Jay Foad [Thu, 11 Jun 2020 16:42:30 +0000 (17:42 +0100)]
SILoadStoreOptimizer: add support for GFX10 image instructions
GFX10 image instructions use one or more address operands starting at
vaddr0, instead of a single vaddr operand, to allow for NSA forms.
Differential Revision: https://reviews.llvm.org/D81675
Jay Foad [Mon, 22 Jun 2020 16:17:17 +0000 (17:17 +0100)]
[AMDGPU] Fix and simplify AMDGPULegalizerInfo::legalizeUDIV_UREM32Impl
Use the algorithm from AMDGPUCodeGenPrepare::expandDivRem32.
Differential Revision: https://reviews.llvm.org/D83383
Jay Foad [Mon, 22 Jun 2020 16:11:58 +0000 (17:11 +0100)]
[AMDGPU] Fix and simplify AMDGPUTargetLowering::LowerUDIVREM
Use the algorithm from AMDGPUCodeGenPrepare::expandDivRem32.
Differential Revision: https://reviews.llvm.org/D83382
Jay Foad [Mon, 22 Jun 2020 14:27:37 +0000 (15:27 +0100)]
[AMDGPU] Fix and simplify AMDGPUCodeGenPrepare::expandDivRem32
Fix the division/remainder algorithm by adding a second quotient
refinement step, which is required in some cases like
0xFFFFFFFFu / 0x11111111u (https://bugs.llvm.org/show_bug.cgi?id=46212).
Also document, rewrite and simplify it by ensuring that we always have a
lower bound on inv(y), which simplifies the UNR step and the quotient
refinement steps.
Differential Revision: https://reviews.llvm.org/D83381
Christopher Tetreault [Wed, 8 Jul 2020 18:00:57 +0000 (11:00 -0700)]
[SVE] Remove calls to VectorType::getNumElements from Scalar
Reviewers: efriedma, fhahn, reames, kmclaughlin, sdesmalen
Reviewed By: sdesmalen
Subscribers: tschuett, hiraditya, rkruppe, psnobl, dantrushin, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D82243
Louis Dionne [Wed, 8 Jul 2020 18:04:45 +0000 (14:04 -0400)]
[libc++] Reimplement platform detection features without running on the test host
It's sufficient to sniff the platform we're running on using the compiler
macros -- we don't need to run any code.
Fangrui Song [Wed, 8 Jul 2020 17:53:30 +0000 (10:53 -0700)]
[test] Add REQUIRES: x86-registered-target to tools/obj2yaml/COFF/bss.s
Fangrui Song [Wed, 8 Jul 2020 17:50:53 +0000 (10:50 -0700)]
[Support] Define llvm::parallel::strategy for -DLLVM_ENABLE_THREADS=off builds after D76885
Michael Park [Wed, 8 Jul 2020 17:46:02 +0000 (10:46 -0700)]
Remove the unnecessary `is_nothrow_swappable` condition in `swap`.
Thanks to @lewissbaker who pointed out the unnecessary condition in
https://reviews.llvm.org/D81954#inline-756872. Since this codepath does not
make use of `swap` anyway (that codepath is a different branch), we can safely
remove this condition and produce better codegen when all types are nothrow
movable but are potentially-throwing swappable.
See codegen in https://gcc.godbolt.org/z/uDFZjz
Reviewed By: ldionne, #libc
Differential Revision: https://reviews.llvm.org/D83274
Craig Topper [Wed, 8 Jul 2020 17:39:56 +0000 (10:39 -0700)]
[X86] Enabled a bunch of 64-bit Interlocked* functions intrinsics on 32-bit Windows to match recent MSVC
This enables _InterlockedAnd64/_InterlockedOr64/_InterlockedXor64/_InterlockedDecrement64/_InterlockedIncrement64/_InterlockedExchange64/_InterlockedExchangeAdd64/_InterlockedExchangeSub64 on 32-bit Windows
The backend already knows how to expand these to a loop using cmpxchg8b on 32-bit targets.
Fixes PR46595
Differential Revision: https://reviews.llvm.org/D83254
Craig Topper [Wed, 8 Jul 2020 17:17:44 +0000 (10:17 -0700)]
hwasan: Don't pass the tagged-globals target-feature to non-aarch64 backends.
The other backends don't know what this feature is and print a
message to stderr.
I recently tried to rework some target feature stuff in X86 and
this unknown feature tripped an assert I added.
Differential Revision: https://reviews.llvm.org/D83369
Fangrui Song [Wed, 8 Jul 2020 17:27:31 +0000 (10:27 -0700)]
[ELF] Rename canRelax to toExecRelax. NFC
In the absence of TLS relaxation (rewrite of code sequences),
there is still an applicable optimization:
[gd]: General Dynamic: resolve DTPMOD to 1 and/or resolve DTPOFF statically
All the other relaxations are only performed when transiting to
executable (`!config->shared`).
Since [gd] is handled differently, we can fold `!config->shared` into canRelax
and simplify its use sites. Rename the variable to reflect to new semantics.
Reviewed By: grimar, psmith
Differential Revision: https://reviews.llvm.org/D83243
Fangrui Song [Wed, 8 Jul 2020 17:22:49 +0000 (10:22 -0700)]
[test] Run llvm/test/**/*.yaml & don't run llvm/test/**/*.cxx (not exist)
This patch extends D58439 (`llvm/test/{yaml2obj,obj2yaml}/**/*.yaml`) and runs all
`llvm/test/**/*.yaml`
Many directories have configured `.yaml` (see the deleted lit.local.cfg
files). Yet still some don't configure .yaml and have caused stale tests:
*
8c5825befb7bbb2e76f7eccedc6d3bf26e9b2a6a test/llvm-readobj
*
bdc3134e237737dd46b51cd1ecd41ecbbe9f921a test/ExecutionEngine
Just hoist .yaml to `llvm/test/lit.cfg.py`. Also delete .cxx which is
not used. The number of tests running on my machine increases from 38304 to 38309.
