Xun Li [Tue, 15 Sep 2020 22:19:57 +0000 (15:19 -0700)]
[TSAN] Handle musttail call properly in EscapeEnumerator (and TSAN)
Call instructions with musttail tag must be optimized as a tailcall, otherwise could lead to incorrect program behavior.
When TSAN is instrumenting functions, it broke the contract by adding a call to the tsan exit function inbetween the musttail call and return instruction, and also inserted exception handling code.
This happend throguh EscapeEnumerator, which adds exception handling code and returns ret instructions as the place to insert instrumentation calls.
This becomes especially problematic for coroutines, because coroutines rely on tail calls to do symmetric transfers properly.
To fix this, this patch moves the location to insert instrumentation calls prior to the musttail call for ret instructions that are following musttail calls, and also does not handle exception for musttail calls.
Differential Revision: https://reviews.llvm.org/D87620
Joseph Huber [Tue, 15 Sep 2020 19:04:37 +0000 (15:04 -0400)]
[OpenMP] Additional Information for Libomptarget Mappings
Summary:
This patch adds additonal support for priting infromation from Libomptarget for
already existing maps and printing the final data mapped on the device at
device destruction.
Reviewers: jdoerfort gkistanova
Subscribers: guansong openmp-commits sstefan1 yaxunl
Tags: #OpenMP
Differential Revision: https://reviews.llvm.org/D87722
Alexey Bataev [Tue, 15 Sep 2020 19:57:11 +0000 (15:57 -0400)]
[OPENMP]Fix codegen for is_device_ptr component, captured by reference.
Need to map the component as TO instead of the literal, because need to
pass a reference to a component if the pointer is overaligned.
Reviewed By: jdoerfert
Differential Revision: https://reviews.llvm.org/D84887
Diego Caballero [Fri, 4 Sep 2020 18:44:32 +0000 (11:44 -0700)]
[mlir] Rename 'setInsertionPointAfter' to avoid ambiguity
Rename 'setInsertionPointAfter(Value)' API to avoid ambiguity with
'setInsertionPointAfter(Operation *)' for SingleResult operations which
implicitly convert to Value (see D86756).
Differential Revision: https://reviews.llvm.org/D87155
Guillaume Chatelet [Tue, 15 Sep 2020 20:48:08 +0000 (20:48 +0000)]
[libc] Remove special case for 8 and 16 bytes
They don't seem to gain much in real apps and its better to favor less branches and smaller code.
Saleem Abdulrasool [Tue, 8 Sep 2020 22:49:41 +0000 (22:49 +0000)]
Sema: add support for `__attribute__((__swift_bridged_typedef__))`
Extend the semantic attributes that clang processes for Swift to include
`swift_bridged_typedef`. This attribute enables typedefs to be bridged
into Swift with a bridged name.
This is based on the work of the original changes in
https://github.com/llvm/llvm-project-staging/commit/
8afaf3aad2af43cfedca7a24cd817848c4e95c0c
Differential Revision: https://reviews.llvm.org/D87396
Reviewed By: Aaron Ballman
Guillaume Chatelet [Tue, 15 Sep 2020 20:09:50 +0000 (20:09 +0000)]
[libc] use stddef instead of string header
Huihui Zhang [Tue, 15 Sep 2020 20:09:56 +0000 (13:09 -0700)]
[SLPVectorizer][SVE] Skip scalable-vector instructions before vectorizeSimpleInstructions.
For scalable type, the aggregated size is unknown at compile-time.
Skip instructions with scalable type to ensure the list of instructions
for vectorizeSimpleInstructions does not contains any scalable-vector instructions.
Reviewed By: RKSimon
Differential Revision: https://reviews.llvm.org/D87550
Guillaume Chatelet [Tue, 15 Sep 2020 20:03:59 +0000 (20:03 +0000)]
[libc] remove useless headers
Stephen Hines [Tue, 15 Sep 2020 19:50:42 +0000 (12:50 -0700)]
Implement __isOSVersionAtLeast for Android
Add the implementation of __isOSVersionAtLeast for Android. Currently,
only the major version is checked against the API level of the platform
which is an integer. The API level is retrieved by reading the system
property ro.build.version.sdk (and optionally ro.build.version.codename
to see if the platform is released or not).
Patch by jiyong@google.com
Bug:
150860940
Bug:
134795810
Test: m
Reviewed By: srhines
Differential Revision: https://reviews.llvm.org/D86596
Ta-Wei Tu [Tue, 15 Sep 2020 19:38:06 +0000 (15:38 -0400)]
[TableGen] Fix invalid comparison function `SizeOrder` in `getMatchingSubClassWithSubRegs`
Building LLVM with -DEXPENSIVE_CHECKS fails with the following error
message with libstdc++ in debug mode:
Error: comparison doesn't meet irreflexive requirements,
assert(!(a < a)).
The patch fixes the comparison function SizeOrder by returning false
when comparing two equal items.
Matt Arsenault [Tue, 15 Sep 2020 17:46:23 +0000 (13:46 -0400)]
InferAddressSpaces: Fix assert with unreachable code
Invalid IR in unreachable code is technically valid IR. In this case,
the address space of the value was never inferred, and we tried to
rewrite it with an invalid address space value which would assert.
