Guillaume Chatelet [Fri, 3 Feb 2023 16:28:56 +0000 (16:28 +0000)]
[NFC] Migrate aarch64 alignment to Align
David Sherwood [Fri, 3 Feb 2023 13:33:18 +0000 (13:33 +0000)]
[AArch64][SVE2p1] Add IntrNoMem to int_aarch64_sve_psel intrinsic
Differential Revision: https://reviews.llvm.org/D143262
Guillaume Chatelet [Fri, 3 Feb 2023 16:05:45 +0000 (16:05 +0000)]
[NFC] Remove dead code
Viktoriia Bakalova [Fri, 3 Feb 2023 13:00:51 +0000 (14:00 +0100)]
[include-mapping] Implement language separation in stdlib recognizer library
Differential Revision: https://reviews.llvm.org/D142992
Teresa Johnson [Thu, 2 Feb 2023 15:01:16 +0000 (07:01 -0800)]
[MemProf] Add helper to access the back (last) call stack id
This is split out of D140908 as suggested.
Differential Revision: https://reviews.llvm.org/D143184
Timm Bäder [Tue, 3 Jan 2023 09:04:27 +0000 (10:04 +0100)]
[clang][Interp] Support pointers in compound assignment operators
Differential Revision: https://reviews.llvm.org/D140874
Guillaume Chatelet [Fri, 3 Feb 2023 15:21:02 +0000 (15:21 +0000)]
[NFC] Simplify logic in ConstantFold
Differential Revision: https://reviews.llvm.org/D143271
Jonas Hahnfeld [Fri, 3 Feb 2023 15:33:02 +0000 (16:33 +0100)]
[Orc] XFAIL test with Comdats on macOS
The bots report:
LLVM ERROR: MachO doesn't support COMDATs, 'f' cannot be lowered.
Janek van Oirschot [Fri, 3 Feb 2023 15:23:29 +0000 (15:23 +0000)]
[AMDGPU] Modify adjustInliningThreshold to also consider the cost of passing function arguments through the stack
A regression from when new PM got enabled as default. Functions with a big number of instructions will elide getting inlined but do not consider the cost of passing arguments over stack if there are a lot of function arguments. This patch attempts to add a heuristic for AMDGPU's function calling convention that also considers function arguments passed through the stack.
Reviewed By: #amdgpu, arsenm
Differential Revision: https://reviews.llvm.org/D140242
Tobias Gysi [Fri, 3 Feb 2023 15:20:30 +0000 (16:20 +0100)]
[mlir][llvm] Use tablegen for enum conversion.
The revision uses tablegen to convert multiple atomic and comparison
related enums automatically rather than using hand coded functions
in the import and export from and to LLVM IR.
The revision also adds additional binary operation cases to the
AtomicBinOp enum that have not been supported till now. It also
introduces the possibility to define unsupported enum cases that exist
only in LLVM IR and that are not imported into MLIR. These unsupported
cases are helpful to handle sentinel values such as BAD_BINOP that
LLVM commonly uses to terminate its enums.
Reviewed By: Dinistro
Differential Revision: https://reviews.llvm.org/D143189
Timm Bäder [Fri, 3 Feb 2023 15:23:39 +0000 (16:23 +0100)]
Revert "[clang][Interp] Only generate disassembly in debug builds"
This reverts commit
60dcc70e48fa9814a4e0d5e12856bc88b7f927eb.
This breaks builders, e.g.
https://lab.llvm.org/buildbot/#/builders/36/builds/30036
Haojian Wu [Fri, 3 Feb 2023 15:08:06 +0000 (16:08 +0100)]
[Tooling/Inclusions] ensure the Mapping is inititalized in
Symbol/Header::all() calls
Timm Bäder [Thu, 22 Dec 2022 16:43:50 +0000 (17:43 +0100)]
[clang][Interp][NFCI] Handle DiscardResult for ExprWithCleanups
Just pop the pointer.
Sander de Smalen [Fri, 3 Feb 2023 09:38:45 +0000 (09:38 +0000)]
[LoopVectorize][TTI] NFCI: Clarify enum for the tail folding style.
This NFC (intended) patch has several small changes:
* It renames PredicationStyle to TailFoldingStyle.
* It renames TTI.emitActiveLaneMask() to TTI.getPreferredTailFoldingStyle()
* Simplifies some of its uses in the LoopVectorizer
Rationale: To my surprise PredicationStyle::None did not mean 'no
predication', but rather 'no active lane mask intrinsic', such that the
predicate is created using a splat + compare with stepvector. The enum is
also highly specific to tail folding, so it seems better to name this
around that feature, i.e. 'tail folding style'.
This also makes it more amenable to extend it to other tail folding styles,
such as the one added in D142109.
Reviewed By: david-arm
Differential Revision: https://reviews.llvm.org/D142887
Sanjay Patel [Fri, 3 Feb 2023 14:31:31 +0000 (09:31 -0500)]
[SDAG] fix miscompiles caused by using ValueTracking matchSelectPattern to create FMINIMUM/FMAXIMUM
ValueTracking attempts to match compare+select patterns to FP min/max
operations, but it was created before the newer IEEE-754-2019
minimum/maximum ops were defined. Ie, matchSelectPattern() does not
account for the -0.0/+0.0 behavior that is specified in the newer
standard.