The list of new tests:
```
ExecutionEngine/RuntimeDyld/X86/ELF_x86-64_none.yaml
Object/archive-error-tmp.txt
tools/llvm-ar/coff-weak.yaml
tools/llvm-readobj/ELF/verneed-flags.yaml
tools/obj2yaml/COFF/bss.s
```
Reviewed By: grimar, jhenderson, rupprecht
Differential Revision: https://reviews.llvm.org/D83350
Fangrui Song [Wed, 8 Jul 2020 17:10:43 +0000 (10:10 -0700)]
[ELF] Add -z dead-reloc-in-nonalloc=<section_glob>=<value>
... to customize the tombstone value we use for an absolute relocation
referencing a discarded symbol. This can be used as a workaround when
some debug processing tool has trouble with current -1 tombstone value
(https://bugs.chromium.org/p/chromium/issues/detail?id=1102223#c11 )
For example, to get the current built-in rules (not considering the .debug_line special case for ICF):
```
-z dead-reloc-in-nonalloc='.debug_*=0xffffffffffffffff'
-z dead-reloc-in-nonalloc=.debug_loc=0xfffffffffffffffe
-z dead-reloc-in-nonalloc=.debug_ranges=0xfffffffffffffffe
```
To get GNU ld (as of binutils 2.35)'s behavior:
```
-z dead-reloc-in-nonalloc='*=0'
-z dead-reloc-in-nonalloc=.debug_ranges=1
```
This option has other use cases. For example, if we want to check
whether a non-SHF_ALLOC section has dead relocations.
With this patch, we can run a regular LLD and run another with a special
-z dead-reloc-in-nonalloc=, then compare their output.
Reviewed By: thakis
Differential Revision: https://reviews.llvm.org/D83264
Pete Steinfeld [Wed, 8 Jul 2020 17:05:04 +0000 (10:05 -0700)]
[flang] Fix a crash when cosubscript list is empty
Summary:
When there are errors in the evaluation of every cosubscript expression in a
coindexed object, the compiler would crash. I fixed this by just checking to
see if there were errors in the evaluation of the cosubscripts before
constructing the `DataRef` for the coindexed object.
Reviewers: klausler, tskeith, DavidTruby
Subscribers: llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D83410
Yaxun (Sam) Liu [Fri, 17 Apr 2020 03:51:40 +0000 (23:51 -0400)]
[CUDA][HIP] Let lambda be host device by default
This patch let lambda be host device by default and adds diagnostics for
capturing host variable by reference in device lambda.
Differential Revision: https://reviews.llvm.org/D78655
shafik [Wed, 8 Jul 2020 17:02:13 +0000 (10:02 -0700)]
[RecordLayout] Fix ItaniumRecordLayoutBuilder so that is grabs the correct bases class offsets from the external source
Currently the ItaniumRecordLayoutBuilder when laying out base classes has the virtual
and non-virtual bases mixed up when pulling the base class layouts from the external source.
This came up in an LLDB bug where on arm64 because of differences in how it deals with
tail padding would layout the bases differently without the correct layout from the
external source (LLDB). This would result in some fields being off by 4 bytes.
Differential Revision: https://reviews.llvm.org/D83008
Louis Dionne [Wed, 8 Jul 2020 17:04:17 +0000 (13:04 -0400)]
[libc++][NFC] Remove outdated TODO item
Sanjay Patel [Wed, 8 Jul 2020 15:57:45 +0000 (11:57 -0400)]
[DAGCombiner] fix function-name formatting; NFC
Sanjay Patel [Wed, 8 Jul 2020 15:48:41 +0000 (11:48 -0400)]
[DAGCombiner] add enum for store source value; NFC
This removes existing code duplication and allows us to
assert that we are handling the expected cases.
We have a list of outstanding bugs that could benefit by
handling truncated source values, so that's a possible
addition going forward.
Simon Pilgrim [Wed, 8 Jul 2020 16:07:41 +0000 (17:07 +0100)]
[X86][SSE] Pull out PACK(SHUFFLE(),SHUFFLE()) folds into its own function. NFC.
Future patches will extend this so declutter combineVectorPack before we start.
Simon Pilgrim [Wed, 8 Jul 2020 15:54:47 +0000 (16:54 +0100)]
[X86] Fix copy+paste typo in combineVectorPack assert message. NFC.
Arthur Eubanks [Wed, 8 Jul 2020 16:27:57 +0000 (09:27 -0700)]
[NewPM] Add PredicateInfoPrinterPass to PassRegistry.def
Fixes tests under NPM in Transforms/Util/PredicateInfo.
Wei Mi [Tue, 30 Jun 2020 21:32:46 +0000 (14:32 -0700)]
[SampleFDO] Enable sample-profile-top-down-load and sample-profile-merge-inlinee
by default.
sample-profile-top-down-load is an internal option which can enable top-down
order of inlining and profile annotation in sample profile load pass. It was
found to be beneficial for better profile annotation.
Recently we found it could also solve some build time issue. Suppose function
A has many callsites in function B. In the last release binary where sample
profile was collected, the outline copy of A is large because there are many
other functions inlined into A. However although all the callsites calling A
in B are inlined, but every inlined body is small (A was inlined into B
before other functions are inlined into A), there is no build time issue in
last release.
