Samuel Parker [Thu, 26 Jan 2023 11:52:24 +0000 (11:52 +0000)]
[DAGCombine] fp_to_sint isSaturatingMinMax
Check for single smax pattern against zero when converting from a
small enough float.
Differential Revision: https://reviews.llvm.org/D142481
Paul Walker [Thu, 26 Jan 2023 12:33:10 +0000 (12:33 +0000)]
[SVE] Move isel for casting between NEON and SVE vector types into tablegen.
This is purely refactoring to remove some unecessary C++ code.
Differential Revision: https://reviews.llvm.org/D142516
David Spickett [Wed, 14 Dec 2022 13:46:27 +0000 (13:46 +0000)]
[LLVM][TableGen] Notebook describing how to write a Python backend
This tutorial uses the dump json option to write a backend for
SQL queries.
It is based on the work of Min-Yih Hsu:
* https://github.com/mshockwave/SQLGen
* https://www.youtube.com/watch?v=UP-LBRbvI_U
I hope that having the same concepts in 3 forms will allow people
to choose the style that fits them.
The main drawback here being that it's in Python. C++ can be used
in a notebook (https://github.com/jupyter-xeus/xeus-cling) but I
decided against it for a few reasons:
* Python is the default for Jupyter, no extra installs needed.
* Having the code in a second language may help people who
know one or the other.
* There is no upstream example of a JSON powered backend.
(and although we would be unlikely to accept one upstream,
I think it's a great option for quick development before
translating into C++)
Reviewed By: myhsu
Differential Revision: https://reviews.llvm.org/D142364
Paul Walker [Wed, 18 Jan 2023 16:33:42 +0000 (16:33 +0000)]
[SVE] Fix incorrect lowering of predicate permute builtins.
When lowering predicate permute builtins we incorrectly assume only
the typically "active" bits for the specified element type play a
role with all other bits zero'd. This is not the case because all
bits are significant, with the element type specifying how they
are grouped:
b8 - permute using a block size of 1 bit
b16 - permute using a block size of 2 bits
b32 - permute using a block size of 4 bits
b64 - permute using a block size of 8 bits
The affected builtins are svrev, svtrn1, svtrn2, svuzp1, svuzp2,
svzip1 and svzip2.
This patch adds new intrinsics to support these operations and
changes the builtin lowering code to emit them. The b8 case remains
unchanged because for that operation the existing intrinsics work
as required and their support for other predicate types has been
maintained as useful if only as a way to test the correctness of
their matching ISD nodes that code generation relies on.
Differential Revision: https://reviews.llvm.org/D142065
Florian Hahn [Thu, 26 Jan 2023 12:04:20 +0000 (12:04 +0000)]
[SCCP] Adjust test after
cdeaf5f28c3dca.
Matt Devereau [Thu, 26 Jan 2023 12:02:16 +0000 (12:02 +0000)]
Revert "[LoopVectorize] Enable integer Mul and Add as select reduction patterns"
This reverts commit
f90103851f9a381bbf7ed6da250217577afd00d2.
Florian Hahn [Thu, 26 Jan 2023 11:45:10 +0000 (11:45 +0000)]
Recommit "[SCCP] Use range info to prove AddInst has NUW flag."
This reverts commit
531756b9544b3e164b9ab998292fce3ebbc7c746.
The recommitted version fixes a crash when one of the operands is a
constant other than a ConstantInt. Test for that case have been added
in
5b16cd97b8e1c273.
It splits off the new logic into a separate function because setting the
flags is quite different compared to the other cases handled in replaceSignedInst
which all require replacing an existing instruction.
It also guards makeGuaranteedNoWrapRegion by `if (!Inst.hasNoUnsignedWrap())`
as discussed in the review.
Fixes #60280.
Fixes #60278.
Original message:
This patch updates SCCP to use the value ranges of AddInst operands to
try to prove the AddInst does not overflow in the unsigned sense and
adds the NUW flag. The reasoning is done with
makeGuaranteedNoWrapRegion (thanks @nikic for point it out!).
Follow-ups will include adding NSW and extension to more
OverflowingBinaryOperators.
Reviewed By: nikic
Differential Revision: https://reviews.llvm.org/D142387
Jannik Silvanus [Thu, 26 Jan 2023 11:42:08 +0000 (12:42 +0100)]
[Coroutine] Fix typos in comment
Vlad Serebrennikov [Thu, 26 Jan 2023 11:07:30 +0000 (14:07 +0300)]
[clang] Add test for CWG1960
P1787: "CWG1960 (currently closed as NAD) is resolved by removing the rule in question (which is widely ignored by implementations and gives subtle interactions between using-declarations)."
Wording: "In a using-declarator that does not name a constructor, every declaration named shall be accessible."
Reviewed By: #clang-language-wg, erichkeane, shafik
Differential Revision: https://reviews.llvm.org/D142381
Luke Lau [Wed, 25 Jan 2023 16:19:11 +0000 (16:19 +0000)]
[RISCV][NFC] Rename narrowing patterns to use W suffix
To match up with the pseudo instruction names
Reviewed By: craig.topper
Differential Revision: https://reviews.llvm.org/D142553
Kerry McLaughlin [Thu, 26 Jan 2023 10:55:06 +0000 (10:55 +0000)]
[AArch64][SME2] Add multi-vector saturating rounding shift right intrinsics
Adds IR intrinsics for the following SME2 instructions:
- sqrshr (2 and 4 vector)
- uqrshr (2 and 4 vector)
- sqrshru (2 and 4 vector)
- sqrshrn (4 vector)
- uqrshrn (4 vector)
- sqrshrun (4 vector)
Also adds intrinsics for the following SVE2p1 instructions:
- sqrshrn (2 vector)
- uqrshrn (2 vector)
- sqrshrun (2 vector)
NOTE: These intrinsics are still in development and are subject to future changes.
