The hoistRedundantVectorTransfers function does not verification of loop
bounds when hoisting vector transfers. This is not safe in general,
since it is possible that the loop will have zero trip count. This PR
uses ValueBounds to verify that the lower bound is less than the upper
bound of the loop before hoisting. Trip count verification is currently
behind an option `verifyNonZeroTrip`, which is false by default.
Zero trip count loops can arise in GPU code generation, where a loop
bound can be dependent on a thread id. If not all threads execute the
loop body, then hoisting out of the loop can cause these threads to
execute the transfers when they are not supposed to.
---------
Signed-off-by: Max Dawkins <max.dawkins@gmail.com>
In the insert_slice bufferization interface implementation, the
destination tensor is not considered read if the full tensor is
overwritten by the slice. This PR adds the same check for
tensor.parallel_insert_slice.
Adds two new StaticValueUtils:
- `isAllConstantIntValue` checks if an array of `OpFoldResult` are all
equal to a passed `int64_t` value.
- `areConstantIntValues` checks if an array of `OpFoldResult` are all
equal to a passed array of `int64_t` values.
fixes https://github.com/llvm/llvm-project/issues/112435
---------
Signed-off-by: Max Dawkins <max.dawkins@gmail.com>
Enabled initial support for max safe distance in DataWithEVL mode. If
max safe distance is required, need to emit special code:
CMP = icmp ult AVL, MAX_SAFE_DISTANCE
SAFE_AVL = select CMP, AVL, MAX_SAFE_DISTANCE
EVL = call i32 @llvm.experimental.get.vector.length(i64 SAFE_AVL)
while vectorize the loop in DataWithEVL tail folding mode.
Reviewers: fhahn
Reviewed By: fhahn
Pull Request: https://github.com/llvm/llvm-project/pull/102897
Enables initial non-power-of-2 support (but still requires number of
elements, forming whole registers) for reductions.
Enables extra vectorization for
MultiSource/Benchmarks/7zip/7zip-benchmark, CINT2006/464.h264ref and
CFP2017rate/526.blender_r (checked for SSE2)
Reviewers: RKSimon
Reviewed By: RKSimon
Pull Request: https://github.com/llvm/llvm-project/pull/112361
Adds a canonicalization pattern for scf.forall that replaces constant
induction variables with a constant index. There is a similar
canonicalization that completely removes constant induction variables
from the loop, but that pattern does not apply on foralls with mappings,
so this one is necessary for those cases.
---------
Signed-off-by: Max Dawkins <max.dawkins@gmail.com>
Fix 'sed' spacing to ensure compatibility with all platforms.
Original failure:
https://lab.llvm.org/buildbot/#/builders/190/builds/7903
```
RUN: at line 33: sed -E '/^__OBJC_\$_CATEGORY_MyBaseClass_\$_Category01:/ { n; s/^[ \t]*\.quad[ \t]+l_OBJC_CLASS_NAME_$/\t.quad\tL_OBJC_IMAGE_INFO+3/ }' merge_cat_minimal.s > merge_cat_minimal_bad_name.s
+ sed -E '/^__OBJC_\$_CATEGORY_MyBaseClass_\$_Category01:/ { n; s/^[ \t]*\.quad[ \t]+l_OBJC_CLASS_NAME_$/\t.quad\tL_OBJC_IMAGE_INFO+3/ }' merge_cat_minimal.s
sed: 1: "/^__OBJC_\$_CATEGORY_My ...": bad flag in substitute command: '}'
```
Encountered several testcase failures when running `ninja check-all`. It
was due to the full path name were shown in the error message instead of
the binary name, and therefore causing the check string mismatch.
The machine was running CentOS 9 with binfmt_misc setup that uses
qemu-aarch64 (8.1.2). Built and ran the unittest as aarch64 host
(through qemu user).
Co-authored-by: h2h <h2h@meta.com>
If any iterator in the concatenation returns by value, the result must
return by value otherwise it'll produce dangling references.
(some context that may or may not be relevant to this part of the code
may be in
981ce8fa15
)
An alternative to #112441
This patch enhances the robustness of lld's Objective-C category
merging. Currently, the category merger assumes it can fully parse and
understand the format of all categories in the input, triggering an
assert if any invalid category data is encountered.
This will end up causing asserts in certain rare corner cases that are
difficult to reproduce in small test cases. The proposed changes modify
the behavior so that if invalid category data is detected, category
merging is skipped for that specific class and all other categories
sharing the same base class. This approach allows the linker to continue
processing other categories without failing entirely due to a single
problematic input.
We also add a LIT test to where we corrupt category data and check that
category merging for that class was skipped but the link was successful.
Moves the existing `llvm_unreachable` statement to the bottom of the
function and changes the case statement to deliberately fall through to
it.
