Craig Topper [Wed, 9 Nov 2016 07:48:51 +0000 (07:48 +0000)]
[AVX-512] Add lowering to cvttpd2udq/cvttps2udq for fptoui v2f64/2f32 to 2i32
This patch adds support for fptoui to 2i32 from both 2f64 and 2f32, building on Simon's change for the signed version in r284459 and using AVX-512 instructions.
If we don't have VLX support we need to use a 512-bit operation for v2f64->v2i32 and extract the result.
It also recognises that cvttpd2udq zeroes the upper 64-bits of the xmm result.
Differential Revision: https://reviews.llvm.org/D26331
llvm-svn: 286345
Craig Topper [Wed, 9 Nov 2016 07:31:32 +0000 (07:31 +0000)]
[X86] Lower AVX512 and SSE intrinsics for CVTTPD2DQ to X86ISD::CVTTPD2DQ.
Summary: This allows the SSE intrinsic to use the EVEX instruction when available. It also fixes EVEX to not use a weird (v4i32 (fp_to_sint v2f64)) node and it merges some isel patterns. This also fixes some cases that weren't combining vzmovl with cvttpd2dq to remove extra moves.
Reviewers: delena, zvi, RKSimon
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D26330
llvm-svn: 286344
Craig Topper [Wed, 9 Nov 2016 05:38:47 +0000 (05:38 +0000)]
[AVX-512] Add more varied alignments to tests for storing the lower 128-bits of a 256 or 512-bit subvector extract.
llvm-svn: 286343
Craig Topper [Wed, 9 Nov 2016 05:31:57 +0000 (05:31 +0000)]
[AVX-512] Use alignedstore256 in patterns that look for stores of the lower 256-bits of a 512-bit vector to use a 256-bit aligned store.
Previously we were only checking for 16 byte alignment instead of 32 byte alignment. Fixes PR30947.
llvm-svn: 286342
Craig Topper [Wed, 9 Nov 2016 05:31:53 +0000 (05:31 +0000)]
[AVX-512] Add test cases to demonstrate PR30947. We accidentally use 32 byte aligned store instructions when the original store was only 16 byte aligned if the store is from the lower bits of a subvector extract.
llvm-svn: 286341
Craig Topper [Wed, 9 Nov 2016 04:51:03 +0000 (04:51 +0000)]
[AVX-512] Make VBMI instruction set enabling imply that the BWI instruction set is also enabled.
Summary: This is needed to make the v64i8 and v32i16 types legal for the 512-bit VBMI instructions. Fixes PR30912.
Reviewers: delena, zvi
Subscribers: RKSimon, cfe-commits
Differential Revision: https://reviews.llvm.org/D26306
llvm-svn: 286340
Craig Topper [Wed, 9 Nov 2016 04:50:48 +0000 (04:50 +0000)]
[AVX-512] Make VBMI instruction set enabling imply that the BWI instruction set is also enabled.
Summary:
This is needed to make the v64i8 and v32i16 types legal for the 512-bit VBMI instructions. Fixes PR30912.
Reviewers: delena, zvi
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D26322
llvm-svn: 286339
Tobias Grosser [Wed, 9 Nov 2016 04:24:49 +0000 (04:24 +0000)]
ScopInfo: only run code needed for ASSERT in DEBUG mode
Suggested-by: Johannes Doerfert
llvm-svn: 286338
Igor Kudrin [Wed, 9 Nov 2016 04:14:31 +0000 (04:14 +0000)]
[libc++abi] Remove the test for checking using of fallback malloc in case of dynamic memory exhaustion.
This test is too fragile and doesn't add significant value. See https://reviews.llvm.org/D26150 for some details.
llvm-svn: 286337
Ekaterina Romanova [Wed, 9 Nov 2016 03:58:30 +0000 (03:58 +0000)]
Doxygen comments for avxintrin.h.
Added doxygen comments to avxintrin.h's intrinsics. As of now, around 75% of the
intrinsics in this file are documented here. The patches for the other 25% will be se
nt out later.
Removed extra spaces in emmitrin.h.
Note: The doxygen comments are automatically generated based on Sony's intrinsics document.
I got an OK from Eric Christopher to commit doxygen comments without prior code
review upstream.
llvm-svn: 286336
Jason Molenda [Wed, 9 Nov 2016 03:42:12 +0000 (03:42 +0000)]
When deciding whether to use the source remapping dictionary from
a dSYM per-uuid plist, only use it when the DBGVersion key has a
value of 2 or greater.
<rdar://problem/
28889578>
<rdar://problem/
29131339>
llvm-svn: 286335
Petr Hosek [Wed, 9 Nov 2016 03:38:21 +0000 (03:38 +0000)]
Revert "[CMake] Check runtimes subdir when looking for libcxx and libuwind"
This reverts commit
eecb79506d88b268fb0d00cce178213b4aa17933.
llvm-svn: 286334
Petr Hosek [Wed, 9 Nov 2016 03:22:28 +0000 (03:22 +0000)]
[CMake] Check runtimes subdir when looking for libcxxabi
The runtimes subdir is the new location for runtimes, we should
include it when looking for libcxxabi headers.