FMINIMUM/FMAXIMUM nodes were created to map to the newer standard:
/// FMINIMUM/FMAXIMUM - NaN-propagating minimum/maximum that also treat -0.0
/// as less than 0.0. While FMINNUM_IEEE/FMAXNUM_IEEE follow IEEE 754-2008
/// semantics, FMINIMUM/FMAXIMUM follow IEEE 754-2018 draft semantics.
We could adjust ValueTracking to deal with signed zero, but it seems like
a moot point given the divergent NaN behavior discussed in D143056, so just
delete this possibility to avoid bugs when converting IR to SDAG.
Differential Revision: https://reviews.llvm.org/D143106
Felipe de Azevedo Piovezan [Tue, 31 Jan 2023 17:23:11 +0000 (12:23 -0500)]
[lldb] Enable arm64 target for entry values test
This test is supposed to work in arm64.
Differential Revision: https://reviews.llvm.org/D143265
David Spickett [Fri, 3 Feb 2023 10:22:35 +0000 (10:22 +0000)]
[libc] Fix quotation marks in overlay mode instructions
These were using ” ("right double quotation mark") instead of
the usual ". This means CMake took the value literally and you
got:
```
CMake Error at CMakeLists.txt:139 (MESSAGE):
”libc” isn't a known project:
bolt;clang;clang-tools-extra;compiler-rt;cross-project-tests;libc;libclc;lld;lldb;mlir;openmp;polly;pstl;flang.
Did you mean to enable it as a runtime in LLVM_ENABLE_RUNTIMES?
```
`="libc"` works fine.
Reviewed By: lntue
Differential Revision: https://reviews.llvm.org/D143254
Timm Bäder [Fri, 27 Jan 2023 08:14:31 +0000 (09:14 +0100)]
[clang][Interp] Only generate disassembly in debug builds
With the current set of opcodes, this saves 3460 lines in the generated
Opcodes.inc in release builds (-17%).
Differential Revision: https://reviews.llvm.org/D142694
Timm Bäder [Mon, 30 Jan 2023 08:09:22 +0000 (09:09 +0100)]
[clang][Interp][NFC] Make VariableScope::getParent() const
Marco Elver [Fri, 3 Feb 2023 14:34:42 +0000 (15:34 +0100)]
[SanitizerBinaryMetadata] Treat constant globals and non-escaping addresses specially
For atomics metadata, we can make data race analysis more efficient by
entirely ignoring functions that include memory accesses but which only
access non-escaping (non-shared) and/or non-mutable memory. Such
functions will not be considered to be covered by "atomics" metadata,
resulting in the following benefits:
1. reduces "covered" metadata; and
2. allows data race analysis to skip such functions.
Reviewed By: dvyukov
Differential Revision: https://reviews.llvm.org/D143159
Sanjay Patel [Thu, 2 Feb 2023 21:16:48 +0000 (16:16 -0500)]
[InstCombine] add tests for shuffle-of-fabs; NFC
Adapted from the existing shuffle-of-fneg tests.
Alex Zinenko [Tue, 31 Jan 2023 09:37:09 +0000 (09:37 +0000)]
[mlir] properly fix concurrent transform interpreter pass base
The original implementation of the transform interpreter pass base was
cloning the entire transform IR in presence of PDL-related operations to
avoid concurrency issues when running the pass with the same transform
IR on multiple operations of the payload IR. The root cause of those
issues is the `transform.pdl_match` operation that was moving the PDL
pattern definition operation into a new module, consumed by the PDL
interpreter and leading to a race. Clone the pattern operation instead.
This avoids the race as well as the cost for transform IR that doesn't
use PDL.
Depends on D142729.
Reviewed By: nicolasvasilache
Differential Revision: https://reviews.llvm.org/D142962
Alex Zinenko [Fri, 27 Jan 2023 16:01:26 +0000 (16:01 +0000)]
[mlir] provide a base class for transform interpreter passes
The transform dialect infrastructure does not provide a default
interpreter pass and instead expects users to create their own to ensure
all relevant extensions and dependent dialects are loaded. Provide a
base class for implementing such passes that includes the additional
facilities for debugging and is aware of the multithreaded nature of
pass execution.
Reviewed By: pifon2a, nicolasvasilache
Differential Revision: https://reviews.llvm.org/D142729
Timm Bäder [Sat, 15 Oct 2022 12:40:23 +0000 (14:40 +0200)]
[clang][Interp] Materializing primitive temporaries
Implement MaterializeTemporaryExpr for primitive types.
Differential Revision: https://reviews.llvm.org/D136017
Jean Perier [Fri, 3 Feb 2023 14:02:26 +0000 (15:02 +0100)]
[flang][hlfir] Handle intrinsic subroutines
The code did not propagate the result optionality for subroutine.
Make the result of genIntrinsicRef optional.
