessarily required.
David Blaikie [Wed, 23 Feb 2022 01:49:38 +0000 (17:49 -0800)]
Remove redundant word word in AsmParser DIFlag parsing
Philip Reames [Wed, 23 Feb 2022 01:25:06 +0000 (17:25 -0800)]
[instcombine] Avoid binops for comparison consistency tests
It turns out that instcombine is smarter than I am, and several of these ended up folded for the wrong reasons.
Amara Emerson [Wed, 23 Feb 2022 00:33:18 +0000 (16:33 -0800)]
Revert "Revert "[AArch64][GlobalISel] Optimize conjunctions of compares to conditional compares.""
This reverts commit
55c181a6c786cfbfa8b7aabe0a8ba721a65b1445.
The original commit I made was an old patch, mea culpa. Committing the right
implementation with test case for the reported crash.
Arthur Eubanks [Wed, 23 Feb 2022 00:54:09 +0000 (16:54 -0800)]
[clang] Remove Address::deprecated() calls in CGVTables.cpp
Alina Sbirlea [Wed, 23 Feb 2022 00:43:33 +0000 (16:43 -0800)]
[Docs] Add self to credits
Alina Sbirlea [Wed, 23 Feb 2022 00:38:57 +0000 (16:38 -0800)]
[Docs]Add office hours.
Philip Reames [Wed, 23 Feb 2022 00:36:08 +0000 (16:36 -0800)]
[NFC] Add a bit more coverage for an upcoming patch
Fangrui Song [Wed, 23 Feb 2022 00:30:02 +0000 (16:30 -0800)]
[sancov] Refactor getPreviousInstructionPc
Note: on some architectures lik AArch64, the PC does not match
compiler-rt/lib/sanitizer_common/sanitizer_coverage_libcdep_new.cpp:`__sanitizer_cov_trace_pc_guard`
Stanislav Mekhanoshin [Wed, 23 Feb 2022 00:20:09 +0000 (16:20 -0800)]
[AMDGPU] Pre-commit load/store combine tests. NFC.
Fangrui Song [Wed, 23 Feb 2022 00:25:57 +0000 (16:25 -0800)]
[libFuzzer] Refactor GetNextInstructionPc/GetPreviousInstructionPc
Port the change to compiler-rt/lib/fuzzer/FuzzerTracePC.cpp .
Update RISCV to use PC-2: this is coarse (C extension may be disabled) but
sufficient for pure symbolization purpose.
The commit is separate from D120362 so that bisecting/reverting is easier.
Aart Bik [Tue, 22 Feb 2022 20:21:07 +0000 (12:21 -0800)]
[mlir][sparse] refactor sparse compiler pipeline to single place
Reviewed By: bixia
Differential Revision: https://reviews.llvm.org/D120347
Philip Reames [Wed, 23 Feb 2022 00:18:39 +0000 (16:18 -0800)]
[instcombine] Add coverage for consistent use of unescaped malloc case
Fangrui Song [Wed, 23 Feb 2022 00:20:40 +0000 (16:20 -0800)]
[sanitizer] Refactor GetNextInstructionPc/GetPreviousInstructionPc
x86 uses offset 1 while most RISC architectures use offset 4.
Check x86 first to prevent changes for new RISC architectures.
Reviewed By: #sanitizers, vitalybuka
Differential Revision: https://reviews.llvm.org/D120362
Arthur Eubanks [Wed, 23 Feb 2022 00:19:06 +0000 (16:19 -0800)]
[clang] Remove an Address::deprecated() call in CGClass.cpp
Arthur Eubanks [Wed, 23 Feb 2022 00:11:49 +0000 (16:11 -0800)]
[NFC] Remove dead code (try 2)
This is causing
../../llvm/include/llvm/Object/MachO.h:379:13: warning: private field 'Kind' is not used [-Wunused-private-field]
FixupKind Kind;
Previous attempt in
a23f7c0cb6b42a06bc9707fdf46ce2a90080f61f.