Differential Revision: https://reviews.llvm.org/D26363
llvm-svn: 286333
Petr Hosek [Wed, 9 Nov 2016 03:22:19 +0000 (03:22 +0000)]
[CMake] Check runtimes subdir when looking for libcxx and libuwind
The runtimes subdir is the new location for runtimes, we should
include it when looking for libcxx and libunwind headers.
Differential Revision: https://reviews.llvm.org/D26362
llvm-svn: 286332
Argyrios Kyrtzidis [Wed, 9 Nov 2016 02:47:07 +0000 (02:47 +0000)]
[index] Fix issue with protocol name locations in conformance list of an ObjC class when they come from a typedef.
The ObjC class protocol list assumes there is an associated location for each protocol but no location is provided
when the protocol list comes from a typedef, and we end up with a buffer overflow when trying to get locations for the protocol names.
Fixes crash of rdar://
28980278.
llvm-svn: 286331
Dean Michael Berris [Wed, 9 Nov 2016 02:12:13 +0000 (02:12 +0000)]
[XRay][docs] Fix llvm snippets to be well-formed
llvm-svn: 286330
Mehdi Amini [Wed, 9 Nov 2016 01:45:13 +0000 (01:45 +0000)]
Revert "[ThinLTO] Prevent exporting of locals used/defined in module level asm"
This reverts commit r286297.
Introduces a dependency from libAnalysis to libObject, which I missed
during the review.
llvm-svn: 286329
Mehdi Amini [Wed, 9 Nov 2016 01:44:42 +0000 (01:44 +0000)]
[doc] Remove explicit CMake version requirement for MSVC
The global minimum one is way past this version.
llvm-svn: 286328
Rafael Espindola [Wed, 9 Nov 2016 01:42:41 +0000 (01:42 +0000)]
Split Header into individual fields.
This is similar to what was done for InputSection.
With this the various fields are stored in host order and only
converted to target order when writing.
llvm-svn: 286327
Peter Collingbourne [Wed, 9 Nov 2016 01:09:11 +0000 (01:09 +0000)]
Bitcode: Remove the remnants of the BitcodeDiagnosticInfo class.
The BitcodeReader no longer produces BitcodeDiagnosticInfo diagnostics.
The only remaining reference was in the gold plugin; the code there has been
dead since we stopped producing InvalidBitcodeSignature error codes in r225562.
While at it remove the InvalidBitcodeSignature error code.
llvm-svn: 286326
Dehao Chen [Wed, 9 Nov 2016 00:58:19 +0000 (00:58 +0000)]
Enable Loop Sink pass for functions that has profile.
Summary: For functions with profile data, we are confident that loop sink will be optimal in sinking code.
Reviewers: davidxl, hfinkel
Subscribers: mehdi_amini, mzolotukhin, llvm-commits
Differential Revision: https://reviews.llvm.org/D26155
llvm-svn: 286325
Hans Wennborg [Wed, 9 Nov 2016 00:56:42 +0000 (00:56 +0000)]
clang-cl: Pass /Zc:threadSafeInit through to MSVC with /fallback (PR30948)
llvm-svn: 286324
Peter Collingbourne [Wed, 9 Nov 2016 00:51:04 +0000 (00:51 +0000)]
Bitcode: Change the BitcodeReader to use llvm::Error internally.
Differential Revision: https://reviews.llvm.org/D26430
llvm-svn: 286323
Adrian Prantl [Wed, 9 Nov 2016 00:42:03 +0000 (00:42 +0000)]
Emit debug info for global constants whose address is taken exactly once.
Add a check to the DeclCache before emitting debug info for a
GlobalVariable a second time and just attach the previsously created one to it.
<rdar://problem/
26721101>
llvm-svn: 286322
Evgeniy Stepanov [Wed, 9 Nov 2016 00:33:43 +0000 (00:33 +0000)]
[asan] Ignore host LD_LIBRARY_PATH in device tests.
I'm not sure why is it there, but it is breaking tests on Android N
because of unexpected linker output about an empty LD_LIBRARY_PATH
entry.
llvm-svn: 286321
Evgeniy Stepanov [Wed, 9 Nov 2016 00:33:41 +0000 (00:33 +0000)]
[asan] Disable unexpected passing test.
This test fails on some versions of Android and passes on other versions.
Replace XFAIL with UNSUPPORTED.
llvm-svn: 286320
Dean Michael Berris [Wed, 9 Nov 2016 00:24:58 +0000 (00:24 +0000)]
[XRay][Docs] Add documentation for XRay in LLVM
Summary:
This is the initial version of the documentation for how to use XRay as
it stands in LLVM, Clang, and compiler-rt. We leave some room for later
expansion mentioining what is work in progress and what could be
expected moving forward.
We also give a high level overview of future work that's both ongoing
and planned.
