OSDN Git Service
Lang Hames [Tue, 17 Jan 2017 04:07:48 +0000 (04:07 +0000)]
[Orc][RPC] Return unsupported rpc function errors from the non-retry cases in
negotiateFunction.
These cases were accidentally left out of r292055, resulting in a less
descriptive ECError being returned on these paths.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@292193
91177308-0d34-0410-b5e6-
96231b3b80d8
Ahmed Bougacha [Tue, 17 Jan 2017 03:31:00 +0000 (03:31 +0000)]
Revert "[TLI] Robustize SDAG proto checking by merging it into TLI."
This reverts commit r292189, as it causes issues on SystemZ bots.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@292191
91177308-0d34-0410-b5e6-
96231b3b80d8
Ahmed Bougacha [Tue, 17 Jan 2017 03:10:06 +0000 (03:10 +0000)]
[TLI] Robustize SDAG proto checking by merging it into TLI.
SelectionDAGBuilder recognizes libfuncs using some homegrown
parameter type-checking.
Use TLI instead, removing another heap of redundant code.
This isn't strictly NFC, as the SDAG code was too lax.
Concretely, this means changes are required to two tests:
- calling a non-variadic function via a variadic prototype isn't OK;
it just happens to work on x86_64 (but not on, e.g., aarch64).
- mempcpy has a size_t parameter; the SDAG code accepts any integer
type, which meant using i32 on x86_64 worked.
I don't think it's worth supporting either of these (IMO) broken
testcases. Instead, fix them to be more correct.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@292189
91177308-0d34-0410-b5e6-
96231b3b80d8
Ahmed Bougacha [Tue, 17 Jan 2017 03:10:02 +0000 (03:10 +0000)]
[TLI] Add prototype checking for all remaining LibFuncs.
This is another step towards unifying all LibFunc prototype checks.
This work started in r267758 (D19469); add the remaining checks.
Also add a unittest that checks each libfunc declared with a known-valid
and known-invalid prototype. New libfuncs added in the future are
required to have prototype checking in place; the known-valid test will
fail otherwise.
Differential Revision: https://reviews.llvm.org/D28030
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@292188
91177308-0d34-0410-b5e6-
96231b3b80d8
Ahmed Bougacha [Tue, 17 Jan 2017 03:10:00 +0000 (03:10 +0000)]
[TLI] Alphabetize some of the prototype check switch. NFC.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@292187
91177308-0d34-0410-b5e6-
96231b3b80d8
Ahmed Bougacha [Tue, 17 Jan 2017 03:09:55 +0000 (03:09 +0000)]
[unittests] Alphabetize cmake file list. NFC.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@292186
91177308-0d34-0410-b5e6-
96231b3b80d8
Alexei Starovoitov [Tue, 17 Jan 2017 01:05:17 +0000 (01:05 +0000)]
[bpf] error when BPF stack size exceeds 512 bytes
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@292180
91177308-0d34-0410-b5e6-
96231b3b80d8
David Majnemer [Tue, 17 Jan 2017 00:45:57 +0000 (00:45 +0000)]
[InstCombine] Fold ((C1-zext(X)) & C2) -> zext((C1-X) & C2)
This is valid if C2 fits within the bitwidth of X thanks to two's
complement modulo arithmetic.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@292179
91177308-0d34-0410-b5e6-
96231b3b80d8
Matt Arsenault [Tue, 17 Jan 2017 00:35:28 +0000 (00:35 +0000)]
Add comment to test file I forgot to save
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@292178
91177308-0d34-0410-b5e6-
96231b3b80d8
Matt Arsenault [Tue, 17 Jan 2017 00:30:31 +0000 (00:30 +0000)]
SimplifyLibCalls: Remove checks for fabs
Use the intrinsic instead of emitting the libcall which
will be replaced by the intrinsic.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@292176
91177308-0d34-0410-b5e6-
96231b3b80d8
Matt Arsenault [Tue, 17 Jan 2017 00:10:43 +0000 (00:10 +0000)]
AMDGPU: Remove dead pattern
This is the unsafe conversion pattern, but not guarded by
an unsafe math check. It is also already done in LegalizeDAG.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@292173
91177308-0d34-0410-b5e6-
96231b3b80d8
Matt Arsenault [Tue, 17 Jan 2017 00:10:40 +0000 (00:10 +0000)]
SimplifyLibCalls: Replace fabs libcalls with intrinsics
Add missing fabs(fpext) optimzation that worked with the call,
and also fixes it creating a second fpext when there were multiple
uses.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@292172
91177308-0d34-0410-b5e6-
96231b3b80d8
Davide Italiano [Mon, 16 Jan 2017 23:28:58 +0000 (23:28 +0000)]
[Object] Fixup permissions of input files.
