OSDN Git Service
Craig Topper [Mon, 30 Nov 2015 02:28:19 +0000 (02:28 +0000)]
Revert r254279 "[X86] Use ArrayRef. NFC". It seems to have upset an MSVC build bot.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254280
91177308-0d34-0410-b5e6-
96231b3b80d8
Craig Topper [Mon, 30 Nov 2015 02:08:05 +0000 (02:08 +0000)]
[X86] Use ArrayRef. NFC
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254279
91177308-0d34-0410-b5e6-
96231b3b80d8
Sanjoy Das [Mon, 30 Nov 2015 01:24:17 +0000 (01:24 +0000)]
[ADT] Fix typo in comment
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254278
91177308-0d34-0410-b5e6-
96231b3b80d8
Craig Topper [Mon, 30 Nov 2015 00:13:24 +0000 (00:13 +0000)]
[AVX512] The vpermi2 instructions require an integer vector for the index vector. This is reflected correctly in the intrinsics, but was not refelected in the isel patterns.
For the floating point types, this requires adding a bitcast to the index vector when its passed through to the output.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254277
91177308-0d34-0410-b5e6-
96231b3b80d8
Sanjoy Das [Sun, 29 Nov 2015 23:40:57 +0000 (23:40 +0000)]
[SCEV] Use lambda instead of std::bind; NFC
The lambda is more readable.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254276
91177308-0d34-0410-b5e6-
96231b3b80d8
Sanjoy Das [Sun, 29 Nov 2015 23:40:53 +0000 (23:40 +0000)]
[SCEV] Use range version of all_of; NFC
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254275
91177308-0d34-0410-b5e6-
96231b3b80d8
Craig Topper [Sun, 29 Nov 2015 23:18:32 +0000 (23:18 +0000)]
[X86] Remove duplicate entries from intrinsics tables and add asserts to verify there are no others.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254274
91177308-0d34-0410-b5e6-
96231b3b80d8
Sanjoy Das [Sun, 29 Nov 2015 23:15:43 +0000 (23:15 +0000)]
Fix out of bounds access in hasStructRetAttr
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254273
91177308-0d34-0410-b5e6-
96231b3b80d8
Dan Gohman [Sun, 29 Nov 2015 23:09:41 +0000 (23:09 +0000)]
[WebAssembly] Delete an obsolete TODO comment.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254272
91177308-0d34-0410-b5e6-
96231b3b80d8
Dan Gohman [Sun, 29 Nov 2015 22:59:19 +0000 (22:59 +0000)]
[WebAssembly] Set several MCInstrDesc flags.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254271
91177308-0d34-0410-b5e6-
96231b3b80d8
Craig Topper [Sun, 29 Nov 2015 22:53:22 +0000 (22:53 +0000)]
[X86] int_x86_avx2_permps and X86ISD::VPERMV should take an integer vector for its shuffle indices.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254269
91177308-0d34-0410-b5e6-
96231b3b80d8
Dan Gohman [Sun, 29 Nov 2015 22:48:57 +0000 (22:48 +0000)]
[WebAssembly] Delete unused functions. NFC.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254268
91177308-0d34-0410-b5e6-
96231b3b80d8
Dan Gohman [Sun, 29 Nov 2015 22:32:02 +0000 (22:32 +0000)]
[WebAssembly] Minor clang-format and selected clang-tidy cleanups. NFC.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254267
91177308-0d34-0410-b5e6-
96231b3b80d8
Sanjay Patel [Sun, 29 Nov 2015 22:09:34 +0000 (22:09 +0000)]
fix typos in comments; NFC
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254266
91177308-0d34-0410-b5e6-
96231b3b80d8
Davide Italiano [Sun, 29 Nov 2015 21:58:56 +0000 (21:58 +0000)]
[SimplifyLibCalls] Don't crash if the function doesn't have a name.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254265
91177308-0d34-0410-b5e6-
96231b3b80d8
Davide Italiano [Sun, 29 Nov 2015 21:00:43 +0000 (21:00 +0000)]
[SimplifyLibCalls] Cross out implemented transformations.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254264
91177308-0d34-0410-b5e6-
96231b3b80d8
Davide Italiano [Sun, 29 Nov 2015 20:58:04 +0000 (20:58 +0000)]
[SimplifyLibCalls] Tranform log(pow(x, y)) -> y*log(x).
