Amara Emerson [Mon, 10 Feb 2020 23:41:53 +0000 (15:41 -0800)]
[GlobalISel][CallLowering] Use stripPointerCasts().
A downstream test exposed a simple logic bug with the manual pointer
stripping code, fix that by just using stripPointerCasts() on the value.
I don't think there's a way to expose this issue upstream.
Davide Italiano [Mon, 10 Feb 2020 23:17:31 +0000 (15:17 -0800)]
[TestKernVerStrLCNOTE] Check the *right* architecture.
Eric Christopher [Mon, 10 Feb 2020 23:06:32 +0000 (15:06 -0800)]
Fix you->your typo.
Peter Collingbourne [Wed, 5 Feb 2020 22:49:19 +0000 (14:49 -0800)]
scudo: Add a dump of primary allocation sizes to malloc_info output.
This will be useful for optimizing the size class map.
Differential Revision: https://reviews.llvm.org/D74098
Peter Collingbourne [Wed, 5 Feb 2020 03:50:25 +0000 (19:50 -0800)]
scudo: Table driven size classes for Android allocator.
Add an optional table lookup after the existing logarithm computation
for MidSize < Size <= MaxSize during size -> class lookups. The lookup is
O(1) due to indexing a precomputed (via constexpr) table based on a size
table. Switch to this approach for the Android size class maps.
Other approaches considered:
- Binary search was found to have an unacceptable (~30%) performance cost.
- An approach using NEON instructions (see older version of D73824) was found
to be slightly slower than this approach on newer SoCs but significantly
slower on older ones.
By selecting the values in the size tables to minimize wastage (for example,
by passing the malloc_info output of a target program to the included
compute_size_class_config program), we can increase the density of allocations
at a small (~0.5% on bionic malloc_sql_trace as measured using an identity
table) performance cost.
Reduces RSS on specific Android processes as follows (KB):
Before After
zygote (median of 50 runs) 26836 26792 (-0.2%)
zygote64 (median of 50 runs) 30384 30076 (-1.0%)
dex2oat (median of 3 runs) 375792 372952 (-0.8%)
I also measured the amount of whole-system idle dirty heap on Android by
rebooting the system and then running the following script repeatedly until
the results were stable:
for i in $(seq 1 50); do grep -A5 scudo: /proc/*/smaps | grep Pss: | cut -d: -f2 | awk '{s+=$1} END {print s}' ; sleep 1; done
I did this 3 times both before and after this change and the results were:
Before: 365650, 356795, 372663
After: 344521, 356328, 342589
These results are noisy so it is hard to make a definite conclusion, but
there does appear to be a significant effect.
On other platforms, increase the sizes of all size classes by a fixed offset
equal to the size of the allocation header. This has also been found to improve
density, since it is likely for allocation sizes to be a power of 2, which
would otherwise waste space by pushing the allocation into the next size class.
Differential Revision: https://reviews.llvm.org/D73824
Peter Collingbourne [Mon, 10 Feb 2020 21:36:49 +0000 (13:36 -0800)]
scudo: Instead of exporting a pointer to the allocator, export the allocator directly. NFCI.
This lets us remove two pointer indirections (one by removing the pointer,
and another by making the AllocatorPtr declaration hidden) in the C++ wrappers.
Differential Revision: https://reviews.llvm.org/D74356
Dimitry Andric [Mon, 10 Feb 2020 22:43:12 +0000 (23:43 +0100)]
[Sanitizers] Get link map on FreeBSD and NetBSD via documented API
Summary:
Instead of hand-crafting an offset into the structure returned by
dlopen(3) to get at the link map, use the documented API. This is
described in dlinfo(3): by calling it with `RTLD_DI_LINKMAP`, the
dynamic linker ensures the right address is returned.
This is a recommit of
92e267a94dc4272511be674062f8a3e8897b7083, with
dlinfo(3) expliclity being referenced only for FreeBSD, non-Android
Linux, NetBSD and Solaris. Other OSes will have to add their own
implementation.
Reviewers: devnexen, emaste, MaskRay, krytarowski
Reviewed By: krytarowski
Subscribers: krytarowski, vitalybuka, #sanitizers, llvm-commits
Tags: #sanitizers, #llvm
Differential Revision: https://reviews.llvm.org/D73990
Vedant Kumar [Mon, 10 Feb 2020 22:33:44 +0000 (14:33 -0800)]
Revert "[Host.mm] Check for the right macro instead of inlining it"
This breaks macOS, because TARGET_OS_EMBEDDED is always defined. Thanks
to Jason Molenda for pointing this out.
Revert "Do not define AcceptPIDFromInferior when it will not be used"
This reverts commit
d23c15a687ff15327b88fa64da3184395012c2dc.
This reverts commit
936d1427da1432d724dfa5851097347bcdf7c521.
Dimitry Andric [Mon, 10 Feb 2020 22:24:26 +0000 (23:24 +0100)]
Revert "[Sanitizers] Get link map on FreeBSD via documented API"
This reverts commit
92e267a94dc4272511be674062f8a3e8897b7083, as it
appears Android is missing dlinfo(3).
Sanjay Patel [Mon, 10 Feb 2020 22:13:26 +0000 (17:13 -0500)]
[EarlyCSE] avoid crashing when detecting min/max/abs patterns (PR41083)
As discussed in PR41083:
https://bugs.llvm.org/show_bug.cgi?id=41083
...we can assert/crash in EarlyCSE using the current hashing scheme and
instructions with flags.