They just need to be read/dumped, so no need to set the exec
bit on any of them. NFCI, I guess.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@292171
91177308-0d34-0410-b5e6-
96231b3b80d8
Davide Italiano [Mon, 16 Jan 2017 23:13:46 +0000 (23:13 +0000)]
[llvm-objdump] Dump PT_NOTE as part of -p.
PR: 31641
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@292170
91177308-0d34-0410-b5e6-
96231b3b80d8
Davide Italiano [Mon, 16 Jan 2017 22:58:26 +0000 (22:58 +0000)]
[llvm-objdump] Dump PT_GNU_RELRO as part of -p.
PR: 31641
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@292169
91177308-0d34-0410-b5e6-
96231b3b80d8
Davide Italiano [Mon, 16 Jan 2017 22:01:41 +0000 (22:01 +0000)]
[llvm-objdump] Dump PT_OPENBSD_{BOOTDATA,RANDOMIZE,WXNEEDED}.
PR: 31641
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@292167
91177308-0d34-0410-b5e6-
96231b3b80d8
David Blaikie [Mon, 16 Jan 2017 21:40:08 +0000 (21:40 +0000)]
Add missing header to see if that clears up the build
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@292166
91177308-0d34-0410-b5e6-
96231b3b80d8
Simon Pilgrim [Mon, 16 Jan 2017 21:34:22 +0000 (21:34 +0000)]
[InstCombine][AVX] Tests showing missed opportunities to pass demanded elts through a permilpd/permilps shuffle mask
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@292165
91177308-0d34-0410-b5e6-
96231b3b80d8
Sanjay Patel [Mon, 16 Jan 2017 21:24:41 +0000 (21:24 +0000)]
[InstCombine] use m_APInt instead of faking it
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@292164
91177308-0d34-0410-b5e6-
96231b3b80d8
David Blaikie [Mon, 16 Jan 2017 21:20:51 +0000 (21:20 +0000)]
Attempt to fix the MSVC build by using llvm::errc instead of std::errc
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@292163
91177308-0d34-0410-b5e6-
96231b3b80d8
Jan Vesely [Mon, 16 Jan 2017 21:20:13 +0000 (21:20 +0000)]
ADMGPU/EG,CM: Implement _noret global atomics
_RTN versions will be a lot more complicated
Differential Revision: https://reviews.llvm.org/D28067
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@292162
91177308-0d34-0410-b5e6-
96231b3b80d8
David Blaikie [Mon, 16 Jan 2017 20:36:26 +0000 (20:36 +0000)]
[XRay] Implement the `llvm-xray graph` subcommand
Here we define the `graph` subcommand which generates a graph from the function
call information and uses it to present the call information graphically with
additional annotations.
Reviewers: dblaikie, dberris
Differential Revision: https://reviews.llvm.org/D27243
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@292156
91177308-0d34-0410-b5e6-
96231b3b80d8
David Blaikie [Mon, 16 Jan 2017 20:28:59 +0000 (20:28 +0000)]
Attempt to workaround MSVC build issue where I suspect an enum class constant 0 is considered a possible null pointer
I can't reproduce this so far with web compilers, so throwing this at
the bots to see if it sticks.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@292155
91177308-0d34-0410-b5e6-
96231b3b80d8
Tony Jiang [Mon, 16 Jan 2017 20:12:26 +0000 (20:12 +0000)]
[PowerPC] Expand ISEL instruction into if-then-else sequence.