This one is enabled only under -ffast-math. There are cases where the
difference between the value computed and the correct value is huge
even for ffast-math, e.g. as Steven pointed out:
x = -1, y = -4
log(pow(-1), 4) = 0
4*log(-1) = NaN
I checked what GCC does and apparently they do the same optimization
(which result in the dramatic difference). Future work might try to
make this (slightly) less worse.
Differential Revision: http://reviews.llvm.org/D14400
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254263
91177308-0d34-0410-b5e6-
96231b3b80d8
Diego Novillo [Sun, 29 Nov 2015 18:23:26 +0000 (18:23 +0000)]
SamplePGO - Do not use std::to_string in diagnostics.
This fixes buildbots in systems that std::to_string is not present. It
also tidies the output of the diagnostic to render doubles a bit better
(thanks Ben Kramer for help with string streams and format).
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254261
91177308-0d34-0410-b5e6-
96231b3b80d8
Craig Topper [Sun, 29 Nov 2015 18:05:22 +0000 (18:05 +0000)]
Use a lambda instead of std::bind and std::mem_fn I introduced in r254242. NFC
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254260
91177308-0d34-0410-b5e6-
96231b3b80d8
Simon Pilgrim [Sun, 29 Nov 2015 16:41:04 +0000 (16:41 +0000)]
[X86][SSE] Added support for lowering to ADDSUBPS/ADDSUBPD with commuted inputs
We could already recognise shuffle(FSUB, FADD) -> ADDSUB, this allow us to recognise shuffle(FADD, FSUB) -> ADDSUB by commuting the shuffle mask prior to matching.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254259
91177308-0d34-0410-b5e6-
96231b3b80d8
Rafael Espindola [Sun, 29 Nov 2015 15:52:12 +0000 (15:52 +0000)]
Add a passing test.
When a comdat is discarded, any globals defined in it become undefined.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254258
91177308-0d34-0410-b5e6-
96231b3b80d8
Rafael Espindola [Sun, 29 Nov 2015 15:22:49 +0000 (15:22 +0000)]
Don't depend on the order the IR is copied.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254257
91177308-0d34-0410-b5e6-
96231b3b80d8
Rafael Espindola [Sun, 29 Nov 2015 15:08:39 +0000 (15:08 +0000)]
Don't depend on the order the IR is copied.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254256
91177308-0d34-0410-b5e6-
96231b3b80d8
Rafael Espindola [Sun, 29 Nov 2015 14:53:06 +0000 (14:53 +0000)]
Make this test less strict.
We just want to test what is copied, no the order.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254255
91177308-0d34-0410-b5e6-
96231b3b80d8
Rafael Espindola [Sun, 29 Nov 2015 14:33:06 +0000 (14:33 +0000)]
Simplify. NFC.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254254
91177308-0d34-0410-b5e6-
96231b3b80d8
Igor Breger [Sun, 29 Nov 2015 07:41:26 +0000 (07:41 +0000)]
AVX512:Implemented encoding for the vmovq.s instruction.
Differential Revision: http://reviews.llvm.org/D14810
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254248
91177308-0d34-0410-b5e6-
96231b3b80d8
Craig Topper [Sun, 29 Nov 2015 05:38:08 +0000 (05:38 +0000)]
Remove an intermediate lambda. NFC
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254246
91177308-0d34-0410-b5e6-
96231b3b80d8
Xinliang David Li [Sun, 29 Nov 2015 04:52:34 +0000 (04:52 +0000)]
Minor code cleanups
- Add const keyword
- fix code comments
- move forward decl to the common file
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254244
91177308-0d34-0410-b5e6-
96231b3b80d8
Craig Topper [Sun, 29 Nov 2015 04:37:14 +0000 (04:37 +0000)]
Remove unnecessary intermediate lambda. NFC
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254243
91177308-0d34-0410-b5e6-
96231b3b80d8
Craig Topper [Sun, 29 Nov 2015 04:37:11 +0000 (04:37 +0000)]
[SelectionDAG] Use std::any_of instead of a manually coded loop. NFC
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254242
91177308-0d34-0410-b5e6-
96231b3b80d8
Rafael Espindola [Sun, 29 Nov 2015 03:29:42 +0000 (03:29 +0000)]
Correctly handle llvm.global_ctors merging.
