Sam McCall [Thu, 10 Mar 2022 22:09:46 +0000 (23:09 +0100)]
Remove redundant lit config already handled in llvm_config. NFCI
This logic duplicates lit.llvm.initialize, which we're already calling
(in lit.site.cfg.py.in).
The equivalent logic was removed from clang in
d4401d354a938dd366bf but
never cleaned up here.
Dave Lee [Thu, 10 Mar 2022 20:11:57 +0000 (12:11 -0800)]
[lldb] Call ThreadPlan::DiscardPlan from Thread::DiscardPlan
Correct `Thread::DiscardPlan` to call `DiscardPlan` instead of `PopPlan`.
Differential Revision: https://reviews.llvm.org/D96655
Sylvestre Ledru [Thu, 10 Mar 2022 21:40:18 +0000 (22:40 +0100)]
llvm-ifs doc: Replace a utf-8 char by a more classic one
Fails on old python (like on ubuntu bionic) otherwise with:
UnicodeDecodeError: 'ascii' codec can't decode byte 0xe2 in position 2130: ordinal not in range(128)
Kirill Stoimenov [Thu, 10 Mar 2022 21:29:55 +0000 (21:29 +0000)]
[ASan] Added back @PLT to __asan_report_* calls to avoid compile errors.
Reviewed By: vitalybuka
Differential Revision: https://reviews.llvm.org/D121409
Saleem Abdulrasool [Thu, 10 Mar 2022 20:55:46 +0000 (20:55 +0000)]
lldb/ObjectFile,Disassembler: read some state from the executable
Add support to inspect the ELF headers for RISCV targets to determine if
RVC or RVE are enabled and the floating point support to enable. As per
the RISCV specification, d implies f, q implies d implies f, which gives
us the cascading effect that is used to enable the features when setting
up the disassembler. With this change, it is now possible to attach the
debugger to a remote process and be able to disassemble the instruction
stream.
~~~
$ bin/lldb tmp/reduced
(lldb) target create "reduced"
Current executable set to '/tmp/reduced' (riscv64).
(lldb) gdb-remote localhost:1234
(lldb) Process 5737 stopped
* thread #1, name = 'reduced', stop reason = signal SIGTRAP
frame #0: 0x0000003ff7fe1b20
-> 0x3ff7fe1b20: mv a0, sp
0x3ff7fe1b22: jal 1936
0x3ff7fe1b26: mv s0, a0
0x3ff7fe1b28: auipc a0, 27
~~~
Siva Chandra Reddy [Thu, 10 Mar 2022 21:35:43 +0000 (21:35 +0000)]
[libc][Obvious] Destroy the block store var in block store test.
owenca [Thu, 10 Mar 2022 04:38:43 +0000 (20:38 -0800)]
[clang-format] Handle "// clang-format off" for RemoveBracesLLVM
Differential Revision: https://reviews.llvm.org/D121352
Siva Chandra Reddy [Wed, 9 Mar 2022 20:54:06 +0000 (20:54 +0000)]
[libc] Add a resizable container with constexpr constructor and destructor.
The new container is used to store atexit callbacks. This way, we avoid
the possibility of the destructor of the container itself getting added
as an at exit callback.
Reviewed By: abrachet
Differential Revision: https://reviews.llvm.org/D121350
Zahira Ammarguellat [Thu, 10 Mar 2022 16:08:12 +0000 (08:08 -0800)]
Diagnose when `#pragma clang fp eval_method` doesn't have a supported value.
Currently the compiler is crashing.
Differential Revision: https://reviews.llvm.org/D121380
Kirill Stoimenov [Thu, 10 Mar 2022 20:38:37 +0000 (20:38 +0000)]
[ASan] Moved optimized callbacks out of asan_static to avoid DSO size increase.
Reviewed By: vitalybuka
Differential Revision: https://reviews.llvm.org/D121405
Florian Hahn [Thu, 10 Mar 2022 21:02:41 +0000 (21:02 +0000)]
[FunctionAttrs] Add initial test for pass statistics.
Simon Pilgrim [Thu, 10 Mar 2022 20:59:31 +0000 (20:59 +0000)]
[X86] Add (u)comiss/sd tests showing eflags handling (PR38960)
Louis Dionne [Mon, 7 Mar 2022 13:43:14 +0000 (08:43 -0500)]
[libc++] Add tests for the content of <cstddef>
As discussed in D114786.
Differential Revision: https://reviews.llvm.org/D121110
Louis Dionne [Thu, 27 Jan 2022 18:57:49 +0000 (13:57 -0500)]
[libc++] Remove operator-> from iterator archetypes that don't need it
operator-> is not a requirement for most iterators, so remove it. To
account for this change, the `common_iterator.operator->` test needs to
be refactored quite a bit -- improve test coverage while we're at it.
Differential Revision: https://reviews.llvm.org/D118400
Jonas Devlieghere [Thu, 10 Mar 2022 20:48:46 +0000 (12:48 -0800)]
[lldb] Always listen for progress events in the default event handler
We came to the conclusion that this doesn't matter for VSCode/Xcode
because they don't use the default event loop and that other clients
who might care should use the setting.