Wouter van Oortmerssen [Tue, 22 Feb 2022 23:45:49 +0000 (15:45 -0800)]
[WebAssembly] Fixed AsmPrinter not emitting .functype for intrinsics
Intrinsics like `memset` were not emitted as `.functype` because
WebAssemblyAsmPrinter::emitExternalDecls explicitly skips symbols
that are isIntrinsic. Removing that check doesn't work, since the symbol
from the module refers to a 4-argument `llvm.memset.p0i8.i32` rather
than the 3-argument `memset` symbol referenced in the call.
Our `WebAssemblyMCLowerPrePass` however does collect the
`memset` symbol, so the current solution is as simple as emitting
`.functype` for those.
Fixes: https://github.com/llvm/llvm-project/issues/53712
Differential Revision: https://reviews.llvm.org/D120365
Philip Reames [Wed, 23 Feb 2022 00:02:27 +0000 (16:02 -0800)]
[instcombine] Autogen a test for ease of update
Philip Reames [Wed, 23 Feb 2022 00:01:39 +0000 (16:01 -0800)]
[instcombine] Extend test coverage for a tricky bit of reasoning about unescaped mallocs
Philip Reames [Tue, 22 Feb 2022 23:52:10 +0000 (15:52 -0800)]
[instcombine] Add test coverage for a tricky bit of reasoning about unescaped mallocs
Owen Anderson [Tue, 28 Dec 2021 19:44:49 +0000 (11:44 -0800)]
Set std::numeric_limits<>::tinyness_before to true for floating point types on ARM platforms.
Set std::numeric_limits<>::tinyness_before to true for floating point types on ARM platforms.
Section E1.3.5 in the ARMv8 Architecture Reference Manual specifies:
Underflow. The bit is set to 1 if the absolute value of the result
of an operation, produced before rounding, is less than the minimum
positive normalized number for the destination precision, and the
rounded result is inexact.
Reviewed By: #libc, majnemer, EricWF
Differential Revision: https://reviews.llvm.org/D116338
Rahman Lavaee [Tue, 22 Feb 2022 22:39:08 +0000 (14:39 -0800)]
Encode address offsets of basic blocks relative to the end of the previous basic blocks.
Conceptually, the new encoding emits the offsets and sizes as label differences between each two consecutive basic block begin and end label. When decoding, the offsets must be aggregated along with basic block sizes to calculate the final relative-to-function offsets of basic blocks.
This encoding uses smaller values compared to the existing one (offsets relative to function symbol).
Smaller values tend to occupy fewer bytes in ULEB128 encoding. As a result, we get about 25% reduction
in the size of the bb-address-map section (reduction from about 9MB to 7MB).
Reviewed By: tmsriram, jhenderson
Differential Revision: https://reviews.llvm.org/D106421
David Green [Tue, 22 Feb 2022 23:37:22 +0000 (23:37 +0000)]
[AArch64] Alter mull shuffle(ext(..)) combine to work on buildvectors
We have a combine for converting mul(dup(ext(..)), ...) into
mul(ext(dup(..)), ..), for allowing more uses of smull and umull
instructions. Currently it looks for vector insert and shuffle vectors
to detect the element that we can convert to a vector extend. Not all
cases will have a shufflevector/insert element though.
This started by extending the recognition to buildvectors (with elements
that may be individually extended). The new method seems to cover all
the cases that the old method captured though, as the shuffle will
eventually be lowered to buildvectors, so the old method has been
removed to keep the code a little simpler. The new code detects legal
build_vector(ext(a), ext(b), ..), converting them to ext(build_vector(a,
b, ..)) providing all the extends/types match up.
Differential Revision: https://reviews.llvm.org/D120018
Jez Ng [Tue, 22 Feb 2022 23:14:47 +0000 (18:14 -0500)]
[lld] Require C++14 in LLD standalone build
This is what the Clang standalone build does too. And setting this
seems to be required to get the standalone build to work on my Mac.
