Artifact combiner is not able to access individual elements after using
LCMTy style merge/unmerge, extract and insert to change vector number of
elements (pad with undef or split to sub-vector instructions).
Use unmerge to individual elements instead and then merge elements into
requested types.
Change argument lowering for vectors and moreElementsVector to use
buildPadVectorWithUndefElements and buildDeleteTrailingVectorElements.
FewerElementsVector had a few helpers that had different behavior,
introduce new helper for most of the opcodes.
FewerElementsVector helper is more flexible since it can create leftover
instruction smaller then requested type (useful in case target wants to
avoid pad with undef and use fewer registers). If target does not want
leftover of different type it should call more elements first.
Some helpers were performing more elements first to have split without
leftover. Opcodes that used this helper use clampMaxNumElementsStrict
(does more elements first) in LegalizerInfo to avoid test changes.
Fixes failures caused by failing to combine artifacts created during
more/fewer elements vector.
Differential Revision: https://reviews.llvm.org/D114198
This is similar to what we have already done to some other tests.
See D102643, D102710, D102754.
Differential Revision: https://reviews.llvm.org/D116108
dwarfgen::Generator cannot be created if there is no asm backend for a
target. For example, if the default target triple is nvptx-nvidia-cuda,
some tests fail even after D98400, which added checks for most cases.
This patch extends the approach to the remaining cases.
Differential Revision: https://reviews.llvm.org/D116107
If Generator::create() returns an error, tests should fail gracefully
and report the cause, for example:
[ RUN ] DebugLineBasicFixture.ParserSkipsCorrectly
.../llvm/unittests/DebugInfo/DWARF/DWARFDebugLineTest.cpp:47: Failure
Value of: llvm::detail::TakeExpected(ExpectedGenerator)
Expected: succeeded
Actual: failed (no asm backend for target nvptx64-nvidia-cuda)
Differential Revision: https://reviews.llvm.org/D116106
Preserve the invariant that offset reported in the case of a
`PartialAlias` between `Loc1` and `Loc2`, is such that
`Loc1 + Offset = Loc2`, where `Loc1` and `Loc2` are the first and
the second argument, respectively, in alias queries.
Differential Revision: https://reviews.llvm.org/D115927
When reading location lists in dwo files the addresses cannot be
resolved, but that's not a problem.
Long term this probably should be fixed with a different API that
exposes location expressions without the need to resolve the address
ranges, since that's all the verifier (in its current state) requires.
(though the verifier should probably also eventually verify the address
ranges in location lists are a subset of the enclosing scope's address
range)
The directory returned from `cache_directory` may not be writable if the unit tests are run with limited permissions (e.g. if $HOME is not writable).
This fixes a unit test of the debuginfod client to set the cache path to the system temporary directory to avoid this issue.
Reviewed By: estewart08
Differential Revision: https://reviews.llvm.org/D115813
Similar versions of these already exist, this effectively just just
factors them out into STLExtras. I plan to use these in future patches.
Reviewed By: dblaikie
Differential Revision: https://reviews.llvm.org/D100672
This patch adds lowering from omp.sections and omp.section (simple lowering along with the nowait clause) to LLVM IR.
Tests for the same are also added.
Reviewed By: ftynse, kiranchandramohan
Differential Revision: https://reviews.llvm.org/D115030
This makes a bunch of these call sites independent of a follow-up change
I'm making to have getAsCString return Expected<const char*> for more
descriptive error messages so that the failures there can be
communicated up to DWARFVerifier (or other callers who want to provide
more verbose diagnostics) so DWARFVerifier doesn't have to re-implement
the string lookup logic and error checking.
The LangRef incorrectly says that if no exact match is found when
seeking alignment for a vector type, the largest vector type smaller
than the sought-after vector type. This is incorrect as vector types
require an exact match, else they fall back to reporting the natural
alignment.
The corrected rule was not added in its place, as rules for other types
(e.g., floating-point types) aren't documented.
A unit test was added to demonstrate this.
Reviewed By: aeubanks
Differential Revision: https://reviews.llvm.org/D112463
The vp.load and vp.gather intrinsics require the intrinsic return
type to determine the correct function signature. With opaque pointers,
it cannot be derived from the parameter pointee types.
Differential Revision: https://reviews.llvm.org/D115632
Make the reduction handling in OpenMPIRBuilder compatible with
opaque pointers by explicitly storing the element type in ReductionInfo,
and also passing it to the atomic reduction callback, as at least
the ones in the test need the type there.
This doesn't make things fully compatible yet, there are other
uses of element types in this class. I also left one
getPointerElementType() call in mlir, because I'm not familiar
with that area.
Differential Revison: https://reviews.llvm.org/D115638
Change getELFRelativeRelocationType() to return R_VE_RELATIVE
as a preparation of lld for VE.
Reviewed By: simoll
Differential Revision: https://reviews.llvm.org/D115592
This allows easier access to the induction descriptor from VPlan,
without needing to go through Legal. VPReductionPHIRecipe already
contains a RecurrenceDescriptor in a similar fashion.
Reviewed By: Ayal
Differential Revision: https://reviews.llvm.org/D115111
Most of `MemoryBuffer` interfaces expose a `RequiresNullTerminator` parameter that's being used to:
* determine how to open a file (`mmap` vs `open`),
* assert newly initialized buffer indeed has an implicit null terminator.
This patch adds the paramater to the `SmallVectorMemoryBuffer` constructors, meaning:
* null terminator can now be added to `SmallVector`s that didn't have one before,
* `SmallVectors` that had a null terminator before keep it even after the move.
In line with existing code, the new parameter is defaulted to `true`. This patch makes sure all calls to the `SmallVectorMemoryBuffer` constructor set it to `false` to preserve the current semantics.