ValueTracking's matchSelectPattern() may rely on overflow (nsw, etc) or
other flags when detecting patterns such as min/max/abs composed of
compare+select. But the value numbering / hashing mechanism used by
EarlyCSE intersects those flags to allow more CSE.
Several alternatives to solve this are discussed in the bug report.
This patch avoids the issue by doing simple matching of min/max/abs
patterns that never requires instruction flags. We give up some CSE
power because of that, but that is not expected to result in much
actual performance difference because InstCombine will canonicalize
these patterns when possible. It even has this comment for abs/nabs:
/// Canonicalize all these variants to 1 pattern.
/// This makes CSE more likely.
(And this patch adds PhaseOrdering tests to verify that the expected
transforms are still happening in the standard optimization pipelines.
I left this code to use ValueTracking's "flavor" enum values, so we
don't have to change the callers' code. If we decide to go back to
using the ValueTracking call (by changing the hashing algorithm
instead), it should be obvious how to replace this chunk.
Differential Revision: https://reviews.llvm.org/D74285
Vedant Kumar [Mon, 10 Feb 2020 21:10:47 +0000 (13:10 -0800)]
[ubsan] Null-check and adjust TypeLoc before using it
Null-check and adjut a TypeLoc before casting it to a FunctionTypeLoc.
This fixes a crash in -fsanitize=nullability-return, and also makes the
location of the nonnull type available when the return type is adjusted.
rdar://
59263039
Differential Revision: https://reviews.llvm.org/D74355
Ted Woodward [Mon, 10 Feb 2020 20:42:48 +0000 (14:42 -0600)]
Remove lit feature object-emission
Summary: The lit feature object-emission was added because Hexagon did not support the integrated assembler, so some tests needed to be turned off with a Hexagon target. Hexagon now supports the integrated assembler, so this feature can be removed.
Reviewers: bcain, kparzysz, jverma, whitequark, JDevlieghere
Reviewed By: JDevlieghere
Subscribers: mehdi_amini, hiraditya, steven_wu, dexonsmith, arphaman, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D73568
LLVM GN Syncbot [Mon, 10 Feb 2020 21:43:41 +0000 (21:43 +0000)]
[gn build] Port
bb383ae6120
Hiroshi Yamauchi [Thu, 30 Jan 2020 21:05:31 +0000 (13:05 -0800)]
[CallPromotionUtils] Add tryPromoteCall.
Summary: It attempts to devirtualize a call on alloca through vtable loads.
Reviewers: davidxl
Subscribers: mgorny, Prazek, hiraditya, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D71308
Davide Italiano [Mon, 10 Feb 2020 20:43:59 +0000 (12:43 -0800)]
Revert "[lldb] Fix+re-enable Assert StackFrame Recognizer on Linux"
This reverts commit
1a39f1b966a8d8f15ed0d5a832d5097cccefe93b as
it breaks macOS.
Lei Zhang [Wed, 5 Feb 2020 01:58:10 +0000 (20:58 -0500)]
[mlir][spirv] Use spv.entry_point_abi in GPU to SPIR-V conversions
We have spv.entry_point_abi for specifying the local workgroup size.
It should be decorated onto input gpu.func ops to drive the SPIR-V
CodeGen to generate the proper SPIR-V module execution mode. Compared
to using command-line options for specifying the configuration, using
attributes also has the benefits that 1) we are now able to use
different local workgroup for different entry points and 2) the
tests contains the configuration directly.
Differential Revision: https://reviews.llvm.org/D74012
Alexey Bataev [Mon, 10 Feb 2020 20:49:05 +0000 (15:49 -0500)]
[OPENMP50]Add support for 'release' clause.
Added full support for 'release' clause in flush|atomic directives.
Hanhan Wang [Sat, 8 Feb 2020 00:30:55 +0000 (19:30 -0500)]
[mlir][Linalg] Add a roundtrip test for indexed_generic op with tensors.
Summary:
After D72555 has been landed, `linalg.indexed_generic` also accepts ranked
tensor as input and output. Add a test for it.
Differential Revision: https://reviews.llvm.org/D74267
Martin Storsjö [Wed, 5 Feb 2020 11:53:56 +0000 (13:53 +0200)]
[test] Disable the Passes/PluginsTest cases on windows with BUILD_SHARED_LIBS
The plugin expects to have undefined references to symbols exported
by the loading process, which isn't supported by shared libraries
on windows.
Differential Revision: https://reviews.llvm.org/D74042
Nico Weber [Mon, 10 Feb 2020 20:47:46 +0000 (15:47 -0500)]
git bisect docs: formatting tweaks
Xiangling Liao [Mon, 10 Feb 2020 18:52:08 +0000 (13:52 -0500)]
[AIX] Enable frame pointer for AIX and add related test suite
This patch:
- enable frame pointer for AIX;
- update some of red zone comments;
- add/update testcases;
Differential Revision: https://reviews.llvm.org/D72454
Matt Arsenault [Mon, 10 Feb 2020 20:18:30 +0000 (15:18 -0500)]
RegisterCoalescer: Add LaneMask to debug printing
Nico Weber [Mon, 10 Feb 2020 20:33:20 +0000 (15:33 -0500)]
add GitBisecting to toctrees to try and placate the sphinx bot
Sterling Augustine [Mon, 10 Feb 2020 20:19:35 +0000 (12:19 -0800)]
[DebugInfo] Support file-level include directories when generating DWARFv5 LineTable prologues.