Snehasish Kumar [Wed, 2 Sep 2020 18:00:46 +0000 (11:00 -0700)]
[clang] Add a command line flag for the Machine Function Splitter.
This patch adds a command line flag for the machine function splitter
(added in rG94faadaca4e1).
-fsplit-machine-functions
Split machine functions using profile information (x86 ELF). On
other targets an error is emitted. If profile information is not
provided a warning is emitted notifying the user that profile
information is required.
Differential Revision: https://reviews.llvm.org/D87047
Muhammad Asif Manzoor [Tue, 15 Sep 2020 19:20:55 +0000 (15:20 -0400)]
[AArch64][SVE] Add lowering for llvm fsqrt
Add the functionality to lower fsqrt for passthru variant
Reviewed By: paulwalker-arm
Differential Revision: https://reviews.llvm.org/D87707
Albion Fung [Tue, 15 Sep 2020 19:18:54 +0000 (15:18 -0400)]
[PowerPC] Implement __int128 vector divide operations
This patch implements __int128 vector divide operations for ISA3.1.
Differential Revision: https://reviews.llvm.org/D85453
Zequan Wu [Mon, 14 Sep 2020 17:57:23 +0000 (10:57 -0700)]
[CodeGen][typeid] Emit typeinfo directly if type is known at compile-time
Differential Revision: https://reviews.llvm.org/D87425
Jonas Toth [Sun, 13 Sep 2020 17:30:56 +0000 (19:30 +0200)]
[ASTMatchers] extract public matchers from const-analysis into own patch
The analysis for const-ness of local variables required a view generally useful
matchers that are extracted into its own patch.
They are decompositionDecl and forEachArgumentWithParamType, that works
for calls through function pointers as well.
This is a reupload of https://reviews.llvm.org/D72505, that already landed,
but had to be reverted due to a GCC crash on powerpc
(https://reviews.llvm.org/rG4c48ea68e491cb42f1b5d43ffba89f6a7f0dadc4)
Because this took a long time to adress, i decided to redo this patch and
have a clean workflow.
I try to coordinate with someone that has a PPC to apply this patch and
test for the crash. If everything is fine, I intend to just commit.
If the crash is still happening, i hope to at least find the cause.
Differential Revision: https://reviews.llvm.org/D87588
Louis Dionne [Tue, 15 Sep 2020 18:59:27 +0000 (14:59 -0400)]
[libc++] Fix broken test for std::any and allocators
The test was not allocating the right number of bytes. This is my fault,
not Marshall's, as I was the one to write the tests for
39c879514170.
Arthur Eubanks [Tue, 15 Sep 2020 18:59:00 +0000 (11:59 -0700)]
[Dominators][NewPM] Pin tests with -analyze to legacy PM
-analyze isn't supported in NPM. All affected tests have corresponding
NPM RUN line.
Arthur Eubanks [Tue, 15 Sep 2020 18:55:10 +0000 (11:55 -0700)]
[DemandedBits][NewPM] Pin some tests to legacy PM
All tests have corresponding NPM RUN lines.
-analyze doesn't work under NPM.
Greg Clayton [Mon, 14 Sep 2020 19:20:45 +0000 (12:20 -0700)]
Fix .debug_aranges parsing.
Code was added that used llvm error checking to parse .debug_aranges, but the error check after parsing the DWARFDebugArangesSet was reversed and was causing no error to be returned with no valid address ranges being actually used. This meant we always would fall back onto creating out own address ranges by parsing the compile unit's ranges. This was causing problems for cases where the DW_TAG_compile_unit had a single address range by using a DW_AT_low_pc and DW_AT_high_pc attribute pair (not using a DW_AT_ranges attribute), but the .debug_aranges had correct split ranges. In this case we would end up using the single range for the compile unit that encompassed all of the ranges from the .debug_aranges section and would cause address resolving issues in LLDB where address lookups would fail for certain addresses.
Differential Revision: https://reviews.llvm.org/D87626
LLVM GN Syncbot [Tue, 15 Sep 2020 18:32:17 +0000 (18:32 +0000)]
[gn build] Port
3d42d549554
Florian Hahn [Tue, 15 Sep 2020 13:47:23 +0000 (14:47 +0100)]
[ConstraintElimination] Add constraint elimination pass.
This patch is a first draft of a new pass that adds a more flexible way
to eliminate compares based on more complex constraints collected from
dominating conditions.
In particular, it aims at simplifying conditions of the forms below
using a forward propagation approach, rather than instcomine-style
ad-hoc backwards walking of def-use chains.
if (x < y)
if (y < z)
if (x < z) <- simplify
or
if (x + 2 < y)
if (x + 1 < y) <- simplify assuming no wraps
The general approach is to collect conditions and blocks, sort them by
dominance and then iterate over the sorted list. Conditions are turned
into a linear inequality and add it to a system containing the linear
inequalities that hold on entry to the block. For blocks, we check each
compare against the system and see if it is implied by the constraints
in the system.
We also keep a stack of processed conditions and remove conditions from
the stack and the constraint system once they go out-of-scope (= do not
dominate the current block any longer).