In an optimized build using the sample profile collected from last release,
without top-down inlining, we saw a case that A got very large because of
inlining, and then multiple callsites of A got inlined into B, and that led
to a huge B which caused significant build time issue besides profile
annotation issue.
To solve that problem, the patch enables the flag
sample-profile-top-down-load by default. sample-profile-top-down-load can
have better performance when it is enabled together with
sample-profile-merge-inlinee so in this patch we also enable
sample-profile-merge-inlinee by default.
Differential Revision: https://reviews.llvm.org/D82919
Ulrich Weigand [Wed, 8 Jul 2020 16:17:03 +0000 (18:17 +0200)]
[SystemZ] Allow specifying integer registers as part of the address calculation
Revision
e1de2773a534957305d7a559c6d88c4b5ac354e2 provided support for
accepting integer registers in inline asm i.e.
__asm("lhi %r0, 5") -> lhi %r0, 5
__asm("lhi 0, 5") -> lhi 0,5
This patch aims to extend this support to instructions which compute
addresses as well. (i.e instructions of type BDMem and BD[X|R|V|L]Mem)
Author: anirudhp
Differential Revision: https://reviews.llvm.org/D83251
Nicolai Hähnle [Thu, 2 Jul 2020 18:36:30 +0000 (20:36 +0200)]
DomTree: remove explicit use of DomTreeNodeBase::iterator
Summary:
Almost all uses of these iterators, including implicit ones, really
only need the const variant (as it should be). The only exception is
in NewGVN, which changes the order of dominator tree child nodes.
Change-Id: I4b5bd71e32d71b0c67b03d4927d93fe9413726d4
Reviewers: arsenm, RKSimon, mehdi_amini, courbet, rriddle, aartbik
Subscribers: wdng, Prazek, hiraditya, kuhar, rogfer01, rriddle, jpienaar, shauheen, antiagainst, nicolasvasilache, arpith-jacob, mgester, lucyrfox, aartbik, liufengdb, stephenneuendorffer, Joonsoo, grosul1, vkmr, Kayjukh, jurahul, msifontes, cfe-commits, llvm-commits
Tags: #clang, #mlir, #llvm
Differential Revision: https://reviews.llvm.org/D83087
serge-sans-paille [Wed, 8 Jul 2020 16:13:59 +0000 (18:13 +0200)]
Revert "Double check that passes correctly set their Modified status"
This reverts commit
37afd99c768b29c7df7c5f2eb645362fb61f9915.
Sjoerd Meijer [Wed, 8 Jul 2020 16:04:23 +0000 (17:04 +0100)]
[matrix] Add some more Verifier negative tests. NFC.
Louis Dionne [Wed, 8 Jul 2020 15:09:03 +0000 (11:09 -0400)]
[libc++] Mark some tests as unsupported on GCC 5
There used to be a workaround where we'd pretend that GCC 5 didn't support
C++14 because it doesn't implement it properly. Since that workaround has
been removed (in
1eb211ada17a), we need to mark a few individual tests as
failing with GCC 5.
Arthur Eubanks [Wed, 8 Jul 2020 15:56:27 +0000 (08:56 -0700)]
[NewPM][opt] Translate "-O#" to NPM's "default<O#>"
Fixes 52 check-llvm tests under NPM.
Reviewed By: hans
Differential Revision: https://reviews.llvm.org/D83367
Aleksandr Platonov [Wed, 8 Jul 2020 15:46:30 +0000 (16:46 +0100)]
[clang] Fix a crash when passing a C structure of incompatible type to a function with a reference parameter.
__builtin_va_*() and __builtin_ms_va_*() are declared as functions with a
parameter of reference type.
This patch fixes a crash when using these functions in C where an argument
of structure type is incompatible with the parameter type.
Differential Revision: https://reviews.llvm.org/D82805
Reviewed By: riccibruno
Patch by: Aleksandr Platonov <platonov.aleksandr@huawei.com>
Chris Gyurgyik [Wed, 8 Jul 2020 15:56:21 +0000 (11:56 -0400)]
[libc] Fix typographical error in math_utils.h.
Arthur Eubanks [Wed, 8 Jul 2020 15:50:22 +0000 (08:50 -0700)]
[NewPM][opt] Share -disable-loop-unrolling between pass managers
There's no reason to introduce a new option for the NPM.
The various PGO options are shared in this manner.
Reviewed By: echristo
Differential Revision: https://reviews.llvm.org/D83368
Evgeny Leviant [Wed, 8 Jul 2020 15:50:00 +0000 (18:50 +0300)]
[MIR] Speedup parsing of function with large number of basic blocks
Patch eliminates string length calculation when lexing a token. Speedup can be up to
1000x.
Differential revision: https://reviews.llvm.org/D83389
Arthur Eubanks [Wed, 8 Jul 2020 15:48:44 +0000 (08:48 -0700)]
[Preallocated] Add @llvm.call.preallocated.teardown
This cleans up the stack allocated by a @llvm.call.preallocated.setup.
Should either call the teardown or the preallocated call to clean up the
stack. Calling both is UB.
Add LangRef.
Add verifier check that the token argument is a @llvm.call.preallocated.setup.
Reviewed By: efriedma
Differential Revision: https://reviews.llvm.org/D83354
Eric Schweitz [Wed, 8 Jul 2020 15:47:13 +0000 (08:47 -0700)]
partial revert on RTBuilder.h to avoid warnings.