Reviewers: echristo, dblaikie, chandlerc
Subscribers: mehdi_amini, llvm-commits
Differential Revision: https://reviews.llvm.org/D26386
llvm-svn: 286319
Sanjay Patel [Wed, 9 Nov 2016 00:24:44 +0000 (00:24 +0000)]
[ValueTracking] recognize obfuscated variants of umin/umax
The smallest tests that expose this are codegen tests (because SelectionDAGBuilder::visitSelect() uses matchSelectPattern
to create UMAX/UMIN nodes), but it's also possible to see the effects in IR alone with folds of min/max pairs.
If these were written as unsigned compares in IR, InstCombine canonicalizes the unsigned compares to signed compares.
Ie, running the optimizer pessimizes the codegen for this case without this patch:
define <4 x i32> @umax_vec(<4 x i32> %x) {
%cmp = icmp ugt <4 x i32> %x, <i32
2147483647, i32
2147483647, i32
2147483647, i32
2147483647>
%sel = select <4 x i1> %cmp, <4 x i32> %x, <4 x i32> <i32
2147483647, i32
2147483647, i32
2147483647, i32
2147483647>
ret <4 x i32> %sel
}
$ ./opt umax.ll -S | ./llc -o - -mattr=avx
vpmaxud LCPI0_0(%rip), %xmm0, %xmm0
$ ./opt -instcombine umax.ll -S | ./llc -o - -mattr=avx
vpxor %xmm1, %xmm1, %xmm1
vpcmpgtd %xmm0, %xmm1, %xmm1
vmovaps LCPI0_0(%rip), %xmm2 ## xmm2 = [
2147483647,
2147483647,
2147483647,
2147483647]
vblendvps %xmm1, %xmm0, %xmm2, %xmm0
Differential Revision: https://reviews.llvm.org/D26096
llvm-svn: 286318
Mehdi Amini [Wed, 9 Nov 2016 00:23:20 +0000 (00:23 +0000)]
[cmake] Fix handling compiler-rt in LLVM_ENABLE_PROJECTS by turning any "-" into "_"
llvm-svn: 286317
Greg Clayton [Wed, 9 Nov 2016 00:15:54 +0000 (00:15 +0000)]
Added the ability to dump hex bytes easily into a raw_ostream.
Unit tests were added to verify this functionality keeps working correctly.
Example output for raw hex bytes:
llvm::ArrayRef<uint8_t> Bytes = ...;
llvm::outs() << format_hex_bytes(Bytes);
554889e5 4881ec70 04000048 8d051002
00004c8d 05fd0100 004c8b0d d0020000
Example output for raw hex bytes with offsets:
llvm::outs() << format_hex_bytes(Bytes, 0x100000d10);
0x0000000100000d10:
554889e5 4881ec70 04000048 8d051002
0x0000000100000d20:
00004c8d 05fd0100 004c8b0d d0020000
Example output for raw hex bytes with ASCII with offsets:
llvm::outs() << format_hex_bytes_with_ascii(Bytes, 0x100000d10);
0x0000000100000d10:
554889e5 4881ec70 04000048 8d051002 |UH.?H.?p...H....|
0x0000000100000d20:
00004c8d 05fd0100 004c8b0d d0020000 |..L..?...L..?...|
The default groups bytes into 4 byte groups, but this can be changed to 1 byte:
llvm::outs() << format_hex_bytes(Bytes, 0x100000d10, 16 /*NumPerLine*/, 1 /*ByteGroupSize*/);
0x0000000100000d10: 55 48 89 e5 48 81 ec 70 04 00 00 48 8d 05 10 02
0x0000000100000d20: 00 00 4c 8d 05 fd 01 00 00 4c 8b 0d d0 02 00 00
llvm::outs() << format_hex_bytes(Bytes, 0x100000d10, 16 /*NumPerLine*/, 2 /*ByteGroupSize*/);
0x0000000100000d10: 5548 89e5 4881 ec70 0400 0048 8d05 1002
0x0000000100000d20: 0000 4c8d 05fd 0100 004c 8b0d d002 0000
llvm::outs() << format_hex_bytes(Bytes, 0x100000d10, 8 /*NumPerLine*/, 1 /*ByteGroupSize*/);
0x0000000100000d10: 55 48 89 e5 48 81 ec 70
0x0000000100000d18: 04 00 00 48 8d 05 10 02
0x0000000100000d20: 00 00 4c 8d 05 fd 01 00
0x0000000100000d28: 00 4c 8b 0d d0 02 00 00
https://reviews.llvm.org/D26405
llvm-svn: 286316
Sanjay Patel [Wed, 9 Nov 2016 00:13:11 +0000 (00:13 +0000)]
[InstCombine] fix profitability equation for max-of-nots transform
As the test change shows, we can increase the critical path by adding
a 'not' instruction, so make sure that we're actually removing an
instruction if we do this transform.
This transform could also cause us to miss folds of min/max pairs.
llvm-svn: 286315
Sanjay Patel [Tue, 8 Nov 2016 23:49:15 +0000 (23:49 +0000)]
[InstCombine] reduce indentation; NFC
llvm-svn: 286314
Justin Lebar [Tue, 8 Nov 2016 23:45:51 +0000 (23:45 +0000)]
[CUDA] Use only the GVALinkage on function definitions.
Summary:
Previously we'd look at the GVALinkage of whatever FunctionDecl you
happened to be calling.