Reviewed By: #lld-macho, MaskRay, Ericson2314, smeenai
Differential Revision: https://reviews.llvm.org/D120269
Ben Barham [Tue, 22 Feb 2022 05:13:38 +0000 (21:13 -0800)]
[VFS] Use generic_category for errors generated from the VFS
Errors are generally checked in clients by comparing to the portable
error condition in `std::errc`, which will have the `generic_category`
(eg. `std::errc::no_such_file_or_directory`). While in practice these
are usually equivalent for the standard errno's, they are not in *all*
implementations. One such example is CentOS 7.
Differential Revision: https://reviews.llvm.org/D120299
Nico Weber [Tue, 22 Feb 2022 22:47:26 +0000 (17:47 -0500)]
[gn build] bump fmsc-version to 1926
This is needed to pick up the workaround in
fb1aa286c1
when building with a modern MSVC (like LLVM now requires).
Brendon Cahoon [Mon, 21 Feb 2022 00:18:26 +0000 (18:18 -0600)]
[SLP] Fix assert from non-constant index in insertelement
A call to getInsertIndex() in getTreeCost() is returning None,
which causes an assert because a non-constant index value for
insertelement was not expected. This case occurs when the
insertelement index value is defined with a PHI.
Differential Revision: https://reviews.llvm.org/D120223
Alexey Bataev [Tue, 22 Feb 2022 21:46:02 +0000 (13:46 -0800)]
[SLP][NFC]Add a test for bottom to top reordering.
Michael Kruse [Tue, 22 Feb 2022 21:37:50 +0000 (15:37 -0600)]
[polly] Remove trailing whitespace from tests. NFC.
Okwan Kwon [Fri, 18 Feb 2022 18:07:36 +0000 (18:07 +0000)]
Fold Tensor.extract_slice into a constant splat.
Fold arith.extract_slice into arith.constant when the source is a constant
splat and the result type is statically shaped.
Alexander Yermolovich [Tue, 22 Feb 2022 21:26:38 +0000 (13:26 -0800)]
[BOLT][DWARF] Remove patchLowHigh unused function.
Cleanup after removing caching mechanims for ranges/abbrevs.
Reviewed By: rafauler, yota9
Differential Revision: https://reviews.llvm.org/D120174
Nico Weber [Tue, 22 Feb 2022 21:25:51 +0000 (16:25 -0500)]
fix comment typo to cycle bots
Björn Schäpers [Mon, 21 Feb 2022 21:03:55 +0000 (22:03 +0100)]
[clang-format] Don't break semi after requires clause ...
..regardless of the chosen style.
Fixes https://github.com/llvm/llvm-project/issues/53818
Differential Revision: https://reviews.llvm.org/D120278
Björn Schäpers [Sun, 2 Jan 2022 20:53:02 +0000 (21:53 +0100)]
[clang-format][NFC] Simplify if in ContinuationIndenter::addTokenOCL
Setting a boolean within an if and only using it in the very next if is
a bit confusing. Merge it into one if.
Differential Revision: https://reviews.llvm.org/D120237
Reid Kleckner [Tue, 22 Feb 2022 20:56:45 +0000 (12:56 -0800)]
Revert "Remove dead code."
This reverts commit
a23f7c0cb6b42a06bc9707fdf46ce2a90080f61f.
Breaks the build.
Adrian Prantl [Tue, 22 Feb 2022 20:47:39 +0000 (12:47 -0800)]
Disable test on big endian machines. Yaml2obj has problems there.
Adrian Prantl [Tue, 22 Feb 2022 20:47:29 +0000 (12:47 -0800)]
Remove dead code.