We were not handling the case where an entry must be dropped and the
destination module has no llvm.global_ctors.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254241
91177308-0d34-0410-b5e6-
96231b3b80d8
Rafael Espindola [Sun, 29 Nov 2015 03:21:30 +0000 (03:21 +0000)]
Fix a crash when writing merged bitcode.
Playing with mutateType in here was making getValueType and getType
incompatible.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254240
91177308-0d34-0410-b5e6-
96231b3b80d8
Davide Italiano [Sat, 28 Nov 2015 22:27:48 +0000 (22:27 +0000)]
[SimplifyLibCalls] Use any_of(). Suggested by David Blaikie!
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254239
91177308-0d34-0410-b5e6-
96231b3b80d8
Benjamin Kramer [Sat, 28 Nov 2015 21:43:12 +0000 (21:43 +0000)]
[SimplifyLibCalls] Fix inverted condition that lead to an uninitialized memory read below.
Found by msan!
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254238
91177308-0d34-0410-b5e6-
96231b3b80d8
Simon Pilgrim [Sat, 28 Nov 2015 19:20:49 +0000 (19:20 +0000)]
[X86][AVX] Regenerate ADDSUB tests
Tidied up triple and regenerate tests using update_llc_test_checks.py
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254237
91177308-0d34-0410-b5e6-
96231b3b80d8
Xinliang David Li [Sat, 28 Nov 2015 19:07:09 +0000 (19:07 +0000)]
[PGO] Move value profile format related structures and APIs to common file
This is the last step to enable profile runtime to share the same value prof
data format and reader/writer code with llvm host tools. The VP related
data structures are moved to a section in InstrProfData.inc enabled with macro
INSTR_PROF_VALUE_PROF_DATA, and common API implementations are enabled with
INSTR_PROF_COMMON_API_IMPL. There should be no functional change.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254235
91177308-0d34-0410-b5e6-
96231b3b80d8
Renato Golin [Sat, 28 Nov 2015 17:23:46 +0000 (17:23 +0000)]
Revert "[ARM] Generate ABI_optimization_goals build attribute, as described in the ARM ARM."
This reverts commit r254201 and r254202, as it broke test-suite,
self-hosting and sanitizer tests on ARM buildbots.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254234
91177308-0d34-0410-b5e6-
96231b3b80d8
Simon Pilgrim [Sat, 28 Nov 2015 16:04:24 +0000 (16:04 +0000)]
[X86][FMA] Added 512-bit tests to match 128/256-bit tests coverage
As discussed on D14909
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254233
91177308-0d34-0410-b5e6-
96231b3b80d8
Simon Pilgrim [Sat, 28 Nov 2015 14:28:44 +0000 (14:28 +0000)]
[X86][FMA] More thorough FMA tests
Added FMADD/FMSUB/FNMADD/FNMSUB tests for all types
Added load folding tests for 512-bit vectors
NOTE: Many of the AVX512 FMA instructions don't yet commute/fold correctly
As discussed on D14909
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254232
91177308-0d34-0410-b5e6-
96231b3b80d8
Simon Pilgrim [Sat, 28 Nov 2015 14:15:40 +0000 (14:15 +0000)]
[X86][AVX2] Tidied up PBROADCAST tests
Tidied up triple and regenerate tests using update_llc_test_checks.py
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254231
91177308-0d34-0410-b5e6-
96231b3b80d8
NAKAMURA Takumi [Sat, 28 Nov 2015 13:05:49 +0000 (13:05 +0000)]
llvm/test/CodeGen/SystemZ/alloca-04.ll REQUIRES asserts due to -debug-pass.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254230
91177308-0d34-0410-b5e6-
96231b3b80d8
Jonas Paulsson [Sat, 28 Nov 2015 11:02:32 +0000 (11:02 +0000)]
[Stack realignment] Handling of aligned allocas.