Currently there still are the least the following areas for improvements
* Currently large unsigned constants cannot be added to the system
(coefficients must be represented as integers)
* The way constraints are managed currently is not very optimized.
Reviewed By: spatel
Differential Revision: https://reviews.llvm.org/D84547
Arthur Eubanks [Tue, 15 Sep 2020 00:17:32 +0000 (17:17 -0700)]
[Bugpoint][NewPM] Pin bugpoint to legacy PM
Bugpoint has lots of assumptions and hacks around the legacy PM, put off migrating it to NPM until later.
Fixes tests under BugPoint under NPM.
Reviewed By: arsenm
Differential Revision: https://reviews.llvm.org/D87655
Craig Topper [Tue, 15 Sep 2020 18:07:52 +0000 (11:07 -0700)]
[X86] Use Align in reduceMaskedLoadToScalarLoad/reduceMaskedStoreToScalarStore. Correct pointer info.
If we offset the pointer, we also need to offset the pointer info
Differential Revision: https://reviews.llvm.org/D87593
Louis Dionne [Tue, 15 Sep 2020 18:19:06 +0000 (14:19 -0400)]
[libc++] Add missing friend keyword
Otherwise, we're declaring a non-static member function, and that
gives errors in C++11 because of the change of semantics between
C++11 and C++14 for non-const constexpr member functions.
This was always intended to be a friend declaration.
Arthur Eubanks [Tue, 15 Sep 2020 01:11:09 +0000 (18:11 -0700)]
[PostDominators][NewPM] Fix tests to work under NPM
Each test has a legacy PM pinned to legacy PM and a NPM RUN line.
Reviewed By: asbirlea
Differential Revision: https://reviews.llvm.org/D87660
Arthur Eubanks [Tue, 15 Sep 2020 01:35:12 +0000 (18:35 -0700)]
[NewPM][opt] Fix -globals-aa not being recognized as alias analysis in NPM
Was missing MODULE_ALIAS_ANALYSIS, previously only FUNCTION_ALIAS_ANALYSIS was taken into account.
Reviewed By: asbirlea
Differential Revision: https://reviews.llvm.org/D87664
Arthur Eubanks [Tue, 15 Sep 2020 00:49:58 +0000 (17:49 -0700)]
[RegionInfo][NewPM] Fix RegionInfo tests to work under NPM
Pin RUN lines with -analyze to legacy PM, add corresponding NPM RUN line if missing.
Reviewed By: asbirlea
Differential Revision: https://reviews.llvm.org/D87658
Arthur Eubanks [Tue, 15 Sep 2020 01:45:30 +0000 (18:45 -0700)]
[DependenceAnalysis][NewPM] Fix tests to work under NPM
All tests had corresponding NPM lines, simply pin non-NPM lines to legacy PM.
Reviewed By: fhahn
Differential Revision: https://reviews.llvm.org/D87665
Arthur Eubanks [Tue, 15 Sep 2020 01:23:08 +0000 (18:23 -0700)]
[LoopAccessAnalysis][NewPM] Fix tests to work under NPM
Pin RUN lines with -analyze to legacy PM, add corresponding NPM RUN lines.
Reviewed By: fhahn
Differential Revision: https://reviews.llvm.org/D87662
Alexey Bataev [Tue, 15 Sep 2020 15:21:47 +0000 (11:21 -0400)]
[OPENMP]Add support for allocate vars in untied tasks.
Local vars, marked with pragma allocate, mustbe allocate by the call of
the runtime function and cannot be allocated as other local variables.
Instead, we allocate a space for the pointer in private record and store
the address, returned by kmpc_alloc call in this pointer.
So, for untied tasks
```
#pragma omp task untied
{
S s;
#pragma omp allocate(s) allocator(allocator)
s = x;
}
```
compiler generates something like this:
```
struct task_with_privates {
S *ptr;
};
void entry(task_with_privates *p) {
S *s = p->s;
switch(partid) {
case 1:
p->s = (S*)kmpc_alloc();
kmpc_omp_task();
br exit;
case 2:
*s = x;
kmpc_omp_task();
br exit;
case 2:
~S(s);
kmpc_free((void*)s);
br exit;
}
exit:
}
```
Reviewed By: jdoerfert
Differential Revision: https://reviews.llvm.org/D86558
Nemanja Ivanovic [Tue, 15 Sep 2020 17:33:31 +0000 (12:33 -0500)]
Fix bot failure after
ccb4124a4172
The test case has a check line for the option on a line that includes
the string lld surrounded by any characters. This causes failures
when said string is in the build path. What the test case presumably
means to test is the actual invocation of the LLD linker (i.e. a
linker that has that string as a suffix). This patch simply removes
the erroneous wildcard after the string.
Fangrui Song [Tue, 15 Sep 2020 17:20:08 +0000 (10:20 -0700)]
[MemoryBuffer] Revert unintended MemoryBuffer change from D86996
Fixes SupportsTest MemoryBufferTest.mmapVolatileNoNull
Florian Hahn [Tue, 15 Sep 2020 16:52:50 +0000 (17:52 +0100)]
Revert "[DSE] Switch to MemorySSA-backed DSE by default."
This reverts commit
fb109c42d91c30c8c7497ef1fd7aff6f2969c6e7.