Dmitry Vassiliev [Tue, 22 Feb 2022 20:11:20 +0000 (00:11 +0400)]
[Transforms] Enhance CorrelatedValuePropagation to handle both values of select
The "Correlated Value Propagation" pass was missing a case when handling select instructions. It was only handling the "false" constant value, while in NVPTX the select may have the condition (and thus the branches) inverted, for example:
```
loop:
%phi = phi i32* [ %sel, %loop ], [ %x, %entry ]
%f = tail call i32* @f(i32* %phi)
%cmp1 = icmp ne i32* %f, %y
%sel = select i1 %cmp1, i32* %f, i32* null
%cmp2 = icmp eq i32* %sel, null
br i1 %cmp2, label %return, label %loop
```
But the select condition can be inverted:
```
%cmp1 = icmp eq i32* %f, %y
%sel = select i1 %cmp1, i32* null, i32* %f
```
The fix is to enhance "Correlated Value Propagation" to handle both branches of the select instruction.
Reviewed By: nikic, lebedev.ri
Differential Revision: https://reviews.llvm.org/D119643
Dmitry Vassiliev [Tue, 22 Feb 2022 20:10:05 +0000 (00:10 +0400)]
[Transforms] Pre-commit test cases for CorrelatedValuePropagation to handle both values of select
This is a pre-commit of test cases relevant for D119643.
CorrelatedValuePropagation should handle inverted select condition, but it does not yet.
Kai Nacke [Tue, 22 Feb 2022 18:54:16 +0000 (13:54 -0500)]
[SystemZ/z/OS] Add va intrinsics for XPLINK
Add support for va intrinsics for the XPLINK ABI.
Only the extended vararg variant, which uses a pointer to next
argument, is supported. The standard variant will build on this.
Reviewed By: uweigand
Differential Revision: https://reviews.llvm.org/D120148
Zarko Todorovski [Tue, 22 Feb 2022 19:34:10 +0000 (14:34 -0500)]
[libc++][AIX] Add AIX error message as expected output
AIX's libc generates "Error -1 occurred" instead of the "Unknown Error"
expected by these test cases. Add this as expected output for AIX only.
Reviewed By: daltenty, #powerpc, #libc, zibi, Quuxplusone
Differential Revision: https://reviews.llvm.org/D119982
Rainer Orth [Tue, 22 Feb 2022 19:18:22 +0000 (20:18 +0100)]
[sanitizer_common] Use GetStaticTlsBoundary on Solaris 11.4
This is a restricted alternative to D91605
<https://reviews.llvm.org/D91605> which only works on Solaris 11.4 SRU 10+,
but would break the build on Solaris 11.3 and Illumos which lack
`dlpi_tls_modid`.
Apart from that, the patch is trivial. One caveat is that the
`sanitizer_common` and `asan` tests need to be linked explicitly with `ld
-z relax=transtls` on Solaris/amd64 since the archives with calls to
`__tls_get_addr` are linked in directly.
Tested on `amd64-pc-solaris2.11`, `sparcv9-sun-solaris2.11`, and
`x86_64-pc-linux-gnu`.
Differential Revision: https://reviews.llvm.org/D120048
Rainer Orth [Tue, 22 Feb 2022 19:14:33 +0000 (20:14 +0100)]
[Driver] Support Solaris/amd64 GetTls
This is the driver part of D91605 <https://reviews.llvm.org/D91605>, a
workaround to allow direct calls to `__tls_get_addr` on Solaris/amd64.
Tested on `amd64-pc-solaris2.11` and `sparcv9-sun-solaris2.11`.
Differential Revision: https://reviews.llvm.org/D119829
Adrian Prantl [Thu, 11 Nov 2021 00:25:26 +0000 (16:25 -0800)]
Add support for chained fixup load commands to MachOObjectFile
This is part of a series of patches to upstream support for Mach-O chained fixups.
This patch adds support for parsing the chained fixup load command and
parsing the chained fixups header. It also puts into place the
abstract interface that will be used to iterate over the fixups.
Differential Revision: https://reviews.llvm.org/D113630
Adrian Prantl [Wed, 10 Nov 2021 00:00:29 +0000 (16:00 -0800)]
Add a (nonfunctional) -dyld_info flag to llvm-objdump.