Pavel Labath [Wed, 8 Jul 2020 15:33:58 +0000 (17:33 +0200)]
[lldb] Modernize/clean up ValueObject::GetChildMemberWithName
Paul Walker [Wed, 8 Jul 2020 15:12:22 +0000 (15:12 +0000)]
[SelectionDAG] Fix incorrect offset when expanding CONCAT_VECTORS.
ExpandVectorBuildThroughStack is also used for CONCAT_VECTORS.
However, when calculating the offsets for each of the operands we
incorrectly use the element size rather than actual size and thus
the stores overlap.
Differential Revision: https://reviews.llvm.org/D83303
Bruno Ricci [Wed, 8 Jul 2020 15:36:48 +0000 (16:36 +0100)]
[clang][NFC] Also test for serialization in test/AST/ast-dump-APValue-*
This does not actually exercise the serialization of APValue, but it
will at least prevent a regression in the future. NFC.
serge-sans-paille [Thu, 28 May 2020 10:40:17 +0000 (12:40 +0200)]
Double check that passes correctly set their Modified status
The approach is simple: if a pass reports that it's not modifying a
Function/Module, compute a loose hash of that Function/Module and compare it
with the original one. If we report no change but there's a hash change, then we
have an error.
This approach misses a lot of change but it's not super intrusive and can
detect most of the simple mistakes.
Differential Revision: https://reviews.llvm.org/D80916
Sam McCall [Wed, 8 Jul 2020 15:30:24 +0000 (17:30 +0200)]
[clangd] Improve serialization error messages. NFC
sstefan1 [Sun, 5 Jul 2020 13:24:36 +0000 (15:24 +0200)]
[OpenMPIRBuilder][Fix] Move llvm::omp::types to OpenMPIRBuilder.
Summary:
D82193 exposed a problem with global type definitions in
`OMPConstants.h`. This causes a race when running in thinLTO mode.
Types now live inside of OpenMPIRBuilder to prevent this from happening.
Reviewers: jdoerfert
Subscribers: yaxunl, hiraditya, guansong, dexonsmith, aaron.ballman, cfe-commits, llvm-commits
Tags: #clang, #llvm
Differential Revision: https://reviews.llvm.org/D83176
Stanislav Mekhanoshin [Fri, 19 Jun 2020 18:48:37 +0000 (11:48 -0700)]
SLP: honor requested max vector size merging PHIs
At the moment this place does not check maximum size set
by TTI and just creates a maximum possible vectors.
Differential Revision: https://reviews.llvm.org/D82227
Sam McCall [Mon, 6 Jul 2020 15:23:08 +0000 (17:23 +0200)]
[clangd] Enable reading config from files behind a flag
Reviewers: kadircet, hokein
Subscribers: ilya-biryukov, MaskRay, jkorous, arphaman, usaxena95, cfe-commits
Tags: #clang
Differential Revision: https://reviews.llvm.org/D83233
Eric Schweitz [Tue, 7 Jul 2020 22:39:09 +0000 (15:39 -0700)]
[flang] Upstreaming intrinsic call lowering.
This module implements the lowering of Fortran intrinsics to the
corresponding calls in support libraries (the Fortran runtime, math
libraries, etc.)
This revision is a tad larger because there are a large number of Fortran
intrinsics and this adds lowering for a fair number of them.
Differential revision: https://reviews.llvm.org/D83355
Pete Steinfeld [Tue, 7 Jul 2020 19:15:39 +0000 (12:15 -0700)]
[flang] Support for image selectors
Summary:
This change implements support for image selectors and image selector
specifications as described in section 9.6.
In check-coarray[.h,cpp] I changed the `Leave()` function for
`parser::ImageSelectorSpec` to take a `parser::ImageSelector`, which
contains a list of image selector specifications. This allows us to
detect when the same specification is used more than once. I also added
code to analyze the expressions for the image selector specifications to
expression.cpp and a test for all of the conditions to check at
compile-time.
Note that we do not check at compile-time to see if the value of the
cosubscripts are within the specified cobounds. We also do not check anything
related to selecting a valid team. We also do not check that the denotation of
the `stat-variable` is not dependent on the evaluation of an entity in the
same statement.
Reviewers: klausler, tskeith, DavidTruby
Subscribers: llvm-commits
Tags: #llvm, #flang
Differential Revision: https://reviews.llvm.org/D83336
Ties Stuij [Wed, 8 Jul 2020 12:12:30 +0000 (13:12 +0100)]
[CodeGen] Don't combine extract + concat vectors with non-legal types
Summary:
The following combine currently breaks in the DAGCombiner:
```
extract_vector_elt (concat_vectors v4i16:a, v4i16:b), x
-> extract_vector_elt a, x
```
This happens because after we have combined these nodes we have inserted nodes
that use individual instances of the vector element type. In the above example
i16. However this isn't a legal type on all backends, and when the combining pass calls
the legalizer it breaks as it expects types to already be legal. The type legalizer has
already been run, and running it again would make a mess of the nodes.
In the example code at least, the generated code is still efficient after the change.
Reviewers: miyuki, arsenm, dmgreen, lebedev.ri
Reviewed By: miyuki, lebedev.ri
Subscribers: lebedev.ri, wdng, hiraditya, steven.zhang, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D83231
Eduardo Caldas [Wed, 24 Jun 2020 15:11:31 +0000 (15:11 +0000)]
Fix crash on overloaded postfix unary operators due to invalid sloc
Subscribers: cfe-commits
Tags: #clang
Differential Revision: https://reviews.llvm.org/D82954
Alexey Bataev [Wed, 8 Jul 2020 13:09:57 +0000 (09:09 -0400)]
[NVPTX]Add a test for debug info for packed bitfields, NFC.