Temporarily revert due to a mis-compile pointed out at D87163.
Petr Hosek [Tue, 15 Sep 2020 17:06:15 +0000 (10:06 -0700)]
Revert "[DebugInfo] Remove dots from getFilenameByIndex return value"
This is failing on Windows bots due to path separator normalization.
This reverts commit
042c23506869b4ae9a49d2c4bc5ea6e6baeabe78.
Fangrui Song [Mon, 14 Sep 2020 22:11:55 +0000 (15:11 -0700)]
[VectorCombine] Don't vectorize scalar load under asan/hwasan/memtag/tsan
Similar to the tsan suppression in
`Utils/VNCoercion.cpp:getLoadLoadClobberFullWidthSize` (rL175034; load widening used by GVN),
the D81766 optimization should be suppressed under tsan due to potential
spurious data race reports:
struct A {
int i;
const short s; // the load cannot be vectorized because
int modify; // it overlaps with bytes being concurrently modified
long pad1, pad2;
};
// __tsan_read16 does not know that some bytes are undef and accessing is safe
Similarly, under asan, users can mark memory regions with
`__asan_poison_memory_region`. A widened load can lead to a spurious
use-after-poison error. hwasan/memtag should be similarly suppressed.
`mustSuppressSpeculation` suppresses asan/hwasan/tsan but not memtag, so
we need to exclude memtag in `vectorizeLoadInsert`.
Note, memtag suppression can be relaxed if the load is aligned to the
its granule (usually 16), but that is out of scope of this patch.
Reviewed By: spatel, vitalybuka
Differential Revision: https://reviews.llvm.org/D87538
Jonas Devlieghere [Tue, 15 Sep 2020 16:36:28 +0000 (09:36 -0700)]
[lldb] Add -l/--language option to script command
Make it possible to run the script command with a different language
than currently selected.
$ ./bin/lldb -l python
(lldb) script -l lua
>>> io.stdout:write("Hello, World!\n")
Hello, World!
When passing the language option and a raw command, you need to separate
the flag from the script code with --.
$ ./bin/lldb -l python
(lldb) script -l lua -- io.stdout:write("Hello, World!\n")
Hello, World!
Differential revision: https://reviews.llvm.org/D86996
Simon Pilgrim [Tue, 15 Sep 2020 16:08:45 +0000 (17:08 +0100)]
[X86][AVX] lowerShuffleWithSHUFPS - handle missed canonicalization cases.
PR47534 exposes a case where calling lowerShuffleWithSHUFPS directly from a derived repeated mask (found by is128BitLaneRepeatedShuffleMask) results in us using an non-canonicalized mask.
The missed canonicalization in this case is trivial - just commute the mask so we have more (swapped) LHS than RHS references so lowerShuffleWithSHUFPS can handle it.
Louis Dionne [Tue, 15 Sep 2020 16:29:41 +0000 (12:29 -0400)]
NFC: Add whitespace change to .git-blame-ignore-revs
Guozhi Wei [Tue, 15 Sep 2020 16:18:18 +0000 (09:18 -0700)]
[MachineBasicBlock] Fix a typo in function copySuccessor
The condition used to decide if need to copy probability should be reversed.
Differential Revision: https://reviews.llvm.org/D87417
Mark de Wever [Tue, 15 Sep 2020 15:08:13 +0000 (11:08 -0400)]
[libc++] Add a benchmark for std::map operations
Before tackling http://llvm.org/PR38722, make sure there is a baseline
benchmark.
Differential Revision: https://reviews.llvm.org/D62778
jasonliu [Tue, 15 Sep 2020 15:50:26 +0000 (15:50 +0000)]
[XCOFF] Run resource intense test only on platforms where it makes sense
This is a follow up commit for the issue raised in
https://reviews.llvm.org/D86879
Kristóf Umann [Tue, 15 Sep 2020 15:43:02 +0000 (17:43 +0200)]
[analyzer][Liveness][NFC] Get rid of statement liveness, because such a thing doesn't exist
The summary and very short discussion in D82122 summarizes whats happening here.
In short, liveness talks about variables, or expressions, anything that
has a value. Well, statements just simply don't have a one.
Differential Revision: https://reviews.llvm.org/D82598
Valentin Clement [Tue, 15 Sep 2020 15:41:50 +0000 (11:41 -0400)]
[mlir][openacc] Add loop op verifier
Add a verifier for the loop op in the OpenACC dialect. Check basic restriction
from 2.9 Loop construct from the OpenACC 3.0 specs.
Reviewed By: ftynse
Differential Revision: https://reviews.llvm.org/D87546
Simon Pilgrim [Tue, 15 Sep 2020 14:23:19 +0000 (15:23 +0100)]
SLPVectorizer.h - remove unnecessary AliasAnalysis.h include. NFCI.
Forward declare AAResults instead of the (old) AliasAnalysis type.
Remove includes from SLPVectorizer.cpp that are already included in SLPVectorizer.h.
Marshall Clow [Tue, 15 Sep 2020 13:56:03 +0000 (09:56 -0400)]
[libc++] Use allocator_traits to consistently allocate/deallocate/construct/destroy objects in std::any
https://llvm.org/PR45099 notes (correctly) that we're inconsistent in memory
allocation in `std::any`. We allocate memory with `std::allocator<T>::allocate`,
construct with placement new, destroy by calling the destructor directly, and
deallocate by calling `delete`. Most of those are customizable by the user,
but in different ways.
