Fangrui Song [Fri, 25 Feb 2022 19:55:56 +0000 (19:55 +0000)]
[libunwind][test] remember_state_leak.pass.sh.s: link with -no-pie
The no-pic large code model style `movabsq $callback, %rsi` does not work with -pie.
Hanhan Wang [Fri, 25 Feb 2022 18:52:08 +0000 (10:52 -0800)]
[mlir][Linalg] Add support for tileFuseAndDistribute on tensors.
This extends TileAndFuse to handle distribution on tensors.
Reviewed By: gysit
Differential Revision: https://reviews.llvm.org/D120441
Craig Topper [Fri, 25 Feb 2022 19:36:23 +0000 (11:36 -0800)]
[RISCV] Remove tab character from test. Autogenerate CHECK lines. NFC
This was a test for an infinite loop so the CHECK lines don't really
matter, but they'd get generated the next time someone runs the script
on the file so might as well do it while I'm touching it.
Jameson Nash [Fri, 25 Feb 2022 19:24:42 +0000 (14:24 -0500)]
mark getTargetTransformInfo and getTargetIRAnalysis as const
Seems like this can be const, since Passes shouldn't modify it.
Reviewed By: wsmoses
Differential Revision: https://reviews.llvm.org/D120518
Diego Caballero [Fri, 25 Feb 2022 18:27:43 +0000 (18:27 +0000)]
[mlir][Vector] Prevent AVX2 lowering for non-f32 transpose ops
The AVX2 lowering for transpose operations is only applicable to f32 vector types.
Reviewed By: aartbik
Differential Revision: https://reviews.llvm.org/D120427
Diego Caballero [Fri, 25 Feb 2022 18:26:37 +0000 (18:26 +0000)]
[mlir][Vector] Generalize AVX2 transpose lowering to n-D vectors
The existing AVX2 lowering patterns for the transpose op only triggers if the
input vector is 2-D. This patch extends the patterns to trigger for n-D vectors
which are effectively 2-D vectors (e.g., vector<1x4x1x8x1). The main constraint
for the generalized AVX2 patterns to be applicable to these vectors is that the
dimensions that are greater than one must be transposed. Otherwise, the existing
patterns are not applicable.
Reviewed By: nicolasvasilache
Differential Revision: https://reviews.llvm.org/D119505
Fangrui Song [Fri, 25 Feb 2022 19:25:18 +0000 (19:25 +0000)]
[ELF] Support some absolute/PC-relative relocation types for REL format
ctfconvert seems to use REL-format `.rel.SUNW_dof` for 32-bit architectures.
```
Binary file usr/ports/lang/perl5.32/work/perl-5.32.1/dtrace_mini.o matches
[alfredo.junior@dell-a ~/tmp/llvm-bug]$ readelf -r dtrace_mini.o
Relocation section (.rel.SUNW_dof):
r_offset r_info r_type st_value st_name
00000184 0000281a R_PPC_REL32
00000000 $dtrace1772974259.Perl_dtrace_probe_load
```
Support R_PPC_REL32 to fix `ld.lld: error: drti.c:(.SUNW_dof+0x4E4): internal linker error: cannot read addend for relocation R_PPC_REL32`.
While here, add some common relocation types for AArch64, PPC, and PPC64.
We perform minimum tests.
Reviewed By: adalava, arichardson
Differential Revision: https://reviews.llvm.org/D120535
Chia-hung Duan [Fri, 25 Feb 2022 18:17:30 +0000 (18:17 +0000)]
[mlir] Support verification order (2/3)
This change gives explicit order of verifier execution and adds
`hasRegionVerifier` and `verifyWithRegions` to increase the granularity
of verifier classification. The orders are as below,
1. InternalOpTrait will be verified first, they can be run independently.
2. `verifyInvariants` which is constructed by ODS, it verifies the type,
attributes, .etc.
3. Other Traits/Interfaces that have marked their verifier as
`verifyTrait` or `verifyWithRegions=0`.
4. Custom verifier which is defined in the op and has marked
`hasVerifier=1`
If an operation has regions, then it may have the second phase,
5. Traits/Interfaces that have marked their verifier as
`verifyRegionTrait` or
`verifyWithRegions=1`. This implies the verifier needs to access the
operations in its regions.
6. Custom verifier which is defined in the op and has marked
`hasRegionVerifier=1`
Note that the second phase will be run after the operations in the
region are verified. Based on the verification order, you will be able to
avoid verifying duplicate things.
Reviewed By: Mogball
Differential Revision: https://reviews.llvm.org/D116789
Alexey Bataev [Thu, 1 Jul 2021 11:16:56 +0000 (04:16 -0700)]
[OPENMP]Fix PR50347: Mapping of global scope deep object fails.
Changed the we handle llvm::Constants in sizes arrays. ConstExprs and
GlobalValues cannot be used as initializers, need to put them at the
runtime, otherwise there wight be the compilation errors.
Differential Revision: https://reviews.llvm.org/D105297
Fangrui Song [Fri, 25 Feb 2022 18:50:39 +0000 (18:50 +0000)]
Revert "[lsan][test] Temporarily disable ppc64 and ppc64le to appease clang-ppc64le-rhel"
This reverts commit
cb76c4d71c41bbbae47852d7980e74b57c5a28df.
The failures were in test/sanitizer_common, not in test/lsan.