Darwin otool implements this flag as a one-stop solution for
displaying bind and rebase info. As I am working on upstreaming
chained fixup support this command will be useful to write testcases.
Differential Revision: https://reviews.llvm.org/D113573
Reid Kleckner [Tue, 22 Feb 2022 18:48:25 +0000 (10:48 -0800)]
Reland "[mlir][pdl] NFC re-add NoSideEffect to Result and Results Op"
This reverts commit
9865c3f28aa812364584b55629eef9b52bb1230e.
Looks like our commits raced and Jeff fixed the build issue.
Mogball [Tue, 22 Feb 2022 18:47:07 +0000 (18:47 +0000)]
[mlir][pdl] NFC fix missing include
Reid Kleckner [Tue, 22 Feb 2022 18:46:49 +0000 (10:46 -0800)]
Revert "[mlir][pdl] NFC re-add NoSideEffect to Result and Results Op"
This reverts commit
63eb963e58663541d6feb58f53a1bd4903e3dabf.
Breaks MLIR build.
Reid Kleckner [Tue, 22 Feb 2022 18:31:09 +0000 (10:31 -0800)]
Revert "[AArch64] Alter mull shuffle(ext(..)) combine to work on buildvectors"
This reverts commit
9fc1a0dcb79afb31470751651c30e843c12e9ca5.
We have bisected a compiler crash to this revision and will provide a
test case soon.
Mogball [Tue, 22 Feb 2022 18:26:52 +0000 (18:26 +0000)]
[mlir][pdl] NFC re-add NoSideEffect to Result and Results Op
Philip Reames [Tue, 22 Feb 2022 18:27:00 +0000 (10:27 -0800)]
[SLP] Use isInSchedulingRegion consistently [NFC]
Wouter van Oortmerssen [Mon, 14 Feb 2022 23:55:24 +0000 (15:55 -0800)]
[WebAssembly] Allow .data shorthand for .section .data,"",@
Shubham Sandeep Rastogi [Tue, 22 Feb 2022 00:44:38 +0000 (16:44 -0800)]
Updated reflection-dump.test for mpenum section
With
1c1e2cce9a50ac9fe6b884b79925d71914cf5a30 a new swift5 reflection section for multi-payload enum mask information was added, which is called mpenum. This change simply adds a check to make sure dsymutil can dump out information in that section into the dSYM bundle.
Differential Revision: https://reviews.llvm.org/D120291
Mogball [Mon, 21 Feb 2022 00:28:21 +0000 (00:28 +0000)]
[pdl] Remove `NoSideEffect` from all PDL ops
This trait results in PDL ops being erroneously CSE'd. These ops are side-effect free in the rewriter but not in the matcher (where unused values aren't allowed anyways). These ops should have a more nuanced side-effect modeling, this is fixing a bug introduced by a previous change.
Reviewed By: rriddle
Differential Revision: https://reviews.llvm.org/D120222
Philip Reames [Tue, 22 Feb 2022 18:13:14 +0000 (10:13 -0800)]
[SLP] Schedule only sub-graph of vectorizable instructions
SLP currently schedules all instructions within a scheduling window which stretches from the first instruction potentially vectorized to the last. This window can include a very large number of unrelated instructions which are not being considered for vectorization. This change switches the code to only schedule the sub-graph consisting of the instructions being vectorized and their transitive users.
This has the effect of greatly reducing the amount of work performed in large basic blocks, and thus greatly improves compile time on degenerate examples. To understand the effects, I added some statistics (not planned for upstream contribution). Here's an illustration from my motivating example:
Before this patch:
704357 SLP - Number of calcDeps actions
699021 SLP - Number of schedule calls
5598 SLP - Number of ReSchedule actions
59 SLP - Number of ReScheduleOnFail actions
10084 SLP - Number of schedule resets
8523 SLP - Number of vector instructions generated
After this patch:
102895 SLP - Number of calcDeps actions
161916 SLP - Number of schedule calls
5637 SLP - Number of ReSchedule actions
55 SLP - Number of ReScheduleOnFail actions
10083 SLP - Number of schedule resets
8403 SLP - Number of vector instructions generated
I do want to highlight that there is a small difference in number of generated vector instructions. This example is hitting the bailout due to maximum window size, and the change in scheduling is slightly perturbing when and how we hit it. This can be seen in the RescheduleOnFail counter change. Given that, I think we can safely ignore.