This is not right. In the absence of the gnu_inline attribute, to be
handled separately, the function definition determines the function's
linkage. So we need to wait until we get a def before we can know
whether something is known-emitted.
Reviewers: tra
Subscribers: cfe-commits, rsmith
Differential Revision: https://reviews.llvm.org/D26268
llvm-svn: 286313
Jim Ingham [Tue, 8 Nov 2016 23:43:36 +0000 (23:43 +0000)]
"thread backtrace" should use the thread-stop-format.
<rdar://problem/
28273697>
llvm-svn: 286312
Petr Hosek [Tue, 8 Nov 2016 23:02:49 +0000 (23:02 +0000)]
[CMake] Get libunwind building under LLVM/runtimes
The new LLVM runtimes directory requires the same conventions to be
followed across the runtime projects. These changes make libunwind
build under the runtimes subdirectory.
This patch contains the following changes:
* Rename LLVM_CONFIG to LLVM_CONFIG_PATH
* Check if compiler supports C++11 (required by libunwind)
Differential Revision: https://reviews.llvm.org/D26360
llvm-svn: 286308
Eric Liu [Tue, 8 Nov 2016 22:44:17 +0000 (22:44 +0000)]
[change-namespace] shorten namespace qualifier based on UsingDecl and UsingDirectiveDecl.
Summary:
when replacing symbol references in moved namespaces, trying to make the replace
name as short as possible by considering UsingDecl (i.e. UsingShadow) and
UsingDirectiveDecl (i.e. using namespace decl).
Reviewers: hokein
Subscribers: cfe-commits
Differential Revision: https://reviews.llvm.org/D25771
llvm-svn: 286307
Dominic Chen [Tue, 8 Nov 2016 22:39:14 +0000 (22:39 +0000)]
[analyzer] Provide Contains() on ImmutableMap program state partial trait.
Reviewers: zaks.anna, dcoughlin
Subscribers: cfe-commits
Differential Revision: https://reviews.llvm.org/D26373
llvm-svn: 286306
Zachary Turner [Tue, 8 Nov 2016 22:30:11 +0000 (22:30 +0000)]
Fix some size_t / uint32_t ambiguity errors.
llvm-svn: 286305
Zachary Turner [Tue, 8 Nov 2016 22:24:53 +0000 (22:24 +0000)]
[CodeView] Hook up CodeViewRecordIO to type serialization path.
Previously support had been added for using CodeViewRecordIO
to read (deserialize) CodeView type records. This patch adds
support for writing those same records. With this patch,
reading and writing of CodeView type records finally uses a single
codepath.
Differential Revision: https://reviews.llvm.org/D26253
llvm-svn: 286304
Zachary Turner [Tue, 8 Nov 2016 22:23:50 +0000 (22:23 +0000)]
Fix some cases where we were printf'ing StringRefs.
llvm-svn: 286303
Adrian Prantl [Tue, 8 Nov 2016 22:11:38 +0000 (22:11 +0000)]
Emit the DW_AT_type for a C++ static member definition
if it is more specific than the one in its DW_AT_specification.
If a static member is an array, the translation unit containing the
member definition may have a more specific type (including its length)
than TUs only seeing the class declaration. This patch adds a
DW_AT_type to the member's DW_TAG_variable in addition to the
DW_AT_specification in these cases. The member type in the
DW_AT_specification still shows the more generic type (without the
length) to avoid defeating type uniquing.
The DWARF standard discourages “duplicating” a DW_AT_type in a member
variable definition but doesn’t explicitly forbid it. Having the more
specific type (with the array length) available is what allows the
debugger to print the contents of a static array member variable.
https://reviews.llvm.org/D26368
rdar://problem/
28706946
llvm-svn: 286302
Jim Ingham [Tue, 8 Nov 2016 22:05:29 +0000 (22:05 +0000)]
Fix up the formats.html for the addition of the thread-stop-format.
<rdar://problem/
28273697>
llvm-svn: 286301
David L. Jones [Tue, 8 Nov 2016 22:03:23 +0000 (22:03 +0000)]
GlobalISel: make sure debugging variables are appropriately elided in release builds.
Summary:
There are two variables here that break. This change constrains both of them to
debug builds (via DEBUG() or #ifndef NDEBUG).
Reviewers: bkramer, t.p.northover
Subscribers: mehdi_amini, vkalintiris
Differential Revision: https://reviews.llvm.org/D26421
llvm-svn: 286300
Kostya Serebryany [Tue, 8 Nov 2016 21:57:37 +0000 (21:57 +0000)]
[libFuzzer] minor docs update
llvm-svn: 286299
Stephen Hines [Tue, 8 Nov 2016 21:54:49 +0000 (21:54 +0000)]
Revert "Define __ANDROID_API__ for all Android builds."
Summary: This reverts commit
a8804ddd9fe71304b28e5b834d134fe93e568ee0.