Differential Revision: https://reviews.llvm.org/D143251
Joe Loser [Wed, 1 Feb 2023 17:19:27 +0000 (10:19 -0700)]
[clang] Change AMX macros to match names from GCC
The current behavior for AMX macros is:
```
gcc -march=native -dM -E - < /dev/null | grep TILE
clang -march=native -dM -E - < /dev/null | grep TILE
```
which is not ideal. Change `__AMXTILE__` and friends to `__AMX_TILE__` (i.e.
have an underscore in them). This makes GCC and Clang agree on the naming of
these AMX macros to simplify downstream user code.
Fix this for `__AMXTILE__`, `__AMX_INT8__`, `__AMX_BF16__`, and `__AMX_FP16__`.
Differential Revision: https://reviews.llvm.org/D143094
Johannes Doerfert [Fri, 3 Feb 2023 13:32:10 +0000 (07:32 -0600)]
[OpenMP][FIX] Do not overalign mapped structures
While we potentially need to align partially mapped structs more than
the first member, we do not need to align past the struct itself. This
prevents us from moving the base pointer past the struct beginning too.
See https://reviews.llvm.org/D142508 for a discussion.
Reviewed By: pavelkopyl, grokos, jhuber6
Differential Revision: https://reviews.llvm.org/D142586
Hassnaa Hamdi [Fri, 3 Feb 2023 13:45:00 +0000 (13:45 +0000)]
Revert "[AArch64][CostModel]: Add costs for zero/sign extend."
This reverts commit
d65c3bf39aa49eda0f5a6bff02c9292895a37e26.
Timm Bäder [Mon, 16 Jan 2023 16:58:22 +0000 (17:58 +0100)]
[clang][Interp] Fix Pointer::toAPValue() for expressions
Differential Revision: https://reviews.llvm.org/D141858
Simon Pilgrim [Thu, 2 Feb 2023 17:14:56 +0000 (17:14 +0000)]
[X86] Add some very basic test coverage for the few @llvm.vp.* intrinsics that correctly expand
Inspired by Issue #60464 - wel can add/enable additional intrinsics when we add expansion support
Guray Ozen [Fri, 3 Feb 2023 11:11:38 +0000 (12:11 +0100)]
[mlir] Add nontemporal field to memref.load/store and convey to llvm.load/store
`llvm.load` op has nonTemporal field which is missing for `memref.load` and `memref.store`. This revision first adds nonTemporal field to memref's load/store op, then it lowers the field to llvm.load/store ops.
Reviewed By: nicolasvasilache
Differential Revision: https://reviews.llvm.org/D142616
Sam McCall [Fri, 3 Feb 2023 12:47:13 +0000 (13:47 +0100)]
Revert unintended debug things :-(
Sam McCall [Tue, 24 Jan 2023 14:40:35 +0000 (15:40 +0100)]
[Tooling] Add stdlib::Symbol::all() and stdlib::Symbol::qualified_name()
These address some remaining reasons to #include StdSymbolMap.inc directly.
Differential Revision: https://reviews.llvm.org/D142467
Sergey Kachkov [Thu, 2 Feb 2023 14:49:54 +0000 (17:49 +0300)]
[NFC] PHITransAddr refactoring - use range-based loops and standard algorithms
Differential Revision: https://reviews.llvm.org/D143179
David Sherwood [Wed, 1 Feb 2023 13:48:48 +0000 (13:48 +0000)]
[AArch64][SVE2p1] Add 2-way SVE2p1 dot product intrinsics
This patch adds the LLVM IR intrinsics for the following:
* sdot (2-way, vectors + indexed)
* udot (2-way, vectors + indexed)
* fdot (vectors + indexed)
Differential Revision: https://reviews.llvm.org/D143066
Matt Arsenault [Fri, 27 Jan 2023 14:11:55 +0000 (10:11 -0400)]
AMDGPU: Add more tests to fneg modifier with casting tests
Valentin Clement [Fri, 3 Feb 2023 11:21:59 +0000 (12:21 +0100)]
[flang] Avoid double finalization when intrinsic assignment is done in the runtime
genRecordAssignment is emitting code to call Assign in the runtime for some cases.
In these cases, the finalization is done by the runtime so we do not need to do it in
a separate cal to avoid multiple finalization..
Also refactor the code in Bridge so the actual finalization of allocatable
is done before any reallocation. We might need to push this into ReallocIfNeeded.
It is not clear if the allocatable lhs needs to be finalized in any cases or only if it is
reallocated.
Reviewed By: jeanPerier
Differential Revision: https://reviews.llvm.org/D143186
David Sherwood [Fri, 3 Feb 2023 11:00:35 +0000 (11:00 +0000)]
Fix up tests committed in
57a6bb34729df30df18a133ee2083b3323a936dc
I committed an older version of the patch without the test updates.