Differential Revision: https://reviews.llvm.org/D74249
Nico Weber [Mon, 10 Feb 2020 20:18:15 +0000 (15:18 -0500)]
git bisect docs: try to make commit ascii art show up
Jan Korous [Mon, 10 Feb 2020 20:11:47 +0000 (12:11 -0800)]
Reland "[clangd][test] Disable a particular testcase in FindExplicitReferencesTest when LLVM_ENABLE_EXPENSIVE_CHECKS""
The test got re-enabled after
d54d71b67e60 landed.
However it seems that the order is still not deterministic as it
currently passes with -DLLVM_ENABLE_EXPENSIVE_CHECKS=OFF but randomly
fails with expensive checks ON.
Sanjay Patel [Mon, 10 Feb 2020 20:10:47 +0000 (15:10 -0500)]
[Transforms] add phase ordering tests for min/max/abs; NFC
Test that instcombine and early-cse can cooperate
to reduce sequences of select patterns that are not
composed of the same underlying instructions.
There's a bug in EarlyCSE (PR41083), and we can test
how much a possible fix (D74285) may affect optimization.
Vedant Kumar [Mon, 10 Feb 2020 20:06:46 +0000 (12:06 -0800)]
Do not define AcceptPIDFromInferior when it will not be used
Nicolas Vasilache [Sat, 8 Feb 2020 18:51:10 +0000 (13:51 -0500)]
[mlir][VectorOps][EDSC] Add EDSC for VectorOps
Summary:
This revision adds EDSC support for VectorOps to enable the creation of a `vector_matmul` declaratively. The `vector_matmul` is a simple configuration
of the `vector.contract` op that follows the StructuredOps abstraction.
Differential Revision: https://reviews.llvm.org/D74284
Sanjay Patel [Mon, 10 Feb 2020 19:50:51 +0000 (14:50 -0500)]
[InstCombine] fix use check when canonicalizing abs/nabs
We were checking for extra uses of the negated operand even
if we were not going to create it as part of this canonicalization.
This was showing up as a regression when we limit EarlyCSE as
proposed in D74285.
Sanjay Patel [Mon, 10 Feb 2020 19:42:50 +0000 (14:42 -0500)]
[InstCombine] add tests for abs with extra use of operand; NFC
Alexey Bataev [Mon, 10 Feb 2020 19:30:39 +0000 (14:30 -0500)]
[OPENMP50]Support for acquire clause.
Added full support for acquire clause in flush|atomic directives.
diggerlin [Mon, 10 Feb 2020 19:45:54 +0000 (14:45 -0500)]
[AIX][XCOFF] Support Mergeable2ByteCString and Mergeable4ByteCString
SUMMARY:
The patch is enable to support Mergeable2ByteCString and Mergeable4ByteCString
Reviewers: daltenty
Subscribers: wuzish, nemanjai, hiraditya
Differential Revision: https://reviews.llvm.org/D74164
Ted Woodward [Mon, 10 Feb 2020 19:40:17 +0000 (13:40 -0600)]
Don't fail step out if remote server doesn't implement qMemoryRegionInfo
Summary:
The return address validation in D71372 will fail if the memory permissions can't be determined. Many embedded stubs either don't implement the qMemoryRegionInfo packet, or don't have memory permissions at all.
Remove the return from the if clause that calls GetLoadAddressPermissions, so this call failing doesn't cause the step out to abort. Instead, assume that the memory permission check doesn't apply to this type of target.
Reviewers: labath, jingham, clayborg, mossberg
Reviewed By: labath, jingham
Subscribers: lldb-commits
Tags: #lldb
Differential Revision: https://reviews.llvm.org/D72513
Nico Weber [Tue, 4 Feb 2020 20:44:09 +0000 (15:44 -0500)]
Add documentation on git bisecting across the MLIR merge
Differential Revision: https://reviews.llvm.org/D73988
serge-sans-paille [Thu, 6 Feb 2020 14:58:29 +0000 (15:58 +0100)]
Prefer __vector over vector keyword for altivec
`vector' uses the keyword-and-predefine mode from gcc, while __vector is
reliably supported.
As a side effect, it also makes the code consistent in its usage of __vector.
Differential Revision: https://reviews.llvm.org/D74129
David Goldman [Mon, 3 Feb 2020 20:14:49 +0000 (15:14 -0500)]
[clang] Add `forceReload` clangd extension to 'textDocument/didChange'
Summary:
- This option forces a preamble rebuild to handle the odd case
of a missing header file being added
Reviewers: sammccall
Subscribers: ilya-biryukov, javed.absar, MaskRay, jkorous, arphaman, jfb, kadircet, usaxena95, cfe-commits
Tags: #clang
Differential Revision: https://reviews.llvm.org/D73916
Nico Weber [Mon, 10 Feb 2020 18:51:23 +0000 (13:51 -0500)]
fix some typos to cycle bots
Eric Christopher [Mon, 10 Feb 2020 18:31:19 +0000 (10:31 -0800)]
Continue removing llgo.
Max Moroz [Mon, 10 Feb 2020 18:19:03 +0000 (10:19 -0800)]
[compiler-rt] Follow up fix for the refactoring in https://reviews.llvm.org/D74137.
Summary:
The refactoring has caused a failure in
http://lab.llvm.org:8011/builders/clang-ppc64le-linux/builds/29265
The idea of failing the symbolization when the symbolizer bufer is too small
was incorrect. The symbolizer can be invoked for other frames that may fit into
the buffer and get symbolized.