Differential revision: https://reviews.llvm.org/D120972
Zequan Wu [Fri, 11 Feb 2022 02:18:59 +0000 (18:18 -0800)]
[LLDB][NativePDB] Add support for S_DEFRANGE_REGISTER and S_DEFRANGE_SUBFIELD_REGISTER
Differential Revision: https://reviews.llvm.org/D119508
Thomas Raoux [Thu, 10 Mar 2022 01:21:07 +0000 (01:21 +0000)]
[mlir][NVVM] Add ldmatrix op to NVVM dialect
Differential Revision: https://reviews.llvm.org/D121347
Stanislav Mekhanoshin [Thu, 10 Mar 2022 20:32:38 +0000 (12:32 -0800)]
[AMDGPU] Updated some tests to run on gfx940. NFC.
Sam McCall [Thu, 10 Mar 2022 20:18:37 +0000 (21:18 +0100)]
[cmake] Remove unused OUTPUT_MAPPING param to configure_lit_site_cfg(). NFC
Looks like it never did anything.
Added in
71deeee593c40ff01e529a6a4b30fa5b55494b82
Jez Ng [Thu, 10 Mar 2022 20:04:31 +0000 (15:04 -0500)]
[lld-macho] Align cstrings less conservatively
Previously, we aligned every cstring to 16 bytes as a temporary hack to
deal with https://github.com/llvm/llvm-project/issues/50135. However, it
was highly wasteful in terms of binary size.
To recap, in contrast to ELF, which puts strings that need different
alignments into different sections, `clang`'s Mach-O backend puts them
all in one section. Strings that need to be aligned have the .p2align
directive emitted before them, which simply translates into zero padding
in the object file. In other words, we have to infer the alignment of
the cstrings from their addresses.
We differ slightly from ld64 in how we've chosen to align these
cstrings. Both LLD and ld64 preserve the number of trailing zeros in
each cstring's address in the input object files. When deduplicating
identical cstrings, both linkers pick the cstring whose address has more
trailing zeros, and preserve the alignment of that address in the final
binary. However, ld64 goes a step further and also preserves the offset
of the cstring from the last section-aligned address. I.e. if a cstring
is at offset 18 in the input, with a section alignment of 16, then both
LLD and ld64 will ensure the final address is 2-byte aligned (since
`18 == 16 + 2`). But ld64 will also ensure that the final address is of
the form 16 * k + 2 for some k (which implies 2-byte alignment).
Note that ld64's heuristic means that a dedup'ed cstring's final address is
dependent on the order of the input object files. E.g. if in addition to the
cstring at offset 18 above, we have a duplicate one in another file with a
`.cstring` section alignment of 2 and an offset of zero, then ld64 will pick
the cstring from the object file earlier on the command line (since both have
the same number of trailing zeros in their address). So the final cstring may
either be at some address `16 * k + 2` or at some address `2 * k`.
I've opted not to follow this behavior primarily for implementation
simplicity, and secondarily to save a few more bytes. It's not clear to me
that preserving the section alignment + offset is ever necessary, and there
are many cases that are clearly redundant. In particular, if an x86_64 object
file contains some strings that are accessed via SIMD instructions, then the
.cstring section in the object file will be 16-byte-aligned (since SIMD
requires its operand addresses to be 16-byte aligned). However, there will
typically also be other cstrings in the same file that aren't used via SIMD
and don't need this alignment. They will be emitted at some arbitrary address
`A`, but ld64 will treat them as being 16-byte aligned with an offset of
`16 % A`.
I have verified that the two repros in https://github.com/llvm/llvm-project/issues/50135
work well with the new alignment behavior.
Fixes https://github.com/llvm/llvm-project/issues/54036.
Reviewed By: #lld-macho, oontvoo
Differential Revision: https://reviews.llvm.org/D121342
Elvina Yakubova [Thu, 10 Mar 2022 20:04:03 +0000 (23:04 +0300)]
[BOLT] Divide RegularPageSize for X86 and AArch64 cases
For AArch64 in some cases/some distributions ld uses 64K alignment of LOAD segments by default.
Reviewed By: yota9, maksfb
Differential Revision: https://reviews.llvm.org/D119267
Dave Lee [Thu, 10 Mar 2022 20:01:15 +0000 (12:01 -0800)]
[lldb] Remove extra space in step logging message
When seeing the extra space in the log, it wasn't clear if there was a missing
printf argument. Removing the extra space removes the potential confusion.
wren romano [Wed, 9 Mar 2022 23:53:04 +0000 (15:53 -0800)]
[mlir][sparse] Improving error messages for openSparseTensorCOO
Reviewed By: aartbik
Differential Revision: https://reviews.llvm.org/D121333
Craig Topper [Thu, 10 Mar 2022 19:26:05 +0000 (11:26 -0800)]
[RISCV] Remove performANY_EXTENDCombine.
This doesn't appear to be needed any more. I did some inspecting
of the gcc torture suite and SPEC2006 with this removed and didn't
find any meaningful changes.