The downside of this change can be seen in the large test diff. We group all vectorizable instructions together at the bottom of the scheduling region. This means that vector instructions can move quite far from their original point in code. While maybe undesirable, I don't see this as being a major problem as this pass is not intended to be a general scheduling pass.
For context, it's worth noting that the pre-scheduling that SLP does while building the vector tree is exactly the sub-graph scheduling implemented by this patch.
Differential Revision: https://reviews.llvm.org/D118538
Valentin Clement [Tue, 22 Feb 2022 18:08:51 +0000 (19:08 +0100)]
[flang] Update PFTBuilder
This patch update the PFTBuilder to be able to lower
the construct present in semantics.
This is a building block for other lowering patches that will be posted soon.
This patch is part of the upstreaming effort from fir-dev branch.
Reviewed By: PeteSteinfeld, schweitz
Differential Revision: https://reviews.llvm.org/D120336
Co-authored-by: Jean Perier <jperier@nvidia.com>
Co-authored-by: V Donaldson <vdonaldson@nvidia.com>
Fangrui Song [Tue, 22 Feb 2022 18:07:58 +0000 (10:07 -0800)]
[ELF] Move duplicate symbol check after input file parsing
https://discourse.llvm.org/t/parallel-input-file-parsing/60164
To decouple symbol initialization and section initialization, `Defined::section`
assignment should be postponed after input file parsing. To avoid spurious
duplicate definition error due to two definitions in COMDAT groups of the same
signature, we should postpone the duplicate symbol check.
The function is called postScan instead of a more specific name like
checkDuplicateSymbols, because we may merge Symbol::mergeProperties into
postScan. It is placed after compileBitcodeFiles to apply to ET_REL files
produced by LTO. This causes minor diagnostic regression
for skipLinkedOutput configurations: ld.lld --thinlto-index-only a.bc b.o
(bitcode definition prevails) won't detect duplicate symbol error. I think this
is an acceptable compromise. The important cases where (a) both files are
bitcode or (b) --thinlto-index-only is unused are still detected.
Reviewed By: ikudrin
Differential Revision: https://reviews.llvm.org/D119908
Shilei Tian [Tue, 22 Feb 2022 17:59:56 +0000 (12:59 -0500)]
[Clang][OpenMP] Add the codegen support for `atomic compare`
This patch adds the codegen support for `atomic compare` in clang.
Reviewed By: ABataev
Differential Revision: https://reviews.llvm.org/D118632
Aaron Ballman [Tue, 22 Feb 2022 17:50:39 +0000 (12:50 -0500)]
Jay Foad [Fri, 18 Feb 2022 11:24:05 +0000 (11:24 +0000)]
[StableHashing] Hash machine basic blocks and functions
This adds very basic support for hashing MachineBasicBlock
and MachineFunction, for use in MachineFunctionPass to
detect passes that modify the MachineFunction wrongly.
Differential Revision: https://reviews.llvm.org/D120122
Jay Foad [Tue, 22 Feb 2022 09:49:19 +0000 (09:49 +0000)]
[StructurizeCFG] Fix boolean not bug
D118623 added code to fold not-of-compare into a compare
with the inverted predicate, if the compare had no other
uses. This relies on accurate use lists in the IR but it
was run before setPhiValues, when some phi inputs are still
stored in a data structure on the side, instead of being
real uses in the IR. The effect was that a phi that should
be using the original compare result would now get an
inverted result instead.
Fix this by moving simplifyConditions after setPhiValues.