This patch uses the latest versions on
https://reviews.llvm.org/D142904
David Sherwood [Mon, 30 Jan 2023 16:10:22 +0000 (16:10 +0000)]
[AArch64][SME2] Add multi-vector zip/uzp intrinsics
This patch adds the LLVM IR intrinsics for the following:
* zip (2 and 4 vectors, 8 to 64-bit elements) - aarch64.sve.zip.*
* zip (2 and 4 vectors, 128-bit elements) - aarch64.sve.zipq.*
* uzp (2 and 4 vectors, 8 to 64-bit elements) - aarch64.sve.uzp.*
* uzp (2 and 4 vectors, 128-bit elements) - aarch64.sve.uzpq.*
I have created separate intrinsics for the 128-bit variants
in a similar way to what was done for int_aarch64_sme_readq_horiz.
This permits us to use any vector type (<vscale x 16 x i8>, etc.)
for the 128-bit versions.
I have also named the tests sve2p1-intrinsics-* because although
the instructions are added as part of the SME2 feature they only
operate on SVE vectors.
NOTE: These intrinsics are still in development and are subject to future changes.
Differential Revision: https://reviews.llvm.org/D142904
Christian Sigg [Fri, 3 Feb 2023 10:47:28 +0000 (11:47 +0100)]
[MLIR] NFC: fully scope use FastMathFlag.
Second instance.
Piotr Sobczak [Thu, 2 Feb 2023 12:37:46 +0000 (13:37 +0100)]
[AMDGPU] Make offset a 32-bit operand
Promote offset to 32-bit, similarily to what D142549 did for flat_offset.
Differential Revision: https://reviews.llvm.org/D143174
Sergey Kachkov [Fri, 3 Feb 2023 10:26:30 +0000 (13:26 +0300)]
[NFC] Remove isSafeToSpeculativelyExecute checks for casts
isSafeToSpeculativelyExecute always return true for casts, so remove
this redundant checks.
Differential Revision: https://reviews.llvm.org/D143255
LLVM GN Syncbot [Fri, 3 Feb 2023 10:26:23 +0000 (10:26 +0000)]
[gn build] Port
1e72920c8859
Max Kazantsev [Fri, 3 Feb 2023 10:21:29 +0000 (17:21 +0700)]
[Test] Add tests with narrow checks for SCEVLoopGuardRewriter
Jonas Hahnfeld [Tue, 24 Jan 2023 09:54:55 +0000 (10:54 +0100)]
[ORC] Drop Comdat when discarding IR symbol
According to the IR verifier, "Declaration[s] may not be in a Comdat!"
This is a re-commit of
76b3f0b4d5a0b8c54147c4c73a30892bbca76467 with
updates to the test:
* Force emission of the extra-module, to trigger the bug after D138264,
by providing a second symbol @g, and making the comdat nodeduplicate.
(Technically only one is needed, but two should be safer.)
* Name the comdat $f to avoid failure on Windows:
LLVM ERROR: Associative COMDAT symbol 'c' does not exist.
Differential Revision: https://reviews.llvm.org/D142443
Alexey Lapshin [Mon, 30 Jan 2023 14:05:53 +0000 (15:05 +0100)]
[dsymutil] dsymutil produces broken lines info (probably) with LTO on mac
This patch fixes #60307 issue. The 8bb4451 introduces the possibility
to unite overlapped or adjacent address ranges to keep address ranges
in an unambiguous state. The AddressRangesMap is used to normalize
address ranges. The AddressRangesMap keeps address ranges and the value
of the relocated address. For intersected range, it creates a united
range that keeps the last inserted mapping value. The same for adjusted ranges.
While it is OK to use the last inserted mapping value for intersected ranges
(as there is no way how to resolve ambiguity) It is not OK to use the
last inserted value for adjacent address ranges. Currently, two following
address ranges are united into a single one:
{0,24,17e685c} {24,d8,55afe20} -> {0,d8,55afe20}
To avoid the problem, the AddressRangesMap should not unite adjacent address ranges
with different relocated addresses. Instead, it should leave adjacent address ranges
as separate ranges. So, the ranges should look like this:
{0,24,17e685c} {24,d8,55afe20}
Differential Revision: https://reviews.llvm.org/D142936
Anton Sidorenko [Wed, 1 Feb 2023 15:58:24 +0000 (18:58 +0300)]
[llvm-objcopy] Fix strip-all-gnu test
Originally the test has a bug: in the input YAML `.symtab.dyn` section has a type
repeated twice with different values (SHT_SYMTAB and SHT_NOBITS). YAML parser
took SHT_NOBITS as a type of the section, so the test wasn't checking the desired
case.
This patch changes the test to verify that a section with SHT_SYMTAB type and
SHF_ALLOC flag is not removed when --strip-all-gnu flag present.
Differential Revision: https://reviews.llvm.org/D143086
Max Kazantsev [Fri, 3 Feb 2023 10:06:37 +0000 (17:06 +0700)]
[Test] Add signed counterparts of unsigned tests on SCEVLoopGuardRewriter
Max Kazantsev [Fri, 3 Feb 2023 09:00:24 +0000 (16:00 +0700)]
[Test] Add motivating test for umin support in SCEVLoopGuardRewriter
This test is equivalent to another one, where we can infer a
reasonable iteration count. The only difference is that umin and zext
here are swapped, and rewriter fails here.