Reviewed By: dexonsmith
Differential Revision: https://reviews.llvm.org/D115331
This prepares it for the regalloc work. Part of it is making model
evaluation accross 'development' and 'release' scenarios more reusable.
This patch:
- extends support to tensors of any shape (not just scalars, like we had
in the inliner -Oz case). While the tensor shape can be anything, we
assume row-major layout and expose the tensor as a buffer.
- exposes the NoInferenceModelRunner, which we use in the 'development'
mode to keep the evaluation code path consistent and simplify logging,
as we'll want to reuse it in the regalloc case.
Differential Revision: https://reviews.llvm.org/D115306
Add `SmallVectorImpl::truncate()`, a variant of `resize()` that cannot
increase the size.
- Compared to `resize()`, this has no code path for growing the
allocation and can be better optimized.
- Compared to `set_size()`, this formally calls destructors, and does
not skip any constructors.
- Compared to `pop_back_n()`, this takes the new desired size, which in
many contexts is more intuitive than the number of elements to remove.
The immediate motivation is to pair this with `resize_for_overwrite()`
to remove uses of `set_size()`, which can then be made private.
Differential Revision: https://reviews.llvm.org/D115383
MaterializationUnit::Interface holds the values that make up the interface
(for ORC's purposes) of a materialization unit: the symbol flags map and
initializer symbol.
Having a type for this will make functions that build materializer interfaces
more readable and maintainable.
Add the calculation of a score, which will be used during ML training. The
score qualifies the quality of a regalloc policy, and is independent of
what we train (currently, just eviction), or the regalloc algo itself.
We can then use scores to guide training (which happens offline), by
formulating a reward based on score variation - the goal being lowering
scores (currently, that reward is percentage reduction relative to
Greedy's heuristic)
Currently, we compute the score by factoring different instruction
counts (loads, stores, etc) with the machine basic block frequency,
regardless of the instructions' provenance - i.e. they could be due to
the regalloc policy or be introduced previously. This is different from
RAGreedy::reportStats, which accummulates the effects of the allocator
alone. We explored this alternative but found (at least currently) that
the more naive alternative introduced here produces better policies. We
do intend to consolidate the two, however, as we are actively
investigating improvements to our reward function, and will likely want
to re-explore scoring just the effects of the allocator.
In either case, we want to decouple score calculation from allocation
algorighm, as we currently evaluate it after a few more passes after
allocation (also, because score calculation should be reusable
regardless of allocation algorithm).
We intentionally accummulate counts independently because it facilitates
per-block reporting, which we found useful for debugging - for instance,
we can easily report the counts indepdently, and then cross-reference
with perf counter measurements.
Differential Revision: https://reviews.llvm.org/D115195
Internally `__Sddd` function-local parent symbols are used to solve ambiguities on symbols in
the same scope with the same mangled name.
Reviewed By: dblaikie
Differential Revision: https://reviews.llvm.org/D114309
Following the discussion in D112753, this moves the HTTPClient from Support to Debuginfod library so that tools depending on Support do not automatically depend on Curl as well. This also removes `HTTPClient::initialize()` and `HTTPClient::cleanup()` from `InitLLVM` so these steps should be implemented by user tools instead.
Reviewed By: phosek
Differential Revision: https://reviews.llvm.org/D115131
This adds a Debuginfod client library which queries servers specified by the `DEBUGINFOD_URLS` environment variable for the debuginfo, executable, or a specified source file associated with a given build id.
Reviewed By: dblaikie
Differential Revision: https://reviews.llvm.org/D112758
This adds a Debuginfod client library which queries servers specified by the `DEBUGINFOD_URLS` environment variable for the debuginfo, executable, or a specified source file associated with a given build id.
Reviewed By: dblaikie
Differential Revision: https://reviews.llvm.org/D112758
This adds a Debuginfod library containing the `fetchDebuginfo` function which queries servers specified by the `DEBUGINFOD_URLS` environment variable for the debuginfo, executable, or a specified source file associated with a given build id.
This diff was split out from D111252.
Reviewed By: dblaikie
Differential Revision: https://reviews.llvm.org/D112758
As mentioned in D106585, this causes non-determinism, which can also be
shown by this test case being flaky without this patch.
We were using the APSInt's bit width for hashing, but not for checking
for equality. APInt::isSameValue() does not check bit width.
Reviewed By: rnk
Differential Revision: https://reviews.llvm.org/D115054
We needed a stricter version of m_Not for D114462, but I wasn't
sure if that was going to be required anywhere else, so I didn't bother
to make that reusable.
It turns out we have one more existing simplification that needs
this (currently miscompiles):
https://alive2.llvm.org/ce/z/9-nTKi
And there's at least one more fold in that family that we could add.
Differential Revision: https://reviews.llvm.org/D114882
Deprecate LLVMAddAlias in favor of LLVMAddAlias2, which accepts a
value type and an address space. Previously these were extracted
from the pointer type.
Differential Revision: https://reviews.llvm.org/D114860
This allows JITDylibs to be removed from the ExecutionSession. Calling
ExecutionSession::removeJITDylib will disconnect the JITDylib from the
ExecutionSession and clear it (removing all trackers associated with it). The
JITDylib object will then be destroyed as soon as the last JITDylibSP pointing
at it is destroyed.
This patch implements a small HTTP client library consisting primarily of the `HTTPRequest`, `HTTPResponseHandler`, and `BufferedHTTPResponseHandler` classes. Unit tests of the `HTTPResponseHandler` and `BufferedHTTPResponseHandler` are included.
Reviewed By: dblaikie
Differential Revision: https://reviews.llvm.org/D112751
Petar Avramovic