Subscribers: cfe-commits, pirama, eugenis, tberghammer, danalbert
Differential Revision: https://reviews.llvm.org/D26422
llvm-svn: 286298
Teresa Johnson [Tue, 8 Nov 2016 21:53:35 +0000 (21:53 +0000)]
[ThinLTO] Prevent exporting of locals used/defined in module level asm
Summary:
This patch uses the same approach added for inline asm in r285513 to
similarly prevent promotion/renaming of locals used or defined in module
level asm.
All static global values defined in normal IR and used in module level asm
should be included on either the llvm.used or llvm.compiler.used global.
The former were already being flagged as NoRename in the summary, and
I've simply added llvm.compiler.used values to this handling.
Module level asm may also contain defs of values. We need to prevent
export of any refs to local values defined in module level asm (e.g. a
ref in normal IR), since that also requires renaming/promotion of the
local. To do that, the summary index builder looks at all values in the
module level asm string that are not marked Weak or Global, which is
exactly the set of locals that are defined. A summary is created for
each of these local defs and flagged as NoRename.
This required adding handling to the BitcodeWriter to look at GV
declarations to see if they have a summary (rather than skipping them
all).
Finally, added an assert to IRObjectFile::CollectAsmUndefinedRefs to
ensure that an MCAsmParser is available, otherwise the module asm parse
would silently fail. Initialized the asm parser in the opt tool for use
in testing this fix.
Fixes PR30610.
Reviewers: mehdi_amini
Subscribers: johanengelen, krasin, llvm-commits
Differential Revision: https://reviews.llvm.org/D26146
llvm-svn: 286297
Kuba Brecka [Tue, 8 Nov 2016 21:30:41 +0000 (21:30 +0000)]
[asan] Speed up compilation of large C++ stringmaps (tons of allocas) with ASan
This addresses PR30746, <https://llvm.org/bugs/show_bug.cgi?id=30746>. The ASan pass iterates over entry-block instructions and checks each alloca whether it's in NonInstrumentedStaticAllocaVec, which is apparently slow. This patch gathers the instructions to move during visitAllocaInst.
Differential Revision: https://reviews.llvm.org/D26380
llvm-svn: 286296
Stephen Hines [Tue, 8 Nov 2016 21:23:26 +0000 (21:23 +0000)]
Define __ANDROID_API__ for all Android builds.
Summary:
Bug: https://llvm.org/bugs/show_bug.cgi?id=30940
This macro (along with __ANDROID__) should always be defined for Android
targets. We set it to the major (only) version of the Android API being
compiled for. The Android version is able to be set as an integer suffix
for any valid Android target.
Reviewers: danalbert, eugenis
Subscribers: cfe-commits, pirama, eugenis, tberghammer, danalbert
Differential Revision: https://reviews.llvm.org/D26385
llvm-svn: 286295
Andrew Kaylor [Tue, 8 Nov 2016 21:07:42 +0000 (21:07 +0000)]
[BasicAA] Teach BasicAA to handle the inaccessiblememonly and inaccessiblemem_or_argmemonly attributes
Differential Revision: https://reviews.llvm.org/D26382
llvm-svn: 286294
Matthias Braun [Tue, 8 Nov 2016 20:59:03 +0000 (20:59 +0000)]
AArch64DeadRegisterDefinitionsPass: Fix Changed flag
Fix a bug in the calculation of the changed flag introduced in r285488.
llvm-svn: 286293
Adrian Prantl [Tue, 8 Nov 2016 20:48:38 +0000 (20:48 +0000)]
Use a default constructor. (NFC)
Thanks to David Blaikie for suggesting this.
llvm-svn: 286292
Sanjoy Das [Tue, 8 Nov 2016 20:46:01 +0000 (20:46 +0000)]
[TBAA] Drop support for "old style" scalar TBAA tags
Summary:
We've had support for auto upgrading old style scalar TBAA access
metadata tags into the "new" struct path aware TBAA metadata for 3 years
now. The only way to actually generate old style TBAA was explicitly
through the IRBuilder API. I think this is a good time for dropping
support for old style scalar TBAA.
I'm not removing support for textual or bitcode upgrade -- if you have
IR with the old style scalar TBAA tags that go through the AsmParser orf
the bitcode parser before LLVM sees them, they will keep working as
usual.
Note:
%val = load i32, i32* %ptr, !tbaa !N
!N = < scalar tbaa node >
is equivalent to
%val = load i32, i32* %ptr, !tbaa !M
!N = < scalar tbaa node >
!M = !{!N, !N, 0}
Reviewers: manmanren, chandlerc, sunfish
Subscribers: mcrosier, llvm-commits, mgorny
Differential Revision: https://reviews.llvm.org/D26229
llvm-svn: 286291
Reid Kleckner [Tue, 8 Nov 2016 20:45:45 +0000 (20:45 +0000)]
[asan/win] Add init hooks to .CRT$XLAB
Summary:
User applications may register hooks in the .CRT$XL* callback list,
which is called very early by the loader. This is very common in
Chromium:
https://cs.chromium.org/search/?q=CRT.XL&sq=package:chromium&type=cs
This has flown under the radar for a long time because the loader
appears to catch exceptions originating from these callbacks. It's a
real problem when you're debugging an asan application, though, since it
makes the program crash early.