I think we're more aggressive about forming ADDIW now using
sign_extend_inreg during type legalization and hasAllWUsers in isel.
This probably helps catch the cases this helped with before.
Pavel Labath [Wed, 9 Mar 2022 17:19:58 +0000 (18:19 +0100)]
[lldb/gdb-remote] Remove ancient debugserver workaround
This workaround is the source of an awkwared Process->Platform
dependency. While this could be solved in various ways (the only thing
we really use is the plugin name), it may be better to just remove it --
the workaround was added 10 years ago (
43c555dfc), and the affected
debugservers were "old" even then, so hopefully they are not in use
anymore.
Differential Revision: https://reviews.llvm.org/D121305
Valentin Clement [Thu, 10 Mar 2022 19:19:57 +0000 (20:19 +0100)]
[flang] Lower pointer component in derived type
This patch lowers pointer component part of derived types to
FIR.
This patch is part of the upstreaming effort from fir-dev branch.
Depends on D121383
Reviewed By: PeteSteinfeld, schweitz
Differential Revision: https://reviews.llvm.org/D121384
Co-authored-by: V Donaldson <vdonaldson@nvidia.com>
Co-authored-by: Jean Perier <jperier@nvidia.com>
Co-authored-by: Eric Schweitz <eschweitz@nvidia.com>
Med Ismail Bennani [Thu, 10 Mar 2022 19:06:42 +0000 (11:06 -0800)]
[lldb/crashlog] Reformat module loading logs (NFC)
Signed-off-by: Med Ismail Bennani <medismail.bennani@gmail.com>
Med Ismail Bennani [Wed, 9 Mar 2022 19:11:49 +0000 (11:11 -0800)]
[lldb/Plugin] Test that a scripted process blueprint can be loaded for the dSYM
This patch ensures that lldb can automatically load a scripted process
blueprint from a dSYM bundle and launch a scripted process with it.
rdar://
74502750
Differential Revision: https://reviews.llvm.org/D121316
Signed-off-by: Med Ismail Bennani <medismail.bennani@gmail.com>
Med Ismail Bennani [Wed, 9 Mar 2022 00:34:31 +0000 (16:34 -0800)]
[lldb/crashlog] Make interactive mode display more user-friendly
This patch makes the crashlog interactive mode show the scripted process
status with the crashed scripted thread backtrace after launching it.
rdar://
89634338
Differential Revision: https://reviews.llvm.org/D121038
Signed-off-by: Med Ismail Bennani <medismail.bennani@gmail.com>
Yuanfang Chen [Thu, 10 Mar 2022 18:50:30 +0000 (10:50 -0800)]
[JMCInstrument] infer proper path style based on debug info
By default, the path style is decided by the host. This patch makes JMC
uses the path style used by the SP directory. This makes JMC output
host-independent.
Fixes: https://github.com/llvm/llvm-project/issues/54219
Reviewed By: rnk
Differential Revision: https://reviews.llvm.org/D121236
Valentin Clement [Thu, 10 Mar 2022 18:43:11 +0000 (19:43 +0100)]
[flang] Lower general forall statement
This patch lowers general forall statements. The forall
are lowered to nested loops.
This patch is part of the upstreaming effort from fir-dev branch.
Depends on D121385
Reviewed By: PeteSteinfeld, schweitz
Differential Revision: https://reviews.llvm.org/D121386
Co-authored-by: V Donaldson <vdonaldson@nvidia.com>
Co-authored-by: Jean Perier <jperier@nvidia.com>
Co-authored-by: Eric Schweitz <eschweitz@nvidia.com>
Petr Hosek [Wed, 9 Mar 2022 08:48:49 +0000 (00:48 -0800)]
[CMake] Include runtimes test suites in check-all
Prior to this change, we would make check-all depend on check-runtimes
which is a target that runs tests in the runtimes build. This means that
the runtimes tests are going to run prior to other test suites in
check-all, and if one of them fails, we won't run the other test suites
at all.
To address this issue, we instead collect the list of test suites and
their dependencies from the runtimes subbuild, and include them in
check-all, so a failure of runtimes test suite doesn't prevent other
test suites from being executed.
This addresses https://github.com/llvm/llvm-project/issues/54154.
Differential Revision: https://reviews.llvm.org/D121276
Joe Nash [Thu, 10 Mar 2022 17:25:11 +0000 (12:25 -0500)]
[AMDGPU] Use subreg encoding instead of reassign
The HWEncoding for these 64 bit registers should be the same as as the
encoding for the previously defined low halves of the registers. So
reuse that value instead of repeating the assignment. NFC.
Reviewed By: foad
Differential Revision: https://reviews.llvm.org/D121391
Aaron Ballman [Thu, 10 Mar 2022 18:08:08 +0000 (13:08 -0500)]
Fix the dates for the C and C++ language working group
It's 1st and 3rd Wed of the month instead of every other.
Craig Topper [Thu, 10 Mar 2022 17:37:03 +0000 (09:37 -0800)]
[RISCV] Add isel patterns for masked RISCVISD::FMA_VL with RISCVISD::FNEG_VL.
This helps us form vfnmsub, vfnmadd, and vfmusb from masked VP
intrinsics.
