* Mostly imported from experimental repo as-is with cosmetic changes.
* Temporarily left out emission code (for building ops at runtime) to keep review size down.
* Documentation and lit tests added fresh.
* Sample op library that represents current Linalg named ops included.
Differential Revision: https://reviews.llvm.org/D97995
The runtimes build uses variables set by add_lit_testsuite to collect
testsuites from all the runtimes.
Differential Revision: https://reviews.llvm.org/D97913
This patch exposes the getter and setter methods for the command
interpreter `print_errors` run option.
rdar://74816984
Differential Revision: https://reviews.llvm.org/D98001
Signed-off-by: Med Ismail Bennani <medismail.bennani@gmail.com>
This patch introduces a new interpreter setting to prevent LLDB from
re-executing the previous command when passing an empty command.
This can be very useful when performing actions that requires a long
time to complete.
To preserve the original behaviour, the setting defaults to `true`.
rdar://74983516
Differential Revision: https://reviews.llvm.org/D97999
Signed-off-by: Med Ismail Bennani <medismail.bennani@gmail.com>
[llvm-exegesis] Disable the LBR check on AMD
https://bugs.llvm.org/show_bug.cgi?id=48918
The bug reported a hang (or very very slow runtime) on a Zen2. Unfortunately, we don't have the hardware right now to debug it and I was not able to reproduce the bug on a HSW.
Theory we've got is that the lbr-checking code could be confused on AMD.
Differential Revision: https://reviews.llvm.org/D97504
New change:
- Surround usages of x86 helper in llvm-exegesis/X86/Target.cpp with ifdef
- Fix bug which caused the caller of getVendorSignature to not have a copy of EAX that it expected.
The last two operands to a gc.relocate represent indices into the associated gc.statepoint's gc bundle list. (Effectively, gc.relocates are projections from the gc.statepoints multiple return values.)
We can use this to recognize when two gc.relocates are equivalent (and can be CSEd), even when the indices are non-equal. This is particular useful when considering a chain of multiple statepoints as it lets us eliminate all duplicate gc.relocates in a single pass.
Differential Revision: https://reviews.llvm.org/D97974
(Note: Part of the reviewed change was split and landed as f352463a)
For some reason, we had been marking gc.relocates as reading memory. There's no known reason for this, and I suspect it to be a legacy of very early implementation conservatism. gc.relocate and gc.result are simply projections of the return values from the associated statepoint. Note that the LangRef has always declared them readnone.
The EarlyCSE change is simply moving the special casing from readonly to readnone handling.
As noted by the test diffs, this does allow some additional CSE when relocates are separated by stores, but since we generate gc.relocates in batches, this is unlikely to help anything in practice.
This was reviewed as part of https://reviews.llvm.org/D97974, but split at reviewer request before landing. The motivation is to enable the GVN changes in that patch.
If we have a value live over a call which is used for deopt at the call, we know that the value must be a base pointer. We can avoid potentially inserting IR to materialize a base for this value.
In it's current form, this is mostly a compile time optimization. Building the base pointer graph (and then optimizing it away again) is a relatively expensive operation. We also sometimes end up with better codegen in practice - due to failures in optimizing away the inserted base pointer propogation - but those are optimization bugs we're fixing concurrently.
The alternative to this would be to extend the base pointer inference with the ability to generally reuse multiple-base input instructions (phis and selects). That's somewhat invasive and complicated, so we're defering it a bit longer.
Differential Revision: https://reviews.llvm.org/D97885
This patch adds support for the default AltiVec ABI for AIX.
Vector registers 20 through 31 are marked as reserved and cannot
be used in the default ABI. This patch adds handling for this case
and also remove the default AltiVec ABI errors.
Reviewed By: sfertile
Differential Revision: https://reviews.llvm.org/D96351
If cross testing on Windows via WSL (at least with WSL 1), the Windows
executables can't be executed if they are in WSL specific directories
(like /tmp).
Differential Revision: https://reviews.llvm.org/D98028
Add ifdefs to the test reference tables for cases where paths are
interpreted differently (paths that contain a root name).
Fix test assumptions regarding has_root_name() and is_absolute() and
add logic to verify the results of is_absolute() for the test cases in
the table.