Generally, the ISEL is expanded into if-then-else sequence, in some
cases (like when the destination register is the same with the true
or false value register), it may just be expanded into just the if
or else sequence.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@292154
91177308-0d34-0410-b5e6-
96231b3b80d8
Sanjay Patel [Mon, 16 Jan 2017 20:05:26 +0000 (20:05 +0000)]
[InstCombine] fix names in canEvaluateShiftedShift(); NFC
It's not clear what 'First' and 'Second' mean, so use 'Inner' and 'Outer'
to match foldShiftedShift() and add comments with formulas, so it's easier
to see what's going on.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@292153
91177308-0d34-0410-b5e6-
96231b3b80d8
Sanjay Patel [Mon, 16 Jan 2017 19:35:45 +0000 (19:35 +0000)]
[InstCombine] use m_APInt to allow shift-shift folds for vectors with splat constants
Some existing 'FIXME' tests are still not folded because of splat holes in value tracking.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@292151
91177308-0d34-0410-b5e6-
96231b3b80d8
Sanjay Patel [Mon, 16 Jan 2017 19:23:34 +0000 (19:23 +0000)]
[InstCombine] add tests to show missed vector folds; NFC
The shift-shift possibilities became easier to see after:
https://reviews.llvm.org/rL292145
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@292150
91177308-0d34-0410-b5e6-
96231b3b80d8
David Blaikie [Mon, 16 Jan 2017 18:48:52 +0000 (18:48 +0000)]
PR31650: Refer to enum constant when initializing llvm::None constant
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@292149
91177308-0d34-0410-b5e6-
96231b3b80d8
Justin Lebar [Mon, 16 Jan 2017 18:39:15 +0000 (18:39 +0000)]
[NVPTX] Add blank line to NVPTXUsage.rst to appease the Sphinx.
Fixes:
Warning, treated as error:
/home/buildbot/llvm-build-dir/llvm-sphinx-docs/llvm/src/docs/NVPTXUsage.rst:333:
ERROR: Error in "code-block" directive:
maximum 1 argument(s) allowed, 17 supplied.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@292148
91177308-0d34-0410-b5e6-
96231b3b80d8
Sanjay Patel [Mon, 16 Jan 2017 17:27:50 +0000 (17:27 +0000)]
[InstCombine] refactor shift-of-shift folds; NFCI
Reduces code duplication and makes it easier to extend these folds for vectors.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@292145
91177308-0d34-0410-b5e6-
96231b3b80d8
Simon Pilgrim [Mon, 16 Jan 2017 17:26:23 +0000 (17:26 +0000)]
[InstCombine][SSE] Tests showing missed opportunities to pass demanded elts through a packss/packus truncation
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@292144
91177308-0d34-0410-b5e6-
96231b3b80d8
Pavel Labath [Mon, 16 Jan 2017 16:38:23 +0000 (16:38 +0000)]
[llvm-xray] Fix android build
std::to_string is not available in the android ndk. Using llvm::to_string
instead.
Committing as obvious.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@292143
91177308-0d34-0410-b5e6-
96231b3b80d8
Chad Rosier [Mon, 16 Jan 2017 16:28:43 +0000 (16:28 +0000)]
[AArch64] Falkor supports Rounding Double Multiply Add/Subtract instructions.
Falkor only partially implements the ARMv8.1a extensions, so this patch
refactors the support for the SQRDML[A|S]H instruction into a separate
feature.
Differential Revision: https://reviews.llvm.org/D28681
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@292142
91177308-0d34-0410-b5e6-
96231b3b80d8
Daniel Sanders [Mon, 16 Jan 2017 15:34:43 +0000 (15:34 +0000)]
Revert r292132: [globalisel] Tablegen-erate current Register Bank Information'...
Several buildbots encountered a crash in tablegen when building this commit.
Reverting while I investigate the cause.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@292136
91177308-0d34-0410-b5e6-
96231b3b80d8
Hal Finkel [Mon, 16 Jan 2017 15:22:01 +0000 (15:22 +0000)]
Fix use-after-free bug in AffectedValueCallbackVH::allUsesReplacedWith
When transferring affected values in the cache from an old value, identified by
the value of the current callback, to the specified new value we might need to
insert a new entry into the DenseMap which constitutes the cache. Doing so
might delete the current callback object. Move the copying logic into a new
function, a member of the assumption cache itself, so that we don't run into UB
should the callback handle itself be removed mid-copy.
Differential Revision: https://reviews.llvm.org/D28749
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@292133
91177308-0d34-0410-b5e6-
96231b3b80d8
Daniel Sanders [Mon, 16 Jan 2017 15:20:43 +0000 (15:20 +0000)]
[globalisel] Tablegen-erate current Register Bank Information
Summary:
Adds a RegisterBank tablegen class that can be used to declare the register
banks and an associated tablegen pass to generate the necessary code.
Reviewers: t.p.northover, ab, rovka, qcolombet
Subscribers: aditya_nandakumar, rengolin, kristof.beyls, vkalintiris, mgorny, dberris, llvm-commits, rovka
Differential Revision: https://reviews.llvm.org/D27338
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@292132
91177308-0d34-0410-b5e6-
96231b3b80d8
Tony Jiang [Mon, 16 Jan 2017 15:01:07 +0000 (15:01 +0000)]
Revert "[PowerPC] Expand ISEL instruction into if-then-else sequence."