Alexander Belyaev [Wed, 8 Jul 2020 11:44:36 +0000 (13:44 +0200)]
[mlir] Add ViewLikeOpInterface to std.memref_cast.
Summery: It's needed for correct work of BufferPlacement.
Differential Revision: https://reviews.llvm.org/D83385
Isuru Fernando [Wed, 8 Jul 2020 12:19:01 +0000 (07:19 -0500)]
[flang] Fix setting mxcsr on MSVC
Reviewers: sscalpone, jdoerfert, #flang, DavidTruby, jeanPerier
Reviewed By: jeanPerier
Subscribers: richard.barton.arm, jeanPerier, ro, llvm-commits
Tags: #llvm, #flang
Differential Revision: https://reviews.llvm.org/D77815
Sanjay Patel [Wed, 8 Jul 2020 12:20:49 +0000 (08:20 -0400)]
[x86] improve codegen for non-splat bit-masked vector compare and select (PR46531)
vselect ((X & Pow2C) == 0), LHS, RHS --> vselect ((shl X, C') < 0), RHS, LHS
Follow-up to D83073 - the non-splat mask cases where we actually see an
improvement are quite limited from what I can tell. AVX1 needs multiply
and blend capabilities and AVX2 needs vector shift and blend capabilities.
The intersection of those 2 constraints is only vectors with 32-bit or
64-bit elements.
XOP is/was better.
Differential Revision: https://reviews.llvm.org/D83181
Med Ismail Bennani [Wed, 8 Jul 2020 12:01:54 +0000 (14:01 +0200)]
[lldb/Core] Update comment to make it more explicit (NFC)
Signed-off-by: Med Ismail Bennani <medismail.bennani@gmail.com>
Sylvestre Ledru [Wed, 8 Jul 2020 11:33:34 +0000 (13:33 +0200)]
clang: Don't show a trailing space with --version when not built from the repo
Reported here:
https://bugs.llvm.org/show_bug.cgi?id=38998#c15
Reviewers: hans
Differential Revision: https://reviews.llvm.org/D83386
Pavel Labath [Wed, 8 Jul 2020 11:54:33 +0000 (13:54 +0200)]
[lldb/Utility] Simplify Scalar float accessors
Make use of APFloat conversion methods to avoid needing to switch based
on the stored value type.
Haojian Wu [Wed, 8 Jul 2020 11:53:47 +0000 (13:53 +0200)]
[AST][RecoveryExpr] Fix the value category for recovery expr.
RecoveryExpr was always lvalue, but it is wrong if we use it to model
broken function calls, function call expression has more compliated rules:
- a call to a function whose return type is an lvalue reference yields an lvalue;
- a call to a function whose return type is an rvalue reference yields an xvalue;
- a call to a function whose return type is nonreference type yields a prvalue;
This patch makes the recovery-expr align with the function call if it is
modeled a broken call.
Differential revision: https://reviews.llvm.org/D83201
Pavel Labath [Mon, 6 Jul 2020 14:44:37 +0000 (16:44 +0200)]
[lldb/API] Overwrite variables with SBLaunchInfo::SetEnvironment(append=true)
Summary:
This function was documented to overwrite entries with D76111, which was
adding a couple of similar functions. However, this function (unlike the
functions added in that patch) was/is not actually overwriting variables
-- any pre-existing variables would get ignored.
This behavior does not seem to be intentional. In fact, before the refactor in
D41359, this function could introduce duplicate entries, which could
have very surprising effects both inside lldb and on other applications
(some applications would take the first value, some the second one; in
lldb, attempting to unset a variable could make the second variable
become active, etc.).
Overwriting seems to be the most reasonable behavior here, so change the
code to match documentation.
Reviewers: clayborg, wallace, jingham
Subscribers: lldb-commits
Tags: #lldb
Differential Revision: https://reviews.llvm.org/D83306
Michael Forney [Wed, 8 Jul 2020 10:42:39 +0000 (12:42 +0200)]
[cmake] Use CMAKE_GENERATOR to determine if Ninja is used
The name of the make program does not necessarily match "ninja",
especially if an alternative implementation like samurai is used.
Using CMAKE_GENERATOR is a more robust detection method, and is
already used elsewhere in this file.
Differential revision: https://reviews.llvm.org/D77091
Simon Pilgrim [Wed, 8 Jul 2020 10:47:43 +0000 (11:47 +0100)]
[X86][AVX] SimplifyDemandedVectorEltsForTargetShuffle - ensure mask is same size as constant size
Fixes test regression reported on D81791
Alex Richardson [Wed, 8 Jul 2020 10:38:13 +0000 (11:38 +0100)]
Add missing REQUIRES: x86-registered-target
This should fix build bot failures after
a80afc032859ebe65af283f76b38a0f5921b683f
Raphael Isemann [Wed, 8 Jul 2020 10:32:07 +0000 (12:32 +0200)]
[lldb][NFC] Fix indentation in expect_expr
Petar Avramovic [Wed, 8 Jul 2020 10:32:17 +0000 (12:32 +0200)]
[GlobalISel][InlineAsm] Fix matching input constraints to mem operand
Mark matching input constraint to mem operand as not supported.
Differential Revision: https://reviews.llvm.org/D83235
Simon Pilgrim [Wed, 8 Jul 2020 10:26:16 +0000 (11:26 +0100)]
[X86][AVX] Add SimplifyDemandedVectorEltsForTargetShuffle test for v32i8->v16i8 PSHUFB
On SKX targets we end up loading a v16i8 PSHUFB mask from a v32i8 constant and scaling incorrectly indexes the demanded elts mask - we're missing a check that the constant pool is the same size as the loaded mask.
