The subview verifier in the rank-reduced case is plainly skipping verification
when the resulting type is a memref with empty affine map. This is generally incorrect.
Instead, form the actual expected rank-reduced MemRefType that takes into account the projections of 1's dimensions. Then, check the canonicalized expected rank-reduced type against the canonicalized candidate type.
Differential Revision: https://reviews.llvm.org/D95316
The header would include OrcJIT headers in OrcTargetProcess, which is not desired. All common declarations should be in OrcShared.
Reviewed By: lhames
Differential Revision: https://reviews.llvm.org/D95606
This patch created a new header file `target_interface.h` for declarations of all target dependent functions. All future targets can get things work by simply implementing all functions declared in the header and macros/data same as each `target_impl.h`.
Reviewed By: JonChesterfield
Differential Revision: https://reviews.llvm.org/D95300
In the past `-O1` was used when building NVPTX bitcode libraries. After
we switched to OpenMP, `-O1` was missing by mistake, leading to a huge performance
regression.
Reviewed By: JonChesterfield
Differential Revision: https://reviews.llvm.org/D95545
`[[clang::loader_uninitialized]]` is in macro `SHARED` but it doesn't
work for array like `parallelLevel`, so the variable will be zero initialized.
There is also a similar issue for `omptarget_nvptx_device_State` which is in
global address space. Its c'tor is also generated, which was not in the past when
building the `deviceRTLs` with CUDA. In this patch, we added the attribute to
the two variables explicitly.
Reviewed By: jdoerfert
Differential Revision: https://reviews.llvm.org/D95550
The MVE VLD2/4 and VST2/4 instructions require the pointer to be aligned
to at least the size of the element type. This adds a check for that
into the ARM lowerInterleavedStore and lowerInterleavedLoad functions,
not creating the intrinsics if they are invalid for the alignment of
the load/store.
Unfortunately this is one of those bug fixes that does effect some
useful codegen, as we were able to sometimes do some nice lowering of
q15 types. But they can cause problem with low aligned pointers.
Differential Revision: https://reviews.llvm.org/D95319
This revision adds a layer of SFINAE to the composable codegen strategy so it does
not have to require statically defined ops but instead can also be used with OpInterfaces, Operation* and an op name string.
A linalg.matmul_i8_i8_i32 is added to the .tc spec to demonstrate how all this works end to end.
Differential Revision: https://reviews.llvm.org/D95600
The layout of the stack frame for SVE means that using the frame pointer
rather than the stack pointer for an access to an SVE stack object
removes the need for an additional add to jump over the non-SVE objects.
Likewise the opposite is true for non-SVE stack objects.
This patch allows for the former to be done by having HasFP return true
in the presence of both SVE and non-SVE stack objects, and also fixes a
minor issue whereby the later would not be done for certain offsets.
Link error occurred when time profiling in libomp is enabled by default
because `libomp` is assumed to be a C library but the dependence on
`libLLVMSupport` for profiling is a C++ library. Currently the issue blocks all
OpenMP tests in Phabricator.
This patch set a new CMake option `OPENMP_ENABLE_LIBOMP_PROFILING` to
enable/disable the feature. By default it is disabled. Note that once time
profiling is enabled for `libomp`, it becomes a C++ library.
Reviewed By: jdoerfert
Differential Revision: https://reviews.llvm.org/D95585
Unlike VPERMILPS, VPERMILPD can have non-repeating masks in each 128-bit subvector, we weren't accounting for this when folding vperm2f128(vpermilpd(x,c),vpermilpd(y,c)) -> vpermilpd(vperm2f128(x,y),c).
I'm intending to add support for this but wanted to get a minimal fix in first for merging into 12.xx.
Fixes PR48908
SimplifyCFG is an utility pass, and the fact that it does not
preserve DomTree's, forces it's users to somehow workaround that,
likely by not preserving DomTrees's themselves.
Indeed, simplifycfg pass didn't know how to preserve dominator tree,
it took me just under a month (starting with e1133179587dd895962a2fe4d6eb0cb1e63b5ee2)
do rectify that, now it fully knows how to,
there's likely some problems with that still,
but i've dealt with everything i can spot so far.
I think we now can flip the switch.
Note that this is functionally an NFC change,
since this doesn't change the users to pass in the DomTree,
that is a separate question.
Reviewed By: kuhar, nikic
Differential Revision: https://reviews.llvm.org/D94827
This revision improves the usage of the codegen strategy by adding a few flags that
make it easier to control for the CLI.
Usage of ModuleOp is replaced by FuncOp as this created issues in multi-threaded mode.
A simple benchmarking capability is added for linalg.matmul as well as linalg.matmul_column_major.
This latter op is also added to linalg.
Now obsolete linalg integration tests that also take too long are deleted.
Correctness checks are still missing at this point.
Differential revision: https://reviews.llvm.org/D95531
Fixes a few small issues in the docs. It seems one of the examples was missing
the expected MLIR output due to a copy-paste typo.
Reviewed By: nicolasvasilache
Differential Revision: https://reviews.llvm.org/D95599
Look throught G_PTRTOINT and G_PTR_ADD nodes when looking for constant
offset for buffer stores. This also helps with merging of these instructions
later on.
Differential Revision: https://reviews.llvm.org/D95242
If the APInt returned by BuildVectorSDNode::isConstantSplat() is narrower than
64 bits, the result produced by XXSPLTI32DX is incorrect. The result returned
by the function appears to be incorrect and we'll investigate/fix it in a
follow-up commit. However, since this causes miscompiles, we must
temporarily disable emitting this instruction for such values.