David Sherwood [Wed, 25 Jan 2023 08:58:44 +0000 (08:58 +0000)]
[AArch64][SME2] Add LLVM IR intrinsics for multi-indexed dots
Adds intrinsics for the following SME2 instructions:
* sdot (multi-indexed, 2 and 4 vectors, 32-bit and 64-bit ZA)
* udot (multi-indexed, 2 and 4 vectors, 32-bit and 64-bit ZA)
* usdot (multi-indexed, 2 and 4 vectors)
* sudot (multi-indexed, 2 and 4 vectors)
* fdot (multi-indexed, 2 and 4 vectors)
* bfdot (multi-indexed, 2 and 4 vectors)
NOTE: These intrinsics are still in development and are subject to future changes.
Differential Revision: https://reviews.llvm.org/D142533
ManuelJBrito [Thu, 2 Feb 2023 17:42:31 +0000 (17:42 +0000)]
[Clang] Add builtin_nondeterministic_value
Differential Revision: https://reviews.llvm.org/D142388
Jean Perier [Fri, 3 Feb 2023 09:29:19 +0000 (10:29 +0100)]
[flang] use runRegionDCE instead of a custom DCE in cg-rewrite
The custom DCE in cg-rewrite is meant to get rid of fir.shape, fir.shift,
fir.shape_shift and fir.slice ops as well as their unused operands
before codegen (that does not lower those abstract operation to LLVM).
However, it turned out to be flowed in case some fir.shape operands were
unused outside of fir.shape and appeared several times as operands:
they were erased at the first appearance, causing the further attemp
to erase it to segfault (since the op IR storage was deallocated).
Instead of trying to fixing the custom DCE code, use mlir::runRegionDCE.
Differential Revision: https://reviews.llvm.org/D143247
Valentin Clement [Fri, 3 Feb 2023 09:16:54 +0000 (10:16 +0100)]
[flang] Fix potential null scope when lowering dispatch table op
Similary to D140209, the scope might need to be retrieved
from the typeSymbol. The test code was crashing because the
scope passed to CollectBindings was initially null.
Reviewed By: jeanPerier
Differential Revision: https://reviews.llvm.org/D143188
chenglin.bi [Fri, 3 Feb 2023 09:15:47 +0000 (17:15 +0800)]
[InstCombine] typo fix in the test xor-and-or.ll; NFC
chenglin.bi [Fri, 3 Feb 2023 09:11:41 +0000 (17:11 +0800)]
[InstCombine] Fold pattern xor(and, or) to select
(A & B) ^ (A | C) --> A ? ~B : C
https://alive2.llvm.org/ce/z/KCBfXr
https://alive2.llvm.org/ce/z/Pm-zJN
https://alive2.llvm.org/ce/z/VT8uC2
Reviewed By: spatel
Differential Revision: https://reviews.llvm.org/D141660
Sergey Kachkov [Thu, 2 Feb 2023 12:52:34 +0000 (15:52 +0300)]
[NFC] PHITransAddr refactoring - return translated value directly or nullptr on
failure (instead of bool flag)
Differential Revision: https://reviews.llvm.org/D143171
Jonas Hahnfeld [Fri, 3 Feb 2023 09:04:38 +0000 (10:04 +0100)]
Revert "[ORC] Drop Comdat when discarding IR symbol"
Failure on Windows:
LLVM ERROR: Associative COMDAT symbol 'c' does not exist.
This reverts commit
76b3f0b4d5a0b8c54147c4c73a30892bbca76467 while
I investigate the problem and a solution that still triggers the
original problem.
Matthias Springer [Fri, 3 Feb 2023 08:44:42 +0000 (09:44 +0100)]
[mlir] GreedyPatternRewriteDriver: Ignore scope when rewriting top-level ops
Top-level ModuleOps cannot be transformed with the GreedyPatternRewriteDriver since D141945 because they do not have an enclosing region that could be used as a scope. Make the scope optional inside GreedyPatternRewriteDriver, so that top-level ops can be processed when they are on the initial list of ops.
Note: This does not allow users to bypass the scoping mechanism by setting `config.scope = nullptr`.
Fixes #60462.
Differential Revision: https://reviews.llvm.org/D143151
Martin Storsjö [Tue, 31 Jan 2023 09:09:50 +0000 (11:09 +0200)]
[docs] Rewrite/improve the docs for LLVM_NATIVE_TOOL_DIR
Don't include it among the mandatory options; the automatically built
tools via a nested cmake build work fine these days
(in particular, since
93010544a813dfbfa64dd7cee68785f572f974d1 /
https://reviews.llvm.org/D126313).
Clarify the directory path-to-host-bin into something more verbose,
to avoid ambiguity with LLVM_HOST_TRIPLE.
Differential Revision: https://reviews.llvm.org/D142960
Martin Storsjö [Mon, 23 Jan 2023 12:13:40 +0000 (14:13 +0200)]
[docs] Prefer setting LLVM_HOST_TRIPLE instead of LLVM_DEFAULT_TARGET_TRIPLE and LLVM_TARGET_ARCH
Setting LLVM_HOST_TRIPLE propagates the information to a few more
places than if only setting LLVM_TARGET_ARCH and
LLVM_DEFAULT_TARGET_TRIPLE, while both of those settings get their
defaults implied from LLVM_HOST_TRIPLE if they're not overridden.