Differential Revision: https://reviews.llvm.org/D120312
Simon Atanasyan [Tue, 22 Feb 2022 13:46:28 +0000 (16:46 +0300)]
[MIPS] Add `-no-pie` option to the clang driver's tests depend on it
Stanislav Mekhanoshin [Mon, 21 Feb 2022 22:48:59 +0000 (14:48 -0800)]
[AMDGPU] Extend SILoadStoreOptimizer to handle global saddr loads
This adds handling of the _SADDR forms to the GLOBAL_LOAD combining.
TODO: merge global stores.
TODO: merge flat load/stores.
TODO: merge flat with global promoting to flat.
Differential Revision: https://reviews.llvm.org/D120285
Nikita Popov [Tue, 22 Feb 2022 16:56:01 +0000 (17:56 +0100)]
[InstCombine] Add test for missed select fold due to one use limitation (NFC)
The eq sub zero fold currently has an artificial one-use limitation,
causing us to miss this fold.
Stanislav Mekhanoshin [Mon, 21 Feb 2022 19:17:58 +0000 (11:17 -0800)]
[AMDGPU] Extend SILoadStoreOptimizer to handle global loads
There can be situations where global and flat loads and stores are not
combined by the vectorizer, in particular if their address space
differ in the IR but they end up the same class instructions after
selection. For example a divergent load from constant address space
ends up being the same global_load as a load from global address space.
TODO: merge global stores.
TODO: handle SADDR forms.
TODO: merge flat load/stores.
TODO: merge flat with global promoting to flat.
Differential Revision: https://reviews.llvm.org/D120279
Nikita Popov [Mon, 14 Feb 2022 16:00:41 +0000 (17:00 +0100)]
[Bitcode] Store type IDs for values
This is the next step towards supporting bitcode auto upgrade with
opaque pointers. The ValueList now stores the Value* together with
its associated type ID, which allows inspecting the original pointer
element type of arbitrary values.
This is a largely mechanical change threading the type ID through
various places. I've left TODOTypeID placeholders in a number of
places where determining the type ID is either non-trivial or
requires allocating a new type ID not present in the original
bitcode. For this reason, the new type IDs are also not used for
anything yet (apart from propagation). They will get used once the
TODOs are resolved.
Differential Revision: https://reviews.llvm.org/D119821
serge-sans-paille [Sun, 20 Feb 2022 08:05:46 +0000 (09:05 +0100)]
Restore documentation for __builtin_assume
This got removed by
6cacd420a1d72bca7809e6b516fb1e18ac6056c8, and that was a
mistake.
Differential Revision: https://reviews.llvm.org/D120205
tyb0807 [Tue, 1 Feb 2022 21:41:23 +0000 (21:41 +0000)]
[AArch64] Remove unused feature flags from AArch64TargetInfo
This removes two feature flags from `AArch64TargetInfo` class:
- `HasHBC`: this feature does not involve generating any IR intrinsics,
so clang does not need to know about whether it is set
- `HasCrypto`: this feature is deprecated in favor of finer grained
features such as AES, SHA2, SHA3 and SM4. The associated ACLE macro
__ARM_FEATURE_CRYPTO is thus no longer used.
Differential Revision: https://reviews.llvm.org/D118757
Marek Kurdej [Tue, 22 Feb 2022 14:46:28 +0000 (15:46 +0100)]
[clang-format] Avoid parsing "requires" as a keyword in non-C++-like languages.
Fixes the issue raised post-review in D113319 (cf. https://reviews.llvm.org/D113319#3337485).
Reviewed By: krasimir
Differential Revision: https://reviews.llvm.org/D120324
Nemanja Ivanovic [Tue, 22 Feb 2022 15:52:19 +0000 (09:52 -0600)]
[PowerPC] Allow absolute expressions in relocations
The Linux kernel build uses absolute expressions suffixed with @lo/@ha
relocations. This currently doesn't work for DS/DQ form instructions and
there is no reason for it not to. It also works with GAS.
This patch allows this as long as the value is a multiple of 4/16
for DS/DQ form.