Also add a testcase for the path "//net/", which seemed like an
omission.
Differential Revision: https://reviews.llvm.org/D89943
Added the following attributes to all LLVM syntax files:
* allocsize
* cold
* convergent
* dereferenceable_or_null
* hot
* inaccessiblemem_or_argmemonly
* inaccessiblememonly
* inalloca
* jumptable
* nocallback
* nocf_check
* noduplicate
* nofree
* nomerge
* noprofile
* nosync
* null_pointer_is_valid
* optforfuzzing
* preallocated
* safestack
* sanitize_hwaddress
* sanitize_memtag
* shadowcallstack
* speculative_load_hardening
* swifterror
* syncscope
* tailcc
* willreturn
I generated that list by comparing:
* Attributes.inc (generated from Attributes.td), and
* the Vim syntax file: llvm/utils/vim/syntax/llvm.vim
My original intention was to focus on the Vim syntax file. Since other
syntax files are also out-of-date, I added these attributes (if missing)
to other files as well. Note that in the other sytnax files (i.e. for
Emacs, VScode and Kate), there will be other attributes missing too.
I've also sorted all attributes alphabetically. Otherwise it's really
hard to automate adding new attributes. And I think that it was the
original intent to keep all of them ordered alphabetically.
Differential Revision: https://reviews.llvm.org/D97627
With this, llvm-tblgen no longer tries and fails to allocate 7953 petabyte
when it runs during the build. Instead, `check-llvm` with lld/mac as host
linker now completes without any failures on an m1 mac.
This vector op handling code matches what happens in:
- ld64's OutputFile::applyFixUps() in OutputFile.cpp for kindStoreARM64PageOff12
- lld.ld64.darwinold's offset12KindFromInstruction() in
lld/lib/ReaderWriter/MachO/ArchHandler_arm64.cpp for offset12scale16
- RuntimeDyld's decodeAddend() in
llvm/lib/ExecutionEngine/RuntimeDyld/Targets/RuntimeDyldMachOAArch64.h for
ARM64_RELOC_PAGEOFF12
Fixes PR49444.
Differential Revision: https://reviews.llvm.org/D98053
Implement the promotion rule for SELECT_CC nodes by upcasting all the parameters and downcasting the result.
The AArch64 target makes use of this rule and, since it was not implemented, in some cases the instruction selector would hit an assertion upon encountering the illegal node.
This patch requires D97840, the included test cases hit both problems.
Reviewed By: craig.topper
Differential Revision: https://reviews.llvm.org/D97859
With this, you can set `clang_base_path = "//out/gn1"` in `out/gn2/args.gn` and
the build in out/gn2 will use clang and lld from out/gn1.
Setting `clang_base_path` to an absolute path (with e.g.
`clang_base_path = getenv("HOME") + "/src/..."`) should behave as before.
Differential Revision: https://reviews.llvm.org/D97989
Copy-paste P9 insns were added back in 2016,
however, looks like the opcodes has changed in ISA3.1.
Reviewed By: #powerpc, nemanjai
Differential Revision: https://reviews.llvm.org/D97416
This patch adds a new metadata node, DIArgList, which contains a list of SSA
values. This node is in many ways similar in function to the existing
ValueAsMetadata node, with the difference being that it tracks a list instead of
a single value. Internally, it uses ValueAsMetadata to track the individual
values, but there is also a reasonable amount of DIArgList-specific
value-tracking logic on top of that. Similar to ValueAsMetadata, it is a special
case in parsing and printing due to the fact that it requires a function state
(as it may reference function-local values).
This patch should not result in any immediate functional change; it allows for
DIArgLists to be parsed and printed, but debug variable intrinsics do not yet
recognize them as a valid argument (outside of parsing).
Differential Revision: https://reviews.llvm.org/D88175
Reduction updates should be masked, just like the load and stores.
Note that alternatively, we could use the fact that masked values are
zero of += updates and mask invariants to get this working but that
would not work for *= updates. Masking the update itself is cleanest.
This change also replaces the constant mask with a broadcast of "true"
since this constant folds much better for various folding patterns.
Reviewed By: nicolasvasilache
Differential Revision: https://reviews.llvm.org/D98000
and __kmpc_end_masked. The "master" construct is deprecated. Changed
proc-bind keyword from "master" to "primary". Use of both master
construct and master as proc-bind keyword is still allowed, but
deprecated.