Fangrui Song [Fri, 25 Feb 2022 18:49:10 +0000 (18:49 +0000)]
[compiler-rt][test] Temporarily disable ppc64 and ppc64le test/sanitizer_common and test/crt
to appease clang-ppc64le-rhel: https://github.com/llvm/llvm-project/issues/54084
Changpeng Fang [Fri, 25 Feb 2022 18:45:35 +0000 (10:45 -0800)]
[AMDGPU][NFC]: Emit metadata for hidden_heap_v1 kernarg
Summary:
Emit metadata for hidden_heap_v1 kernarg
Reviewers:
sameerds, b-sumner
Fixes:
SWDEV-307188
Differential Revision:
https://reviews.llvm.org/D119027
Alexander Yermolovich [Fri, 25 Feb 2022 18:31:22 +0000 (10:31 -0800)]
[BOLT][DWARF] Fix how DW_AT_high_pc [DW_FORM_udata] is handled
We were not handling correctly conversion from DW_AT_high_pc into DW_AT_ranges,
when size of DW_AT_high_pc is not 4/8 bytes.
Reviewed By: maksfb
Differential Revision: https://reviews.llvm.org/D120528
Fangrui Song [Fri, 25 Feb 2022 18:25:45 +0000 (18:25 +0000)]
[lsan][test] Temporarily disable ppc64 and ppc64le to appease clang-ppc64le-rhel
Seems that ppc64 lsan doesn't work with default PIE (see D120305):
https://lab.llvm.org/buildbot/#/builders/57/builds/15506
Arthur Eubanks [Fri, 25 Feb 2022 17:46:05 +0000 (09:46 -0800)]
[NFC] Remove unnecessary function pass managers
Florian Hahn [Fri, 25 Feb 2022 17:59:44 +0000 (17:59 +0000)]
[AArch64] Add tests for tbl + cmp splitting.
Additional tests showing potential for follow-ups after
D120571.
Kiran Chandramohan [Fri, 25 Feb 2022 17:30:44 +0000 (17:30 +0000)]
Lower Fortran intrinsic to a runtime call/llvm intrinsic
This patch brings in code which can lower a Fortran intrinsic to
a runtime call or an llvm intrinsic. For math intrinsics the
runtime call is to the `math` or `pgmath` library. Non-math
intrinsics are covered by the Flang runtime. A distance computation
mechanism is introduced to find the runtime function that closely
matches the types of the intrinsic call.
In this patch, the `abs` intrinsic is lowered in the following way,
-> Integer version is lowered as a group of MLIR/FIR operations
-> Real version is lowered to llvm intrinsics
-> Complex version is lowered to the `math_hypot` runtime function
This patch is part of upstreaming from the fir-dev branch of https://github.com/flang-compiler/f18-llvm-project
Reviewed By: clementval
Differential Revision: https://reviews.llvm.org/D120403
Co-authored-by: Jean Perier <jperier@nvidia.com>
Co-authored-by: Eric Schweitz <eschweitz@nvidia.com>
Co-authored-by: zacharyselk <zrselk@gmail.com>
Co-authored-by: V Donaldson <vdonaldson@nvidia.com>
Co-authored-by: Valentin Clement <clementval@gmail.com>
Rong Xu [Fri, 25 Feb 2022 17:05:06 +0000 (09:05 -0800)]
[Sample-PGO] Emit FS discriminators only when -fdebug-info-for-profiling is set
IR level addDiscriminator pass is guarded by DebugInfoForProfiling
(set by option -fdebug-info-for-profiling).
This patch syncs the logic for the MIR and IR level implementations.
Differential Revision: https://reviews.llvm.org/D120536
Stefan Pintilie [Fri, 25 Feb 2022 17:16:04 +0000 (11:16 -0600)]
[PowerPC][NFC] Split out the MMA instructions from the P10 instructions.
Currently all of the MMA instructions as well as the MMA related register info
is bundled with the Power 10 instructions. This patch just splits them out.
Reviewed By: lei
Differential Revision: https://reviews.llvm.org/D120515
Aakanksha [Fri, 25 Feb 2022 17:30:12 +0000 (17:30 +0000)]
Avoid comparisons between types of different widths in a loop condition to prevent the loop from behaving unexpectedly
This change fixes the code violations flagged in AMD compute CodeQL scan -
Query Description: "Comparisons between types of different widths in a loop condition can cause the loop to behave unexpectedly."
Differential Revision: https://reviews.llvm.org/D120355
Valentin Clement [Fri, 25 Feb 2022 17:21:44 +0000 (18:21 +0100)]
[flang] Lower simple character return
Handles function with character return.
Character scalar results are passed as arguments in lowering so
that an assumed length character function callee can access the result
length.
This patch is part of the upstreaming effort from fir-dev branch.
Reviewed By: PeteSteinfeld, schweitz
Differential Revision: https://reviews.llvm.org/D120558
Co-authored-by: Eric Schweitz <eschweitz@nvidia.com>
Co-authored-by: Jean Perier <jperier@nvidia.com>
Co-authored-by: V Donaldson <vdonaldson@nvidia.com>
David Green [Fri, 25 Feb 2022 17:11:24 +0000 (17:11 +0000)]
[AArch64] Add scalar min/max costs. NFC
The vector costs were already added, this adds scalar variants to
complete the test coverage.
Paul Walker [Thu, 24 Feb 2022 16:50:54 +0000 (16:50 +0000)]
[SVE] Add missing splat patterns for bfloat vectors.
Differential Revision: https://reviews.llvm.org/D120496
Nikita Popov [Fri, 25 Feb 2022 16:36:29 +0000 (17:36 +0100)]
[MergeICmps] Don't require GEP
With opaque pointers, the zero-offset load will generally not use
a GEP. Allow a direct load without GEP, which is treated the same
way as a zero-offset GEP.