The standard is silent on how these things are to be accomplished.
This patch makes it so we use `allocator_traits<allocator<T>>` for all
of these operations (allocate, construct, destruct, deallocate).
This is, at least, consistent.
Fixes https://llvm.org/PR45099.
Differential Revision: https://reviews.llvm.org/D81133
Sanjay Patel [Tue, 15 Sep 2020 14:30:35 +0000 (10:30 -0400)]
[InstSimplify] add limit folds for fmin/fmax
If the constant operand is the opposite of the min/max value,
then the result must be the other value.
This is based on the similar codegen transform proposed in:
D87571
Kristóf Umann [Tue, 25 Aug 2020 11:49:41 +0000 (13:49 +0200)]
[analyzer] Add documentation for alpha.fuchsia.Lock and alpha.core.C11Lock
Differential Revision: https://reviews.llvm.org/D86532
Georgii Rymar [Tue, 1 Sep 2020 08:11:34 +0000 (11:11 +0300)]
[llvm-readobj/elf] - Don't crash when the size of s dynamic symbol table, inferred from the hash table, is broken.
Currently we might derive the dynamic symbol table size from the DT_HASH hash table (using its `nchain` field).
It is possible to crash dumpers with a broken relocation that refers to a symbol with an index
that is too large. To trigger it, the inferred size of the dynamic symbol table should go past the end of the object.
This patch adds a size validation + warning.
Differential revision: https://reviews.llvm.org/D86923
Gabor Marton [Thu, 3 Sep 2020 11:23:49 +0000 (13:23 +0200)]
[analyzer][StdLibraryFunctionsChecker] Elaborate the summary of fread and fwrite
Add the BufferSize argument constraint to fread and fwrite. This change
itself makes it possible to discover a security critical case, described
in SEI-CERT ARR38-C.
We also add the not-null constraint on the 3rd arguments.
In this patch, I also remove those lambdas that don't take any
parameters (Fwrite, Fread, Getc), thus making the code better
structured.
Differential Revision: https://reviews.llvm.org/D87081
Guillaume Chatelet [Tue, 15 Sep 2020 14:26:04 +0000 (14:26 +0000)]
[libc] Add missing TableGen dependency
Differential Revision: https://reviews.llvm.org/D87689
Guillaume Chatelet [Tue, 15 Sep 2020 14:25:34 +0000 (14:25 +0000)]
[libc] Add missing LibcFPTestHelpers library
Differential Revision: https://reviews.llvm.org/D87690
Guillaume Chatelet [Tue, 15 Sep 2020 14:25:00 +0000 (14:25 +0000)]
[libc] Fix typo in platform_defs.h.inc
Differential Revision: https://reviews.llvm.org/D87687
Stefan Pintilie [Tue, 15 Sep 2020 13:23:58 +0000 (08:23 -0500)]
[LLD][PowerPC] Add support for R_PPC64_TPREL34 used in TLS Local Exec
Add Thread Local Storage Local Exec support to LLD. This is to support PC Relative addressing of Local Exec.
The patch teaches LLD to handle:
```
paddi r9, r13, x1@tprel
```
The relocation is:
```
R_PPC_TPREL34
```
Reviewed By: NeHuang, MaskRay
Differential Revision: https://reviews.llvm.org/D86608
Qiu Chaofan [Tue, 15 Sep 2020 14:03:50 +0000 (22:03 +0800)]
Revert "[SelectionDAG] Remove unused FP constant in getNegatedExpression"
2508ef01 doesn't totally fix the issue since we did not handle the case
when unused temporary negated result is the same with the result, which
is found by address sanitizer.
Simon Pilgrim [Tue, 15 Sep 2020 13:48:40 +0000 (14:48 +0100)]
AMDGPUPrintfRuntimeBinding.cpp - drop unnecessary casts/dyn_casts. NFCI.
GetElementPtrInst::Create returns a GetElementPtrInst* so we don't need to cast. Similarly IntegerType inherits from the Type base class.
Also, I've used auto* in a few places to cleanup the code.
Helps fix some clang-tidy warnings which saw the dyn_casts and warned that these can return null.
Simon Pilgrim [Tue, 15 Sep 2020 12:48:40 +0000 (13:48 +0100)]
[Utils] isLegalToPromote - Fix missing null check before writing to FailureReason.
The FailureReason input parameter maybe null, we check this in all other cases in the method but this one was missed somehow.
Fixes clang-tidy warning.
Valentin Clement [Tue, 15 Sep 2020 13:48:24 +0000 (09:48 -0400)]
[mlir][openacc] Add missing print of vector_length in parallel op
This patch adds the missing print for the vector_length in the parallel operation.