I've used "srcvalue" for the mask parameter in the fneg nodes. We
can't match "V0" because that doesn't ensure the mask the is the same.
Instead it matches two different nodes and generates two copies to
V0 of those separate values.
Reviewed By: rogfer01
Differential Revision: https://reviews.llvm.org/D120287
Sanjay Patel [Thu, 10 Mar 2022 17:37:36 +0000 (12:37 -0500)]
[InstCombine] replace negated operand in fcmp with 0.0
X (any pred) -X --> X (any pred) 0.0
This works with all FP values and preserves FMF.
Alive2 examples:
https://alive2.llvm.org/ce/z/dj6jhp
This can also create one of the patterns that we match as "fabs"
as shown in one of the test diffs.
Sanjay Patel [Thu, 10 Mar 2022 16:35:23 +0000 (11:35 -0500)]
[InstCombine] add tests for fcmp with fneg common operand; NFC
Lorenzo Albano [Thu, 10 Mar 2022 15:39:12 +0000 (16:39 +0100)]
[VP] Strided loads/stores
This patch introduces two new experimental IR intrinsics and SDAG nodes
to represent vector strided loads and stores.
Reviewed By: simoll
Differential Revision: https://reviews.llvm.org/D114884
Valentin Clement [Thu, 10 Mar 2022 17:43:40 +0000 (18:43 +0100)]
[flang] Lower where statement
This patch lowers where statement to FIR.
The where statement is lowered to a conbination of
loops and if conditions.
This patch is part of the upstreaming effort from fir-dev branch.
Reviewed By: PeteSteinfeld
Differential Revision: https://reviews.llvm.org/D121385
Co-authored-by: Jean Perier <jperier@nvidia.com>
Co-authored-by: Eric Schweitz <eschweitz@nvidia.com>
Bixia Zheng [Thu, 10 Mar 2022 17:08:41 +0000 (09:08 -0800)]
[mlir][linalg] Add a few unary operations.
Add operations abs, ceil, floor, and neg to the C++ API and Python API.
Add test cases.
Reviewed By: gysit
Differential Revision: https://reviews.llvm.org/D121339
Shraiysh Vaishay [Thu, 10 Mar 2022 17:10:23 +0000 (22:40 +0530)]
[flang] Added basic connect to lower OpenMP constructs
Reviewed By: clementval
Differential Revision: https://reviews.llvm.org/D121382
Valentin Clement [Thu, 10 Mar 2022 17:06:20 +0000 (18:06 +0100)]
[flang] Lower basic derived types
This patch lowers basic derived type to FIR.
This patch is part of the upstreaming effort from fir-dev branch.
Reviewed By: PeteSteinfeld
Differential Revision: https://reviews.llvm.org/D121383
Co-authored-by: V Donaldson <vdonaldson@nvidia.com>
Co-authored-by: Jean Perier <jperier@nvidia.com>
Co-authored-by: Eric Schweitz <eschweitz@nvidia.com>
Groverkss [Thu, 10 Mar 2022 16:48:44 +0000 (22:18 +0530)]
[MLIR][Presburger] Remove `spaceKind` from PresburgerSpace
This patch remove `spaceKind` from PresburgerSpace, making PresburgerSpace only
a space supporting relations.
Sets are still implemented in the same way, i.e. with a zero domain but instead
the asserts to check if the space is still set are added to users of
PresburgerSpace which treat it as a Set space.
Reviewed By: arjunp
Differential Revision: https://reviews.llvm.org/D121357
Vladislav Khmelevsky [Wed, 9 Mar 2022 16:28:06 +0000 (19:28 +0300)]
[BOLT] LongJmp: Use per-function alignment values
The per-function alignment values must be used in order to create
tentative layout.
Differential Revision: https://reviews.llvm.org/D121298
Emilio Cota [Wed, 9 Mar 2022 15:08:17 +0000 (10:08 -0500)]
[mlir] ExecutionEngine: default enableObjectCache to false
The enableObjectCache option was added in
https://reviews.llvm.org/rG06e8101034e, defaulting to false. However,
the init code added there got its logic reversed
(cache(enableObjectCache ? nullptr : new SimpleObjectCache()), which was
fixed in https://reviews.llvm.org/rGd1186fcb04 by setting the default to
true, thereby preserving the existing behavior even if it was
unintentional.
Default now the object cache to false as it was originally intended.
While at it, mention in enableObjectCache's documentation how the
cache can be dumped.
Reviewed-by: mehdi_amini
Differential Revision: https://reviews.llvm.org/D121291
serge-sans-paille [Wed, 9 Mar 2022 21:27:35 +0000 (22:27 +0100)]
Cleanup includes: WindowsDriver & WindowsManifest
Discourse thread: https://discourse.llvm.org/t/include-what-you-use-include-cleanup
Differential Revision: https://reviews.llvm.org/D121330
Dmitry Vyukov [Thu, 10 Mar 2022 09:35:32 +0000 (10:35 +0100)]
tsan: fix another false positive related to open/close
The false positive fixed by commit
f831d6fc80
("tsan: fix false positive during fd close") still happens episodically
on the added more stressful test which does just open/close.