Reviewers: vitalybuka, eugenis
Subscribers: dberris, #sanitizers, llvm-commits
Tags: #sanitizers, #llvm
Differential Revision: https://reviews.llvm.org/D74343
Florian Hahn [Mon, 10 Feb 2020 18:19:13 +0000 (18:19 +0000)]
[ValueLattice] Remove obsolete getConstantInt (NFC).
ConstantInt values are always represented as constant ranges with a
single element. getConstantInt is obsolete, as pointed out by @nikic
during D60581.
Reviewers: nikic
Reviewed By: nikic
Differential Revision: https://reviews.llvm.org/D74329
Stephen Neuendorffer [Mon, 10 Feb 2020 18:13:46 +0000 (10:13 -0800)]
[MLIR] Fix lib/Dialect/Linalg/EDSC for BUILD_SHARED_LIBS=on
Stephen Neuendorffer [Mon, 10 Feb 2020 16:55:27 +0000 (08:55 -0800)]
[MLIR] Fix lib/ExecutionEngine for BUILD_SHARED_LIBS=on
Stephen Neuendorffer [Tue, 10 Dec 2019 19:59:13 +0000 (11:59 -0800)]
[MLIR][Standard] Implement constant folding for IndexCast
Differential Revision: https://reviews.llvm.org/D73672
Stephen Neuendorffer [Thu, 5 Dec 2019 00:15:10 +0000 (16:15 -0800)]
[MLIR][Standard] Add folding for indexCast(indexCast(x)) -> x
Allow this only if the types are the same. e.g.:
i16 -> index -> i16 or
index -> i16 -> index
Differential Revision: https://reviews.llvm.org/D73671
Stephen Neuendorffer [Mon, 9 Dec 2019 22:27:11 +0000 (14:27 -0800)]
[MLIR] Allow non-binary operations to be commutative
NFC for binary operations.
Differential Revision: https://reviews.llvm.org/D73670
Rachel Craik [Mon, 10 Feb 2020 18:14:59 +0000 (13:14 -0500)]
[LoopCacheAnalysis]: Add support for negative stride
LoopCacheAnalysis currently assumes the loop will be iterated over in
a forward direction. This patch addresses the issue by using the
absolute value of the stride when iterating backwards.
Note: this patch will treat negative and positive array access the
same, resulting in the same cost being calculated for single and
bi-directional access patterns. This should be improved in a
subsequent patch.
Reviewed By: jdoerfert
Differential Revision: https://reviews.llvm.org/D73064
Dimitry Andric [Thu, 6 Feb 2020 22:18:19 +0000 (23:18 +0100)]
[Sanitizers] Get link map on FreeBSD via documented API
Summary:
Instead of hand-crafting an offset into the structure returned by
dlopen(3) to get at the link map, use the documented API. This is
described in dlinfo(3): by calling it with `RTLD_DI_LINKMAP`, the
dynamic linker ensures the right address is returned.
Reviewers: devnexen, emaste, MaskRay, krytarowski
Reviewed By: krytarowski
Subscribers: krytarowski, vitalybuka, #sanitizers, llvm-commits
Tags: #sanitizers, #llvm
Differential Revision: https://reviews.llvm.org/D73990
Vedant Kumar [Mon, 10 Feb 2020 18:10:10 +0000 (10:10 -0800)]
debugserver: Rely on mig architectures being specified externally
Look up the -arch flags to pass to the mig invocation from an
optionally-defined MIG_ARCHS variable. We can't use CMAKE_OSX_ARCHS
because the {i,tv,watch}OS builds don't use this mechanism to achieve
fat builds (they build each slice separately & then lipo them together).
This supercedes the mig -arch/-isysroot fix from
510758dae2a8fa4b0b26dea89d4d1efd576b8ad6.
Jonas Paulsson [Wed, 5 Feb 2020 13:43:35 +0000 (14:43 +0100)]
[SystemZ] Add a subtarget cache like some other targets already have.
Each function is with this compiled with the SystemZSubtarget initialized
from the functions attributes.
Review: Ulrich Weigand.
Differential Revision: https://reviews.llvm.org/D74086
Eric Christopher [Mon, 10 Feb 2020 18:07:37 +0000 (10:07 -0800)]
Remove llgo per discussion on llvm-dev:
http://lists.llvm.org/pipermail/llvm-dev/2020-February/139058.html
Approved by dblaikie, pcc.
Vedant Kumar [Mon, 10 Feb 2020 18:06:14 +0000 (10:06 -0800)]
[Host.mm] Check for the right macro instead of inlining it
Differential Revision: https://reviews.llvm.org/D73938
Matt Arsenault [Sun, 9 Feb 2020 21:38:56 +0000 (16:38 -0500)]
AMDGPU: Move R600 test compatability hack
Instead of handling the r600 intrinsics on amdgcn, handle the amdgcn
intrinsics on r600.
Simon Pilgrim [Mon, 10 Feb 2020 17:47:46 +0000 (17:47 +0000)]
[X86] combineConcatVectorOps - combine X86ISD::PACKSS ops
Simon Pilgrim [Mon, 10 Feb 2020 17:32:58 +0000 (17:32 +0000)]
[X86] combineConcatVectorOps - combine X86ISD::VPERMI ops
Michael Kruse [Mon, 10 Feb 2020 17:13:00 +0000 (11:13 -0600)]
Silence compiler warning. NFC.