Stefan Pintilie [Fri, 25 Feb 2022 16:19:19 +0000 (10:19 -0600)]
[PowerPC][NFC] Add file info and license that was missing from this file.
Added the license info as well as description about how classes should be named
based on existing documentation.
Reviewed By: lei, #powerpc
Differential Revision: https://reviews.llvm.org/D120530
Evgeny Shulgin [Fri, 25 Feb 2022 15:55:08 +0000 (16:55 +0100)]
[Clang][Sema] Do not evaluate value-dependent immediate invocations
Value-dependent ConstantExprs are not meant to be evaluated.
There is an assert in Expr::EvaluateAsConstantExpr that ensures this condition.
But before this patch the method was called without prior check.
Fixes https://github.com/llvm/llvm-project/issues/52768
Reviewed By: erichkeane
Differential Revision: https://reviews.llvm.org/D119375
Paul Walker [Thu, 24 Feb 2022 16:19:59 +0000 (16:19 +0000)]
[NFC][SVE] Refactor SelectSVEAddSubImm to match SelectSVECpyDupImm.
They're equivalent other than one is signed and the other unsigned.
Paul Walker [Wed, 23 Feb 2022 18:46:21 +0000 (18:46 +0000)]
[SVE] Refactor complex immediate pattern used by CPY/DUP.
SelectSVE8BitLslImm didn't account for constant values that have a
larger bit width than the result vector's element type. This only
seems to affect a single corner case when lowering fixed length
vectors but the code itself is also not consistent with how other
related complex patterns are implemented so I've taken the
opportunity to refactor the code.
Differential Revision: https://reviews.llvm.org/D120440
Stefan Pintilie [Fri, 25 Feb 2022 15:26:25 +0000 (09:26 -0600)]
[libcxx] String format class marked as packed
This patch marks the class _Flags as packed because the design assumes that it
is packed and a number of tests also assume that it is packed. However on AIX
the class is not packed unless it is marked as such.
Reviewed By: hubert.reinterpretcast, #libc, Mordante, ldionne, Quuxplusone
Differential Revision: https://reviews.llvm.org/D119567
Kristóf Umann [Tue, 22 Feb 2022 14:30:47 +0000 (15:30 +0100)]
[analyzer] Don't crash if the analyzer-constraint is set to Z3, but llvm is not built with it
Exactly what it says on the tin! We had a nasty crash with the following incovation:
$ clang --analyze -Xclang -analyzer-constraints=z3 test.c
fatal error: error in backend: LLVM was not compiled with Z3 support, rebuild with -DLLVM_ENABLE_Z3_SOLVER=ON
... <stack trace> ...
Differential Revision: https://reviews.llvm.org/D120325
Bixia Zheng [Fri, 25 Feb 2022 00:24:29 +0000 (16:24 -0800)]
[mlir][sparse][taco] Use np.array_equal to compare integer values.
Fix MLIR-PyTACO and some tests to use np.array_equal to compare integer
values.
Reviewed By: aartbik
Differential Revision: https://reviews.llvm.org/D120526
Florian Hahn [Fri, 25 Feb 2022 15:36:31 +0000 (15:36 +0000)]
[AArch64] Add test cases where zext can be lowered to series of tbl.
Add a set of tests for upcoming patches that allow lowering vector zext
using AArch64 tbl instructions instead of shifts.
gysit [Fri, 25 Feb 2022 15:11:52 +0000 (15:11 +0000)]
[mlir][OpDSL] Split arithmetic functions.
Split arithmetic function into unary and binary functions. The revision prepares the introduction of unary and binary function attributes that work similar to type function attributes.
Depends On D120108
Reviewed By: aartbik
Differential Revision: https://reviews.llvm.org/D120109
Whisperity [Fri, 25 Feb 2022 13:21:18 +0000 (14:21 +0100)]
[clang-tidy] Fix `readability-suspicious-call-argument` crash for arguments without name-like identifier
As originally reported by @steakhal in
http://github.com/llvm/llvm-project/issues/54074, the name extraction logic of
`readability-suspicious-call-argument` crashes if the argument passed to a
function was a function call to a non-trivially named entity (e.g. an operator).
Fixed this crash case by ignoring such constructs and considering them as having
no name.
Reviewed By: aaron.ballman, steakhal
Differential Revision: http://reviews.llvm.org/D120555
Bixia Zheng [Thu, 24 Feb 2022 23:26:36 +0000 (15:26 -0800)]
[mlir][sparse][taco] Add support for scalar tensors.
This change allows the use of scalar tensors with index 0 in tensor index
expressions. In this case, the scalar value is broadcast to match the
dimensions of other tensors in the same expression.
Using scalar tensors as a destination in tensor index expressions is not
supported in the PyTACO DSL.
Add a PyTACO test to show the use of scalar tensors.
Reviewed By: aartbik
Differential Revision: https://reviews.llvm.org/D120524
Sam Clegg [Thu, 24 Feb 2022 06:23:05 +0000 (22:23 -0800)]
[MC][WebAssembly] Fix crash when relocation addend underlows U32
For the object file writer we need to allow the underflow (ar write
zero), but for the final linker output we should probably generate an
error (I've left that as a TODO for now).