I don't have a coherent explanation as to what exactly happens
but the fix fixes the false positive on this test as well.
The issue may be related to lost writes during asynchronous MADV_DONTNEED.
I've debugged similar unexplainable false positive related to freed and
reused memory and at the time the only possible explanation I found is that
an asynchronous MADV_DONTNEED may lead to lost writes. That's why commit
302ec7b9bc ("tsan: add memory_limit_mb flag") added StopTheWorld around
the memory flush, but unfortunately the commit does not capture these findings.
Reviewed By: melver
Differential Revision: https://reviews.llvm.org/D121363
Shraiysh Vaishay [Thu, 10 Mar 2022 15:55:05 +0000 (21:25 +0530)]
[MLIR][OpenMP][NFC] Move the modifier testcases to correct directory
This patch moves the testcases from
`mlir/test/Target/LLVMIR/openmp-llvm-bad-schedule-modifier.mlir` to
`mlir/test/Dialect/OpenMP/invalid.mlir` as they test the verifier
(not the translation to LLVM IR).
Reviewed By: NimishMishra
Differential Revision: https://reviews.llvm.org/D120877
John Demme [Thu, 10 Mar 2022 15:43:43 +0000 (07:43 -0800)]
[MLIR] [CMake/Python] Fix ignored Python3_EXECUTABLE
On Windows (at least), cmake ignores Python3_EXECUTABLE unless the
'Interpreter' component is being found. If the user is specifying a
different version than the latest installed (say, 3.8 vs 3.9) with the
Python3_EXECUTABLE, cmake was using a combination of the newest version
and the desired version. Mitigated by adding 'Interpreter' in the first
invocation like the second one.
Sanjay Patel [Thu, 10 Mar 2022 15:22:22 +0000 (10:22 -0500)]
Revert "[InstCombine] fold fcmp with lossy casted constant"
This reverts commit
9397bdc67eb2b9eedc247a89bef01c2484b48b89.
This optimization is likely to surprise programmers as seen
in post-commit comments, so we should add a clang warning
first (that is proposed in D121306).
David Green [Thu, 10 Mar 2022 15:03:27 +0000 (15:03 +0000)]
[AArch64] Fix type in comment. NFC
Nikita Popov [Thu, 10 Mar 2022 15:01:09 +0000 (16:01 +0100)]
[GlobalOpt] Handle undef global_ctors gracefully
If there are no ctors, then this can have an arbirary zero-sized
value. The current code checks for null, but it could also be
undef or poison.
Replacing the specific null check with a check for
non-ConstantArray.
Nikita Popov [Thu, 10 Mar 2022 14:47:58 +0000 (15:47 +0100)]
[BitcodeReader] Support GEP without indices
LLVM considers these to be legal, so make sure the bitcode reader
can read them. I broke this when implementing opaque pointer
auto upgrade support.
David Green [Thu, 10 Mar 2022 14:45:13 +0000 (14:45 +0000)]
[AArch64] TBL uses zero for out of range elements.
A TBL instruction will use zero for any out of range values. We can use
this in GenerateTBL to help turn a TBL2 into a TBL1, avoiding the need
to materialise the zero.
Differential Revision: https://reviews.llvm.org/D121139
Tue Ly [Thu, 10 Mar 2022 14:25:54 +0000 (09:25 -0500)]
[libc] Initial support for darwin-aarch64.
Add initial support for darwin-aarch64 (macOS M1).
Some differences compared to linux-aarch64:
- `math.h` defined `math_errhandling` by the compiler builtin `__math_errhandling()` but Apple Clang 13.0.0 on M1 does not support `__math_errhandling()` builtin as a macro function or a constexpr function.
- `math.h` defines `UNDERFLOW` and `OVERFLOW` macros.
- Besides 5 usual floating point exceptions: `FE_INEXACT`, `FE_UNDERFLOW`, `FE_OVERFLOW`, `FE_DIVBYZERO`, and `FE_INVALID`, `fenv.h` also has another floating point exception: `FE_FLUSHTOZERO`. The corresponding trap for `FE_FLUSHTOZERO` in the control register is at the different location compared to the status register.
- `FE_FLUSHTOZERO` exception flag cannot be raised with the default CPU floating point operation mode.
Reviewed By: sivachandra
Differential Revision: https://reviews.llvm.org/D120914
Louis Dionne [Wed, 2 Feb 2022 18:08:52 +0000 (13:08 -0500)]
[libc++] Bump minimum compiler requirements
Now that we've branched for the LLVM 14 release, our support window
moves to clang-13 and clang-14. Similarly, AppleClang 13 has been
released for some time now, so that should be the oldest compiler
we support, per our policy.
A possible follow-up would be to remove _LIBCPP_HAS_NO_CONCEPTS, since
I don't think we support any compiler that doesn't support concepts
anymore.
Differential Revision: https://reviews.llvm.org/D118831
David Green [Thu, 10 Mar 2022 13:51:04 +0000 (13:51 +0000)]
[AArch64] Extra tests for tbl with zero elements. NFC
Louis Dionne [Wed, 9 Mar 2022 22:06:16 +0000 (17:06 -0500)]
[libc++] Add test coverage for std::shared_ptr<const T>
Those tests were extracted from D120996.