Reviewed By: CarolineConcatto
Differential Revision: https://reviews.llvm.org/D142466
Akash Banerjee [Tue, 24 Jan 2023 11:29:50 +0000 (11:29 +0000)]
[MLIR][OpenMP] Add Lowering support for OpenMP Target Data, Exit Data and Enter Data directives
This patch adds Fortran Lowering support for the OpenMP Target Data, Target Exit Data and Target Enter Data constructs.
operation.
Differential Revision: https://reviews.llvm.org/D142357
Jay Foad [Wed, 25 Jan 2023 12:02:54 +0000 (12:02 +0000)]
[SPIRV] Fix build warnings
Fix these build warnings:
SPIRVBuiltins.cpp:1590:30: warning: 'getPrefTypeAlignment' is deprecated: use getPrefTypeAlign instead [-Wdeprecated-declarations]
SPIRVUtils.cpp:209:3: warning: default label in switch which covers all enumeration values [-Wcovered-switch-default]
Differential Revision: https://reviews.llvm.org/D142532
Kerry McLaughlin [Thu, 26 Jan 2023 09:56:39 +0000 (09:56 +0000)]
[AArch64][SME2] Add multi-vector saturating doubling multiply high intrinsics
Adds intrinsics for the following SME2 instructions:
- sqdmulh, 2 vector (single & multi)
- sqdmulh, 4 vector (single & multi)
NOTE: These intrinsics are still in development and are subject to future changes.
Reviewed By: david-arm
Differential Revision: https://reviews.llvm.org/D142368
Nikita Popov [Thu, 26 Jan 2023 09:10:01 +0000 (10:10 +0100)]
[SROA] Break typed pointer support
This removes typed pointer support in a prominent place in the
optimization pipeline, to ensure that any non-trivial consumers
of tip-of-tree LLVM are aware that this is no longer a supported
configuration.
Haojian Wu [Thu, 26 Jan 2023 09:46:51 +0000 (10:46 +0100)]
Revert "[clang] Build UsingType for elaborated type specifiers."
This reverts commit
e70ca7b35319a3621f9d9c6475926428f8c5c000 and the
followup patch "[clang] Fix the location of UsingTypeLoc"
(
ebbeb164c25a40cb6ba9c6b18dce5dcd06c0bb07).
The patch causes an incorrect lookup result:
```
namespace ns { struct Foo { };}
using ns::Foo;
void test() {
struct Foo {
} k; // the type of k refers to ns::Foo, rather than the local Foo!
}
```
Nikita Popov [Thu, 26 Jan 2023 09:18:42 +0000 (10:18 +0100)]
[Clang] Convert some tests to opaque pointers (NFC)
These are all tests that end up running SROA.
Samuel Parker [Thu, 26 Jan 2023 10:26:24 +0000 (10:26 +0000)]
[NFC][WebAssembly] Updated tests
Run update_llc_test_checks on a number of codegen tests.
Dana Jansens [Thu, 26 Jan 2023 08:53:59 +0000 (09:53 +0100)]
Allow getRawCommentForDecl to find comments in macros
The key part of getRawCommentForDecl() required to find a comment
is determining where to look for it. The location of the decl
itself is usually right, except when macros get involved. The
comment in the macro is stored in RawCommentList at the spelling
location of the decl, not at the place where the decl comes into
being as the macro is instantiated.
getDeclLocForCommentSearch() already contained to branches to try
handle comments inside macros, and we are able to replace them
and handle more cases as well, by returning the spelling location
of the decl's begin location. That is:
SourceMgr.getSpellingLoc(D->getBeginLoc())
Reviewed By: gribozavr2
Differential Revision: https://reviews.llvm.org/D142560
Nikolas Klauser [Wed, 25 Jan 2023 09:02:04 +0000 (10:02 +0100)]
[libc++][NFC] Rename multidimentional uninitialized algorithms
Reviewed By: Mordante, #libc
Spies: libcxx-commits
Differential Revision: https://reviews.llvm.org/D142526
Jay Foad [Wed, 25 Jan 2023 14:31:09 +0000 (14:31 +0000)]
[AMDGPU] sendmsg(MSG_ORDERED_PS_DONE) is not supported on GFX11
Differential Revision: https://reviews.llvm.org/D142537
Matthias Springer [Thu, 26 Jan 2023 08:15:49 +0000 (09:15 +0100)]
[mlir] GreedyPatternRewriteDriver: Keep track of surviving ops
This change adds `allErased` to the `applyOpPatternsAndFold(ArrayRef<Operation *>, ...)` overload. This overload now supports all functionality that is also supported by `applyOpPatternsAndFold(Operation *, ...)` and can be used as a replacement.
This change has no performance implications when `allErased = nullptr`.
The single-operation overload is removed in a subsequent NFC change.