Fixes: https://github.com/llvm/llvm-project/issues/54012
Differential Revision: https://reviews.llvm.org/D120522
Sam Clegg [Thu, 24 Feb 2022 22:37:40 +0000 (14:37 -0800)]
[WebAssembly] Covert llvm/test/MC/WebAssembly/reloc-code.ll to asm. NFC
Also increase coverage of call_indirect via explict function table
(enabled when reference types is enabled) in
llvm/test/CodeGen/WebAssembly/call-indirect.ll (I believe this
was an oversight that it was not added in https://reviews.llvm.org/D90948)
Differential Revision: https://reviews.llvm.org/D120521
gysit [Fri, 25 Feb 2022 15:04:38 +0000 (15:04 +0000)]
[mlir][OpDSL] Refactor function handling.
Prepare the OpDSL function handling to introduce more function classes. A follow up commit will split ArithFn into UnaryFn and BinaryFn. This revision prepares the split by adding a function kind enum to handle different function types using a single class on the various levels of the stack (for example, there is now one TensorFn and one ScalarFn).
Depends On D119718
Reviewed By: aartbik
Differential Revision: https://reviews.llvm.org/D120108
Stefan Pintilie [Fri, 25 Feb 2022 14:48:58 +0000 (08:48 -0600)]
[libcxx] Fix the type in __estimate_column_width
It seems that we are using wchar_t in __estimate_column_width and assume that
it is a 32 bit type. However, on AIX 32 the size of wchar_t is only 16 bits.
Changed wchar_t to uint32_t since the variable is being passed to a function
that uses uint32_t anyway.
Reviewed By: hubert.reinterpretcast, daltenty, Mordante, #libc, Quuxplusone
Differential Revision: https://reviews.llvm.org/D119770
LLVM GN Syncbot [Fri, 25 Feb 2022 14:48:36 +0000 (14:48 +0000)]
[gn build] Port
53dcd9efd16f
Stanislav Gatev [Mon, 21 Feb 2022 23:43:55 +0000 (23:43 +0000)]
[clang][dataflow] Add SAT solver interface and implementation
This is part of the implementation of the dataflow analysis framework.
See "[RFC] A dataflow analysis framework for Clang AST" on cfe-dev.
Reviewed-by: ymandel, xazax.hun
Differential Revision: https://reviews.llvm.org/D120289
Simon Pilgrim [Fri, 25 Feb 2022 14:42:31 +0000 (14:42 +0000)]
[IPO] AAFunctionReachabilityFunction.updateImpl - reduce AAReachability scope. NFCI.
We already have a check for !InstQueries.empty(), so move the for-range over InstQueries inside to avoid the AAReachability uninitialized variable static analysis warnings.
Aaron Ballman [Fri, 25 Feb 2022 14:07:32 +0000 (09:07 -0500)]
Use function prototypes when appropriate; NFC
This prepares the clang-tools-extra project for -Wstrict-prototypes
being enabled by default.
Arjun P [Fri, 25 Feb 2022 13:48:49 +0000 (13:48 +0000)]
[MLIR][Presburger] coalesce fixups: inline comments /// -> //, i++ -> ++i (NFC)
Also use empty() instead of size() == 0.
Simon Pilgrim [Fri, 25 Feb 2022 13:52:34 +0000 (13:52 +0000)]
[X86] Regenerate x86-cmov-converter.ll checks
Pawe Bylica [Fri, 25 Feb 2022 13:17:42 +0000 (14:17 +0100)]
[X86] Combine ADC(ADD(X,Y),0,Carry) -> ADC(X,Y,Carry)
Reviewed By: RKSimon
Differential Revision: https://reviews.llvm.org/D120435
Xiaodong Liu [Fri, 25 Feb 2022 13:08:47 +0000 (21:08 +0800)]
[sanitizer] Disable pc guard coverage test on PPC64/s390x
Reviewed By: benshi001, uweigand
Differential Revision: https://reviews.llvm.org/D120541
Michael Forster [Fri, 25 Feb 2022 13:07:49 +0000 (13:07 +0000)]
Revert "[lldb/test] Fix TestProgressReporting.py race issue with the event listener"
This reverts commit
3e3e79a9e4c378b59f5f393f556e6a84edcd8898.
MemorySanitizer: use-of-uninitialized-value
Michel Weber [Fri, 25 Feb 2022 12:22:48 +0000 (12:22 +0000)]
[MLIR][Presburger] Refactor looping strategy in coalesce
This patch refactors the looping strategy of coalesce for future patches. The new strategy works in-place and uses IneqType to organize inequalities into vectors of the same type. Future coalesce cases will pattern match on this organization. E.g. the contained case needs all inequalities and equalities to be redundant, so this case becomes checking whether the respective vectors are empty. For other cases, the patterns consider the types of all inequalities of both sets making it wasteful to only consider whether a can be coalesced with b in one step, as inequalities would need to be typed again for the opposite case. Therefore, the new strategy tries to coalesce a with b and b with a in a single step.
Reviewed By: Groverkss, arjunp
Differential Revision: https://reviews.llvm.org/D120392
Haocong.Lu [Fri, 25 Feb 2022 12:37:07 +0000 (12:37 +0000)]
[RISCV] Fix a mistake in PostprocessISelDAG
With the condition N->use_empty(), the root node of DAG always
misses peephole optimization. So a dummy node is needed.
Reviewed By: craig.topper
Differential Revision: https://reviews.llvm.org/D119934
LLVM GN Syncbot [Fri, 25 Feb 2022 12:21:23 +0000 (12:21 +0000)]
[gn build] Port
f9e8e92cf586
Aaron Ballman [Fri, 25 Feb 2022 12:15:01 +0000 (07:15 -0500)]
Revert "[clang][analyzer] Add modeling of 'errno'."