This reverts commit
1d0e0374438ca6e153844c683826ba9b82486bb1.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@292131
91177308-0d34-0410-b5e6-
96231b3b80d8
Simon Pilgrim [Mon, 16 Jan 2017 14:49:26 +0000 (14:49 +0000)]
[SelectionDAG] Add knownbits support for BITREVERSE
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@292130
91177308-0d34-0410-b5e6-
96231b3b80d8
Tony Jiang [Mon, 16 Jan 2017 14:43:12 +0000 (14:43 +0000)]
[PowerPC] Expand ISEL instruction into if-then-else sequence.
Generally, the ISEL is expanded into if-then-else sequence, in some
cases (like when the destination register is the same with the true
or false value register), it may just be expanded into just the if
or else sequence.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@292128
91177308-0d34-0410-b5e6-
96231b3b80d8
NAKAMURA Takumi [Mon, 16 Jan 2017 14:33:37 +0000 (14:33 +0000)]
DWARFDebugInfoTest.cpp: Don't use ArrayRef with initializer. It was allocated locally.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@292127
91177308-0d34-0410-b5e6-
96231b3b80d8
Simon Pilgrim [Mon, 16 Jan 2017 13:59:42 +0000 (13:59 +0000)]
[X86][SSE] Test showing missing BITREVERSE knownbits support
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@292118
91177308-0d34-0410-b5e6-
96231b3b80d8
Simon Dardis [Mon, 16 Jan 2017 13:55:58 +0000 (13:55 +0000)]
[mips] Correct c.cond.fmt instruction definition.
Permit explicit $fcc<X> operand in c.cond.fmt instruction.
Add c.cond.fmt to the MIPS to microMIPS instruction mapping table.
Check that $fcc1 - $fcc7 are unusable for MIPS-I to MIPS-III for
c.cond.fmt, bc1t, bc1f.
Reviewers: seanbruno, zoran.jovanovic, vkalintiris
Differential Revision: https://reviews.llvm.org/D24510
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@292117
91177308-0d34-0410-b5e6-
96231b3b80d8
Simon Pilgrim [Mon, 16 Jan 2017 13:39:00 +0000 (13:39 +0000)]
[SelectionDAG] Add support for BITREVERSE constant folding
We were relying on constant folding of the legalized instructions to do what constant folding we had previously
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@292114
91177308-0d34-0410-b5e6-
96231b3b80d8
Simon Pilgrim [Mon, 16 Jan 2017 13:18:07 +0000 (13:18 +0000)]
[X86][SSE] Tests showing missing BITREVERSE constant folding
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@292111
91177308-0d34-0410-b5e6-
96231b3b80d8
Simon Pilgrim [Mon, 16 Jan 2017 11:30:41 +0000 (11:30 +0000)]
[InstCombine][SSE] Add DemandedElts support for PSHUFB instructions
Simplify a pshufb shuffle mask based on the elements of the mask that are actually demanded.
Differential Revision: https://reviews.llvm.org/D28745
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@292101
91177308-0d34-0410-b5e6-
96231b3b80d8
Craig Topper [Mon, 16 Jan 2017 06:49:09 +0000 (06:49 +0000)]
[X86] Merge the disassemblers handling of the different TYPE_RELs by getting the size information from the ENCODING field. NFCI
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@292096
91177308-0d34-0410-b5e6-
96231b3b80d8
Craig Topper [Mon, 16 Jan 2017 06:49:03 +0000 (06:49 +0000)]
[X86] Reduce the number of operand 'types' the disassembler needs to deal with. NFCI
We were frequently checking for a list of types and the different types
conveyed no real information. So lump them together explicitly.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@292095
91177308-0d34-0410-b5e6-
96231b3b80d8
Craig Topper [Mon, 16 Jan 2017 05:44:33 +0000 (05:44 +0000)]
[AVX-512] Teach the disassembler about all of the EVEX gather and scatter instructions.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@292094
91177308-0d34-0410-b5e6-
96231b3b80d8
Craig Topper [Mon, 16 Jan 2017 05:44:25 +0000 (05:44 +0000)]
[AVX-512] Begin giving the disassembler a way to recognize that VSIB is a different encoding than regular addressing modes.