This patch adds support for the full range of vector int-to-float,
float-to-int, and float-to-float conversions on legal types.
Many conversions are supported natively in RVV so are lowered with
patterns. These include conversions between (element) types of the same
size, and those that are half/double the size of the input. When
conversions take place between types that are less than half or more
than double the size we must lower them using sequences of instructions
which go via intermediate types.
Reviewed By: craig.topper
Differential Revision: https://reviews.llvm.org/D95447
A simple refactoring patch which let us use `DataExtractor::getSLEB128` rather than using a lambda function.
Differential Revision: https://reviews.llvm.org/D95158
This patch makes all macros forwarding to `PARSE_OPTION_WITH_MARSHALLING` and `GENERATE_OPTION_WITH_MARSHALLING` variadic.
Sice we will be splitting up all CompilerInvocation parts, this will allow us to avoid a lot of boilerplate code.
The local macros prefix forwarded arguments with local variables required by the main macros. The `{THIS,NO}_PREFIX` macros make it possible for forwarding macros in member functions (`parseSimpleArgs`, `generateCC1CommandLine`) to prefix keypaths with `this->`. (Some build bots seem to require that.)
Reviewed By: dexonsmith
Differential Revision: https://reviews.llvm.org/D95532
https://bugs.llvm.org/show_bug.cgi?id=48232
When PrologEpilogInserter writes callee-saved registers to the stack, LR is not reloaded but is instead loaded directly into PC.
This was not taken into account when determining if each callee-saved register was liveout for the block.
When frame elimination inserts virtual registers, and the register scavenger tries to scavenge LR, it considers it liveout and tries to spill again.
However there is no emergency spill slot to use, and it fails with an error:
fatal error: error in backend: Error while trying to spill LR from class GPR: Cannot scavenge register without an emergency spill slot!
This patch pervents any callee-saved registers which are not reloaded (including LR) from being marked liveout.
They are therefore available to scavenge without requiring an extra spill.
If an initial value is given for a bitfield that does not fit in the
bitfield, the value should be truncated. Constant folding for
expressions did not account for this truncation in the case of union
member functions, despite a warning being emitted. In some contexts,
evaluation of expressions was not enabled unless C++11, ROPI or RWPI
was enabled.
Differential Revision: https://reviews.llvm.org/D93101
Currently we don't allow the following definition:
```
Sections:
- Type: SectionHeaderTable
- Name: .foo
Type: SHT_PROGBITS
```
We report an error: "SectionHeaderTable can't be empty. Use 'NoHeaders' key to drop the section header table".
It was implemented in this way earlier, when `SectionHeaderTable`
was a dedicated key outside of the `Sections` list. And we did not
allow to select where the table is written.
Currently it makes sense to allow it, because a user might
want to place the default section header table at an arbitrary position,
e.g. before other sections. In this case it is not convenient and error prone
to require specifying all sections:
```
Sections:
- Type: SectionHeaderTable
Sections:
- Name: .foo
- Name: .strtab
- Name: .shstrtab
- Name: .foo
Type: SHT_PROGBITS
```
This patch allows empty SectionHeaderTable definitions.
Differential revision: https://reviews.llvm.org/D95341
Before the patch it was possible to trigger a constant bus
violation when folding immediates into a shrunk instruction.
The patch adds a check to enforce the legality of the new operand.
Differential Revision: https://reviews.llvm.org/D95527
NativeEnumInjectedSources.h needs PDBFile but relies on a
forward declaration of PDBFile in InjectedSourceStream.h.
This patch adds a forward declaration right in
NativeEnumInjectedSources.h.
While we are at it, this patch removes the one in
InjectedSourceStream.h, where it is unnecessary.
This change brings up support of context-sensitive profiles in the format of extended binary. Existing sample profile reader/writer/merger code is being tweaked to reflect the fact of bracketed input contexts, like (`[...]`). The paired brackets are also needed in extbinary profiles because we don't yet have an otherwise good way to tell calling contexts apart from regular function names since the context delimiter `@` can somehow serve as a part of the C++ mangled names.
Reviewed By: wmi, wenlei
Differential Revision: https://reviews.llvm.org/D95547
In d2927f786e877410d90c1e6f0e0c7d99524529c5, I added patterns
to remove (and X, 31) from sllw/srlw/sraw shift amounts.
There is code in SelectionDAGISel.cpp that knows to use
computeKnownBits to fill in bits of the mask that were removed
by SimplifyDemandedBits based on bits being known zero.
The non-W shift patterns use immbottomxlenset which allows the
mask to have more than log2(xlen) trailing ones, but doesn't
have a call to computeKnownBits to fill in bits of the mask that may
have been cleared by SimplifyDemandedBits.
This patch copies code from X86 to handle more than log2(xlen)
bottom bits set and uses computeKnownBits to fill in missing bits
before counting.
Reviewed By: luismarques
Differential Revision: https://reviews.llvm.org/D95422
We need at least 252 UniqAttributes now, which will soon overflow.
Actually with downstream backends we can easily use up the last few values.
So bump to uint16_t.
This change fixes two issues with building LLVM on Haiku. The first issue is
that LLVM requires wait4(), which on Haiku is hidden behind the _BSD_SOURCE
feature flag when using the --std=c++14 flag. Additionally, the wait4()
function is only available in libbsd.so, so this is now a dependency.
The other fix is that Haiku does not have the (non-standard) rusage.maxrss
member, so by default the used memory info will be set to 0 on this platform.
Reviewed By: sepavloff
Differential Revision: https://reviews.llvm.org/D87920
Patch by Niels Sascha Reedijk.