The solution is to add our own callback to this list, and sort it very
early in the list like we do elsewhere. Also add a test with such an
instrumented callback, and test that it gets called with asan.
Reviewers: etienneb
Subscribers: llvm-commits, kubabrecka
Differential Revision: https://reviews.llvm.org/D26404
llvm-svn: 286290
Tim Northover [Tue, 8 Nov 2016 20:39:03 +0000 (20:39 +0000)]
GlobalISel: allow CodeGen to fallback on VReg type/class issues.
After instruction selection we perform some checks on each VReg just before
discarding the type information. These checks were assertions before, but that
breaks the fallback path so this patch moves the logic into the main flow and
reports a better error on failure.
llvm-svn: 286289
Jim Ingham [Tue, 8 Nov 2016 20:36:40 +0000 (20:36 +0000)]
Clean up the stop printing header lines.
I added a "thread-stop-format" to distinguish between the form
that is just the thread info (since the stop printing immediately prints
the frame info) and one with more frame 0 info - which is useful for
"thread list" and the like.
I also added a frame.no-debug boolean to the format entities so you can
print frame information differently between frames with source info and those
without.
This closes https://reviews.llvm.org/D26383.
<rdar://problem/
28273697>
llvm-svn: 286288
Rui Ueyama [Tue, 8 Nov 2016 20:30:19 +0000 (20:30 +0000)]
Simplify getLocation() function.
All tests pass without the first parameter, so I guess we don't need it.
Differential Revision: https://reviews.llvm.org/D26411
llvm-svn: 286287
Rafael Espindola [Tue, 8 Nov 2016 20:25:44 +0000 (20:25 +0000)]
Store the size the same way as any other OutputSection. NFC.
llvm-svn: 286286
Ulrich Weigand [Tue, 8 Nov 2016 20:18:41 +0000 (20:18 +0000)]
[SystemZ] Add missing FP extension instructions
This completes assembler / disassembler support for all BFP
instructions provided by the floating-point extensions facility.
The instructions added here are not currently used for codegen.
llvm-svn: 286285
Ulrich Weigand [Tue, 8 Nov 2016 20:17:02 +0000 (20:17 +0000)]
[SystemZ] Add program mask and addressing mode instructions
Add several instructions that operate on the program mask
or the addressing mode. These are not really needed for
code generation under Linux, but are provided for completeness
for the assembler/disassembler.
llvm-svn: 286284
Ulrich Weigand [Tue, 8 Nov 2016 20:15:26 +0000 (20:15 +0000)]
[SystemZ] Model access registers as LLVM registers
Add the 16 access registers as LLVM registers. This allows removing
a lot of special cases in the assembler and disassembler where we
were handling access registers; this can all just use the generic
register code now.
Also add a bunch of instructions to operate on access registers,
for assembler/disassembler use only. No change in code generation
intended.
llvm-svn: 286283
Rui Ueyama [Tue, 8 Nov 2016 20:02:23 +0000 (20:02 +0000)]
Rename a function to avoid function overloading. NFC.
llvm-svn: 286282
Haojian Wu [Tue, 8 Nov 2016 19:55:13 +0000 (19:55 +0000)]
[clang-move] Move all code from old.h/cc directly when moving all class declarations from old.h.
Summary: When moving all code to new.h/cc, these code also will be formatted based on the given code style.
Reviewers: ioeric
Subscribers: cfe-commits
Differential Revision: https://reviews.llvm.org/D26236
llvm-svn: 286281
Davide Italiano [Tue, 8 Nov 2016 19:52:32 +0000 (19:52 +0000)]
[LoopDistribute] Preserve GlobalsAA also in the new Pass Manager.
Differential Revision: https://reviews.llvm.org/D26408
llvm-svn: 286280
Daniel Jasper [Tue, 8 Nov 2016 19:47:19 +0000 (19:47 +0000)]
Remove FormatContext from libClang as it is now unused.
llvm-svn: 286279
Eli Friedman [Tue, 8 Nov 2016 19:43:56 +0000 (19:43 +0000)]
Don't store Twine in a local variable.
Fixes post-commit review comment from r286177.
llvm-svn: 286275
Dan Gohman [Tue, 8 Nov 2016 19:40:38 +0000 (19:40 +0000)]
[WebAssembly] Convert stackified IMPLICIT_DEF into constant 0.
Since IMPLIFIT_DEF instructions are omitted in the output, when the output
of an IMPLICIT_DEF instruction is stackified, the resulting register lacks
an explicit push, leading to a push/pop mismatch. Fix this by converting
such IMPLICIT_DEFs into CONST_I32 0 instructions so that they have explicit
pushes.
llvm-svn: 286274
Ahmed Bougacha [Tue, 8 Nov 2016 19:27:13 +0000 (19:27 +0000)]
[GlobalISel] Dump all instructions inserted by selector.
This is helpful when multiple instructions are inserted.
llvm-svn: 286273
Ahmed Bougacha [Tue, 8 Nov 2016 19:27:10 +0000 (19:27 +0000)]
[GlobalISel] Permit select() to erase.
Erasing reverse_iterators is problematic; iterate manually.
While there, keep track of the range of inserted instructions.