Reviewed By: ftynse
Differential Revision: https://reviews.llvm.org/D87630
Stephan Herhut [Mon, 14 Sep 2020 09:54:55 +0000 (11:54 +0200)]
[mlir][Standard] Add canonicalizer for dynamic_tensor_from_elements
This add canonicalizer for
- extracting an element from a dynamic_tensor_from_elements
- propagating constant operands to the type of dynamic_tensor_from_elements
Differential Revision: https://reviews.llvm.org/D87525
LLVM GN Syncbot [Tue, 15 Sep 2020 13:32:48 +0000 (13:32 +0000)]
[gn build] Port
cd4edf94cd4
LLVM GN Syncbot [Tue, 15 Sep 2020 13:32:47 +0000 (13:32 +0000)]
[gn build] Port
a8058c6f8d1
Sanjay Patel [Tue, 15 Sep 2020 13:27:16 +0000 (09:27 -0400)]
[InstCombine] fix bug in pow expansion
There at least one other bug related to pow -> sqrt transforms:
http://lists.llvm.org/pipermail/llvm-dev/2020-September/145051.html
...but we probably can't solve that without fixing this first.
Sanjay Patel [Tue, 15 Sep 2020 13:21:20 +0000 (09:21 -0400)]
[InstCombine] add RUN to show miscompile of pow expansion; NFC
The code drops the sqrt op instead of bailing out,
so this is very wrong.
Sanjay Patel [Tue, 15 Sep 2020 12:38:51 +0000 (08:38 -0400)]
[InstCombine] improve test names; NFC
This is not a valid transform unless we can prove
that the program does not read errno after the pow
call and before some other function changes it.
Oliver Stannard [Tue, 15 Sep 2020 10:40:05 +0000 (11:40 +0100)]
[libcxx] Disable failing test for no-exceptions build
This test tries to create a 2 GiB std::string, catching the bad_alloc
exception if the allocation fails. However, for no-exceptions builds
there is no way for the error to be reported, so this crashes with a
null pointer dereference.
Differential revision: https://reviews.llvm.org/D87682
Florian Hahn [Fri, 24 Jul 2020 18:36:48 +0000 (19:36 +0100)]
[ConstraintElimination] Add initial tests.
Nico Weber [Tue, 15 Sep 2020 13:25:19 +0000 (09:25 -0400)]
[gn build] (semi-manually) port
380e746bcca
Sam Clegg [Tue, 15 Sep 2020 01:28:26 +0000 (18:28 -0700)]
[lld][WebAssembly] Fix --export-all when __stack_pointer is present
With https://reviews.llvm.org/D87537 we made it an error
to import or export a mutable global with the +mutable-globals
feature present. However the scan was of the entire symbol
table rather than just the imports or exports and the filter
didn't match exaclyt meaning the `__stack_pointer` (a mutable
global) was always triggering with error when the `--export-all`
flag was used.
This also revealed that we didn't have any test coverage for
the `--export-all` flag.
This change fixes the current breakage on the emscripten-releases
roller.
Differential Revision: https://reviews.llvm.org/D87663
sameeran joshi [Sun, 13 Sep 2020 11:54:34 +0000 (17:24 +0530)]
[Flang] Add GettingInvolved documentation page and sidebar.
Adds a new GettingInvolved page to documentation which provides details about
mailing list, chats and calls.
Adds a sidebar page which provides common links on
all documentation pages.
The links include:
- Getting Started
- Getting Involved
- Github Repository
- Bug Reports
- Code Review
Depends on https://reviews.llvm.org/D87242
Reviewed By: richard.barton.arm
Differential Revision: https://reviews.llvm.org/D87270
Florian Hahn [Tue, 15 Sep 2020 12:50:11 +0000 (13:50 +0100)]
[ConstraintSolver] Add isConditionImplied helper.
This patch adds a isConditionImplied function that
takes a constraint and returns true if the constraint
is implied by the current constraints in the system.
Reviewed By: spatel
Differential Revision: https://reviews.llvm.org/D84545
Felix Berger [Tue, 15 Sep 2020 12:44:13 +0000 (08:44 -0400)]
Restrict UnnecessaryCopyInitialization check to variables initialized from free functions without arguments
This restriction avoids cases where an alias is returned to an argument and
which could lead to to a false positive change.
Louis Dionne [Tue, 15 Sep 2020 12:43:08 +0000 (08:43 -0400)]
[libc++] Allow building without threads in standalone builds
Setting _LIBCPP_HAS_NO_THREADS is needed when building libcxxabi without
threads in standalone mode. This is useful when target WASM. Otherwise,
you get an error like "No thread API" when building libcxxabi.
It would be better to link against a properly-configured libc++ headers
CMake target when building libc++abi instead, but we don't generate such
targets yet.
Thanks to Matthew Bauer for the patch.
Differential Revision: https://reviews.llvm.org/D60743
Simon Pilgrim [Tue, 15 Sep 2020 12:34:19 +0000 (13:34 +0100)]
LoopCacheAnalysis.h - remove unnecessary includes. NFCI.
More remaining dependencies down to LoopCacheAnalysis.cpp
Simon Pilgrim [Tue, 15 Sep 2020 12:09:03 +0000 (13:09 +0100)]
ProvenanceAnalysis.h - remove unnecessary AliasAnalysis.h include. NFCI.
Forward declare AAResults instead of the (old) AliasAnalysis type.
Simon Pilgrim [Tue, 15 Sep 2020 11:59:00 +0000 (12:59 +0100)]
AliasSetTracker.cpp - remove unnecessary includes. NFCI.