The idiom
for (auto i = n - n; i < n; i += 1)
was intended to automatically derive the type of i from n
(signed/unsigned int) and avoid the 'mixed signed/unsigned comparison'
warning. However, almost-always-auto was never used in the LLVM coding
style (although we used it in Polly for some time) and I did never
intended to use this idiom upstream.
PVS Studio may warns about this idiom as 'warning: both sides of
operator are equivalent [misc-redundant-expression]'.
Remove the use of auto and directly use unsigned.
Also see http://llvm.org/PR44768
aartbik [Fri, 7 Feb 2020 22:32:08 +0000 (14:32 -0800)]
[mlir] [LLVMIR] add all vector reduction intrinsics to LLVM IR dialect
Summary:
This allows for lowering of VectorOps (and others) into a LLVM IR
that maps directly to efficient implementations on the target machines.
http://llvm.org/docs/LangRef.html#experimental-vector-reduction-intrinsics
Reviewers: ftynse, andydavis1, nicolasvasilache, rriddle
Reviewed By: ftynse, rriddle
Subscribers: jfb, mehdi_amini, rriddle, jpienaar, burmako, shauheen, antiagainst, nicolasvasilache, arpith-jacob, mgester, lucyrfox, liufengdb, Joonsoo, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D74171
Nicolas Vasilache [Mon, 10 Feb 2020 16:55:02 +0000 (11:55 -0500)]
[mlir][EDSC] Almost NFC - Refactor and untangle EDSC dependencies
This CL refactors EDSCs to layer them better and break unnecessary
dependencies. After this refactoring, the top-level EDSC target only
depends on IR but not on Dialects anymore and each dialect has its
own EDSC directory.
This simplifies the layering and breaks cyclic dependencies.
In particular, the declarative builder + folder are made explicit and
are now confined to Linalg.
As the refactoring occurred, certain classes and abstractions that were not
paying for themselves have been removed.
Differential Revision: https://reviews.llvm.org/D74302
Simon Pilgrim [Mon, 10 Feb 2020 16:48:39 +0000 (16:48 +0000)]
[X86] combineConcatVectorOps - combine VSHLI/VSRAI/VSRLI ops
Non-AVX512BW targets failed to concatenate 256-bit shifts back to 512-bits (split during 512-bit shuffle lowering as they don't have v32i16/v64i8 types).
David Stenberg [Mon, 10 Feb 2020 16:14:15 +0000 (17:14 +0100)]
Revert "[InstCombine][DebugInfo] Fold constants wrapped in metadata"
This reverts commit
b54a8ec1bcd3689771c847cb37515b627034e518.
The commit triggered debug invariance (different output with/without
-g). The patch seems to have exposed a pre-existing invariance problem
in GlobalOpt, which I'll write a bug report for.
Kadir Cetinkaya [Mon, 10 Feb 2020 16:54:00 +0000 (17:54 +0100)]
[mlir] Delete unused header
Saleem Abdulrasool [Mon, 10 Feb 2020 16:52:31 +0000 (08:52 -0800)]
unwind: rename `__personality_routine` to `_Unwind_Personality_Fn`
This patch renames `__personality_routine` to `_Unwind_Personality_Fn`
in `unwind.h`. Both `unwind.h` from clang and GCC headers use this name
instead of `__personality_routine`. With this patch one is also able to
build libc++abi with libunwind support on Windows.
Patch by Markus Böck!
Tobias Gysi [Mon, 10 Feb 2020 16:29:50 +0000 (17:29 +0100)]
[mlir] subview op lowering for target memrefs with const offset
The current standard to llvm conversion pass lowers subview ops only if
dynamic offsets are provided. This commit extends the lowering with a
code path that uses the constant offset of the target memref for the
subview op lowering (see Example 3 of the subview op definition for an
example) if no dynamic offsets are provided.
Differential Revision: https://reviews.llvm.org/D74280
Stanislav Mekhanoshin [Fri, 7 Feb 2020 20:08:32 +0000 (12:08 -0800)]
[AMDGPU] Split R600 and GCN subregs
These are generated and do not need to have the same values.
We are defining separate subregs for R600 and GCN but then
using AMDGPU subregs on R600.
Differential Revision: https://reviews.llvm.org/D74248
James Henderson [Mon, 10 Feb 2020 16:21:46 +0000 (16:21 +0000)]
[DebugInfo][test] Fix host endian test issue
The test previously assumed that the host was little endian, which broke
the big endian build bots.
Simon Pilgrim [Mon, 10 Feb 2020 16:16:16 +0000 (16:16 +0000)]
[X86] Add lowerShuffleAsBitRotate (PR44379)
As noted on PR44379, we didn't attempt to lower vector shuffles using bit rotations on XOP/AVX512F targets.
This patch lowers to uniform ISD:ROTL nodes - ROTR isn't supported by XOP and they are interchangeable for constant values anyway.
There might be cases where targets without ISD:ROTL support would benefit from this (expanding to SRL+SHL+OR), which I'll investigate in a future patch.
REAPPLIED rGe82e17d4d4ca after reversion at rG39eade73a567 - fixed offset matching in matchShuffleAsBitRotate.
LLVM GN Syncbot [Mon, 10 Feb 2020 15:56:57 +0000 (15:56 +0000)]
[gn build] Port
0151ddc2e83
Michael Wyman [Fri, 7 Feb 2020 23:08:17 +0000 (16:08 -0700)]
Create a clang-tidy check to warn when -dealloc is implemented inside an ObjC class category.