Build break was introduced by #111203
It was not caught by the builders as they use Visual Studio 2019,
whereas this warning only appears in 2022.
---------
Co-authored-by: Matheus Izvekov <mizvekov@gmail.com>
The -enable-memprof-indirect-call-support meant to guard the recently
added memprof ICP support was not used in enough places. Specifically,
it was not checked in mayHaveMemprofSummary, which is called from the
ThinLTO backend applyImports. This led to failures when checking the
callsite records, as we incorrectly expected records for indirect calls.
Fix the option to be checked in all necessary locations, and add
testing.
This is one of the many PRs to fix errors with LLVM_ENABLE_WERROR=on.
Built by GCC 11.
Fix warnings:
llvm-project/llvm/lib/CodeGen/AsmPrinter/AsmPrinter.cpp: In member
function ‘void llvm::AsmPrinter::emitJumpTableSizesSection(const
llvm::MachineJumpTableInfo*, const llvm::Function&) const’:
llvm-project/llvm/lib/CodeGen/AsmPrinter/AsmPrinter.cpp:2852:31: error:
enumerated and non-enumerated type in conditional expression
[-Werror=extra]
2852 | int Flags = F.hasComdat() ? ELF::SHF_GROUP : 0;
| ~~~~~~~~~~~~~~^~~~~~~~~~~~~~~~~~~~
This is one of the many PRs to fix errors with LLVM_ENABLE_WERROR=on.
Built by GCC 11.
Fix warning
llvm-project/clang/unittests/Basic/DiagnosticTest.cpp:19:6: error: ‘void
clang::DiagnosticsTestHelper(clang::DiagnosticsEngine&)’ has not been
declared within ‘clang’ [-Werror]
19 | void clang::DiagnosticsTestHelper(DiagnosticsEngine &diag) {
| ^~~~~
In file included from
llvm-project/clang/unittests/Basic/DiagnosticTest.cpp:9:
llvm-project/clang/include/clang/Basic/Diagnostic.h:567:15: note: only
here as a ‘friend’
567 | friend void DiagnosticsTestHelper(DiagnosticsEngine &);
This is one of the many PRs to fix errors with LLVM_ENABLE_WERROR=on.
Built by GCC 11.
Fix warnings:
llvm-project/clang/include/clang-c/Index.h:2983:3: error: C++ style
comments are not allowed in ISO C90 [-Werror]
2983 | // HLSL Types
This is one of the many PRs to fix errors with LLVM_ENABLE_WERROR=on.
Built by GCC 11.
Fix warnining:
llvm/unittests/Object/GOFFObjectFileTest.cpp:511:17: error: overflow in
conversion from ‘int’ to ‘char’ changes value from ‘240’ to
‘'\37777777760'’ [-Werror=overflow]
511 | GOFFData[1] = 0xF0;
Currently, libc++'s `bitset`, `forward_list`, and `list` have
non-conforming member typedef name `base`. The typedef is private, but
can cause ambiguity in name lookup.
Some other classes in libc++ that are either implementation details or
not precisely specified by the standard also have member typdef `base`.
I think this can still be conforming.
Follows up #80706 and #111127.
Summary:
This caused an error when copying a file of the same name when multiple
architectures needed the file. The buffer identifier which we use for
the name in `-save-temps` mode would be empty and create in invalid
filename. Copy this correctly now.
Fixes loop comparison condition in the vectorizer.
As that logic is used specifically for vectorising `tensor.extract`, I
also added a test that violates the assumptions made inside
`getTrailingNonUnitLoopDimIdx`, namely that Linalg loops are non-empty.
Vectorizer pre-conditions will capture that much earlier making sure
that `getTrailingNonUnitLoopDimIdx` is only run when all the assumptions
are actually met.
Thank you for pointing this out, @pfusik !
The ATT assembler dialect on SystemZ seems to have been taken from the
existing ATT/Intel code. However, on SystemZ, ATT does not hold any
meaning. In reality, we are splitting the difference between GNU Asm
syntax and HLASM Asm syntax, so it makes sense to rename ATT to GNU
instead.
Co-authored-by: Tony Tao <tonytao@ca.ibm.com>
That macro was removed incorrectly from the clang-format file because it
had a typo in its name. However, the macro with the right name is still
being used in the library (sadly, in a single place).
In IEEE 754 and C standards, when calling `frexp` with Inf/Nan inputs,
the exponent result is unspecified. In this case, FreeBSD libc and musl
just passthrough `exp`, while glibc, FreeBSD libm set exp = 0, and MSVC
set exp = -1.
By default, LLVM libc will passthrough `exp` just as FreeBSD libc and
musl, but we also allow users to explicitly choose the return exp value
in this case for compatibility with other libc.
Notice that, gcc did generate passthrough `exp` for `frexp(NaN/Inf,
exp)`: https://godbolt.org/z/sM8fEej4E