Test case from D81791 post-commit review.
Oliver Stannard [Wed, 8 Jul 2020 10:24:03 +0000 (11:24 +0100)]
[Support] Fix signed/unsigned comparison warning
Alex Richardson [Wed, 8 Jul 2020 09:59:50 +0000 (10:59 +0100)]
[UpdateTestChecks] Add UTC_ARGS support for update_{llc,cc}_test_checks.py
https://reviews.llvm.org/D69701 added support for on-the-fly argument
changes for update scripts. I recently wanted to keep some manual check
lines in a test generated by update_cc_test_checks.py in our CHERI fork, so
this commit adds support for UTC_ARGS in update_cc_test_checks.py. And since
I was refactoring the code to be in common.py, I also added it for
update_llc_test_checks.py.
Reviewed By: jdoerfert, MaskRay
Differential Revision: https://reviews.llvm.org/D78478
Alex Richardson [Tue, 7 Jul 2020 15:38:28 +0000 (16:38 +0100)]
[UpdateTestChecks] Move more update_test_checks.py logic to common.py
I intend to reuse this to add UTC_ARGS support for update_llc_test_checks.py
and update_cc_test_checks.py in D78478.
Reviewed By: jdoerfert
Differential Revision: https://reviews.llvm.org/D78618
Paul Walker [Wed, 8 Jul 2020 09:10:16 +0000 (09:10 +0000)]
[SVE] Custom ISel for fixed length extract/insert_subvector.
We use extact_subvector and insert_subvector to "cast" between
fixed length and scalable vectors. This patch adds custom c++
based ISel for the following cases:
fixed_vector = ISD::EXTRACT_SUBVECTOR scalable_vector, 0
scalable_vector = ISD::INSERT_SUBVECTOR undef(scalable_vector), fixed_vector, 0
Which result in either EXTRACT_SUBREG/INSERT_SUBREG for NEON sized
vectors or COPY_TO_REGCLASS otherwise.
Differential Revision: https://reviews.llvm.org/D82871
Raphael Isemann [Wed, 8 Jul 2020 09:45:30 +0000 (11:45 +0200)]
[lldb] Skip TestIOHandlerResizeNoEditline on Windows
It seems opening the empty file and trying use that file object as an input
stream doesn't work on Windows. Skipping it for now.
George Mitenkov [Wed, 8 Jul 2020 08:33:27 +0000 (11:33 +0300)]
[MLIR][SPIRVToLLVM] Implementation of spv.BitFieldSExtract and spv.BitFieldUExtract patterns
This patch adds conversion patterns for `spv.BitFieldSExtract` and `spv.BitFieldUExtract`.
As in the patch for `spv.BitFieldInsert`, `offset` and `count` have to be broadcasted in
vector case and casted to match the type of the base.
Reviewed By: antiagainst
Differential Revision: https://reviews.llvm.org/D82640
Jeremy Morse [Wed, 8 Jul 2020 09:01:14 +0000 (10:01 +0100)]
[DWARF] Add cuttoff guarding quadratic validThroughout behaviour
Occasionally we see absolutely massive basic blocks, typically in global
constructors that are vulnerable to heavy inlining. When these blocks are
dense with DBG_VALUE instructions, we can hit near quadratic complexity in
DwarfDebug's validThroughout function. The problem is caused by:
* validThroughout having to step through all instructions in the block to
examine their lexical scope,
* and a high proportion of instructions in that block being DBG_VALUEs
for a unique variable fragment,
Leading to us stepping through every instruction in the block, for (nearly)
each instruction in the block.
By adding this guard, we force variables in large blocks to use a location
list rather than a single-location expression, as shown in the added test.
This shouldn't change the meaning of the output DWARF at all: instead we
use a less efficient DWARF encoding to avoid a poor-performance code path.
Differential Revision: https://reviews.llvm.org/D83236
Roman Lebedev [Wed, 8 Jul 2020 09:05:15 +0000 (12:05 +0300)]
[NFCI][llvm-reduce] Cleanup Delta passes to use Oracle abstraction
Summary:
I think, this results in much more understandable/readable flow.
At least the original logic was perhaps the most hard thing for me to grasp when taking an initial look on the delta passes.
Reviewers: nickdesaulniers, dblaikie, diegotf, george.burgess.iv
Reviewed By: nickdesaulniers
Subscribers: llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D83287
Simon Pilgrim [Wed, 8 Jul 2020 09:18:20 +0000 (10:18 +0100)]
Fix MSVC "not all control paths return a value" warnings. NFC.
Simon Pilgrim [Tue, 7 Jul 2020 19:30:16 +0000 (20:30 +0100)]
[X86][AVX] Remove redundant EXTRACT_VECTOR_ELT(VBROADCAST(SCALAR())) fold
Noticed while looking for similar cases to rG931ec74f7a29 - SimplifyDemandedVectorElts and shuffle combining both should handle this now.
Georgii Rymar [Tue, 30 Jun 2020 12:02:28 +0000 (15:02 +0300)]
[DebugInfo/DWARF] - Test invalid CFI opcodes properly and refine related `CFIProgram::parse` code.
There are following issues with `CFIProgram::parse` code:
1) Invalid CFI opcodes were never tested. And currently a test would fail
when the `LLVM_ENABLE_ABI_BREAKING_CHECKS` is enabled. It happens because
the `DataExtractor::Cursor C` remains unchecked when the
"Invalid extended CFI opcode" error is reported:
```
.eh_frame section at offset 0x1128 address 0x0:
Program aborted due to an unhandled Error:
Error value was Success. (Note: Success values must still be checked prior to being destroyed).