Summary: Such implementations may override the class's own implementation, and even be a danger in case someone later comes and adds one to the class itself. Most times this has been encountered have been a mistake.
Reviewers: stephanemoore, benhamilton, dmaclach
Reviewed By: stephanemoore, benhamilton, dmaclach
Subscribers: dmaclach, mgorny, cfe-commits
Tags: #clang-tools-extra, #clang
Differential Revision: https://reviews.llvm.org/D72876
Gabor Marton [Mon, 10 Feb 2020 15:16:44 +0000 (16:16 +0100)]
[analyzer] StdLibraryFunctionsChecker refactor: remove macros
Reviewers: NoQ
Tags: #clang
Differential Revision: https://reviews.llvm.org/D73897
Kadir Cetinkaya [Mon, 10 Feb 2020 13:44:51 +0000 (14:44 +0100)]
Revert "[OpenMP] Fix unused variable"
This breaks under asan, see http://lab.llvm.org:8011/builders/sanitizer-x86_64-linux-fast/builds/38597/steps/check-clang%20asan/logs/stdio
This reverts commit
bb5045429545f47a76980864322a637c31594c7f.
Revert "[FIX] Ordering problem accidentally introduced with D72304"
This reverts commit
08c0a06d8f375e48d4acebac886bfdf19a2276ed.
Revert "[OpenMP][OMPIRBuilder] Add Directives (master and critical) to OMPBuilder."
This reverts commit
e8a436c5ea26f69378e4c1cf3ddb5b647b201e0f.
Florian Hahn [Mon, 10 Feb 2020 15:18:46 +0000 (15:18 +0000)]
[DSE,MSSA] Adjust mda-with-dbg-values.ll to MSSA backed DSE.
-memdep-block-scan-limit is not relevant with MSSA.
James Henderson [Mon, 10 Feb 2020 15:09:16 +0000 (15:09 +0000)]
[DebugInfo][test] Fix(?) build bots due to incorrect type usage
Bill Wendling [Mon, 10 Feb 2020 15:06:45 +0000 (07:06 -0800)]
Revert "Remove redundant "std::move"s in return statements"
The build failed with
error: call to deleted constructor of 'llvm::Error'
errors.
This reverts commit
1c2241a7936bf85aa68aef94bd40c3ba77d8ddf2.
Max Moroz [Thu, 6 Feb 2020 16:44:42 +0000 (08:44 -0800)]
[compiler-rt] Some clean up / refactoring in sanitizer_symbolizer_libcdep.cpp.
Summary:
Nothing critical, just a few potential improvements I've noticed while reading
the code:
- return `false` when symbolizer buffer is too small to read all data
- invert some conditions to reduce indentation
- prefer `nullptr` over `0` for pointers; init some pointers on stack;
- remove minor code duplication
Reviewers: eugenis, vitalybuka
Subscribers: dberris, #sanitizers, llvm-commits, kcc
Tags: #sanitizers, #llvm
Differential Revision: https://reviews.llvm.org/D74137
James Henderson [Mon, 10 Feb 2020 14:17:46 +0000 (14:17 +0000)]
[DebugInfo] Reject line tables of version > 5
If a debug line section with version of greater than 5 is encountered,
prior to this change the parser would accept it and treat it as version
5. This might work to some extent, but then it might not at all, as it
really depends on the format of the unspecified future version, which
will be different (otherwise there would be no point in changing the
version number). Any information we could provide has a good chance of
being invalid, so we should just refuse to parse such tables.
Reviewed by: dblaikie, MaskRay
Differential Revision: https://reviews.llvm.org/D74204
James Henderson [Mon, 10 Feb 2020 14:40:47 +0000 (14:40 +0000)]
[NFC] Fix line endings
Bill Wendling [Mon, 10 Feb 2020 13:18:09 +0000 (05:18 -0800)]
Remove redundant "std::move"s in return statements
Luke Geeson [Mon, 10 Feb 2020 14:29:59 +0000 (14:29 +0000)]
[AArch64] Make Read Write System Registers Read Only
This patch makes the following System Registers Read Only:
- CurrentEL
- ICH_MISR_EL2
- PMBIDR_EL1
- PMSIDR_EL1
as found in:
https://developer.arm.com/docs/ddi0595/e/aarch64-system-registers
Relative line numbers were also added to the tests so we get more
informative error messages on failure.
Change-Id: I963b4f01ca5737b58f9e8e7abe9ca1d99e328758
Richard Smith [Mon, 10 Feb 2020 14:05:59 +0000 (06:05 -0800)]
CWG2445: For function template partial ordering, take reversal of
function arguments into account when forming P/A pairs.
Sebastian Neubauer [Fri, 31 Jan 2020 13:18:18 +0000 (14:18 +0100)]
[SelectionDAG] Optimize build_vector of truncates and shifts
Add a simplification to fuse a manual vector extract with shifts and
truncate into a bitcast.
Unpacking and packing values into vectors is only optimized with
extractelement instructions, not when manually unpacked using shifts
and truncates.
This patch simplifies shifts and truncates into a bitcast if possible.
Simplify (build_vec (trunc $1)
(trunc (srl $1 width))
(trunc (srl $1 (2 * width))) ...)
to (bitcast $1)
Differential Revision: https://reviews.llvm.org/D73892
Alex Zinenko [Mon, 10 Feb 2020 13:12:47 +0000 (14:12 +0100)]
[mlir] use unpacked memref descriptors at function boundaries
The existing (default) calling convention for memrefs in standard-to-LLVM
conversion was motivated by interfacing with LLVM IR produced from C sources.