Differential Revision: https://reviews.llvm.org/D142404
Adrian Kuegel [Fri, 3 Feb 2023 08:42:34 +0000 (09:42 +0100)]
[mlir][Bazel] Update bazel BUILD after
889a11783ec
Owen Pan [Fri, 3 Feb 2023 08:23:46 +0000 (00:23 -0800)]
[clang-format][doc] Fix a typo
Thomas Raoux [Fri, 3 Feb 2023 07:53:30 +0000 (07:53 +0000)]
Fix bazel dependency
Differential Revision: https://reviews.llvm.org/D143244
Thomas Raoux [Fri, 3 Feb 2023 07:49:38 +0000 (07:49 +0000)]
[mlir][linalg] Fix crash in vectorizer when expanding affine apply
Fix the insert point when expanding affine apply and handle cases with
symbols. Also add missing precondition to dynamic shape vectorization.
Differential Revision: https://reviews.llvm.org/D143243
Fangrui Song [Fri, 3 Feb 2023 08:04:50 +0000 (00:04 -0800)]
[ELF] Fix help message for --lto-pgo-warn-mismatch
Christian Ulmann [Fri, 3 Feb 2023 07:33:37 +0000 (08:33 +0100)]
[mlir][llvm] Add structured loop metadata
This commit introduces a structured representation of loop metadata to
the LLVM dialect. This attribute explicitly models all known `!llvm.loop`
metadata fields and groups them by introducing nested attributes for each
namespace.
The new attribute replaces the LoopOptionAttr that could only model a
limited subset of loop metadata.
Reviewed By: gysit
Differential Revision: https://reviews.llvm.org/D143064
Konstantin Varlamov [Fri, 3 Feb 2023 07:54:22 +0000 (23:54 -0800)]
[libc++][format] Fix a missing include in tests.
Jonas Hahnfeld [Tue, 24 Jan 2023 09:54:55 +0000 (10:54 +0100)]
[ORC] Drop Comdat when discarding IR symbol
According to the IR verifier, "Declaration[s] may not be in a Comdat!"
Differential Revision: https://reviews.llvm.org/D142443
Phoebe Wang [Fri, 3 Feb 2023 07:19:21 +0000 (15:19 +0800)]
[X86][FP16] Set Custom action for vector FROUND
Reviewed By: RKSimon
Differential Revision: https://reviews.llvm.org/D143169
varconst [Thu, 2 Feb 2023 18:15:48 +0000 (10:15 -0800)]
[libc++] Make some tests in `math_nodiscard_extensions` Clang-only.
Clang implicitly adds the `[[gnu::const]]` attribute to many math
functions from the C standard library functions. Since this behavior is
Clang-specific, make sure the test only runs on Clang.
Differential Revision: https://reviews.llvm.org/D143033
Chuanqi Xu [Fri, 3 Feb 2023 06:29:14 +0000 (14:29 +0800)]
[NFC] [Serialization] Add static assert for Num*Declbits
This re-commits part of
c79635cce845. It is reverted since it contains
platform-inconsistent constant. Now the patch only contains constant
defined in DeclBase.h so it should be platform-independent. And this
should be still helpful.
Reviewed By: erichkeane
Differential Revision: https://reviews.llvm.org/D141992
gonglingqin [Fri, 3 Feb 2023 01:37:17 +0000 (09:37 +0800)]
[LoongArch] Override TargetLowering::hasAndNotCompare()
Override hasAndNotCompare() to use more `ANDN` instead of using `AND`
and `NOT`.
This patch enables the following transforms:
(X & Y) == Y ---> (~X & Y) == 0
(X & Y) != Y ---> (~X & Y) != 0.
Differential Revision: https://reviews.llvm.org/D143037
Fangrui Song [Fri, 3 Feb 2023 04:03:34 +0000 (20:03 -0800)]
[mlir] print-op-graph: StringMap=>map to stabilize iteration order
Shilei Tian [Fri, 3 Feb 2023 03:54:25 +0000 (22:54 -0500)]
[OpenMP] Guard the code if ITT is not used
`check_loc` is not used if ITT is disabled or debug is off, causing a
compiler warning.
Reviewed By: jlpeyton
Differential Revision: https://reviews.llvm.org/D143004
Changpeng Fang [Fri, 3 Feb 2023 02:57:26 +0000 (18:57 -0800)]
AMDGPU: Use module flag to get code object version at IR level
Summary:
This patch introduces a mechanism to check the code object version from the module flag, This avoids checking from command line.
In case the module flag is missing, we use the current default code object version supported in the compiler.
For tools whose inputs are not IR, we may need other approach (directive, for example) to check the code
object version, That will be in a separate patch later.
For LIT tests update, we directly add module flag if there is only a single code object version associated with all checks in one file.
In cause of multiple code object version in one file, we use the "sed" method to "clone" the checks to achieve the goal.