Differential Revision: https://reviews.llvm.org/D121340
serge-sans-paille [Wed, 9 Mar 2022 21:26:23 +0000 (22:26 +0100)]
Cleanup includes: MCDisassembler
Some extra minor cleanup.
Discourse thread: https://discourse.llvm.org/t/include-what-you-use-include-cleanup
Differential Revision: https://reviews.llvm.org/D121329
Nikita Popov [Thu, 10 Mar 2022 13:18:37 +0000 (14:18 +0100)]
[FuzzMutate] Support opaque pointers
Avoid checks that are irrelevant for opaque pointers, and pick
load/GEP types independently of the pointer type.
The GEP case at least could be done more efficiently by directly
generating a type, but this would require some significant API
changes.
Simon Pilgrim [Thu, 10 Mar 2022 13:30:42 +0000 (13:30 +0000)]
[InstCombine] Add vector support to icmp(add(X,C1),add(Y,C2)) -> icmp(add(X,C1-C2),Y) fold
As discussed on Issue #32161 this fold can be generalized a lot more than it currently is, but this patch at least adds vector support.
Differential Revision: https://reviews.llvm.org/D121358
Simon Pilgrim [Thu, 10 Mar 2022 13:25:06 +0000 (13:25 +0000)]
[X86] Add GFNI shift/rotate/funnel-shift vXi8 by constant splat tests
GFNI can efficiently perform vXi8 shift/rotates by constant splats using X86ISD::GF2P8AFFINEQB
Simon Pilgrim [Thu, 10 Mar 2022 13:18:10 +0000 (13:18 +0000)]
Fix M68k missing header errors.
Sanjay Patel [Thu, 10 Mar 2022 13:08:33 +0000 (08:08 -0500)]
[PhaseOrdering][x86] add TODO comments to test files; NFC
The intended folds were removed because an underlying
analysis was not correct (
b48fe158e0a8 ).
Nico Weber [Thu, 10 Mar 2022 12:54:41 +0000 (07:54 -0500)]
Revert "Cleanup codegen includes"
This reverts commit
7f230feeeac8a67b335f52bd2e900a05c6098f20.
Breaks CodeGenCUDA/link-device-bitcode.cu in check-clang,
and many LLVM tests, see comments on https://reviews.llvm.org/D121169
Shraiysh Vaishay [Thu, 10 Mar 2022 12:58:08 +0000 (18:28 +0530)]
[OpenMP][mlir] Lowering for omp.atomic.update
This patch adds lowering from omp.atomic.update to LLVM IR. Whenever a
special LLVM IR instruction is available for the operation, `atomicrmw`
instruction is emitted, otherwise a compare-exchange loop based update
is emitted.
Depends on D119522
Reviewed By: ftynse, peixin
Differential Revision: https://reviews.llvm.org/D119657
Florian Hahn [Thu, 10 Mar 2022 12:22:05 +0000 (12:22 +0000)]
[AArch64] Move fp16 cast tests.
Move FP16 tests to fp16cast function, as suggested in D113700.
Nico Weber [Thu, 10 Mar 2022 12:14:07 +0000 (07:14 -0500)]
[gn build] (manually) port
af98b0af6705
alex-t [Thu, 10 Mar 2022 09:57:15 +0000 (12:57 +0300)]
[AMDGPU] Enable divergence predicates for negative inline constant subtraction
We have a pattern that undo sub x, c -> add x, -c canonicalization since c is more likely
an inline immediate than -c. This patch enables it to select scalar or vector subtracion by the input node divergence.
Reviewed By: foad
Differential Revision: https://reviews.llvm.org/D121360
Stanislav Gatev [Thu, 10 Mar 2022 08:57:32 +0000 (08:57 +0000)]
[clang][dataflow] Add analysis that detects unsafe accesses to optionals
This commit reverts
e0cc28dfdc67105974924cce42bb8c85bd44925a and moves
UncheckedOptionalAccessModelTest.cpp into clang/unittests/Analysis/FlowSensitive,
to avoid build failures. The test will be moved back into a Models subdir
in a follow up patch that will address the build configuration issues.
Original description:
Adds a dataflow analysis that detects unsafe accesses to values of type
`std::optional`, `absl::optional`, or `base::Optional`.
Reviewed-by: ymandel, xazax.hun
Differential Revision: https://reviews.llvm.org/D121197
Nikita Popov [Thu, 10 Mar 2022 10:55:58 +0000 (11:55 +0100)]
[AMDGPU] Fix regenerated test checks (NFC)
I used the wrong build to generate the checks, sorry :(
Nikita Popov [Thu, 10 Mar 2022 10:53:45 +0000 (11:53 +0100)]
[AMDGPU] Regenerate test checks (NFC)
Nikita Popov [Thu, 10 Mar 2022 10:14:45 +0000 (11:14 +0100)]
[AMDGPU] Regenerate test checks (NFC)
Also rename variables to avoid file check clash.