These are all directly included in AliasSetTracker.h
Sjoerd Meijer [Tue, 15 Sep 2020 12:10:30 +0000 (13:10 +0100)]
[ARM][MVE] Tail-predication: use unsigned SCEV ranges for tripcount
Loop tripcount expressions have a positive range, so use unsigned SCEV ranges
for them.
Differential Revision: https://reviews.llvm.org/D87608
Bjorn Pettersson [Mon, 14 Sep 2020 20:53:54 +0000 (22:53 +0200)]
[Scalarizer] Avoid changing name of non-instructions
The "takeName" logic in ScalarizerVisitor::gather did not consider
that the value vector could refer to non-instructions, such as
global variables. This patch make sure that we avoid changing the
name of a value if it isn't an instruction.
Reviewed By: lebedev.ri
Differential Revision: https://reviews.llvm.org/D87685
Pavel Labath [Tue, 15 Sep 2020 11:20:09 +0000 (13:20 +0200)]
[lldb] Improve qemu interop for aarch64
qemu calls the "fp" and "lr" registers via their generic names
(x29/x30). This mismatch manifested itself as not being able to unwind
or display values of some local variables.
Pavel Labath [Fri, 28 Aug 2020 10:31:16 +0000 (12:31 +0200)]
[lldb/cmake] Fix testing support library dependencies
lldbUtilityHelpers does not depend on lldbSymbolHelpers. Remove that
dependency, and add direct lldbSymbolHelpers dependencies where needed.
Hans Wennborg [Tue, 15 Sep 2020 08:47:02 +0000 (10:47 +0200)]
Revert "RegAllocFast: Record internal state based on register units"
This seems to have caused incorrect register allocation in some cases,
breaking tests in the Zig standard library (PR47278).
As discussed on the bug, revert back to green for now.
> Record internal state based on register units. This is often more
> efficient as there are typically fewer register units to update
> compared to iterating over all the aliases of a register.
>
> Original patch by Matthias Braun, but I've been rebasing and fixing it
> for almost 2 years and fixed a few bugs causing intermediate failures
> to make this patch independent of the changes in
> https://reviews.llvm.org/D52010.
This reverts commit
66251f7e1de79a7c1620659b7f58352b8c8e892e, and
follow-ups
931a68f26b9a3de853807ffad7b2cd0a2dd30922
and
0671a4c5087d40450603d9d26cf239f1a8b1367e. It also adjust some
test expectations.
Simon Pilgrim [Tue, 15 Sep 2020 11:18:06 +0000 (12:18 +0100)]
SpillPlacement.cpp - remove unnecessary includes. NFCI.
These are all directly included in SpillPlacement.h
Simon Pilgrim [Tue, 15 Sep 2020 11:00:21 +0000 (12:00 +0100)]
StatepointLowering.cpp - remove unnecessary includes. NFCI.
These are all directly included in StatepointLowering.h
Simon Pilgrim [Tue, 15 Sep 2020 10:44:47 +0000 (11:44 +0100)]
SelectionDAGBuilder.h - remove unnecessary includes. NFCI.
Reduce to forward declarations and move implicit dependencies down to the cpp files.
Florian Hahn [Tue, 15 Sep 2020 09:28:25 +0000 (10:28 +0100)]
Recommit "[ConstraintSystem] Add helpers to deal with linear constraints."
This patch recommits "[ConstraintSystem] Add helpers to deal with linear constraints."
(it reverts the revert commit
8da6ae4ce1b686c5c13698e4c5ee937811fda6f7).
The reason for the revert was using __builtin_multiply_overflow, which
is not available for all compilers. The patch has been updated to use
MulOverflow from MathExtras.h
Alex Zinenko [Mon, 14 Sep 2020 11:47:27 +0000 (13:47 +0200)]
[mlir] check for failures when packing function sigunatures in std->llvm conversion
When packing function results into a structure during the standard-to-llvm
dialect conversion, do not assume the conversion was successful and propagate
nullptr as error state.
Fixes PR45184.
Reviewed By: nicolasvasilache
Differential Revision: https://reviews.llvm.org/D87605
Jakub Lichman [Tue, 15 Sep 2020 10:07:29 +0000 (10:07 +0000)]
[mlir][integration_test] Linalg Conv folder renamed to CPU
Changing directory name to reflect naming convention discussed here:
https://llvm.discourse.group/t/vectorops-rfc-add-suite-of-integration-tests-for-vector-dialect-operations/1213
Differential Revision: https://reviews.llvm.org/D87678
Benjamin Kramer [Tue, 15 Sep 2020 10:22:47 +0000 (12:22 +0200)]
Revert "[InstCombine] Simplify select operand based on equality condition"
This reverts commit
cfff88c03cf9e9b72906a41fd11e06721d54f293. Sends
instcombine into an infinite loop.