This part first teaches it not to check error if EVEX.V2 is used by a VSIB instruction.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@292093
91177308-0d34-0410-b5e6-
96231b3b80d8
Xin Tong [Mon, 16 Jan 2017 03:41:09 +0000 (03:41 +0000)]
Fix typos. NFC
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@292092
91177308-0d34-0410-b5e6-
96231b3b80d8
Craig Topper [Mon, 16 Jan 2017 00:58:20 +0000 (00:58 +0000)]
[AVX-512] Add more gather/scatter encoding test cases.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@292089
91177308-0d34-0410-b5e6-
96231b3b80d8
Craig Topper [Mon, 16 Jan 2017 00:55:58 +0000 (00:55 +0000)]
[AVX-512] Correct memory operand size for VPGATHERQPS and VPGATHERQD
with ZMM index. Similar for SCATTER and the prefetch gather and scatter
instructions.
Fixes PR31618.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@292088
91177308-0d34-0410-b5e6-
96231b3b80d8
Craig Topper [Mon, 16 Jan 2017 00:55:50 +0000 (00:55 +0000)]
[AVX-512] Fix register class in one of the gather/scatter memory operands so that all 32 bit registers can be allowed.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@292087
91177308-0d34-0410-b5e6-
96231b3b80d8
Sanjay Patel [Sun, 15 Jan 2017 23:45:03 +0000 (23:45 +0000)]
[InstCombine] add tests to show missed vector folds; NFC
Also, add comments and remove bogus comment.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@292082
91177308-0d34-0410-b5e6-
96231b3b80d8
Xin Tong [Sun, 15 Jan 2017 23:32:11 +0000 (23:32 +0000)]
Empty line. NFC.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@292081
91177308-0d34-0410-b5e6-
96231b3b80d8
Xin Tong [Sun, 15 Jan 2017 21:17:52 +0000 (21:17 +0000)]
Use getLoopLatch in place of isLoopSimplifyForm
Summary:
Use getLoopLatch in place of isLoopSimplifyForm. we do not need
to know whether the loop has a preheader nor dedicated exits.
Reviewers: hfinkel, sanjoy, atrick, mkuper
Subscribers: mzolotukhin, llvm-commits
Differential Revision: https://reviews.llvm.org/D28724
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@292078
91177308-0d34-0410-b5e6-
96231b3b80d8
Simon Pilgrim [Sun, 15 Jan 2017 20:44:00 +0000 (20:44 +0000)]
[CostModel][X86] Fix AVX512BW vector shift costs for vXi16 types
We already have patterns in place to support 128/256-bit shifts without AVX512VL
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@292077
91177308-0d34-0410-b5e6-
96231b3b80d8
Simon Pilgrim [Sun, 15 Jan 2017 20:19:28 +0000 (20:19 +0000)]
[CostModel][X86] Drop separate AVX512VL checks - they match existing AVX512 costs
Keep the tests though.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@292076
91177308-0d34-0410-b5e6-
96231b3b80d8
Simon Pilgrim [Sun, 15 Jan 2017 20:10:28 +0000 (20:10 +0000)]
[CostModel][X86] Update vector shift tests to correctly check by non-constant uniform values.
Use shuffle( scslar_to_vector, zeroinitializer) pattern instead of shuffle( vec, zeroinitializer)
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@292075
91177308-0d34-0410-b5e6-
96231b3b80d8
Xin Tong [Sun, 15 Jan 2017 19:53:59 +0000 (19:53 +0000)]
Delete a dead argument. NFC
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@292074
91177308-0d34-0410-b5e6-
96231b3b80d8
Sanjay Patel [Sun, 15 Jan 2017 17:55:35 +0000 (17:55 +0000)]
[InstCombine] fix formatting; NFC
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@292073
91177308-0d34-0410-b5e6-
96231b3b80d8
Simon Pilgrim [Sun, 15 Jan 2017 17:49:04 +0000 (17:49 +0000)]
[InstCombine][SSE] Tests showing missed opportunities to pass demanded elts through a pshufb shuffle mask
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@292072
91177308-0d34-0410-b5e6-
96231b3b80d8
Justin Lebar [Sun, 15 Jan 2017 16:55:54 +0000 (16:55 +0000)]
[NVPTX] Add fptosi tests to convert-fp.ll.