Michel Weber [Thu, 10 Mar 2022 10:12:06 +0000 (15:42 +0530)]
[MLIR][Presburger] fix vector update in coalesce
When `addCoalescedPolyhedron` was called with `j == n - 1`,
the `polyhedrons`-vector was not properly updated (the
`IntegerPolyhedron` at position `n - 2` was "lost"). This patch adds
special handling to that case and a regression testcase.
Reviewed By: Groverkss
Differential Revision: https://reviews.llvm.org/D121356
Groverkss [Thu, 10 Mar 2022 10:09:32 +0000 (15:39 +0530)]
[MLIR][Presburger] Use IdKind for removeIdRange in PresburgerSpace
This patch moves PresburgerSpace::removeIdRange(idStart, idLimit) to
PresburgerSpace::removeIdRange(kind, idStart, idLimit), i.e. identifiers
can only be removed at once for a single kind.
This makes users of PresburgerSpace to not assume any inside ordering of
identifier kinds.
Reviewed By: arjunp
Differential Revision: https://reviews.llvm.org/D121079
Weining Lu [Thu, 10 Mar 2022 08:35:38 +0000 (16:35 +0800)]
[LoongArch] Add basic support to Disassembler
With the addition of disassembler now we can do instructions 'round-trip' test
that assembles `.s` to obj with `llvm-mc` and disassembles it with `llvm-objdump`
to check instruction mnemonics.
Reviewed By: xen0n, MaskRay
Differential Revision: https://reviews.llvm.org/D120477
serge-sans-paille [Thu, 10 Mar 2022 09:05:14 +0000 (10:05 +0100)]
Add missing include in llvm/CodeGen/CodeGenPassBuilder.h
As a follow-up to
7f230feeeac8
Simon Pilgrim [Thu, 10 Mar 2022 09:00:40 +0000 (09:00 +0000)]
[InstCombine] Add vector tests for icmp_eq(add(X,C1),add(Y,C2)) -> icmp_eq(add(X,C1-C2),Y)
As mentioned on Issue #32161 we don't even have uniform vector support for this fold
serge-sans-paille [Mon, 7 Mar 2022 23:25:52 +0000 (00:25 +0100)]
Cleanup codegen includes
after:
1061034926
before:
1063332844
Differential Revision: https://reviews.llvm.org/D121169
serge-sans-paille [Mon, 7 Mar 2022 23:25:11 +0000 (00:25 +0100)]
Cleanup includes: LLVMTarget
Most notably, Pass.h is no longer included by TargetMachine.h
before:
1063570306
after:
1063332844
Differential Revision: https://reviews.llvm.org/D121168
Nikita Popov [Fri, 4 Mar 2022 12:25:05 +0000 (13:25 +0100)]
[Coroutines] Support opaque pointers in solveTypeName()
As far as I can tell, these names are only intended to be
informative, so just use a generic "PointerType" for opaque pointers.
The code in solveDIType() also treats pointers as basic types (and
does not try to encode the pointed-to type further), so I believe
this should be fine.
Differential Revision: https://reviews.llvm.org/D121280
Luke [Sun, 13 Feb 2022 14:22:34 +0000 (22:22 +0800)]
[RISCV] Add fixed-length vector instrinsics for segment load
Inspired by reviews.llvm.org/D107790.
Reviewed By: craig.topper
Differential Revision: https://reviews.llvm.org/D119834
Craig Topper [Thu, 10 Mar 2022 06:44:27 +0000 (22:44 -0800)]
[RISCV] Remove RISCVISD::VLE_VL/VSE_VL. Use intrinsics instead.
Similar to what we do for other loads/stores, use the intrinsic
version that we already have custom isel for.
Reviewed By: rogfer01
Differential Revision: https://reviews.llvm.org/D121166
Craig Topper [Thu, 10 Mar 2022 04:39:14 +0000 (20:39 -0800)]
[RISCV] Add tests showing the optimization pipeline for O0 and O3.
Other targets like ARM, AArch64, and X86 have similar tests.
Reviewed By: asb
Differential Revision: https://reviews.llvm.org/D120840
Amir Ayupov [Thu, 10 Mar 2022 04:27:15 +0000 (20:27 -0800)]
[BOLT][NFC] Fix print-cfg data race
Addresses ThreadSanitizer warning
Reviewed By: rafauler
Differential Revision: https://reviews.llvm.org/D121337
Amir Ayupov [Thu, 10 Mar 2022 04:24:20 +0000 (20:24 -0800)]
[BOLT][NFC] Check errors from Obj.dynamicEntries
Addresses fuzzer crash
Reviewed By: rafauler
Differential Revision: https://reviews.llvm.org/D121336
Siva Chandra Reddy [Wed, 9 Mar 2022 21:30:22 +0000 (21:30 +0000)]
[libc] Use the constexpr constructor to initialize exit handlers mutex.
Reviewed By: abrachet
Differential Revision: https://reviews.llvm.org/D121334
Carl Ritson [Thu, 10 Mar 2022 02:11:20 +0000 (11:11 +0900)]
[MachineSink] Pre-commit test for D121277. NFC.