This reverts commit
29b512ba322cb6dd2c45d5e07645e20db47fad0d.
This broke several build bots:
https://lab.llvm.org/buildbot/#/builders/86/builds/30183
https://lab.llvm.org/buildbot/#/builders/216/builds/488
Nikita Popov [Fri, 25 Feb 2022 12:18:45 +0000 (13:18 +0100)]
[MergeICmps] Add opaque pointer test (NFC)
We fail to merge the icmps here, because the zero-offset access
does not go through a GEP.
Muhammad Omair Javaid [Fri, 25 Feb 2022 12:08:03 +0000 (17:08 +0500)]
[LLDB] XFAIL TestUnambiguousTailCalls.py for Arm/Linux
This patch marks TestUnambiguousTailCalls.py as XFAIL on Arm/Linux.
Test started failing after
3c4ed02698afec021c6bca80740d1e58e3ee019e.
Differential Revision: https://reviews.llvm.org/D120305
Groverkss [Fri, 25 Feb 2022 12:05:09 +0000 (17:35 +0530)]
[MLIR][Presburger] Use Matrix utilities for IntegerPolyhedron
This patch replaces various functions over inequalities/equalities in
IntegerPolyhedron with Matrix functions already implementing them or refactors
them to a Matrix function.
Reviewed By: arjunp
Differential Revision: https://reviews.llvm.org/D120482
LLVM GN Syncbot [Fri, 25 Feb 2022 11:47:20 +0000 (11:47 +0000)]
[gn build] Port
29b512ba322c
Balázs Kéri [Fri, 25 Feb 2022 09:15:06 +0000 (10:15 +0100)]
[clang][analyzer] Add modeling of 'errno'.
Add a checker to maintain the system-defined value 'errno'.
The value is supposed to be set in the future by existing or
new checkers that evaluate errno-modifying function calls.
Reviewed By: NoQ, steakhal
Differential Revision: https://reviews.llvm.org/D120310
Paul Walker [Tue, 22 Feb 2022 18:06:47 +0000 (18:06 +0000)]
[SVE] Don't custom lower constant predicate ISD:SPLAT_VECTOR operations.
Differential Revision: https://reviews.llvm.org/D120340
Momchil Velikov [Fri, 25 Feb 2022 10:50:02 +0000 (10:50 +0000)]
[AArch64] Async unwind - Refactor generation of shadow call stack prologue/epilogue
This patch is in preparation for the async unwind CFI.
Move the emission of the shadow call stack prologue/epilogue
instructions to the `emitPrologue`/`emitEpilogue`. This greatly
simplifies especially epilogue generation and makes unnecessary some
quite fragile code, that tries to skip over those
Reviewed By: MaskRay, efriedma
Differential Revision: https://reviews.llvm.org/D112329
Nikita Popov [Fri, 25 Feb 2022 11:07:49 +0000 (12:07 +0100)]
[IndVars] Use phis() (NFC)
Sven van Haastregt [Fri, 25 Feb 2022 11:05:56 +0000 (11:05 +0000)]
[OpenCL] opencl-c.h: Fix incorrect get_image_width guard
`cl_khr_3d_image_writes` should not guard `read_only image3d_t`.
Muhammad Omair Javaid [Fri, 25 Feb 2022 10:52:55 +0000 (15:52 +0500)]
[LLDB] Remove XFAIL from minidebuginfo-set-and-hit-breakpoint.test
This patch removes XFAIL from minidebuginfo-set-and-hit-breakpoint.test.
Test started passing after
3c4ed02698afec021c6bca80740d1e58e3ee019e.
Differential Revision: https://reviews.llvm.org/D120305
Groverkss [Fri, 25 Feb 2022 10:31:23 +0000 (16:01 +0530)]
[MLIR][Presburger] Move Presburger/ files to presburger namespace
This patch moves the Presburger library to a new `presburger` namespace.
This allows to shorten some names, helps to avoid polluting the mlir namespace,
and also provides some structure.
Reviewed By: arjunp
Differential Revision: https://reviews.llvm.org/D120505
Nikita Popov [Thu, 24 Feb 2022 12:17:55 +0000 (13:17 +0100)]
[InstCombine] Remove SPF min/max canonicalization
Now that we canonicalize SPF min/max to intrinsics, there's no
need to canonicalize the structure of the SPF min/max itself
anymore. This is conceptually NFC, but in practice does slightly
impact results due to folding order differences.
Groverkss [Fri, 25 Feb 2022 10:10:20 +0000 (15:40 +0530)]
[MLIR][Presburger][NFC] Refactor redundant code in fourierMotzkinEliminate
This patch removes redundant code from fourierMotzkinEliminate implementation
using existing functions in IntegerPolyhedron.
Reviewed By: arjunp
Differential Revision: https://reviews.llvm.org/D120502
Andrzej Warzynski [Mon, 21 Feb 2022 11:51:32 +0000 (11:51 +0000)]
[flang][driver] Add support for `--target`/`--triple`
This patch adds support for:
* `--target` in the compiler driver (`flang-new`)
* `--triple` in the frontend driver (`flang-new -fc1`)
The semantics of these flags are inherited from `clangDriver`, i.e.
consistent with `clang --target` and `clang -cc1 --triple`,
respectively.
A new structure is defined, `TargetOptions`, that will hold various
Frontend options related to the target. Currently, this is mostly a
placeholder that contains the target triple. In the future, it will be
used for storing e.g. the CPU to tune for or the target features to
enable.