Differential revision: https://reviews.llvm.org/D115419
Marek Kurdej [Tue, 22 Feb 2022 15:40:30 +0000 (16:40 +0100)]
[clang-format] Use FormatToken::is* functions without passing through `Tok`. NFC.
Timm Bäder [Tue, 22 Feb 2022 15:34:12 +0000 (16:34 +0100)]
Fix docs build after
f8cedc642d9b85720cb7175ef25ddde90a3fbca2
Looks like rst doesn't like '#' in link texts. Just remove it.
Nikita Popov [Tue, 22 Feb 2022 15:31:46 +0000 (16:31 +0100)]
[CodeGen] Add test for PR53990 (NFC)
Timm Bäder [Fri, 11 Feb 2022 07:27:33 +0000 (08:27 +0100)]
[clang] Never wrap a nullptr in CXXNewExpr::getArraySize()
Otherwise callers of these functions have to check both the return value
for and the contents of the returned llvm::Optional.
Fixes #53742
Differential Revision: https://reviews.llvm.org/D119525
Matthias Springer [Tue, 22 Feb 2022 14:58:57 +0000 (23:58 +0900)]
[mlir][bufferize] Fix GCC build
Differential Revision: https://reviews.llvm.org/D120326
Pavel Labath [Mon, 21 Feb 2022 14:38:52 +0000 (15:38 +0100)]
[lldb] Remove HostProcess:GetMainModule
the function is unused, and the posix implementation is only really correct on linux.
Timm Bäder [Tue, 22 Feb 2022 10:07:39 +0000 (11:07 +0100)]
[clang][www] Port make_cxx_dr_status script to Python3
And run it to re-generate the cxx_dr_status.html
Differential Revision: https://reviews.llvm.org/D120313
Krasimir Georgiev [Tue, 22 Feb 2022 14:43:09 +0000 (15:43 +0100)]
[clang-format] Fix preprocessor nesting after commit
529aa4b011c4ae808d658022ef643c44dd9b2c9c
In https://github.com/llvm/llvm-project/commit/
529aa4b011c4ae808d658022ef643c44dd9b2c9c
by setting the identifier info to nullptr, we started to subtly
interfere with the parts in the beginning of the function,
https://github.com/llvm/llvm-project/blob/
529aa4b011c4ae808d658022ef643c44dd9b2c9c/clang/lib/Format/UnwrappedLineParser.cpp#L991
causing the preprocessor nesting to change in some cases. E.g., for the
added regression test, clang-format started incorrectly guessing the
language as C++.
This tries to address this by introducing an internal identifier info
element to use instead.
Reviewed By: curdeius, MyDeveloperDay
Differential Revision: https://reviews.llvm.org/D120315
Sander de Smalen [Tue, 22 Feb 2022 13:48:21 +0000 (13:48 +0000)]
[AArch64][SME] Remove term 'streaming-sve' from assembler diagnostics.
'streaming-sve' is not a feature that users should be able to set,
hence why it shouldn't show up in user-diagnostics. The only
flag that end-users should be able to set is '+sme'.
Reviewed By: paulwalker-arm
Differential Revision: https://reviews.llvm.org/D120256
Egor Zhdan [Mon, 24 Jan 2022 14:18:14 +0000 (14:18 +0000)]
Add DriverKit support
This patch is the first in a series of patches to upstream the support for Apple's DriverKit. Once complete, it will allow targeting DriverKit platform with Clang similarly to AppleClang.
This code was originally authored by JF Bastien.
Differential Revision: https://reviews.llvm.org/D118046
Simon Moll [Tue, 22 Feb 2022 13:29:29 +0000 (14:29 +0100)]
[VE] Split unsupported v512.32 ops
Split v512.32 binary ops into two v256.32 ops using packing support
opcodes (vec_unpack_lo|hi, vec_pack).
Depends on D120053 for packing opcodes.
Reviewed By: kaz7
Differential Revision: https://reviews.llvm.org/D120146