Reviewer: arsenm
Differential Revision:
https://reviews.llvm.org/D14313
Matt Arsenault [Tue, 31 Jan 2023 18:21:17 +0000 (14:21 -0400)]
X86: Add some baseline tests for broken is.fpclass handling with DAZ
If DAZ is enabled, a test against fcZero is not equivalent to a compare
with 0.
Matt Arsenault [Thu, 2 Feb 2023 13:59:08 +0000 (09:59 -0400)]
AMDGPU: Add additional tests for is.fpclass legalization
Matt Arsenault [Thu, 2 Feb 2023 13:49:54 +0000 (09:49 -0400)]
AMDGPU: Regenerate test checks
Use right prefix order to get merging.
Also drop -verify-machineinstrs and add -amdgpu-enable-delay-alu=0
Matt Arsenault [Wed, 14 Dec 2022 22:58:56 +0000 (17:58 -0500)]
AMDGPU: Factor out fneg fold predicate function
Matt Arsenault [Thu, 15 Dec 2022 15:20:01 +0000 (10:20 -0500)]
AMDGPU: Try to unfold fneg source when matching legacy fmin/fmax
This is NFC as it stands, since other combines will effectively
prevent this from being reachable. This will avoid regressions in a
future change which tries to make better use of select source
modifiers.
Didn't bother with the GlobalISel part for now, since the baseline
combine doesn't seem to work on the existing test.
Chuanqi Xu [Fri, 3 Feb 2023 02:27:02 +0000 (10:27 +0800)]
[C++20] [Modules] Pop Expression Evaluation Context when we skip its body during parsing
Close https://github.com/llvm/llvm-project/issues/60275
The root cause of issue 60275 is the imbalance of
PushExpressionEvaluationContext() and PopExpressionEvaluationContext().
See
https://github.com/llvm/llvm-project/blob/
f1c4f927f7c15b5efdc3589c050fd0513bf6b303/clang/lib/Parse/Parser.cpp#L1396-L1437
We will PushExpressionEvaluationContext() in ActOnStartOfFunctionDef()
in line 1396 and we should pop it in ActOnFinishFunctionBody later.
However if we skip the function body in line 1402, the expression
evaluation context will not be popped. Then here is the issue report. I
fix the issue by inserting codes to pop the expression evaluation
context explicitly if the function body is skipped. Maybe this looks
like an ad-hoc fix. But if we want to fix this in a pretty way, we
should refactor the current framework for pushing and popping expression
evaluation contexts. Currently there are 23
PushExpressionEvaluationContext() callsities and 21
PopExpressionEvaluationContext() callsites in the code. And it seems not
easy to balance them well and fast. So I suggest to land this fix first.
At least it can prevent the crash.
Reviewed By: cor3ntin
Differential Revision: https://reviews.llvm.org/D143053
Ting Wang [Fri, 3 Feb 2023 01:52:52 +0000 (20:52 -0500)]
[PowerPC] add a peephole to remove redundant swap instructions after vector splats on P8
Vector store on P8 little endian will have swap instruction added before
the store in PPCISelLowring. If the vector is generated by splat, the
swap instruction can be eliminated.
Reviewed By: shchenz
Differential Revision: https://reviews.llvm.org/D139691
LiaoChunyu [Fri, 3 Feb 2023 01:16:49 +0000 (09:16 +0800)]
[RISCV] Permit tail call to an externally-defined function with weak linkage
As described in D45395 `This has been modeled after ARM's tail call opt.`
ARM's abi seems to limit weak symbol.
I did not find the limitation for RISCV. (Please correct me if I am wrong)
gcc seems to use the tail-call opt: https://godbolt.org/z/bjWE68n5o
Reviewed By: MaskRay
Differential Revision: https://reviews.llvm.org/D143137
Chris Bieneman [Fri, 3 Feb 2023 01:25:28 +0000 (19:25 -0600)]
[DX] Speculative big endian fix
I think this will get it this time.
Chris Bieneman [Fri, 3 Feb 2023 00:21:19 +0000 (18:21 -0600)]
[DX] Fix big-endian... again
Big endian is the bane of my existance today...
Mitch Phillips [Thu, 2 Feb 2023 23:48:50 +0000 (15:48 -0800)]
Revert "Add CFI integer types normalization"
This reverts commit
b1e9ab7438a098a18fecda88fc87ef4ccadfcf1e.
Reason: Looks like it broke the MSan buildbot, more details in the
phabricator review: https://reviews.llvm.org/D139395
Mitch Phillips [Thu, 2 Feb 2023 23:37:33 +0000 (15:37 -0800)]
[GWP-ASan] Remove thread clamping in tests.
It's better and easier for us to just have threads contend against each
other in the tests if it's more than the maximum supported number of
hardware threads available.
Specifically, the recoverable test fails on Android because the
GTEST_SKIP in a called function, and it only properly works from the
TEST_* harness function. Android tests run on cuttlefish, which can be a
single core with two hyperthreads.