In particular, it passes a pointer to the memref descriptor structure when
calling the function. Therefore, the descriptor is allocated on stack before
the call. This convention leads to several problems. PR44644 indicates a
problem with stack exhaustion when calling functions with memref-typed
arguments in a loop. Allocating outside of the loop may lead to concurrent
access problems in case the loop is parallel. When targeting GPUs, the contents
of the stack-allocated memory for the descriptor (passed by pointer) needs to
be explicitly copied to the device. Using an aggregate type makes it impossible
to attach pointer-specific argument attributes pertaining to alignment and
aliasing in the LLVM dialect.
Change the default calling convention for memrefs in standard-to-LLVM
conversion to transform a memref into a list of arguments, each of primitive
type, that are comprised in the memref descriptor. This avoids stack allocation
for ranked memrefs (and thus stack exhaustion and potential concurrent access
problems) and simplifies the device function invocation on GPUs.
Provide an option in the standard-to-LLVM conversion to generate auxiliary
wrapper function with the same interface as the previous calling convention,
compatible with LLVM IR porduced from C sources. These auxiliary functions
pack the individual values into a descriptor structure or unpack it. They also
handle descriptor stack allocation if necessary, serving as an allocation
scope: the memory reserved by `alloca` will be freed on exiting the auxiliary
function.
The effect of this change on MLIR-generated only LLVM IR is minimal. When
interfacing MLIR-generated LLVM IR with C-generated LLVM IR, the integration
only needs to require auxiliary functions and change the function name to call
the wrapper function instead of the original function.
This also opens the door to forwarding aliasing and alignment information from
memrefs to LLVM IR pointers in the standrd-to-LLVM conversion.
James Henderson [Fri, 7 Feb 2020 09:18:43 +0000 (09:18 +0000)]
[DebugInfo][test] Replace pre-canned binary test
The DebugInfo/dwarfdump-invalid-line-table test used a pre-canned binary
generated by a fuzzer to demonstrate a bug fix. Unfortunately, the
binary is rigid and requires hand-editing if we change behaviour, such
as rejecting certain properties within it (as I plan on doing in another
change).
Rather than hand-edit the binary, I have replaced it with two tests. The
first tests the high-level code path from the debug line parser that
produces the same error as this test previously did, and the second is a
set of unit test cases that comprehensively cover the
FormValue::skipValue method, which in turn covers the area that the
original bug fix touched.
Reviewed by: MaskRay, dblaikie
Differential Revision: https://reviews.llvm.org/D74202
Kai Nacke [Wed, 8 Jan 2020 19:26:12 +0000 (14:26 -0500)]
[SystemZ] Add implementation for the intrinsic llvm.read_register
This change implements the llvm intrinsic llvm.read_register for
the SystemZ platform which returns the value of the specified
register
(http://llvm.org/docs/LangRef.html#llvm-read-register-and-llvm-write-register-intrinsics).
This implementation returns the value of the stack register, and
can be extended to return the value of other registers. The
implementation for this intrinsic exists on various other platforms
including Power, x86, ARM, etc. but missing on SystemZ.
Reviewers: uweigand
Differential Revision: https://reviews.llvm.org/D73378
Hans Wennborg [Mon, 10 Feb 2020 13:07:41 +0000 (14:07 +0100)]
Fix an unused variable warning
Georgii Rymar [Wed, 5 Feb 2020 13:53:27 +0000 (16:53 +0300)]
[llvm-readobj] - Change the error to warning when a section name is unknown.
We reported the error in this case.
But it was asked (https://reviews.llvm.org/D73193#inline-665595) to convert it
to a warning. This patch does it.
Differential revision: https://reviews.llvm.org/D74047
Mikael Holmen [Mon, 10 Feb 2020 12:32:44 +0000 (13:32 +0100)]
Fix compiler warning when compiling without asserts [NFC]
Louis Dionne [Mon, 10 Feb 2020 12:47:27 +0000 (13:47 +0100)]
[libc++][span] Add failing tests for span::first and span::last
Both methods have compile time constraints that we should test against.
Patch by Michael Schellenberger Costa
Differential Revision: https://reviews.llvm.org/D71999
Kadir Cetinkaya [Mon, 10 Feb 2020 12:38:58 +0000 (13:38 +0100)]
[OpenMP] Fix unused variable
Nico Weber [Mon, 10 Feb 2020 12:42:27 +0000 (07:42 -0500)]
[gn build] make 'clang' target depend on libcxx/include on mac
On macOS, libc++ headers are distributed with the compiler, not
the sysroot. Without this, compiling a file that includes something
like <string> won't compile with gn-built clang without manual tweaks.
I used to do the manual tweaks, but now that other people are starting
to use this on mac, let's make it Just Work.
(This is marginally nicer than the cmake build now in that you can
just build 'clang' and it'll do the right thing.)
Differential Revision: https://reviews.llvm.org/D74247
Louis Dionne [Mon, 10 Feb 2020 12:38:04 +0000 (13:38 +0100)]
[libc++] Protect <span> against min/max macro
Patch by Corentin Jabot
Differential Revision: https://reviews.llvm.org/D73855
Florian Hahn [Mon, 10 Feb 2020 12:37:22 +0000 (12:37 +0000)]
[DSE,MSSA] Move more passing test cases from todo to simple.ll.