Remove references to "master" in comments and strings, and replace
with "primary" or "primary thread". Function names and variables were
not touched, nor were references to deprecated master construct. These
can be updated over time. No new code should refer to master.
Add diagnostic tests with fir-opt for the diagnostics emitted by the ops verifier
Reviewed By: jeanPerier
Differential Revision: https://reviews.llvm.org/D97996
A few files in libc++abi make use of libc++ headers and a few of those use threading primitives provided by libc++. Since libc++ has multiple threading APIs it may be necessary to override auto-detection.
This patch adds the LIBCXXABI_HAS_WIN32_THREAD_API which does roughly the same as LIBCXXABI_HAS_PTHREAD_API and the similarly named LIBCXX_HAS_WIN32_THREAD_API from libc++. Instead of using autodetection it will force the use of win32 threads instead of pthreads in headers included from libc++.
Without this patch, libc++abi may depend on pthreads if present on the users build environment, even if win32 threading was selected for libc++.
Differential revision: https://reviews.llvm.org/D98021
Previously, lld/mac only ad-hoc codesigned executables on arm64.
Matches ld64 behavior. Part of PR49443. Fixes 14 of 17 failures when running
check-llvm with lld as host linker on an M1 MBP.
Differential Revision: https://reviews.llvm.org/D97994
Some BPF programs compiled on s390 fail to load, because s390
arch-specific linux headers contain float and double types. At the
moment there is no BTF_KIND for floats and doubles, so the release
version of LLVM ends up emitting type id 0 for them, which the
in-kernel verifier does not accept.
Introduce support for such types to libbpf by representing them using
the new BTF_KIND_FLOAT.
Reviewed By: yonghong-song
Differential Revision: https://reviews.llvm.org/D83289
We have no way to reason about the bool returned by try_emplace, so we
simply ignore any std::move()s that happen in a try_emplace argument.
A lot of the time in this situation, the code will be checking the
bool and doing something else if it turns out the value wasn't moved
into the map, and this has been causing false positives so far.
I don't currently have any intentions of handling "maybe move" functions
more generally.
Reviewed By: sammccall
Differential Revision: https://reviews.llvm.org/D98034
A bug was introduced when adding -munsafe-fp-atomics.
By default it should be off.
Reviewed by: Artem Belevich
Differential Revision: https://reviews.llvm.org/D97967
Like nvptx and some other targets, -mconstructor-aliases does not work well with amdgpu,
therefore we disable it in the same approach.
Reviewed by: Artem Belevich
Differential Revision: https://reviews.llvm.org/D97959
`mix` is subtly different from `clamp`: in the overloads where the
last argument is a scalar, the second argument should be a gentype for
`mix`.
As scalars can be implicitly converted to vectors, this cannot be
caught in the Sema test. Hence adding a CodeGen test, where we can
verify the types using the mangled name.
Previously if we couldn't run the clang-format command
for some reason, you'd get an unhelpful error message:
```
OSError: [Errno 2] No such file or directory
```
Which doesn't tell you what was happening to cause this.
Catch the error and add the command we were attempting to run:
```
RuntimeError: Failed to run "<...>/clang-food <...>" - No such file or directory"
RuntimeError: Failed to run "<...>/clang-format <...>" - Permission denied"
```
Reviewed By: krasimir
Differential Revision: https://reviews.llvm.org/D98032
Rewrites test to use correct architecture triple; fixes incorrect
reference in SourceLevelDebugging doc; simplifies `spillReg` behaviour
so as to not be dependent on changes elsewhere in the patch stack.
This reverts commit d2000b45d0.
This patch uses the errno python library to print out the correct error messages instead of hardcoding the error message per platform.
Reviewed By: jhenderson, ASDenysPetrov
Differential Revision: https://reviews.llvm.org/D97472
Same as other memory instructions, ds instructions add latency even if
exec is zero. Jumping over them if exec=0 is cheaper than executing
them.
With this change, the branch instruction that skips over a basic block
if exec=0 is not removed when the block contains a ds instruction.
Differential Revision: https://reviews.llvm.org/D97922
Valentin Clement