Ahmed Bougacha [Wed, 9 Mar 2022 23:52:21 +0000 (15:52 -0800)]
[clang][Driver] Get darwin -Xarch_ working for subtypes, again.
35ca7d9ddf4 broke
471c4f829934 for -arch flags that don't map 1:1
to the triple arch. This has been broken for the many years since.
It hasn't mattered much since then, mostly because few people use it,
but also because it works for x86_64/i386, armv7/armv7s
don't differ much, arm64 is its own arch, and arm64/arm64_32 have
different arches (and it's a rare combination anyway).
But arm64/arm64e exposes this issue again.
Patch by: Justin Bogner <mail@justinbogner.com>
with some added tests.
Yaxun (Sam) Liu [Wed, 9 Mar 2022 16:52:48 +0000 (11:52 -0500)]
[HIP] Fix -fno-gpu-sanitize
Fix a typo about -fno-gpu-sanitize handling and disable warnings when
-fno-gpu-sanitize is specified.
Reviewed by: Artem Belevich
Differential Revision: https://reviews.llvm.org/D121302
Yaxun (Sam) Liu [Wed, 9 Mar 2022 14:10:17 +0000 (09:10 -0500)]
[HIP] Fix HIP include path
The clang compiler prepends the HIP header include paths to the search
list using -internal-isystem when building for the HIP language. This
prevents warnings related to things like reserved identifiers when
including the HIP headers even when ROCm is installed in a non-system
directory, such as /opt/rocm.
However, when HIP is installed in /usr, then the prepended include
path would be /usr/include. That is a problem, because the C standard
library headers are stored in /usr/include and the C++ standard
library headers must come before the C library headers in the search
path list (because the C++ standard library headers use #include_next
to include the C standard library headers).
While the HIP wrapper headers _do_ need to be earlier in the search
than the C++ headers, those headers get their own subdirectory and
their own explicit -internal-isystem argument. This include path is for
<hip/hip_runtime_api.h> and <hip/hip_runtime.h>, which do not require a
particular search ordering with respect to the C or C++ headers. Thus,
HIP include path is added after other system include paths.
With contribution from Cordell Bloor.
Reviewed by: Artem Belevich
Differential Revision: https://reviews.llvm.org/D120132
LLVM GN Syncbot [Thu, 10 Mar 2022 01:35:38 +0000 (01:35 +0000)]
[gn build] Port
c31014322c0b
Brad Smith [Thu, 10 Mar 2022 01:29:49 +0000 (20:29 -0500)]
[Driver][OpenBSD] Disable unwind tables on Arm
Arm is not quite ready for unwinding yet.
Xiang1 Zhang [Wed, 2 Mar 2022 06:47:54 +0000 (14:47 +0800)]
TLS loads opimization (hoist)
Reviewed By: craig.topper
Differential Revision: https://reviews.llvm.org/D120000
Yevgeny Rouban [Wed, 9 Mar 2022 11:10:21 +0000 (18:10 +0700)]
[Support] Try 2: Reset option to its default if its Default field is undefined
opt::setDefaultImpl() is changed to set the option value to the option
type's default if the Default field is not set. This results in option
value reset by Option::reset() or ResetAllOptionOccurrences() even if
the cl::init() is not specified.
Example:
StackOption<std::string> Str("str"); // No cl::init().
Str = "some value";
cl::ResetAllOptionOccurrences();
EXPECT_EQ("", Str); // The Str is reset.
Reviewed By: lattner
Differential Revision: https://reviews.llvm.org/D115433
Vitaly Buka [Thu, 10 Mar 2022 01:20:49 +0000 (01:20 +0000)]
[msan] Fix unittest for
9397bdc6
"1.1" is double which can't be represented precicely as a float.
Craig Topper [Thu, 10 Mar 2022 00:43:17 +0000 (16:43 -0800)]
[RISCV] Support 'generic' as a valid CPU name.
Most other targets support 'generic', but RISCV issues an error.
This can require a special case in tools that use LLVM that aren't
clang.
This patch treats "generic" the same as an empty string and remaps
it to generic-rv/rv64 based on the triple. Unfortunately, it has to
be added to RISCV.td because MCSubtargetInfo is constructed and
parses the CPU before RISCVSubtarget's constructor gets a chance
to remap it. The CPU will then reparsed and the state in the
MCSubtargetInfo subclass will be updated again.
Fixes PR54146.
Reviewed By: khchen
Differential Revision: https://reviews.llvm.org/D121149
Florian Mayer [Thu, 10 Mar 2022 00:16:24 +0000 (16:16 -0800)]
[NFC] [HWASan] document why we tag Size but untag AlignedSize.
Aart Bik [Wed, 9 Mar 2022 23:11:31 +0000 (15:11 -0800)]
[mlir][sparse] add end2end test for linalg.dot sparsification
Reviewed By: bixia
Differential Revision: https://reviews.llvm.org/D121344
Shilei Tian [Wed, 9 Mar 2022 23:48:44 +0000 (18:48 -0500)]
[OpenMP][CUDA] Fix the check of `setContext`