```
2) It is impossible to reach the "Invalid primary CFI opcode" error with the current code.
There are 3 possible primary opcode values and all of them are handled. Hence this error
should be replaced with llvm_unreachable.
3) Errors currently reported are upper-case.
This patch refines the code in the `CFIProgram::parse` method to fix all issues mentioned
and adds unit tests for all possible invalid extended CFI opcodes.
Differential revision: https://reviews.llvm.org/D82868
Georgii Rymar [Tue, 7 Jul 2020 15:13:26 +0000 (18:13 +0300)]
[llvm-readobj] - Refine error reporting in MipsGOTParser<ELFT> helper.
This is a follow-up for D83225. This does the following:
1) Adds missing tests for existent errors.
2) Stops using `unwrapOrError` to propagate errors to caller.
(I am trying to get rid of all `unwrapOrErr` calls in the llvm-readelf code).
3) Improves error messages reported slightly.
Differential revision: https://reviews.llvm.org/D83314
David Sherwood [Fri, 3 Jul 2020 12:32:02 +0000 (13:32 +0100)]
[CodeGen] Fix wrong use of getVectorNumElements() in DAGTypeLegalizer::SplitVecRes_ExtendOp
In DAGTypeLegalizer::SplitVecRes_ExtendOp I have replaced an invalid
call to getVectorNumElements() with a call to getVectorMinNumElements(),
since the code path works for both fixed and scalable vectors.
This fixes up a warning in the following test:
sve-sext-zext.ll
Differential Revision: https://reviews.llvm.org/D83197
David Sherwood [Fri, 3 Jul 2020 12:21:27 +0000 (13:21 +0100)]
[CodeGen] Fix wrong use of getVectorNumElements in PromoteIntRes_EXTRACT_SUBVECTOR
Calling getVectorNumElements() is not safe for scalable vectors and we
should normally use getVectorElementCount() instead. However, for the
code changed in this patch I decided to simply move the instantiation of
the variable 'OutNumElems' lower down to the place where only fixed-width
vectors are used, and hence it is safe to call getVectorNumElements().
Fixes up one warning in this test:
sve-sext-zext.ll
Differential Revision: https://reviews.llvm.org/D83195
Mikael Holmen [Wed, 8 Jul 2020 08:14:53 +0000 (10:14 +0200)]
[analyzer] Silence gcc -Wparentheses warning [NFC]
David Sherwood [Thu, 2 Jul 2020 11:24:45 +0000 (12:24 +0100)]
[CodeGen] Fix warnings in sve-ld1-addressing-mode-reg-imm.ll
For the GetElementPtr case in function
AddressingModeMatcher::matchOperationAddr
I've changed the code to use the TypeSize class instead of relying
upon the implicit conversion to a uint64_t. As part of this we now
check for scalable types and if we encounter one just bail out for
now as the subsequent optimisations doesn't currently support them.
This changes fixes up all warnings in the following tests:
llvm/test/CodeGen/AArch64/sve-ld1-addressing-mode-reg-imm.ll
llvm/test/CodeGen/AArch64/sve-st1-addressing-mode-reg-imm.ll
Differential Revision: https://reviews.llvm.org/D83124
George Mitenkov [Tue, 7 Jul 2020 10:35:45 +0000 (13:35 +0300)]
[MLIR][SPIRVToLLVM] Miscellaneous ops conversion: select, fmul and undef
This patch introduces 3 new direct conversions for SPIR-V ops:
- `spv.Select`
- `spv.Undef`
- `spv.FMul` that was skipped in the patch with arithmetic ops
Differential Revision: https://reviews.llvm.org/D83291
Heejin Ahn [Wed, 8 Jul 2020 07:59:21 +0000 (00:59 -0700)]
[WebAssembly] Generate unreachable after __stack_chk_fail
`__stack_chk_fail` does not return, but `unreachable` was not generated
following `call __stack_chk_fail`. This had a possibility to generate an
invalid binary for functions with a return type, because
`__stack_chk_fail`'s return type is void and `call __stack_chk_fail` can
be the last instruction in the function whose return type is non-void.
Generating `unreachable` after it makes sure CFGStackify's
`fixEndsAtEndOfFunction` handles it correctly.
Reviewed By: tlively
Differential Revision: https://reviews.llvm.org/D83277
Florian Hahn [Wed, 8 Jul 2020 07:42:55 +0000 (08:42 +0100)]
[DSE,MSSA] Eliminate stores by terminators (free,lifetime.end).
This patch adds support for eliminating stores by free & lifetime.end
calls. We can remove stores that are not read before calling a memory
terminator and we can eliminate all stores after a memory terminator
until we see a new lifetime.start. The second case seems to not really
trigger much in practice though.
Reviewers: dmgreen, rnk, efriedma, bryant, asbirlea, Tyker
Reviewed By: asbirlea
Differential Revision: https://reviews.llvm.org/D72410
LLVM GN Syncbot [Wed, 8 Jul 2020 07:52:15 +0000 (07:52 +0000)]
[gn build] Port
20e271a98de
Nithin Vadukkumchery Rajendrakumar [Fri, 26 Jun 2020 00:31:20 +0000 (02:31 +0200)]
[analyzer] Warning for default constructed unique_ptr dereference
Summary: Add support for warning incase of default constructed unique pointer dereferences
Reviewed By: NoQ, Szelethus, vsavchenko, xazax.hun
Tags: #clang
Differential Revision: https://reviews.llvm.org/D81315
serge-sans-paille [Tue, 23 Jun 2020 13:44:29 +0000 (15:44 +0200)]
Upgrade TypePromotionTransaction to be able to report changes in CodeGenPrepare
optimizeMemoryInst was reporting no change while still modifying the IR.