Differential Revision: https://reviews.llvm.org/D141920
Sylvestre Ledru [Thu, 26 Jan 2023 08:18:00 +0000 (09:18 +0100)]
flang: fix a copy and paste mistake
Reviewed By: clementval
Differential Revision: https://reviews.llvm.org/D142488
Timm Bäder [Thu, 26 Jan 2023 06:50:14 +0000 (07:50 +0100)]
Re-apply "[clang][Interp] Fix left-/right-shifting more than sizeof(unsigned)"
Kazu Hirata [Thu, 26 Jan 2023 06:59:51 +0000 (22:59 -0800)]
[AST] Use llvm::bit_width (NFC)
serge-sans-paille [Wed, 25 Jan 2023 19:43:06 +0000 (20:43 +0100)]
[llvm] Replace array allocation pattern by SmallVector in ComputeMappedEditDistance
Differential Revision: https://reviews.llvm.org/D142574
serge-sans-paille [Tue, 24 Jan 2023 13:52:43 +0000 (14:52 +0100)]
[llvm] Do not zero out write_unsigned_impl internal buffer
Current implementation uselessly calls memset on its internal buffer
while it does not read the non overwritten part.
Differential Revision: https://reviews.llvm.org/D142464
Timm Bäder [Thu, 26 Jan 2023 06:49:10 +0000 (07:49 +0100)]
Revert "[clang][Interp] Fix left-/right-shifting more than sizeof(unsigned)"
This reverts commit
00e967f6c2d626d1913f5af5763beab7946978ce.
This breaks builders where long is only 32 bits, e.g.
https://lab.llvm.org/buildbot/#/builders/65/builds/7721
https://lab.llvm.org/buildbot/#/builders/245/builds/3899
Timm Bäder [Sat, 7 Jan 2023 07:10:48 +0000 (08:10 +0100)]
[clang][Interp][NFCI] Don't crash on void builtin functions
classifyPrim() runs into a llvm_unreachable() for those.
Craig Topper [Thu, 26 Jan 2023 06:08:02 +0000 (22:08 -0800)]
[RISCV] Add Zcd and Zcf to RISCVUsage.rst. NFC
Timm Bäder [Mon, 2 Jan 2023 14:11:27 +0000 (15:11 +0100)]
[clang][Interp] Fix left-/right-shifting more than sizeof(unsigned)
We were just casting to `unsigned` before, so that caused problems when
shifting more bits than `unsigned` has.
Differential Revision: https://reviews.llvm.org/D140845
Kazu Hirata [Thu, 26 Jan 2023 06:05:06 +0000 (22:05 -0800)]
[SystemZ] Use llvm::countl_zero and llvm::countr_zero (NFC)
isVectorConstantLegal calls findFirstSet and findLastSet, but we don't
rely on their ability to return std::numeric_limits<T>::max() on input
0.
This patch replaces those calls with calls to llvm::countl_zero and
llvm::countr_zero.
Due to an off-by-one error in the original code, the value of Upper
could change at bit N, where N is the index of the highest set bit in
SplatBitsZ, but the difference doesn't matter at the end. Without
this patch, Upper could have bit N set. With this patch, Upper never
has bit N set. Either way, both calls to tryValue have this bit set
because the argument is ORed with SplatBitsZ.
Timm Bäder [Sat, 31 Dec 2022 19:04:41 +0000 (20:04 +0100)]
[clang][Interp] Implement logical and/or operators
Differential Revision: https://reviews.llvm.org/D140809
Kazu Hirata [Thu, 26 Jan 2023 05:34:09 +0000 (21:34 -0800)]
Use llvm::Log2_32 and llvm::Log2_64 instead of llvm::findLastSet (NFC)
For a nonzero argument, llvm::findLastSet(x) is equivalent to
llvm::Log2_32(x) or llvm::Log2_64(x). None of the calls to
llvm::findLastSet in this patch relies on llvm::findLastSet's ability
to return std::numeric_limits<T>::max() on input 0.
Kazu Hirata [Thu, 26 Jan 2023 05:22:11 +0000 (21:22 -0800)]
[Support] Use llvm::countr_zero and llvm::Log2_64 in APInt.cpp (NFC)
partMSB and partLSB never get 0 as the argument. That is, we don't
rely on find{First,Last}Set's ability to return
std::numeric_limits<T>::max() on input 0.
This patch replaces partLSB and partMSB with llvm::countr_zero and
llvm::Log2_64, respectively.
FWIW, nobody in LLVM (except unit test MathExtrasTest.cpp) relies on
find{First,Last}Set's ability to return std::numeric_limits<T>::max()
on input 0.
Owen Anderson [Sat, 14 Jan 2023 04:40:04 +0000 (21:40 -0700)]
Resolve a FIXME in MachineCopyPropagation by allowig propagation to subregister uses.
Reviewed By: barannikov88
Differential Revision: https://reviews.llvm.org/D141747
Kazu Hirata [Thu, 26 Jan 2023 04:54:43 +0000 (20:54 -0800)]
[Mips] Simplify loadImmediate (NFC)
loadImmediate computes ShiftAmount in an unnecessarily complicated
manner. We just need to know the minimum right shift amount to bring
the immediate down to an unsigned 16-bit value, so
unsigned ShiftAmount = llvm::bit_width((uint64_t)ImmValue) - 16;
is sufficient. In other words, the following are all equivalent:
unsigned ShiftAmount = FirstSet - (15 - (LastSet - FirstSet));
unsigned ShiftAmount = llvm::countr_zero(IV) - (15 - (63 - llvm::countl_zero(IV) - llvm::countr_zero(IV)));
unsigned ShiftAmount = llvm::countr_zero(IV) - 15 + (63 - llvm::countl_zero(IV) - llvm::countr_zero(IV));
unsigned ShiftAmount = 48 - llvm::countl_zero(IV);
unsigned ShiftAmount = 64 - llvm::countl_zero(IV) - 16;
unsigned ShiftAmount = llvm::bit_width(IV) - 16;
where IV represents (uint64)ImmValue. I've also checked the
equivalence empirically up to 2u << 32.