Additionally, the following target/triple related options are enabled
[*]: `-print-effective-triple`, `-print-target-triple`. Definitions in
Options.td are updated accordingly and, to facilated testing,
`-emit-llvm` is added to the list of options available in `flang-new`
(previously it was only enabled in `flang-new -fc1`).
[*] These options were actually available before (like all other options
defined in `clangDriver`), but not included in `flang-new --help`.
Before this change, `flang-new` would just use `native` for defining the
target, so these options were of little value.
Differential Revision: https://reviews.llvm.org/D120246
Iain Sandoe [Sun, 7 Feb 2021 17:08:42 +0000 (17:08 +0000)]
[C++20][Modules][4/8] Handle generation of partition implementation CMIs.
Partition implementations are special, they generate a CMI, but it
does not have an 'export' line, and we cannot export anything from the
it [that is it can only make decls available to other members of the
owning module, not to importers of that].
Add initial testcases for partition handling, derived from the examples in
Section 10 of the C++20 standard, which identifies what should be accepted
and/or rejected.
Differential Revision: https://reviews.llvm.org/D118587
Hsiangkai Wang [Fri, 25 Feb 2022 06:58:31 +0000 (06:58 +0000)]
[NFC][AArch64][SME] Remove '#' prefix in PSEL test cases
There is no need to prefix '#' for element index in PSEL.
This is a follow up of D111213.
Differential Revision: https://reviews.llvm.org/D120543
Nikita Popov [Fri, 25 Feb 2022 09:09:16 +0000 (10:09 +0100)]
[SCEVExpander] Use early returns in FindValueInExprValueMap() (NFC)
Nikita Popov [Fri, 25 Feb 2022 08:59:50 +0000 (09:59 +0100)]
[IR] Use CallBase::getParamElementType() (NFC)
As this method now exists on CallBase, use it rather than the
one on AttributeList.
Simon Pilgrim [Fri, 25 Feb 2022 08:59:47 +0000 (08:59 +0000)]
Revert rG87753cebf5f861eee418d6bce155dfa0b00f9878 "[X86] combineX86ShufflesRecursively - don't both widening inputs before calling combineX86ShuffleChain"
Reverting while we investigate codegen regression reports
Jason Molenda [Fri, 25 Feb 2022 08:55:54 +0000 (00:55 -0800)]
PlatformMacOSX should be activated for lldb built to run on an iOS etc device
In the changes Jonas made in https://reviews.llvm.org/D117340 , a
small oversight was that PlatformMacOSX (despite the name) is active
for any native Darwin operating system, where lldb and the target
process are running on the same system. This patch uses compile-time
checks to return the appropriate OSType for the OS lldb is being
compiled to, so the "host" platform will correctly be selected when
lldb & the inferior are both running on that OS. And a small change
to PlatformMacOSX::GetSupportedArchitectures which adds additional
recognized triples when running on macOS but not other native Darwin
systems.
Differential Revision: https://reviews.llvm.org/D120517
rdar://
89247060
Nikita Popov [Fri, 25 Feb 2022 08:30:28 +0000 (09:30 +0100)]
[SCEV] Return ArrayRef from getSCEVValues() (NFC)
Return a read-only view on this set. For the one internal use,
directly access ExprValueMap.
gysit [Fri, 25 Feb 2022 08:12:34 +0000 (08:12 +0000)]
[mlir][OpDSL] Add type function attributes.
Previously, OpDSL operation used hardcoded type conversion operations (cast or cast_unsigned). Supporting signed and unsigned casts thus meant implementing two different operations. Type function attributes allow us to define a single operation that has a cast type function attribute which at operation instantiation time may be set to cast or cast_unsigned. We may for example, defina a matmul operation with a cast argument:
```
@linalg_structured_op
def matmul(A=TensorDef(T1, S.M, S.K), B=TensorDef(T2, S.K, S.N), C=TensorDef(U, S.M, S.N, output=True),
cast=TypeFnAttrDef(default=TypeFn.cast)):
C[D.m, D.n] += cast(U, A[D.m, D.k]) * cast(U, B[D.k, D.n])
```
When instantiating the operation the attribute may be set to the desired cast function:
```
linalg.matmul(lhs, rhs, outs=[out], cast=TypeFn.cast_unsigned)
```
The revsion introduces a enum in the Linalg dialect that maps one-by-one to the type functions defined by OpDSL.
Reviewed By: aartbik
Differential Revision: https://reviews.llvm.org/D119718
Kristina Bessonova [Fri, 25 Feb 2022 08:17:40 +0000 (10:17 +0200)]
[NVPTX][AsmPrinter] Emit .attribute(.managed) for global variable declarations
Declaration and definition attributes must match,
otherwise it may cause issues on linking.
Reviewed By: tra
Differential Revision: https://reviews.llvm.org/D120493
Nikita Popov [Tue, 22 Feb 2022 09:47:15 +0000 (10:47 +0100)]
[SCEV] Don't try to reuse expressions with offset
SCEVs ExprValueMap currently tracks not only which IR Values
correspond to a given SCEV expression, but additionally stores that
it may be expanded in the form X+Offset. In theory, this allows
reusing existing IR Values in more cases.
In practice, this doesn't seem to be particularly useful (the test
changes are rather underwhelming) and adds a good bit of complexity.
Per https://github.com/llvm/llvm-project/issues/53905, we have an
invalidation issue with these offseted expressions.