Kerry McLaughlin [Mon, 10 Feb 2020 11:33:06 +0000 (11:33 +0000)]
[AArch64][SVE] SVE2 intrinsics for complex integer arithmetic
Summary:
Adds the following SVE2 intrinsics:
- cadd & sqcadd
- cmla & sqrdcmlah
- saddlbt, ssublbt & ssubltb
Reviewers: sdesmalen, dancgr, efriedma, cameron.mcinally, c-rhodes, rengolin
Reviewed By: sdesmalen
Subscribers: tschuett, kristof.beyls, hiraditya, rkruppe, psnobl, cfe-commits, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D73636
Simon Pilgrim [Mon, 10 Feb 2020 11:58:05 +0000 (11:58 +0000)]
Revert rGe82e17d4d4cac8b2df00094e80d5e1cb22795664 - [X86] Add lowerShuffleAsBitRotate (PR44379)
As noted on PR44379, we didn't attempt to lower vector shuffles using bit rotations on XOP/AVX512F targets.
This patch lowers to uniform ISD:ROTL nodes - ROTR isn't supported by XOP and they are interchangeable for constant values anyway.
There might be cases where targets without ISD:ROTL support would benefit from this (expanding to SRL+SHL+OR), which I'll investigate in a future patch.
Also, non-AVX512BW targets fail to concatenate 256-bit rotations back to 512-bits (split during shuffle lowering as they don't have v32i16/v64i8 types).
---
Internal shuffle tests indicate theres a bug somewhere that I haven't been able to track down yet.
Raphael Isemann [Mon, 10 Feb 2020 11:45:19 +0000 (12:45 +0100)]
[lldb][NFC] Don't hide a bool in LibCxxOptional's OptionalFrontend::m_size
m_size can only be 1 or 0 and indicates if the optional has a value. Calling
it 'm_size', giving it a size_t data type and then also comparing indices against
'size' is very confusing. Let's just make this a bool.
Florian Hahn [Mon, 10 Feb 2020 10:44:37 +0000 (10:44 +0000)]
[DSE] Add first version of MemorySSA-backed DSE (Bottom up walk).
This patch adds a first version of a MemorySSA based DSE. It is missing
a lot of features, which will get added as follow-ups, to help to keep
the review manageable.
The patch uses the following general approach: given a MemoryDef, walk
upwards to find clobbering MemoryDefs that may be killed by the
starting def. Then check that there are no uses that may read the
location of the original MemoryDef in between both MemoryDefs. A bit
more concretely:
For all MemoryDefs StartDef:
1. Get the next dominating clobbering MemoryDef (DomAccess) by walking upwards.
2. Check that there no reads between DomAccess and the StartDef by checking
all uses starting at DomAccess and walking until we see StartDef.
3. For each found DomDef, check that:
1. There are no barrier instructions between DomDef and StartDef (like
throws or stores with ordering constraints).
2. StartDef is executed whenever DomDef is executed.
3. StartDef completely overwrites DomDef.
4. Erase DomDef from the function and MemorySSA.
The patch uses a very simple approach to guarantee that no throwing
instructions are between 2 stores: We only allow accesses to stack
objects, access that are in the same basic block if the block does not
contain any throwing instructions or accesses in functions that do
not contain any throwing instructions. This will get lifted later.
Besides adding support for the missing cases, there is plenty of additional
potential for improvements as follow-up work, e.g. the way we visit stores
(could be just a traversal of the MemorySSA, rather than collecting them
up-front), using the alias information discovered during walking to optimize
the MemorySSA.
This is loosely based on D40480 by Dave Green.
Reviewers: dmgreen, rnk, efriedma, bryant, asbirlea, Tyker
Reviewed By: asbirlea
Differential Revision: https://reviews.llvm.org/D72700
Raphael Isemann [Mon, 10 Feb 2020 11:39:42 +0000 (12:39 +0100)]
[lldb][NFC] Don't call call formatv for no reason in LibCxxOptional
Raphael Isemann [Mon, 10 Feb 2020 11:37:44 +0000 (12:37 +0100)]
[lldb][NFC] Fix code style of LibcxxVariantIndexValidity
Enum cases aren't all uppercase.
Raphael Isemann [Mon, 10 Feb 2020 11:23:01 +0000 (12:23 +0100)]
[lldb][NFC] Don't construct a ConstString twice in LibCxxVariant
Kerry McLaughlin [Mon, 10 Feb 2020 11:08:00 +0000 (11:08 +0000)]
[AArch64][SVE] SVE2 intrinsics for character match & histogram generation
Summary:
Implements the following intrinsics:
- @llvm.aarch64.sve.histcnt
- @llvm.aarch64.sve.histseg
- @llvm.aarch64.sve.match
- @llvm.aarch64.sve.nmatch
Reviewers: c-rhodes, sdesmalen, dancgr, efriedma, rengolin
Reviewed By: c-rhodes
Subscribers: tschuett, kristof.beyls, hiraditya, rkruppe, psnobl, cfe-commits, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D74117
Kirill Bobyrev [Mon, 10 Feb 2020 10:53:17 +0000 (11:53 +0100)]
[clangd] Support renaming designated initializers
Summary:
Clangd does not find references of designated iniitializers yet and, as a
result, is unable to rename such references. This patch addresses this issue.
Resolves: https://github.com/clangd/clangd/issues/247
Reviewers: sammccall
Reviewed By: sammccall
Subscribers: merge_guards_bot, ilya-biryukov, MaskRay, jkorous, arphaman, kadircet, usaxena95, cfe-commits
Tags: #clang
Differential Revision: https://reviews.llvm.org/D72867