Kazu Hirata [Thu, 26 Jan 2023 04:44:53 +0000 (20:44 -0800)]
[Mips] Simplify isShiftedUIntAtAnyPosition (NFC)
isShiftedUIntAtAnyPosition never gets zero as the argument because the
caller processes ImmValue satisfying isInt<16>(ImmValue), which
includes zero, long before it calls isShiftedUIntAtAnyPosition.
Given that the argument is always nonzero, findFirstSet is identical
to llvm::countr_zero. Also, x == x >> BitNum << BitNum is always
true, so we are left with:
isUInt<N>(x >> llvm::countr_zero(x))
Just in case the caller changes its behavior and starts passing zero
to us, we can protect the shift from undefined behavior "x << 64" by
adding "x &&".
Douglas Yung [Thu, 26 Jan 2023 03:11:08 +0000 (19:11 -0800)]
Revert "[llvm-cov] Look up object files using debuginfod"
This reverts commit
efbc8bb18eda63007216ad0cb5a8de04963eddd5.
This change is causing failures when detecting curl on several build bots:
- https://lab.llvm.org/buildbot/#/builders/247/builds/884
- https://lab.llvm.org/buildbot/#/builders/231/builds/7688
- https://lab.llvm.org/buildbot/#/builders/121/builds/27389
- https://lab.llvm.org/buildbot/#/builders/230/builds/8464
- https://lab.llvm.org/buildbot/#/builders/57/builds/24209
- https://lab.llvm.org/buildbot/#/builders/127/builds/42722
Douglas Yung [Thu, 26 Jan 2023 03:10:29 +0000 (19:10 -0800)]
Revert "Fix compilation failure in CoverageMapping.cpp"
This reverts commit
46013fc10a6879f4c9b4c9b9fbd43e4dc70f3c8b.
The original commit
efbc8bb18eda63007216ad0cb5a8de04963eddd5 is failing on several bots, so
reverting this follow-up commit as well as the original commit.
Matt Arsenault [Mon, 23 Jan 2023 20:16:44 +0000 (16:16 -0400)]
AMDGPU: Add fneg from integer tests
Matt Arsenault [Mon, 9 Jan 2023 17:17:38 +0000 (12:17 -0500)]
Verifier: Add checks for associated metadata
Also add missing assembler test for the valid cases.
Craig Topper [Thu, 26 Jan 2023 01:08:42 +0000 (17:08 -0800)]
[RISCV] Rename CS_ALU tablegen class to CA_ALU. NFC
The format this uses is CA. I think it may have once shared CS and
this didn't get renamed when that changed in D54302.
Andrew Young [Thu, 26 Jan 2023 01:01:11 +0000 (17:01 -0800)]
[MLIR] Fix python tests for LocationAttr
Follow up to https://reviews.llvm.org/D142522. I forgot to push the
test changes, was caught only in the post-merge builds.
NAKAMURA Takumi [Thu, 26 Jan 2023 00:52:38 +0000 (09:52 +0900)]
GVN.cpp: Suppress a warning in D141712 [-Wunused-variable]
Benjamin Kramer [Thu, 26 Jan 2023 00:49:59 +0000 (01:49 +0100)]
Paul Scoropan [Wed, 25 Jan 2023 23:23:42 +0000 (23:23 +0000)]
[flang] Add driver install directory to AIX toolchain program paths list.
flang-new encounters an issue where its unable to invoke itself because the
install directory was missing from the program paths on AIX.
Reviewed By: w2yehia
Differential Revision: https://reviews.llvm.org/D140628
David Blaikie [Thu, 26 Jan 2023 00:14:30 +0000 (00:14 +0000)]
Migrate away from C++20-deprecated POD type traits
Matt Arsenault [Wed, 25 Jan 2023 19:10:44 +0000 (15:10 -0400)]
AMDGPU: Remove redundant test
implicit-arg-v5-opt.ll already covers this with more cases.
Andrew Young [Thu, 19 Jan 2023 23:01:46 +0000 (15:01 -0800)]
[MLIR] Add LocationAttr to the Python API
This is a follow up to D142182, to expose LocationAttrs through Python.
Reviewed By: ftynse
Differential Revision: https://reviews.llvm.org/D142522
Shilei Tian [Thu, 26 Jan 2023 00:05:53 +0000 (19:05 -0500)]
[NFC] clang-format OpenMPOpt.cpp
Paul Kirth [Wed, 25 Jan 2023 23:55:02 +0000 (23:55 +0000)]
[clang-doc] Removed unused method in the Info class
Reviewed By: brettw
Differential Revision: https://reviews.llvm.org/D142577
Michael Jones [Tue, 24 Jan 2023 22:41:28 +0000 (14:41 -0800)]
[libc] add scanf pointer conversion
This patch adds the last conversion for scanf, %p. It is set up to match
the %p implementation in our printf.
Reviewed By: sivachandra, lntue
Differential Revision: https://reviews.llvm.org/D142510
Michael Jones [Tue, 24 Jan 2023 19:25:54 +0000 (11:25 -0800)]
[libc] add scanf current position conversion
To add the current position (%n) conversion, some reorganization needed
to be done. The "write a number to this pointer using the length
modifier" utilities and a couple other shared parsing functions have
been moved into converter_utils.h. This made implementing
current_pos_converter very simple.