It can miss instructions inserted elsewhere, but those are harder
to track.
Differential Revision: http://reviews.llvm.org/D22924
llvm-svn: 286272
Davide Italiano [Tue, 8 Nov 2016 19:18:20 +0000 (19:18 +0000)]
[LibcallsShrinkWrap] This pass doesn't preserve the CFG.
For example, it invalidates the domtree, causing assertions
in later passes which need dominator infos. Make it preserve
GlobalsAA, as suggested by Eli.
Differential Revision: https://reviews.llvm.org/D26381
llvm-svn: 286271
Chad Rosier [Tue, 8 Nov 2016 19:10:25 +0000 (19:10 +0000)]
Fix typo in comment. NFC.
llvm-svn: 286270
Michael Kuperstein [Tue, 8 Nov 2016 18:44:40 +0000 (18:44 +0000)]
CODE_OWNERS: Take ownership of the loop vectorizer.
llvm-svn: 286269
Ulrich Weigand [Tue, 8 Nov 2016 18:37:48 +0000 (18:37 +0000)]
[SystemZ] Always use semantic instruction classes
Define a couple of additional semantic classes and use them
throughout the .td files to make them more consistent and
more easily readable.
No functional change.
llvm-svn: 286268
Ulrich Weigand [Tue, 8 Nov 2016 18:36:31 +0000 (18:36 +0000)]
[SystemZ] Refactor InstRR* instruction format patterns
This changes the InstRR (and related) patterns to no longer
automatically add an "r" at the end of the mnemonic. This
makes the .td files more obviously understandable, and also
allows using the patterns for those few instructions that
do not follow the *r scheme.
Also add some more sub-formats of the RRF format class, to
match operand names and sequence from the PoP better.
No functional change.
llvm-svn: 286267
Ulrich Weigand [Tue, 8 Nov 2016 18:32:50 +0000 (18:32 +0000)]
[SystemZ] Rename some Inst* instruction format classes
Now that we've added instruction format subclasses like
InstRIb, it makes sense to rename the old InstRI to InstRIa.
Similar for InstRX, InstRXY, InstRS, InstRSY, and InstSS.
No functional change.
llvm-svn: 286266
Nirav Dave [Tue, 8 Nov 2016 18:31:04 +0000 (18:31 +0000)]
[MC][AArch64] Cleanup end-of-line parsing in AArch64 AsmParser.
Reviewers: t.p.northover, rengolin
Subscribers: llvm-commits, aemerson
Differential Revision: https://reviews.llvm.org/D26309
llvm-svn: 286265
Daniel Jasper [Tue, 8 Nov 2016 18:30:52 +0000 (18:30 +0000)]
Revert "Remove now unnecessary FormatRewriterContext."
This reverts commit r286262. I accidentally committed it without all of
the changes.
llvm-svn: 286264
Ulrich Weigand [Tue, 8 Nov 2016 18:30:50 +0000 (18:30 +0000)]
[SystemZ] Refactor branch and conditional instruction patterns
Rework patterns for branches, call & return instructions,
compare-and-branch, compare-and-trap, and conditional move
instructions.
In particular, simplify creation of patterns for the extended
opcodes of instructions that take a CC mask.
Also, use semantical instruction classes for all the instructions
instead of open-coding them in SystemZInstrInfo.td.
Adds a couple of the basic branch instructions (that are unused
for codegen) for the assembler/disassembler.
llvm-svn: 286263
Daniel Jasper [Tue, 8 Nov 2016 18:29:19 +0000 (18:29 +0000)]
Remove now unnecessary FormatRewriterContext.
llvm-svn: 286262
Rafael Espindola [Tue, 8 Nov 2016 18:23:02 +0000 (18:23 +0000)]
Don't add null and discarded sections to the global list.
Avoids having to skip them multiple times.
llvm-svn: 286261
Piotr Padlewski [Tue, 8 Nov 2016 18:20:51 +0000 (18:20 +0000)]
NFC small changes in MemDep
llvm-svn: 286260
Wei Mi [Tue, 8 Nov 2016 18:19:36 +0000 (18:19 +0000)]
[RegAllocGreedy] Another fix about NewVRegs for last chance recoloring after r281783.
About when we should move a vreg from CurrentNewVRegs to NewVRegs,
if the vreg in CurrentNewVRegs was added into RecoloringCandidate and was
evicted, it shouldn't be added to NewVRegs because its physical register
will be restored at the end of tryLastChanceRecoloring after the recoloring
failed. If the vreg in CurrentNewVRegs was not in RecoloringCandidate, i.e.
it was evicted in selectOrSplitImpl inside tryRecoloringCandidates, its
physical register will not be restored even if the recoloring failed. In
that case, we need to add the vreg to NewVRegs.
