This patch implements codegen for __managed__ variable attribute for HIP.
Diagnostics will be added later.
Differential Revision: https://reviews.llvm.org/D94814
On z/OS, the following error message is not matched correctly in lit tests. This patch updates the CHECK expression to match the end period successfully.
```
EDC5129I No such file or directory.
```
Differential Revision: https://reviews.llvm.org/D94239
Simplify vperm2x128(concat(X,Y),concat(Z,W)) folding.
Use collectConcatOps / ISD::INSERT_SUBVECTOR to find the source subvectors instead of hardcoded immediate matching.
The existing code did not deal with atomic loads correctly. Such loads
are represented as MemoryDefs. Bail out on any MemoryAccess that is not
a MemoryUse.
Because we were not looking for the llvm.coro.id.async intrinsic in the
early coro pass which triggers follow-up passes we relied on the
llvm.coro.end intrinsic being present. This might not be the case in
functions that end in unreachable code.
Differential Revision: https://reviews.llvm.org/D95144
[libomptarget][devicertl] Drop templated atomic functions
The five __kmpc_atomic templates are instantiated a total of seven times.
This change replaces the template with explictly typed functions, which
have the same prototype for amdgcn and nvptx, and implements them with
the same code presently in use.
Rolls in the accepted but not yet landed D95085.
The unsigned long long type can be replaced with uint64_t when replacing
the cuda function. Until then, clang warns on casting a pointer to one to
a pointer to the other.
Reviewed By: tianshilei1992
Differential Revision: https://reviews.llvm.org/D95093
Iff we know we can get rid of the inversions in the new pattern,
we can thus get rid of the inversion in the old pattern,
this decreasing instruction count.
Note that we could position this transformation as just hoisting
of the `not` (still, iff y is freely negatible), but the test changes
show a number of regressions, so let's not do that.
Iff we know we can get rid of the inversions in the new pattern,
we can thus get rid of the inversion in the old pattern,
this decreasing instruction count.
I'm intentionally structuring it this way, so that the actual fold only
does the fold, and no legality/correctness checks, all of which must be
done by the caller. This allows for the fold code to be more compact
and more easily grokable.
Hoist the successor updating out of the code that deals with branch
weight updating, and hoist the 'has weights' check from the latter,
making code more consistent and easier to follow.
While we already ignore uncond branches, we could still potentially
end up with a conditional branches with identical destinations
due to the visitation order, or because we were called as an utility.
But if we have such a disguised uncond branch,
we still probably shouldn't deal with it here.
The case where BB ends with an unconditional branch,
and has a single predecessor w/ conditional branch
to BB and a single successor of BB is exactly the pattern
SpeculativelyExecuteBB() transform deals with.
(and in this case they both allow speculating only a single instruction)
Well, or FoldTwoEntryPHINode(), if the final block
has only those two predecessors.
Here, in FoldBranchToCommonDest(), only a weird subset of that
transform is supported, and it's glued on the side in a weird way.
In particular, it took me a bit to understand that the Cond
isn't actually a branch condition in that case, but just the value
we allow to speculate (otherwise it reads as a miscompile to me).
Additionally, this only supports for the speculated instruction
to be an ICmp.
So let's just unclutter FoldBranchToCommonDest(), and leave
this transform up to SpeculativelyExecuteBB(). As far as i can tell,
this shouldn't really impact optimization potential, but if it does,
improving SpeculativelyExecuteBB() will be more beneficial anyways.
Notably, this only affects a single test,
but EarlyCSE should have run beforehand in the pipeline,
and then FoldTwoEntryPHINode() would have caught it.
This reverts commit rL158392 / commit d33f4efbfd.
I may have given bad advice, and skipping sext_inreg when matching SSAT
patterns is not valid on it's own. It at least needs to sext_inreg the
input again, but as far as I can tell is still only valid based on
demanded bits. For the moment disable that part of the combine,
hopefully reimplementing it in the future more correctly.
Instead of using the type llvm::StringMapEntry<{stringified_value_type}>
use only the base class llvm::StringMapEntryBase and calculate the
offsets of the member variables manually. The approach with stringifying
the name of the value type is pretty fragile as it can easily break with
local and dependent types.
Differential Revision: https://reviews.llvm.org/D94431
lto::Config has a field to control whether the build is "freestanding"
(no builtins) or not, but it is not hooked up to the code actually
running the passes.
This patch adds support for the flag to both the code that runs
optimization with the new and old pass managers, by explicitly adding a
TargetLibraryInfo instance. If Freestanding is true, all library functions
are disabled.
Reviewed By: steven_wu
Differential Revision: https://reviews.llvm.org/D94630
This is a step towards allowing CDB behavior to being configurable.
Previously ClangdServer itself created the configs and installed them into
contexts. This was natural as it knows how to deal with resulting diagnostics.
However this prevents config being used in CDB, which must be created before
ClangdServer. So we extract the context provider (config loader) as a separate
object, which publishes diagnostics to a ClangdServer::Callbacks itself.
Now initialization looks like:
- First create the config::Provider
- Then create the ClangdLSPServer, passing config provider
- Next, create the context provider, passing config provider + diagnostic callbacks
- now create the CDB, passing context provider
- finally create ClangdServer, passing CDB, context provider, and diagnostic callbacks
Differential Revision: https://reviews.llvm.org/D95087
Without this patch the old index could be freed, but there still could be tries to access it via the function returned by `SwapIndex::indexedFiles()` call.