Reviewed By: lntue
Differential Revision: https://reviews.llvm.org/D142495
Paul Kirth [Wed, 25 Jan 2023 21:34:01 +0000 (21:34 +0000)]
[llvm][NFC] Rename variables to match style guide in Local.cpp
Reviewed By: aeubanks
Differential Revision: https://reviews.llvm.org/D142579
Arthur Lafrance [Wed, 25 Jan 2023 21:02:16 +0000 (13:02 -0800)]
[mlir-translate] More specific error message for BlockAddress
BlockAddress is currently unimplemented in the LLVM dialect of MLIR;
when converting to LLVM dialect MLIR from LLVM IR, mlir-translate
currently terminates with an "unhandled constant" error message.
Instead, this message could be made more specific, to let the user know
that the specific issue is that BlockAddress is unimplemented in the
LLVM dialect.
Differential Revision: https://reviews.llvm.org/D142337
Daniel Thornburgh [Wed, 25 Jan 2023 22:34:45 +0000 (14:34 -0800)]
Fix compilation failure in CoverageMapping.cpp
Joseph Huber [Wed, 25 Jan 2023 22:20:16 +0000 (16:20 -0600)]
[CUDA] Fix output from `ptxas` being removes as a temporary file
Summary:
The logic here is to add the `.cubin` temporary file if we had to create
a new filename to handle it. Unfortuantely the logic was wrong because
we compare `const char *` values here. This logic seems to have been
wrong for some time, but was never noticed since we never used the
relocatable output.
Fixes https://github.com/llvm/llvm-project/issues/60301
Paul Kirth [Thu, 19 Jan 2023 20:47:43 +0000 (20:47 +0000)]
Reland [pgo] Avoid introducing relocations by using private alias
In many cases, we can use an alias to avoid a symbolic relocations,
instead of using the public, interposable symbol. When the instrumented
function is in a COMDAT, we can use a hidden alias, and still avoid
references to discarded sections.
Previous versions of this patch allowed the compiler to name the
generated alias, but that would only be valid when the functions were
local. Since the alias may be used across TUs we use a more
deterministic naming convention, and add a ".local" suffix to the alias
name just as we do for relative vtables aliases.
https://reviews.llvm.org/rG20894a478da224bdd69c91a22a5175b28bc08ed9
removed an incorrect assertion on Mach-O which caused assertion failures in LLD.
We prevent duplicate symbols under ThinLTO + PGO + CFI by disabling
alias generation when the target function has MD_type metadata used in
CFI.
Reviewed By: phosek
Differential Revision: https://reviews.llvm.org/D137982
Daniel Thornburgh [Mon, 24 Oct 2022 22:35:00 +0000 (15:35 -0700)]
[llvm-cov] Look up object files using debuginfod
Reviewed By: gulfem
Differential Revision: https://reviews.llvm.org/D136702
Paul Robinson [Wed, 25 Jan 2023 21:58:34 +0000 (13:58 -0800)]
[unittests] Use GTEST_SKIP() instead of return when appropriate
Basically NFC: A TEST/TEST_F/etc that bails out early (usually because
setup failed or some other runtime condition wasn't met) generally
should use GTEST_SKIP() to report its status correctly, unless it
takes steps to report another status (e.g., FAIL()).
I did see a handful of tests show up as SKIPPED after this change,
which is not unexpected. The status seemed appropriate in all the new
cases.
Michael Jones [Wed, 25 Jan 2023 00:08:41 +0000 (16:08 -0800)]
[libc] fix strtointmax tests
The strtointmax tests weren't running because they depend on the same
shared logic as strtol tests and the target for that was being defined
after the inttypes targets, causing them to be skipped. This patch moves
the inttypes test directory later in the order so that the shared tests
are defined before inttypes tests try to use them.
Reviewed By: sivachandra, lntue
Differential Revision: https://reviews.llvm.org/D142515
Florian Hahn [Wed, 25 Jan 2023 21:41:14 +0000 (21:41 +0000)]
[SCCP] Add additional tests for Add NUW/NSW flag inference.
Jon Chesterfield [Wed, 25 Jan 2023 21:38:49 +0000 (21:38 +0000)]
[amdgpuarch] Delete stray hsa #include line
Craig Topper [Wed, 25 Jan 2023 21:09:58 +0000 (13:09 -0800)]
[CodeGen] Split some functionality from DetectDeadLanes into its own class to be reused. NFC
This is an alternative to D140382.
This moves the first part of runOnce into a new class that can be
reused by D129735. This encapsulates the ownership of the worklist,
and VRegInfos into the new class. The code for updating the dead
lanes stays in the DetectDeadLanes class.
The new class is created on the stack during runOnMachineFunction
so all the data structures will be deleted after each run.
Previously we only cleared them after each run so the memory might
have stayed allocated across runs. Except for VRegInfo which was
always deleted. Hopefully this allocation change isn't a big deal.
Reviewed By: BeMg
Differential Revision: https://reviews.llvm.org/D141993
Rafael Auler [Wed, 25 Jan 2023 19:38:07 +0000 (11:38 -0800)]
[BOLT][NFC] Remove C-style out of bounds array ref
Old code breaks build with libstdc++ with assertions. Fix it.
Aaron Ballman [Wed, 25 Jan 2023 20:52:49 +0000 (15:52 -0500)]
Fix clang-tools-extra Sphinx build
This addresses the issue found in:
https://lab.llvm.org/buildbot/#/builders/115/builds/41077
Stanislav Mekhanoshin [Wed, 25 Jan 2023 19:58:58 +0000 (11:58 -0800)]
[AMDGPU] Remove predicates from real dot instructions. NFCI.
These are copied from pseudos automatically.