Differential Revision: https://reviews.llvm.org/D120311
Neumann Hon [Fri, 25 Feb 2022 07:37:05 +0000 (02:37 -0500)]
[SystemZ] [z/OS] Add support for generating huge (1 MiB) stack frames in XPLINK64
This patch extends support for generating huge stack frames on 64-bit XPLINK by implementing the ABI-mandated call to the stack extension routine.
Reviewed By: uweigand
Differential Revision: https://reviews.llvm.org/D120450
Michel Weber [Fri, 25 Feb 2022 07:32:22 +0000 (07:32 +0000)]
[MLIR][Presburger] enable copy assignment operator for Simplex
This patch removes the `const` from `usingBigM` to enable the implicit copy assignment operator for Simplex.
Reviewed By: Groverkss
Differential Revision: https://reviews.llvm.org/D120542
Sockke [Fri, 25 Feb 2022 06:44:26 +0000 (14:44 +0800)]
[clang-tidy] Fix `readability-non-const-parameter` for parameter referenced by an lvalue
The checker missed a check for a case when the parameter is referenced by an lvalue and this could cause build breakages.
Reviewed By: aaron.ballman
Differential Revision: https://reviews.llvm.org/D117090
Chenbing Zheng [Fri, 25 Feb 2022 06:44:25 +0000 (14:44 +0800)]
[RISCV] DAG Combine vcpop and vfirst with VL=0 to li imm
vcpop and vfirst are still useful when VL=0.
vcpop equivalents to li 0 and vfirst equivalents to li -1,
since no mask elements are active.
Reviewed By: craig.topper
Differential Revision: https://reviews.llvm.org/D120302
Zakk Chen [Fri, 25 Feb 2022 03:04:24 +0000 (19:04 -0800)]
[RISCV] Update computeTargetABI from llc as well as clang
Clang computes the default ABI if -mabi is empty
and encode it in LLVM IR module flag since D105555.
For correctness, llc need to give the same target-abi
(Options.MCOptions.ABIName) with ABI encoded in IR.
The getSubtargetImpl already has a check for them only if
Options.MCOptions.ABIName is not empty.
In order to get more robustness we could have a check for
explicit ABI, but now we have two different logic to
compute the default ABI.
The front-end ABI is defautl to the ilp32/ilp32e/lp64, and
ilp32d/lp64d when hardware support for extension D.
The backend ABI is default to the ilp32/ilp32e/lp64.
Reviewed by: asb, jrtc27
Differential Revision: https://reviews.llvm.org/D118333
Anton Afanasyev [Fri, 25 Feb 2022 05:04:11 +0000 (08:04 +0300)]
[AggressiveInstCombine] Fix `TruncInstCombine` (fix
f84d732f)
Erase phi-nodes from `InstInfoMap` before erasing themselves
Anton Afanasyev [Wed, 15 Sep 2021 06:19:53 +0000 (09:19 +0300)]
[AggressiveInstCombine] Add `phi` nodes support to `TruncInstCombine`
Expand `TruncInstCombine` to handle loops by adding `phi` nodes
to expression graph.
Reviewed by: RKSimon, lebedev.ri
(recommit of fixed
f84d732f, reverted by 8ad6d5e after sanitizer breakage)
Differential Revision: https://reviews.llvm.org/D109817
Shangwu Yao [Fri, 25 Feb 2022 04:51:43 +0000 (20:51 -0800)]
[CUDA][SPIRV] Assign global address space to CUDA kernel arguments
(resubmit https://reviews.llvm.org/D119207 after fixing the test for
some build settings)
This patch converts CUDA pointer kernel arguments with default address
space to CrossWorkGroup address space (__global in OpenCL). This is
because Generic or Function (OpenCL's private) is not supported as
storage class for kernel pointer types.
Differential revision: https://reviews.llvm.org/D120366
Raúl Peñacoba [Fri, 25 Feb 2022 04:41:02 +0000 (04:41 +0000)]
[Driver] Support GCC detection for GCC compiled with --enable-version-specific-runtime-libs
GCC's compiled with --enable-version-specific-runtime-libs change the paths where includes and libs are found.
This patch adds support for these cases
Reviewed By: MaskRay
Differential Revision: https://reviews.llvm.org/D118700
Vitaly Buka [Fri, 25 Feb 2022 04:18:56 +0000 (20:18 -0800)]
[sanitizer] Don't collect unused info
Fangrui Song [Fri, 25 Feb 2022 04:05:59 +0000 (20:05 -0800)]
[ELF] Parallelize initializeLocalSymbols
ObjFile::parse combines symbol initialization and resolution. Many tasks
unrelated to symbol resolution can be postponed and parallelized. This patch
extracts local symbol initialization and parallelizes it.
Technically the new function initializeLocalSymbols can be merged into
ObjFile::postParse, but functions like getSrcMsg may access the
uninitialized (all nullptr) local part of InputFile::symbols.
Linking chrome: 1.02x as fast with glibc malloc, 1.04x as fast with mimalloc
Reviewed By: ikudrin
Differential Revision: https://reviews.llvm.org/D119909
Vasileios Porpodas [Thu, 24 Feb 2022 02:04:36 +0000 (18:04 -0800)]
[SLP] Fix for the min/max intrinsic cost.
The min/max intrinsic cost is currently too low because in the cost calculation
we subtract the cost of the vector compare as we will not emit it.
For the cost of the vector compare we are currently passing BAD_ICMP_PREDICATE
which returns 3, the worst case cost.
I think we should be passing VecPred instead, since we know the predicates of
the compare instr.