These seem to have been left off by accident.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@292071
91177308-0d34-0410-b5e6-
96231b3b80d8
Justin Lebar [Sun, 15 Jan 2017 16:55:37 +0000 (16:55 +0000)]
[NVPTX] Add codegen tests for llvm.fma.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@292070
91177308-0d34-0410-b5e6-
96231b3b80d8
Justin Lebar [Sun, 15 Jan 2017 16:54:57 +0000 (16:54 +0000)]
[NVPTX] Modernize intrinics.ll test.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@292069
91177308-0d34-0410-b5e6-
96231b3b80d8
Justin Lebar [Sun, 15 Jan 2017 16:54:35 +0000 (16:54 +0000)]
[NVPTX] Let there be One True Way to set NVVMReflect params.
Summary:
Previously there were three ways to inform the NVVMReflect pass whether
you wanted to flush denormals to zero:
* An LLVM command-line option
* Parameters to the NVVMReflect constructor
* Metadata on the module itself.
This change removes the first two, leaving only the third.
The motivation for this change, aside from simplifying things, is that
we want LLVM to be aware of whether it's operating in FTZ mode, so other
passes can use this information. Ideally we'd have a target-generic
piece of metadata on the module. This change moves us in that
direction.
Reviewers: tra
Subscribers: jholewinski, llvm-commits
Differential Revision: https://reviews.llvm.org/D28700
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@292068
91177308-0d34-0410-b5e6-
96231b3b80d8
Michael Zuckerman [Sun, 15 Jan 2017 16:43:14 +0000 (16:43 +0000)]
Fix blend mask by switch the side of the operand since Blend node uses opposite mask then Select NODE.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@292066
91177308-0d34-0410-b5e6-
96231b3b80d8
Daniel Jasper [Sun, 15 Jan 2017 16:42:36 +0000 (16:42 +0000)]
Fix un-initialized error introduced by r291959.
This is uncovered when running tools/dsymutil/X86/empty_range.s.test
with ASAN. Haven't investigate yet, whether that means there is an ODR
violation in that test.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@292065
91177308-0d34-0410-b5e6-
96231b3b80d8
Sanjay Patel [Sun, 15 Jan 2017 16:38:19 +0000 (16:38 +0000)]
[InstCombine] use m_APInt to allow ashr folds for vectors with splat constants
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@292064
91177308-0d34-0410-b5e6-
96231b3b80d8
Sanjay Patel [Sun, 15 Jan 2017 16:22:26 +0000 (16:22 +0000)]
[InstCombine] add explanatory comments to tests; NFC
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@292063
91177308-0d34-0410-b5e6-
96231b3b80d8
Serge Pavlov [Sun, 15 Jan 2017 10:23:18 +0000 (10:23 +0000)]
Reverted: Track validity of pass results
Commits r291882 and related r291887.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@292062
91177308-0d34-0410-b5e6-
96231b3b80d8
Daniel Jasper [Sun, 15 Jan 2017 09:41:49 +0000 (09:41 +0000)]
Revert "[GlobalISel] track predecessor mapping during switch lowering."
This reverts commit r291973.
The test fails in a Release build with LLVM_BUILD_GLOBAL_ISEL enabled.
AFAICT, llc segfaults. I'll add a few more details to the original
commit.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@292061
91177308-0d34-0410-b5e6-
96231b3b80d8
Chandler Carruth [Sun, 15 Jan 2017 09:29:27 +0000 (09:29 +0000)]
[PM] Clean up the testing for IVUsers, especially with the new PM.
First, I've moved a test of IVUsers from the LSR tree to a dedicated
IVUsers test directory. I've also simplified its RUN line now that the
new pass manager's loop PM is providing analyses on their own.
No functionality changed, but it makes subsequent changes cleaner.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@292060
91177308-0d34-0410-b5e6-
96231b3b80d8
Daniel Berlin [Sun, 15 Jan 2017 09:18:41 +0000 (09:18 +0000)]
NewGVN: Change a bunch of densemap find_or_creates to lookups, since they should not be creating new entries
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@292059
91177308-0d34-0410-b5e6-
96231b3b80d8
Chandler Carruth [Sun, 15 Jan 2017 08:20:50 +0000 (08:20 +0000)]
[PM] Teach the optimization remarks emitter to handle invalidation
events.
This pass sometimes has a pointer to BlockFrequencyInfo so it needs
custom invalidation logic. It is also otherwise immutable so we can
reduce the number of invalidations that happen substantially.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@292058
91177308-0d34-0410-b5e6-
96231b3b80d8
Daniel Berlin [Sun, 15 Jan 2017 07:40:51 +0000 (07:40 +0000)]
These two functions should be const. We often could detect it but this just makes it always true.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@292057
91177308-0d34-0410-b5e6-
96231b3b80d8
Daniel Berlin [Sun, 15 Jan 2017 07:40:46 +0000 (07:40 +0000)]
Update update_test_checks so that . is a valid identifier character in addition to _
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@292056
91177308-0d34-0410-b5e6-
96231b3b80d8
Lang Hames [Sun, 15 Jan 2017 06:34:25 +0000 (06:34 +0000)]
[Orc][RPC] Add an RPCFunctionNotSupported error type and return it from
negotiateFunction where appropriate.