Differential Revision: https://reviews.llvm.org/D142575
Elizabeth Andrews [Tue, 10 Jan 2023 12:32:05 +0000 (04:32 -0800)]
[Clang] Fix compilation errors for unsupported __bf16 intrinsics
This patch uses existing deferred diagnostics framework to emit error
for unsupported type __bf16 in device code. Error is not emitted in
host code.
Differential Revision: https://reviews.llvm.org/D141375
rkayaith [Wed, 9 Nov 2022 04:23:28 +0000 (23:23 -0500)]
[mlir][Pass] Make PassManager default to op-agnostic
Currently `PassManager` defaults to being anchored on `builtin.module`.
Switching the default makes `PassManager` consistent with
`OpPassManager` and avoids the implicit dependency on `builtin.module`.
Specifying the anchor op type isn't strictly necessary when using
explicit nesting (existing pipelines will continue to work), but I've
updated most call sites to specify the anchor since it allows for better
error-checking during pipeline construction.
Reviewed By: rriddle
Differential Revision: https://reviews.llvm.org/D137731
Aaron Ballman [Wed, 25 Jan 2023 20:07:14 +0000 (15:07 -0500)]
Fix a dead link in the docs
The link used to go to a page on Chris Lattner's site that has been
gone since at least 2015. Replaced the link with one that has similar
information to what Chris had.
Fixes #60238
Denis Nikitin [Wed, 25 Jan 2023 19:58:38 +0000 (19:58 +0000)]
[clang-tidy] Fix segfault in bugprone-standalone-empty
The check incorrectly identified empty() method call in the template
class definition as a stand-alone function call.
This led to a crash because the checker did not expect empty() function
calls without arguments.
Fixes: https://github.com/llvm/llvm-project/issues/59487
Reviewed By: cjdb
Differential Revision: https://reviews.llvm.org/D142423
Joseph Huber [Wed, 25 Jan 2023 20:04:25 +0000 (14:04 -0600)]
[Libomptarget] Add correct relative path for the nexgen plugin
Summary:
I forgot that this file "borrowed" the source from the other file tree.
Fix that.
Joseph Huber [Wed, 25 Jan 2023 19:39:02 +0000 (13:39 -0600)]
[Libomptarget] Clean up CUDA plugin CMake files
Clean up this file after changing it in D142568.
Depends on D142568
Reviewed By: tianshilei1992
Differential Revision: https://reviews.llvm.org/D142573
Joseph Huber [Wed, 25 Jan 2023 19:19:37 +0000 (13:19 -0600)]
[Libomptarget] Remove find_package(CUDA) as it has been deprecated
Since D137724 and the LLVM 17 release we have updated to CMake version
3.20. This means that `find_package(CUDA)` is officially deprecated and
can be replaced with `find_package(CUDAToolkit)` instead. This patch
does this and also cleans up a bit of the CMake.
Reviewed By: tianshilei1992
Differential Revision: https://reviews.llvm.org/D142568
Joseph Huber [Wed, 25 Jan 2023 19:24:34 +0000 (13:24 -0600)]
[nvptx-arch] Remove `find_package(CUDA)` as it has been deprecated.
Since D137724 and the LLVM 17 release we have updated to CMake version
3.20. This means we can use `find_package(CUDAToolkit)` instead as the
legacy module is deprecated.
Reviewed By: tianshilei1992
Differential Revision: https://reviews.llvm.org/D142570
Guozhi Wei [Wed, 25 Jan 2023 19:45:01 +0000 (19:45 +0000)]
[GVN] Improve PRE on load instructions
This patch implements the enhancement proposed by
https://github.com/llvm/llvm-project/issues/59312.
Suppose we have following code
v0 = load %addr
br %LoadBB
LoadBB:
v1 = load %addr
...
PredBB:
...
br %cond, label %LoadBB, label %SuccBB
SuccBB:
v2 = load %addr
...
Instruction v1 in LoadBB is partially redundant, edge (PredBB, LoadBB) is a
critical edge. SuccBB is another successor of PredBB, it contains another load
v2 which is identical to v1. Current GVN splits the critical edge
(PredBB, LoadBB) and inserts a new load in it. A better method is move the load
of v2 into PredBB, then v1 can be changed to a PHI instruction.
If there are two or more similar predecessors, like the test case in the bug
entry, current GVN simply gives up because otherwise it needs to split multiple
critical edges. But we can move all loads in successor blocks into predecessors.
Differential Revision: https://reviews.llvm.org/D141712
Amir Ayupov [Wed, 25 Jan 2023 19:37:55 +0000 (11:37 -0800)]
[BOLT] Use LTO fuzzy name matching in function-order
Allow partial name matching wrt LTO suffixes in `function-order`
user-supplied function list, the same as permitted by profile matching.
Reviewed By: #bolt, rafauler
Differential Revision: https://reviews.llvm.org/D142269
Amir Ayupov [Wed, 25 Jan 2023 19:37:42 +0000 (11:37 -0800)]
[BOLT] Emit a warning about invalid entries in function-order list
Move individual warnings under verbosity >= 1, print out a warning with
aggregate number.
Reviewed By: #bolt, rafauler
Differential Revision: https://reviews.llvm.org/D142397
wren romano [Tue, 24 Jan 2023 21:23:52 +0000 (13:23 -0800)]
[mlir][sparse] (re)introducing getRankedTensorType/getMemrefType
The bulk of D142074 seems to have gotten overwritten due to some sort of merge conflict (afaict there's no record of it having been reverted intentionally). So this commit redoes those changes. In addition to the original changes, this commit also:
* moves the definition of `getRankedTensorType` (from `Transforms/CodegenUtils.h` to `IR/SparseTensor.h`), so that it can be used by `IR/SparseTensorDialect.cpp`.