Same as r281783, the problem was seen on out-of-tree target and we didn't
have a test case that reproduce the problem with in-tree targets.
llvm-svn: 286259
Chris Bieneman [Tue, 8 Nov 2016 18:14:42 +0000 (18:14 +0000)]
[Test Suite] Properly quote python string
Oops!
llvm-svn: 286258
Alexander Kornienko [Tue, 8 Nov 2016 18:12:56 +0000 (18:12 +0000)]
Remove mentions of clang-analyzer-alpha
llvm-svn: 286257
Sanjay Patel [Tue, 8 Nov 2016 18:12:19 +0000 (18:12 +0000)]
[InstCombine] move min/max tests to min/max test file; NFC
llvm-svn: 286256
Sanjay Patel [Tue, 8 Nov 2016 18:06:14 +0000 (18:06 +0000)]
[InstCombine] update checks; NFC
llvm-svn: 286255
Chris Bieneman [Tue, 8 Nov 2016 17:53:18 +0000 (17:53 +0000)]
[Test Suite] Attempt to fix issue zturner reported to me
Not sure why this didn't explode more massively, but this should fix the issue with the non-framework tests.
llvm-svn: 286254
Tim Northover [Tue, 8 Nov 2016 17:44:07 +0000 (17:44 +0000)]
GlobalISel: support selecting fpext/fptrunc instructions on AArch64.
llvm-svn: 286253
Anton Korobeynikov [Tue, 8 Nov 2016 17:19:59 +0000 (17:19 +0000)]
Fix PR27500: on MSP430 the branch destination offset is measured in words, not bytes.
Summary: In addition, the branch instructions will have proper BB destinations, not offsets, like before.
Reviewers: asl
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D23718
llvm-svn: 286252
Chad Rosier [Tue, 8 Nov 2016 16:51:19 +0000 (16:51 +0000)]
Remove unused include. NFC.
llvm-svn: 286250
Sanjay Patel [Tue, 8 Nov 2016 16:49:24 +0000 (16:49 +0000)]
[docs] fix link to AMD manuals (PR30946)
llvm-svn: 286249
Eugene Leviant [Tue, 8 Nov 2016 16:44:48 +0000 (16:44 +0000)]
[ELF] Attempt to fix buildbot (duplicate _gp symbol)
llvm-svn: 286247
Dehao Chen [Tue, 8 Nov 2016 16:32:32 +0000 (16:32 +0000)]
Use the last 7 bits to represent the discriminator to fit it in 1 byte ULEB128 (NFC).
From experiments, discriminator is rarely greater than 127. Here we enforce it to be no greater than 127 so that it will always fit in 1 byte.
llvm-svn: 286245
Eugene Leviant [Tue, 8 Nov 2016 16:26:32 +0000 (16:26 +0000)]
[ELF] Better error reporting for duplicate symbol
Differential revision: https://reviews.llvm.org/D26397
llvm-svn: 286244
Daniel Jasper [Tue, 8 Nov 2016 16:11:33 +0000 (16:11 +0000)]
[clang-format] Remove (SourceManager, FileID) variants
In Format, remove the reformat() and clean() functions taking a SourceManager
and a FileID. Keep the versions taking StringRef Code.
- there was duplicated functionality
- the FileID versions were harder to use
- the clean() version is dead code anyways
Patch by Krasimir Georgiev. Thank you.
llvm-svn: 286243
Rafael Espindola [Tue, 8 Nov 2016 15:51:00 +0000 (15:51 +0000)]
Remove dead arguments. NFC.
llvm-svn: 286242
Simon Pilgrim [Tue, 8 Nov 2016 15:42:49 +0000 (15:42 +0000)]
[X86][SSE] Regenerate test (just adds missing header)
llvm-svn: 286241
George Rimar [Tue, 8 Nov 2016 15:26:21 +0000 (15:26 +0000)]
[ELF] - Fix mistype. NFC.
Thanks to Malcolm Parsons who pointed on that.
llvm-svn: 286239
Simon Pilgrim [Tue, 8 Nov 2016 15:07:01 +0000 (15:07 +0000)]
[TargetLowering] Fix undef vector element issue with true/false result handling
Fixed an issue with vector usage of TargetLowering::isConstTrueVal / TargetLowering::isConstFalseVal boolean result matching.
The comment said we shouldn't handle constant splat vectors with undef elements. But the the actual code was returning false if the build vector contained no undef elements....
This patch now ignores the number of undefs (getConstantSplatNode will return null if the build vector is all undefs).
The change has also unearthed a couple of missed opportunities in AVX512 comparison code that will need to be addressed.
Differential Revision: https://reviews.llvm.org/D26031
llvm-svn: 286238
Rafael Espindola [Tue, 8 Nov 2016 14:56:27 +0000 (14:56 +0000)]
Delete the InterpSection class.
We can just use a regular InputSection.
llvm-svn: 286237
Pablo Barrio [Tue, 8 Nov 2016 14:53:30 +0000 (14:53 +0000)]
[JumpThreading] Unfold selects that depend on the same condition
Summary:
These are good candidates for jump threading. This enables later opts
(such as InstCombine) to combine instructions from the selects with
instructions out of the selects. SimplifyCFG will fold the select
again if unfolding wasn't worth it.
Patch by James Molloy and Pablo Barrio.
Reviewers: rengolin, haicheng, sebpop
Subscribers: jojo, jmolloy, llvm-commits
Differential Revision: https://reviews.llvm.org/D26391
llvm-svn: 286236