Replacing the old ECError with a custom type allows us to attach the name of
the function that could not be negotiated, enabling better diagnostics for
negotiation failures.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@292055
91177308-0d34-0410-b5e6-
96231b3b80d8
Chandler Carruth [Sun, 15 Jan 2017 06:32:49 +0000 (06:32 +0000)]
[PM] Introduce an analysis set used to preserve all analyses over
a function's CFG when that CFG is unchanged.
This allows transformation passes to simply claim they preserve the CFG
and analysis passes to check for the CFG being preserved to remove the
fanout of all analyses being listed in all passes.
I've gone through and removed or cleaned up as many of the comments
reminding us to do this as I could.
Differential Revision: https://reviews.llvm.org/D28627
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@292054
91177308-0d34-0410-b5e6-
96231b3b80d8
Craig Topper [Sun, 15 Jan 2017 06:15:45 +0000 (06:15 +0000)]
[AVX-512] Remove unnecessary duplicate broadcast patterns. NFC
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@292053
91177308-0d34-0410-b5e6-
96231b3b80d8
Craig Topper [Sun, 15 Jan 2017 05:47:45 +0000 (05:47 +0000)]
[AVX-512] Replicate some broadcast patterns to VLX and disable the AVX2 patterns when VLX is available.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@292051
91177308-0d34-0410-b5e6-
96231b3b80d8
Craig Topper [Sun, 15 Jan 2017 05:21:29 +0000 (05:21 +0000)]
[X86] Remove untested MOVDDUP patterns.
These all involve bitcasts around the memory operands. This isn't
something we normally do for isel patterns. I suspect DAG combine should
convert the load type making this unnecessary.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@292050
91177308-0d34-0410-b5e6-
96231b3b80d8
Mehdi Amini [Sun, 15 Jan 2017 03:21:30 +0000 (03:21 +0000)]
Add a LLVM_USE_LINKER that defines the linker to use when building LLVM
Summary:
This string parameter is passed to -fuse-ld when linking. It can be
an absolute path to your custom linker, otherwise clang will look for
`ld.{name}`.
Reviewers: davide, tejohnson, pcc
Subscribers: llvm-commits, mgorny
Differential Revision: https://reviews.llvm.org/D28738
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@292047
91177308-0d34-0410-b5e6-
96231b3b80d8
Eric Fiselier [Sun, 15 Jan 2017 02:06:44 +0000 (02:06 +0000)]
Give comparator const call operator
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@292043
91177308-0d34-0410-b5e6-
96231b3b80d8
Rui Ueyama [Sun, 15 Jan 2017 00:36:02 +0000 (00:36 +0000)]
PDB: Add a class to create the /names stream contents.
This patch adds a new class NameHashTableBuilder which creates /names streams.
This patch contains a test to confirm that a stream created by
NameHashTableBuilder can be read by NameHashTable reader class.
Differential Revision: https://reviews.llvm.org/D28707
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@292040
91177308-0d34-0410-b5e6-
96231b3b80d8
Chandler Carruth [Sun, 15 Jan 2017 00:26:18 +0000 (00:26 +0000)]
[PM] The assumption cache is fundamentally designed to be self-updating,
mark it as never invalidated in the new PM.
The old PM already required this to work, and after a discussion with
Hal this seems to really be the only sensible answer. The cache
gracefully degrades as the IR is mutated, and most things which do this
should already be incrementally updating the cache.
This gets rid of a bunch of logic preserving and testing the
invalidation of this analysis.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@292039
91177308-0d34-0410-b5e6-
96231b3b80d8
Chandler Carruth [Sat, 14 Jan 2017 23:25:22 +0000 (23:25 +0000)]
[PM] Fix instcombine's analysis preservation in the new pass manager to
cover domtree and alias analysis. These are the pretty clear analyses
that we would always want to survive this pass.
To make these survive, we also need to preserve the assumption cache.