I think this is related to commit
b3b993a7ad817 which introduced the predicate
argument to getCmpSelInstrCost().
https://reviews.llvm.org/rGb3b993a7ad817c3c5801341fa78f34332900eb83
Differential Revision: https://reviews.llvm.org/D120439
Vasileios Porpodas [Thu, 24 Feb 2022 02:03:20 +0000 (18:03 -0800)]
[SLP][NFC] Test for a follow-up fix of the the vector min/max instrinsic cost calculation.
The code in this test should not have been vectorized.
It looks worse than the scalar code.
Differential Revision: https://reviews.llvm.org/D120438
lian wang [Wed, 16 Feb 2022 01:59:50 +0000 (01:59 +0000)]
[RISCV] Add schedule class for Zbt extension
Reviewed By: craig.topper
Differential Revision: https://reviews.llvm.org/D119808
Xiaodong Liu [Fri, 25 Feb 2022 01:42:19 +0000 (09:42 +0800)]
[sanitizer] Enable trace pc guard coverage test on PPC64/s390x/MIPS
Reviewed By: MaskRay
Differential Revision: https://reviews.llvm.org/D120472
Carl Ritson [Fri, 25 Feb 2022 00:42:55 +0000 (09:42 +0900)]
[AMDGPU] Extend pre-emit peephole to redundantly masked VCC
Extend pre-emit peephole for S_CBRANCH_VCC[N]Z to eliminate
redundant S_AND operations against EXEC for V_CMP results in VCC.
These occur after after register allocation when VCC has been
selected as the comparison destination.
Reviewed By: rampitec
Differential Revision: https://reviews.llvm.org/D120202
Harald van Dijk [Fri, 25 Feb 2022 00:58:50 +0000 (00:58 +0000)]
[ADT, CSSPGO] Specify set comparer
In _GLIBCXX_DEBUG builds, potentially implicitly enabled by
LLVM_ENABLE_EXPENSIVE_CHECKS, std::set<A, B>::iterator and
std::set<A, C>::iterator are distinct types that are not
interconvertible. This change aligns the iterator types with the set
types.
Reviewed By: hoy
Differential Revision: https://reviews.llvm.org/D119798
Corentin Jabot [Fri, 25 Feb 2022 00:54:00 +0000 (19:54 -0500)]
[libc++] Make sure calls to std::move are always qualified
This fixes instances of the newly added `-Wunqualified-std-cast-call`.
(Commit
7853371146 removed unqualified `move` from the tests,
but these unqualified `move`s remained undetected in the actual headers.)
Differential Revision: https://reviews.llvm.org/D120509
Arthur Eubanks [Fri, 25 Feb 2022 00:54:18 +0000 (16:54 -0800)]
[OpaquePtr][AArch64] Use load/store value type instead of pointer type for ldnt1/stnt1 alignment
Daniel Thornburgh [Tue, 15 Feb 2022 22:21:14 +0000 (22:21 +0000)]
[Debuginfod] Add BUILD_ID syntax to llvm-symbolizer.
This adds a BUILD_ID prefix to the llvm-symbolizer stdin and argument
syntax. The prefix causes the given binary name to be interpreted as a
build ID instead of an object file path. The semantics are analagous to
the behavior of --obj and --build-id.
Reviewed By: jhenderson
Differential Revision: https://reviews.llvm.org/D119901
Qihan Cai [Thu, 24 Feb 2022 22:13:51 +0000 (09:13 +1100)]
[RISCV] Change rvv version to 1.0 and remove ratify notice
This patch changes the version of V extension from 0.1 to 1.0 in RISCVInstrInfoVPseudos.td, RISCVInstrInfoVSDPatterns.td, RISCVInstrInfoVVLPatterns.td, RISCVInstrInfoV.td
Reviewed By: craig.topper
Differential Revision: https://reviews.llvm.org/D120525
Fangrui Song [Fri, 25 Feb 2022 00:34:05 +0000 (00:34 +0000)]
[Driver][test] Make linux-ld.c work with CLANG_DEFAULT_PIE_ON_LINUX=on
Daniel Thornburgh [Tue, 8 Feb 2022 22:07:14 +0000 (22:07 +0000)]
[Symbolize] LRU cache binaries in llvm-symbolizer.
This change adds a simple LRU cache to the Symbolize class to put a cap
on llvm-symbolizer memory usage. Previously, the Symbolizer's virtual
memory footprint would grow without bound as additional binaries were
referenced.
I'm putting this out there early for an informal review, since there may be
a dramatically different/better way to go about this. I still need to
figure out a good default constant for the memory cap and benchmark the
implementation against a large symbolization workload. Right now I've
pegged max memory usage at zero for testing purposes, which evicts the whole
cache every time.
Unfortunately, it looks like StringRefs in the returned DI objects can
directly refer to the contents of binaries. Accordingly, the cache
pruning must be explicitly requested by the caller, as the caller must
guarantee that none of the returned objects will be used afterwards.
For llvm-symbolizer this a light burden; symbolization occurs
line-by-line, and the returned objects are discarded after each.
Implementation wise, there are a number of nested caches that depend
on one another. I've implemented a simple Evictor callback system to
allow derived caches to register eviction actions to occur when the
underlying binaries are evicted.
Reviewed By: dblaikie
Differential Revision: https://reviews.llvm.org/D119784
Ben Shi [Thu, 24 Feb 2022 11:23:05 +0000 (11:23 +0000)]
[clang][NFC] Move all avr CodeGen tests to avr specific directory
Reviewed By: MaskRay
Differential Revision: https://reviews.llvm.org/D120468