Inspect the status of TypePromotionTransaction to get a better status.
Related to https://reviews.llvm.org/D80916
Differential Revision: https://reviews.llvm.org/D81256
Mehdi Amini [Wed, 8 Jul 2020 04:04:54 +0000 (04:04 +0000)]
Rename `xla_lhlo.terminator` into `return` in SCF parallel loop test (NFC)
It seems that these were inserted here inadvertently instead of using
standard `return`.
Nico Weber [Wed, 8 Jul 2020 02:56:08 +0000 (22:56 -0400)]
Revert "[X86] Add back the assert in getImpliedFeatures that I removed in
ef4cc70f3ed2a91e0a48c6448c517c3ba34c2846"
This reverts commit
91f70675cc6e5c872e0059c11d797b8726eeac67.
It seems to break most (all?) hwasan tests.
clementval [Wed, 8 Jul 2020 02:38:13 +0000 (22:38 -0400)]
[openmp][NFC] Remove duplicate clause defaultmap for target parallel do
Nico Weber [Wed, 8 Jul 2020 02:29:19 +0000 (22:29 -0400)]
[gn build] Port
baca8f977ed
Muhammad Omair Javaid [Wed, 8 Jul 2020 02:02:16 +0000 (07:02 +0500)]
[LLDB] Disable flaky lldb-vscode tests on arm
Summary:
These two tests are flaky on lldb Arm buildbot as well. They are already
being skipped for aarch64. I am going to mark them skipped for Arm.
Tags: #lldb
Differential Revision: https://reviews.llvm.org/D81978
Craig Topper [Wed, 8 Jul 2020 02:01:22 +0000 (19:01 -0700)]
Recommit "[X86] Merge the FEATURE_64BIT and FEATURE_EM64T bits in X86TargetParser.def."
These represent the same thing but 64BIT only showed up from
getHostCPUFeatures providing a list of featuers to clang. While
EM64T showed up from getting the features for a named CPU.
EM64T didn't have a string specifically so it would not be passed
up to clang when getting features for a named CPU. While 64bit
needed a name since that's how it is index.
Merge them by filtering 64bit out before sending features to clang
for named CPUs.
Ben Shi [Wed, 8 Jul 2020 01:54:22 +0000 (18:54 -0700)]
[RISCV] optimize addition with a pair of (addi imm)
For an addition with an immediate in specific ranges, a pair of
addi-addi can be generated instead of the ordinary lui-addi-add serial.
Reviewed By: MaskRay, luismarques
Differential Revision: https://reviews.llvm.org/D82262
Ben Shi [Wed, 8 Jul 2020 01:48:46 +0000 (18:48 -0700)]
[RISCV] Optimize multiplication by constant
... to shift/add or shift/sub.
Do not enable it on riscv32 with the M extension where decomposeMulByConstant
may not be an optimization.
Reviewed By: luismarques, MaskRay
Differential Revision: https://reviews.llvm.org/D82660
HazemAbdelhafez [Wed, 8 Jul 2020 01:32:39 +0000 (21:32 -0400)]
[mlir][spirv] Add MatrixTimesMatrix operation
Add MatrixTimesMatrix operation to SPIRV Dialect and add NoSideEffect trait
to Matrix ops.
Reviewed By: antiagainst
Differential Revision: https://reviews.llvm.org/D82671
Richard Smith [Wed, 8 Jul 2020 01:25:57 +0000 (18:25 -0700)]
PR45521: Preserve the value kind when performing a standard conversion
sequence on a glvalue expression.
If the sequence is supposed to perform an lvalue-to-rvalue conversion,
then one will be specified as the first conversion in the sequence.
Otherwise, one should not be invented.
Rahul Joshi [Tue, 7 Jul 2020 23:15:44 +0000 (16:15 -0700)]
[MLIR] Combine the 2 overloads of FuncOp::build() into one.
- This will eliminate the need to pass an empty `ArrayRef<NamedAttribute>{}` when
no named attributes are required on the function.
Differential Revision: https://reviews.llvm.org/D83356
Craig Topper [Wed, 8 Jul 2020 01:20:07 +0000 (18:20 -0700)]
Revert "[X86] Merge the FEATURE_64BIT and FEATURE_EM64T bits in X86TargetParser.def."
An accidental change snuck in here
This reverts commit
f1d290d81298092b693076725cef4f34e951e974.
Craig Topper [Wed, 8 Jul 2020 00:55:07 +0000 (17:55 -0700)]
[X86] Merge the FEATURE_64BIT and FEATURE_EM64T bits in X86TargetParser.def.
These represent the same thing but 64BIT only showed up from
getHostCPUFeatures providing a list of featuers to clang. While
EM64T showed up from getting the features for a named CPU.
EM64T didn't have a string specifically so it would not be passed
up to clang when getting features for a named CPU. While 64bit
needed a name since that's how it is index.
Merge them by filtering 64bit out before sending features to clang
for named CPUs.
Wouter van Oortmerssen [Wed, 8 Jul 2020 00:47:43 +0000 (17:47 -0700)]
[WebAssembly] fix gcc 10 warning
Sid Manning [Fri, 19 Jun 2020 22:37:44 +0000 (17:37 -0500)]
[compiler-rt][Hexagon] Remove fma/fmin/max code
This code should reside in the c-library.
Differential Revision: https://reviews.llvm.org/D82263