```
define i1 @foo(i32 %arg, i32 %arg1) {
bb:
%tmp = udiv i32 %arg, %arg1
%tmp2 = mul nsw i32 %tmp, %arg1
%tmp3 = icmp eq i32 %tmp2, %arg
%tmp4 = select i1 %tmp3, i32 %tmp, i32 undef
%tmp5 = icmp sgt i32 %tmp4, 255
ret i1 %tmp5
}
```
Simon Pilgrim [Tue, 15 Sep 2020 10:18:44 +0000 (11:18 +0100)]
[X86] Update SSE/AVX integer MINMAX intrinsics to emit llvm.smax.* etc. (PR46851)
We're now getting close to having the necessary analysis/combines etc. for the new generic llvm smax/smin/umax/umin intrinsics.
This patch updates the SSE/AVX integer MINMAX intrinsics to emit the generic equivalents instead of the icmp+select code pattern.
Differential Revision: https://reviews.llvm.org/D87603
Meera Nakrani [Tue, 15 Sep 2020 10:14:30 +0000 (10:14 +0000)]
[ARM] Corrected condition in isSaturatingConditional
Fixed a small error in an if condition to prevent usat/ssat being generated if (upper constant + 1) is not a
power of 2.
Qiu Chaofan [Tue, 15 Sep 2020 09:59:10 +0000 (17:59 +0800)]
[SelectionDAG] Remove unused FP constant in getNegatedExpression
960cbc53 immediately removes nodes that won't be used to avoid
compilation time explosion. This patch adds the removal to constants to
fix PR47517.
Reviewed By: RKSimon, steven.zhang
Differential Revision: https://reviews.llvm.org/D87614
Simon Pilgrim [Tue, 15 Sep 2020 09:37:25 +0000 (10:37 +0100)]
[Transforms][Coroutines] Add missing header path to CMakeLists.txt
Helps Visual Studio check include dependencies.
Sjoerd Meijer [Tue, 15 Sep 2020 09:04:02 +0000 (10:04 +0100)]
[MVE] fix typo in llvm debug message. NFC.
Simon Pilgrim [Tue, 15 Sep 2020 09:06:35 +0000 (10:06 +0100)]
[X86] detectAVGPattern - accept non-pow2 vectors by padding.
Drop the pow2 vector limitation for AVG generation by padding the vector to the next pow2, creating the PAVG nodes and then extracting the final subvector.
Fixes some poor codegen that has been annoying me for years.....
Georgii Rymar [Sat, 5 Sep 2020 16:13:50 +0000 (19:13 +0300)]
[llvm-readobj/elf] - Simplify and refine the implementation which dumps .stack_sizes
Our implementation of stack sizes section dumping heavily uses `ELFObjectFile<ELFT>`,
while the rest of the code uses `ELFFile<ELFT>`.
That APIs are very different. `ELFObjectFile<ELFT>` is very generic
and has `SectionRef`, `RelocationRef`, `SymbolRef` and other generic concepts.
The `ELFFile<ELFT>` class works directly with `Elf_Shdr`, `Elf_Rel[a]`, `Elf_Sym` etc,
what is probably much cleaner for ELF dumper.
Also, `ELFObjectFile<ELFT>` API does not always provide a way to check
for possible errors. E.g. the implementation of `symbol_end()` does not verify the `sh_size`:
```
template <class ELFT>
basic_symbol_iterator ELFObjectFile<ELFT>::symbol_end() const {
const Elf_Shdr *SymTab = DotSymtabSec;
if (!SymTab)
return symbol_begin();
DataRefImpl Sym = toDRI(SymTab, SymTab->sh_size / sizeof(Elf_Sym));
return basic_symbol_iterator(SymbolRef(Sym, this));
}
```
There are many other examples which makes me thing we might win from
switching to `ELFFile<ELFT>` API, where we heavily validate an input data already.
This patch is the first step in this direction. I've converted the large portion of the code
to use `ELFFile<ELFT>`.
Differential revision: https://reviews.llvm.org/D87362
Petr Hosek [Tue, 15 Sep 2020 08:46:58 +0000 (01:46 -0700)]
[NFC][DebugInfo] Use consistent regex group spelling
This is a follow up to
c1f2fb5184ca.
Georgii Rymar [Wed, 9 Sep 2020 14:03:53 +0000 (17:03 +0300)]
[lib/Object] - Refine interface of ELFFile<ELFT>. NFCI.
`ELFFile<ELFT>` has many methods that take pointers,
though they assume that arguments are never null and
hence could take references instead.
This patch performs such clean-up.
Differential revision: https://reviews.llvm.org/D87385
Petar Avramovic [Tue, 15 Sep 2020 08:25:38 +0000 (10:25 +0200)]
GlobalISel/IRTranslator resetTargetOptions based on function attributes
Update TargetMachine.Options with function attributes before we start
to generate MIR instructions. This allows access to correct function
attributes via TargetMachine.Options (it used to access attributes of
the function that was translated first).
This affects some existing tests with "no-nans-fp-math" attribute.
Follow-up on D87456.
Differential Revision: https://reviews.llvm.org/D87511
Sjoerd Meijer [Tue, 15 Sep 2020 08:09:59 +0000 (09:09 +0100)]
[MVE] Rename of tests making them consistent with tail-predication tests. NFC.
Hans Wennborg [Thu, 10 Sep 2020 17:32:45 +0000 (19:32 +0200)]
[Support] Make building with snmalloc work
Differential revision: https://reviews.llvm.org/D87471