Reviewed By: fmayer
Differential Revision: https://reviews.llvm.org/D143221
Ilia Diachkov [Mon, 30 Jan 2023 20:43:53 +0000 (23:43 +0300)]
[SPIRV][NFC] fix build warning and error
The patch fixes gcc's warning in SPIRVUtils.cpp after D142532.
Also it fixes compilation error by MSVC in SPIRVBuiltins.cpp.
Differential Revision: https://reviews.llvm.org/D142937
bixia1 [Thu, 2 Feb 2023 22:54:45 +0000 (14:54 -0800)]
[mlir][sparse] Implement heap sort for sparse_tensor.sort.
Reviewed By: aartbik
Differential Revision: https://reviews.llvm.org/D142913
Med Ismail Bennani [Wed, 1 Feb 2023 22:52:50 +0000 (14:52 -0800)]
[lldb] Fix typo in ScriptedProcess python docstrings (NFC)
Signed-off-by: Med Ismail Bennani <medismail.bennani@gmail.com>
Med Ismail Bennani [Thu, 2 Feb 2023 00:33:53 +0000 (16:33 -0800)]
[lldb/python] Fix scripted_platform python module creation
This patch should fix the creation and addition of the `scripted_platform`
python module into the `lldb.plugins` module.
Previously, we were creating the `plugins` submodule, each time with a
different source file (either `scripted_process` or `scripted_platform`).
The removes the redundant `create_python_package` call and group both
python source files toghether.
Differential Revision: https://reviews.llvm.org/D143122
Signed-off-by: Med Ismail Bennani <medismail.bennani@gmail.com>
Steven Wu [Thu, 2 Feb 2023 23:07:16 +0000 (15:07 -0800)]
[DeclContext] Sort the Decls before adding into DeclContext
Fix a non-deterministic issue in clang module generation, which the
anonymous declaration number from a function context is not
deterministic. This is due to the unstable iteration order for decls in
scope so the order after moving the decls into function decl context is
not deterministic.
From https://reviews.llvm.org/D135118, we can't use a set that preserves
the order without the performance penalty. Fix the issue by sorting the
decls based on raw encoding of their source location.
rdar://
104097976
Reviewed By: akyrtzi, vsapsai
Differential Revision: https://reviews.llvm.org/D141625
Fangrui Song [Thu, 2 Feb 2023 23:13:04 +0000 (15:13 -0800)]
libRemarks.{so,dylib}: remove Remarks.dylib.exports
Remarks.exports is only intended for NOT (BUILD_SHARED_LIBS OR LLVM_LINK_LLVM_DYLIB)
builds.
For (unintended use case) BUILD_SHARED_LIBS OR LLVM_LINK_LLVM_DYLIB (the latter
is used by some Linux distros), the library defines just one symbol on ELF.
There is no need to use a version script.
I think this is a more proper solution than D139932 and fixes `symbol not
defined` errors after lld default change D135402.
Fangrui Song [Thu, 2 Feb 2023 22:51:35 +0000 (14:51 -0800)]
libLTO.{so,dylib}: remove unused non-LTO symbols
These dissembler symbols are not used by LTO (see Apple ld64's use
in check-llvm-tools-lto). On ELF platforms, these symbols are not defined and are
rejected by ld --no-undefined-version.
I think this is a more proper solution than D139932 and this fixes
-DBUILD_SHARED_LIBS=on for ELF as well.
Chris Bieneman [Thu, 2 Feb 2023 19:11:34 +0000 (13:11 -0600)]
[DX] Add support for PSV resource bindings
This patch continues implementing DirectX pipeline state validation
information by adding support for resource binding metadata.
Reviewed By: python3kgae
Differential Revision: https://reviews.llvm.org/D143130
Mircea Trofin [Thu, 2 Feb 2023 22:41:55 +0000 (14:41 -0800)]
[mlgo][nfc] Refactor the log_reader.py utility
Small refactoring in preparation for tests for the interactive mode.
This allows reading the header, and performing observations, as explicit
steps. The latter is in particular necessary because the exit condition
for the interactive host will be that the child process (the compiler)
exited.
Johannes Doerfert [Thu, 2 Feb 2023 21:05:20 +0000 (13:05 -0800)]
[Attributor][NFCI] Avoid spending time resolving kernel reachability queries
We know kernels (generally) cannot be called from within the module. Thus,
for reachability we would need to step back from a kernel which would allow
us to reach anything anyway. Even if a kernel is invoked from another
kernel, values like allocas and shared memory are not accessible. We
implicitly check for this situation to avoid costly lookups.
Johannes Doerfert [Fri, 27 Jan 2023 05:05:30 +0000 (21:05 -0800)]
[Attributor][NCFI] Explicitly state what interfering accesses to look for
We used to check the query instructions for effects but that does not
work well with complex accesses we will probably support in the future.
Now we simply let the user decide what accesses to look for.
Jim Ingham [Thu, 2 Feb 2023 21:46:42 +0000 (13:46 -0800)]
Add usage info for backtick to the lldb tutorial.
Fangrui Song [Thu, 2 Feb 2023 21:46:52 +0000 (13:46 -0800)]
[modularize] StringMap=>map to make iteration order deterministic