This leads to hard to reproduce clangd crashes at code completion.
This patch keeps the old index alive until the function returned by `SwapIndex::indexedFiles()` call is alive.
Reviewed By: sammccall
Differential Revision: https://reviews.llvm.org/D95206
combineX86ShufflesConstants/canonicalizeShuffleMaskWithHorizOp can both handle/earlyout shuffles with inputs of different widths, so delay widening as late as possible to make it easier to match constant folds etc.
The plan is to eventually move the widening inside combineX86ShuffleChain so that we don't create any new nodes unless we successfully combine the shuffles.
Walking the use list of a Constant (particularly, ConstantData)
is not scalable, since a given constant may be used by many
instructinos in many functions in many modules.
Differential Revision: https://reviews.llvm.org/D94713
The class `SymbolOccurrences` can store either a single `SourceRange`
in-place or multiple `SourceRanges` on the heap. In the latter case
the number of source ranges is stored in the internal representation
of the beginning `SourceLocation` of the in-place `SourceRange`
object.
This change gets rid of such hack by placing `SourceRange` in a union
which holds either a valid `SourceRange` or an `unsigned int` (a number
of ranges).
The change also adds `static_assert`s that check that `SourceRange` and
`SourceLocation` are trivially destructible (this is required for the
current patch and for D94237 which has already been committed).
Reviewed By: MarkMurrayARM, simon_tatham
Differential Revision: https://reviews.llvm.org/D94599
This type is used as an aggregate, i.e. it has no member functions.
Starting with C++20 types with deleted default constructors are not
aggregate types anymore which means that aggregate initialization will
not work for this class anymore. This leads to a compile error in
clang::AnalyzerOptions::getDiagOpts() for example.
Also set the boolean flags to false by default to avoid undefined
behavior. Previously this was prevented by deleting the default
constructor, now we explicitly initialize them.
Differential Revision: https://reviews.llvm.org/D92221
CMake 3.19 introduced the `presets`.
Quoting the documentation:
> `CMakePresets.json` may be checked into a version control system, and
> `CMakeUserPresets.json` **should NOT be checked in**.
We will ignore the `CMakeUserPresets.json` file if that is present
at the root of a subproject.
Reviewed By: dblaikie
Differential Revision: https://reviews.llvm.org/D93167
rGbe69e66b1cd8 added the fold, but DAGCombiner.visitVECTOR_SHUFFLE doesn't merge shuffles if the inner shuffle is a splat, so we need to bail.
The non-fast-horiz-ops paths see some minor regressions, we might be able to improve on this after lowering to target shuffles.
Fix PR48823
This replaces the isSaturatingConditional function with
LowerSaturatingConditional that directly returns a new SSAT or
USAT SDValue, instead of returning true and the components of it.
When adding an enum attribute to an AttributeList, avoid going
through an AttrBuilder and instead directly add the attribute to
the correct set. Going through AttrBuilder is expensive, because
it requires all string attributes to be reconstructed.
This can be further improved by inserting the attribute at the
right position and using the AttributeSetNode::getSorted() API.
This recovers the small compile-time regression from D94633.
Allow parsing generated mir with custom pseudo source value tokens.
Also rename pseudo source values to have more meaningful names.
Relands ba7dcd8542, which had memory leaks.
Differential Revision: https://reviews.llvm.org/D95215
This speeds up setLastUser enough to give a 5% to 10% speed up on
trivial invocations of opt and llc, as measured by:
perf stat -r 100 opt -S -o /dev/null -O3 /dev/null
perf stat -r 100 llc -march=amdgcn /dev/null -filetype null
Don't dump last use information unless -debug-pass=Details to avoid
printing lots of spam that will break some existing lit tests. Before
this patch, dumping last use information was broken anyway, because it
used InversedLastUser before it had been populated.
Differential Revision: https://reviews.llvm.org/D92309
I have removed an unnecessary assert in LoopVectorizationCostModel::getInstructionCost
that prevented a cost being calculated for select instructions when using
scalable vectors. In addition, I have changed AArch64TTIImpl::getCmpSelInstrCost
to only do special cost calculations for fixed width vectors and fall
back to the base version for scalable vectors.
I have added a simple cost model test for cmps and selects:
test/Analysis/CostModel/sve-cmpsel.ll
and some simple tests that show we vectorize loops with cmp and select:
test/Transforms/LoopVectorize/AArch64/sve-basic-vec.ll
Differential Revision: https://reviews.llvm.org/D95039
During instruction selection, there is an inconsistency in choosing
the initial soffset value. With certain early passes, this value is
getting modified and that brought additional fixup during
eliminateFrameIndex to work for all cases. This whole transformation
looks trivial and can be handled better.
This patch clearly defines the initial value for soffset and keeps it
unchanged before eliminateFrameIndex. The initial value must be zero
for MUBUF with a frame index. The non-frame index MUBUF forms that
use a raw offset from SP will have the stack register for soffset.
During frame elimination, the soffset remains zero for entry functions
with zero dynamic allocas and no callsites, or else is updated to the
appropriate frame/stack register.
Also, did some code clean up and made all asserts around soffset
stricter to match.
Reviewed By: scott.linder
Differential Revision: https://reviews.llvm.org/D95071
1. Op2 type in vrgather.vx should be XLEN instead of SEW
2. Add double type in vrgather-rv32 cases.
Reviewed By: craig.topper
Differential Revision: https://reviews.llvm.org/D95207
Yaxun (Sam) Liu