* adds `getMemRefType` as another abbreviation.
Reviewed By: aartbik
Differential Revision: https://reviews.llvm.org/D142503
Ingo Müller [Wed, 25 Jan 2023 16:22:11 +0000 (16:22 +0000)]
[mlir][scf] Fix typo in comment in BufferizableOpInterfaceImpl.cpp (NFC).
Reviewed By: ingomueller-net
Differential Revision: https://reviews.llvm.org/D142554
Dave Lee [Wed, 11 Jan 2023 19:37:07 +0000 (11:37 -0800)]
[lldb][test] Replace use of p with expression (NFC)
In API tests, replace use of the `p` alias with the `expression` command.
To avoid conflating tests of the alias with tests of the expression command,
this patch canonicalizes to the use `expression`.
Differential Revision: https://reviews.llvm.org/D141539
Michael Liao [Wed, 25 Jan 2023 15:19:44 +0000 (15:19 +0000)]
[clang][CodeGen][NFC] Fix `llvm-else-after-return`
Rob Suderman [Wed, 25 Jan 2023 17:58:03 +0000 (17:58 +0000)]
[mlir][tosa] Update tosa.avg_pool2d for bit-exact TOSA behavior
The normalization component of average pool has a very specific
rounding behavior for compouting the division for floating
point values. Updated so that the bit-exact version is implemented.
Also includes a fix for computing the stride part of the average pool
operation.
Reviewed By: jpienaar
Differential Revision: https://reviews.llvm.org/D141339
Shafik Yaghmour [Wed, 25 Jan 2023 18:47:21 +0000 (10:47 -0800)]
[Clang] Add sanity check in Sema::getDestructorName to prevent nullptr dereference
Currently in Sema::getDestructorName we call SS.getScopeRep()->getPrefix() but
SS.getScopeRep() can return nullptr because LookupInNestedNameSpec(...) called a
little before can invalidate SS.
This fixes: https://github.com/llvm/llvm-project/issues/59446
Differential Revision: https://reviews.llvm.org/D140598
Matt Arsenault [Wed, 25 Jan 2023 14:04:26 +0000 (10:04 -0400)]
AMDGPU: Convert some tests to opaque pointers
Matt Arsenault [Wed, 25 Jan 2023 13:45:48 +0000 (09:45 -0400)]
LangRef: Fix extra spaces in udec_wrap description
Arthur Eubanks [Tue, 6 Dec 2022 17:20:08 +0000 (09:20 -0800)]
[llvm-c] Remove LLVMContextSetOpaquePointers
Reviewed By: nikic
Differential Revision: https://reviews.llvm.org/D139441
Florian Hahn [Wed, 25 Jan 2023 18:22:24 +0000 (18:22 +0000)]
Revert "Recommit "[SCCP] Use range info to prove AddInst has NUW flag.""
This reverts commit
366e1faa2fffd5b2284e25b09b6a26bcd2aca2b7.
It looks like this exposes another set of crashes on the buildbots.
Revert while I investigate.
Florian Hahn [Wed, 25 Jan 2023 18:07:24 +0000 (18:07 +0000)]
Recommit "[SCCP] Use range info to prove AddInst has NUW flag."
The recommitted version fixes a crash when one of the operands is a
constant other than a ConstantInt. Test for that case have been added
in
5b16cd97b8e1c273.
It splits off the new logic into a separate function because setting the
flags is quite different compared to the other cases handled in replaceSignedInst
which all require replacing an existing instruction. Instructions are
now refined before any replacements are done, which has the advantage
that we should have lattice values for all operands (fixing the crashes
and simplifies the logic) and also allows optimizing more cases where one
of the operands also gets replaced (see improvements in
@sge_with_sext_to_zext_conversion).
It also guards makeGuaranteedNoWrapRegion by `if (!Inst.hasNoUnsignedWrap())`
as discussed in the review.
Fixes #60280.
Fixes #60278.
Original message:
This patch updates SCCP to use the value ranges of AddInst operands to
try to prove the AddInst does not overflow in the unsigned sense and
adds the NUW flag. The reasoning is done with
makeGuaranteedNoWrapRegion (thanks @nikic for point it out!).
Follow-ups will include adding NSW and extension to more
OverflowingBinaryOperators.
Reviewed By: nikic
Differential Revision: https://reviews.llvm.org/D142387
Austin Kerbow [Mon, 23 Jan 2023 05:26:08 +0000 (21:26 -0800)]
[ScheduleDAG] Fix removing edges with weak deps
In SUnit::removePred edges are removed from the Preds and Succs lists before
updating the bookkeeping. This could result in incorrect values for
NumPreds/SuccsLeft and cause WeakPreds/SuccsLeft to underflow, since the
incorrect SDep will be used to update these values.
Reviewed By: rampitec
Differential Revision: https://reviews.llvm.org/D142325
She Dongrui [Wed, 25 Jan 2023 17:52:48 +0000 (09:52 -0800)]
Avoid removing useful loop metadata when stripping debug info
The stripDebugLocFromLoopID() may mistakenly remove useful metadata nodes
when they are operands of a child node, which also has DILocation operands.
It can be reproduced by the output of clang for code similar to the following:
for(int i = 0; i < n; i++)
x[i] = 10;
-strip-debug removes the child node of llvm.loop.vectorize.followup_all,
which contains llvm.loop.isvectorized and llvm.loop.unroll.count.
This patch fixes by checking all children nodes and only remove a metadata
node if all its children are DILocation.