Added a test that verifies the important bits of this preservation.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@292037
91177308-0d34-0410-b5e6-
96231b3b80d8
Sanjay Patel [Sat, 14 Jan 2017 23:13:50 +0000 (23:13 +0000)]
[InstCombine] clean up visitAshr(); NFCI
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@292036
91177308-0d34-0410-b5e6-
96231b3b80d8
Sanjay Patel [Sat, 14 Jan 2017 23:12:29 +0000 (23:12 +0000)]
[InstCombine] add test to show missed vector fold; NFC
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@292035
91177308-0d34-0410-b5e6-
96231b3b80d8
David Majnemer [Sat, 14 Jan 2017 21:54:58 +0000 (21:54 +0000)]
Adding const overloads of operator* and operator-> for DenseSet iterators
This fixes some problems when building ClangDiagnostics.cpp on Visual Studio 2017 RC. As far as I understand, there was a change in the implementation of the constructor for std::vector with two iterator parameters, which in our case causes an attempt to dereference const Iterator objects. Since there was no overload for a const Iterator, the compile would fail.
Patch by Hugo Puhlmann!
Differential Revision: https://reviews.llvm.org/D28726
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@292034
91177308-0d34-0410-b5e6-
96231b3b80d8
Davide Italiano [Sat, 14 Jan 2017 20:44:08 +0000 (20:44 +0000)]
[NewGVN] Fix a warning from GCC.
Patch by Gonsolo.
Differential Revision: https://reviews.llvm.org/D28731
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@292031
91177308-0d34-0410-b5e6-
96231b3b80d8
Davide Italiano [Sat, 14 Jan 2017 20:15:04 +0000 (20:15 +0000)]
[NewGVN] clang-format this file after recent changes.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@292026
91177308-0d34-0410-b5e6-
96231b3b80d8
Davide Italiano [Sat, 14 Jan 2017 20:13:18 +0000 (20:13 +0000)]
[NewGVN] Try to be consistent wit the style used in this file. NFCI.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@292025
91177308-0d34-0410-b5e6-
96231b3b80d8
Davide Italiano [Sat, 14 Jan 2017 20:09:29 +0000 (20:09 +0000)]
[TargetLowering] Simplfiy a bit. NFCI.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@292024
91177308-0d34-0410-b5e6-
96231b3b80d8
Simon Pilgrim [Sat, 14 Jan 2017 19:24:23 +0000 (19:24 +0000)]
[CostModel][X86] Updated vXi64 ASHR costs on AVX512 targets now that D28604 has landed
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@292023
91177308-0d34-0410-b5e6-
96231b3b80d8
Simon Pilgrim [Sat, 14 Jan 2017 18:52:13 +0000 (18:52 +0000)]
[X86][XOP] Added support for VPMADCSWD 'extend+hadd' IFMA patterns
VPMADCSWD act as VPADDD( VPMADDWD( x, y ), z ) - multiply+extend+hadd and add to v4i32 accumulator
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@292021
91177308-0d34-0410-b5e6-
96231b3b80d8
Simon Pilgrim [Sat, 14 Jan 2017 18:08:54 +0000 (18:08 +0000)]
[X86][XOP] Added support for VPMACSDQH/VPMACSDQL 'extension' IFMA patterns
VPMACSDQH/VPMACSDQL act as VPADDQ( VPMULDQ( x, y ), z ) - multiply+extending either the odd/even 4i32 input elements and adding to v2i64 accumulator
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@292020
91177308-0d34-0410-b5e6-
96231b3b80d8
Simon Pilgrim [Sat, 14 Jan 2017 17:13:52 +0000 (17:13 +0000)]
[X86][XOP] Added support for VPMACSWW/VPMACSDD 'lossy' IFMA patterns
VPMACSWW/VPMACSDD act as add( mul( x, y ), z ) - ignoring any upper bits from both the multiply and add stages
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@292019
91177308-0d34-0410-b5e6-
96231b3b80d8
Simon Pilgrim [Sat, 14 Jan 2017 13:07:22 +0000 (13:07 +0000)]
[X86][XOP] Add tests for integer fused multiply add
Tests showing missed opportunities to use XOP's integer fma instructions
Some of these are pretty awkward to match as they often have implicit sext/trunc stages but many just ignore overflow bits which makes things pretty straightforward.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@292017
91177308-0d34-0410-b5e6-
96231b3b80d8
Sylvestre Ledru [Sat, 14 Jan 2017 11:37:01 +0000 (11:37 +0000)]
fix some typos in the doc
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@292014
91177308-0d34-0410-b5e6-
96231b3b80d8