Differential Revision: https://reviews.llvm.org/D141909
Advenam Tacet [Wed, 25 Jan 2023 18:01:24 +0000 (19:01 +0100)]
[ASan][libcxx] Annotating std::vector with all allocators
This revision is a part of a series of patches extending
AddressSanitizer C++ container overflow detection
capabilities by adding annotations, similar to those existing
in std::vector, to std::string and std::deque collections.
These changes allow ASan to detect cases when the instrumented
program accesses memory which is internally allocated by
the collection but is still not in-use (accesses before or
after the stored elements for std::deque, or between the size and
capacity bounds for std::string).
The motivation for the research and those changes was a bug,
found by Trail of Bits, in a real code where an out-of-bounds read
could happen as two strings were compared via a std::equals function
that took iter1_begin, iter1_end, iter2_begin iterators
(with a custom comparison function).
When object iter1 was longer than iter2, read out-of-bounds on iter2
could happen. Container sanitization would detect it.
In revision D132522, support for non-aligned memory buffers (sharing
first/last granule with other objects) was added, therefore the
check for standard allocator is not necessary anymore.
This patch removes the check in std::vector annotation member
function (__annotate_contiguous_container) to support
different allocators.
If you have any questions, please email:
- advenam.tacet@trailofbits.com
- disconnect3d@trailofbits.com
Reviewed By: #libc, #sanitizers, philnik, vitalybuka
Spies: EricWF, philnik, #sanitizers, libcxx-commits
Differential Revision: https://reviews.llvm.org/D136765
Alex Brachet [Wed, 25 Jan 2023 17:58:13 +0000 (17:58 +0000)]
[libc] Implement strcasestr
Differential Revision: https://reviews.llvm.org/D142518
Campbell Suter [Wed, 25 Jan 2023 14:12:19 +0000 (17:12 +0300)]
[RS4GC] Remove the hardcoded GC strategy names (v2)
Previously, RewriteStatepointsForGC had a hardcoded list of GC
strategies for which it would run, and using it with a custom strategy
required patching LLVM.
The logic for selecting the variables that are considered managed was
also hardcoded to use pointers in address space 1, rather than
delegating to GCStrategy::isGCManagedPointer.
This patch fixes both of these flaws: this pass now applies to all
functions whose GCStrategy returns true for useStatepoints, and checking
if a pointer is managed or not is also now done by the strategy.
One potentially questionable design decision in this change: the pass will
be enabled for all GC strategies that use statepoints. It seems unlikely
this would be a problem - consumers that don't use this pass probably
aren't adding it to the pass manager anyway - but if you had two different
GC strategies and only one wants this pass enabled then that'd need a new
flag in GCStrategy, which I can add if anyone thinks it's necessary.
This is an updated version of D140458, rebased to account for LLVM's
changes since D140504 (required by this patch) landed.
Reviewed By: dantrushin
Differential Revision: https://reviews.llvm.org/D141110
Denis Antrushin [Wed, 25 Jan 2023 17:17:06 +0000 (20:17 +0300)]
Fix assertion in GCStrategy.
It meant to check that `UseRS4GC` requires `UseStatepoints`.
Instead it always required `UseStatepoints` when `useRS4GC()`
was called.
Sergei Barannikov [Mon, 23 Jan 2023 13:52:20 +0000 (16:52 +0300)]
[llvm-ml] Disallow '{' and '}' at the start of a statement
'{', '}' and (conditionally) '*' were allowed at the start of a
statement. This behavior was copied from AsmParser, where they were
added to support Hexagon bundles (braces) and BFP memory instructions
(the star).
MASM dialect is x86-specific and does not allow these symbols to be
used at the beginning of an instruction.
Worth noting that '{' is a valid first character in AT&T syntax
(e.g. "{vex} vpmadd52huq ..."); MASM variant is to omit the braces.
Reviewed By: epastor
Differential Revision: https://reviews.llvm.org/D142321
Valentin Clement [Wed, 25 Jan 2023 17:38:34 +0000 (18:38 +0100)]
[flang] Fix element indexing in derived type default initialization
Derived type default initialization was not taking the step into
consideration.
```
module dt_init
type p1
integer :: a
end type
type, extends(p1) :: p2
integer :: b = 10
end type
contains
subroutine init_dt(z)
class(p1), intent(out) :: z(:)
select type(z)
type is (p2)
print*,z
end select
end subroutine
end module
program test
use dt_init
type(p2) :: t(6) = [ p2(1,2),p2(3,4),p2(5,6),p2(7,8),p2(9,10),p2(11,12) ]
print*,t
call init_dt(t(::2))
print*,t
end program
```
Without the fix, the three first elements are initialized
```
1 2 3 4 5 6 7 8 9 10 11 12
1 10 5 10 9 10
1 10 3 10 5 10 7 8 9 10 11 12
```
Where it should be element number 1,3,5
```
1 2 3 4 5 6 7 8 9 10 11 12
1 10 5 10 9 10
1 10 3 4 5 10 7 8 9 10 11 12
```
Reviewed By: klausler
Differential Revision: https://reviews.llvm.org/D142527
Jay Foad [Wed, 25 Jan 2023 11:29:12 +0000 (11:29 +0000)]
[BOLT] Use MCInstrDesc::operands() instead of OpInfo
operands() is the preferred accessor since D142213. OpInfo will be
removed in D142219.
Differential Revision: https://reviews.llvm.org/D142530
Michael Kruse [Wed, 25 Jan 2023 17:07:47 +0000 (11:07 -0600)]
[Polly] Remove failing Polly-ACC test.
The test has been failing since after last update of ISL (
a749e09e).