This patch implements dynamic realignment of stack objects for targets
with a non-realigned stack pointer. Behaviour in FunctionLoweringInfo
is changed so that for a target that has StackRealignable set to
false, over-aligned static allocas are considered to be variable-sized
objects and are handled with DYNAMIC_STACKALLOC nodes.
It would be good to group aligned allocas into a single big alloca as
an optimization, but this is yet todo.
SystemZ benefits from this, due to its stack frame layout.
New tests SystemZ/alloca-03.ll for aligned allocas, and
SystemZ/alloca-04.ll for "no-realign-stack" attribute on functions.
Review and help from Ulrich Weigand and Hal Finkel.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254227
91177308-0d34-0410-b5e6-
96231b3b80d8
Craig Topper [Sat, 28 Nov 2015 08:23:04 +0000 (08:23 +0000)]
Use range-based for loops. NFC
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254222
91177308-0d34-0410-b5e6-
96231b3b80d8
Craig Topper [Sat, 28 Nov 2015 08:23:02 +0000 (08:23 +0000)]
[TableGen] Use SmallString instead of std::string to build up a string to avoid heap allocations. NFC
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254221
91177308-0d34-0410-b5e6-
96231b3b80d8
Xinliang David Li [Sat, 28 Nov 2015 05:47:34 +0000 (05:47 +0000)]
[PGO] Add return code for vp rt record init routine to indicate error condition
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254220
91177308-0d34-0410-b5e6-
96231b3b80d8
Xinliang David Li [Sat, 28 Nov 2015 05:37:01 +0000 (05:37 +0000)]
[PGO] Allow value profile writer interface to allocated target buffer
Raw profile writer needs to write all data of one kind in one continuous block,
so the buffer needs to be pre-allocated and passed to the writer method in
pieces for function profile data. The change adds the support for raw value data
writing.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254219
91177308-0d34-0410-b5e6-
96231b3b80d8
Xinliang David Li [Sat, 28 Nov 2015 05:06:00 +0000 (05:06 +0000)]
Function name cleanup (NFC)
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254218
91177308-0d34-0410-b5e6-
96231b3b80d8
Xinliang David Li [Sat, 28 Nov 2015 04:56:07 +0000 (04:56 +0000)]
[PGO] Extract VP data integrity check code into a helper function (NFC)
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254217
91177308-0d34-0410-b5e6-
96231b3b80d8
Keno Fischer [Sat, 28 Nov 2015 00:54:12 +0000 (00:54 +0000)]
[autoconf] Fix MinGW build
This is the autoconf analog of r251201. I realize autoconf is
deprecated, but while it's in tree, it should at least be kept working.
Also add the deprecation message to configure.ac such that AutoRegen
actually picks ip up.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254215
91177308-0d34-0410-b5e6-
96231b3b80d8
Rafael Espindola [Fri, 27 Nov 2015 23:47:15 +0000 (23:47 +0000)]
Pass .ll directly to llvm-link.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254214
91177308-0d34-0410-b5e6-
96231b3b80d8
Rafael Espindola [Fri, 27 Nov 2015 23:21:45 +0000 (23:21 +0000)]
Pass .ll directly to llvm-link
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254213
91177308-0d34-0410-b5e6-
96231b3b80d8
Diego Novillo [Fri, 27 Nov 2015 23:14:51 +0000 (23:14 +0000)]
SamplePGO - Add initial support for inliner annotations.
This adds two thresholds to the sample profiler to affect inlining
decisions: the concept of global hotness and coldness.
Functions that have accumulated more than a certain fraction of samples at
runtime, are annotated with the InlineHint attribute. Conversely,
functions that accumulate less than a certain fraction of samples, are
annotated with the Cold attribute.
This is very similar to the hints emitted by Clang when using
instrumentation profiles.
Notice that this is a very blunt instrument. A function may have
globally collected a significant fraction of samples, but that does not
necessarily mean that every callsite for that function is hot.
Ideally, we would annotate each callsite with the samples collected at
that callsite. This way, the inliner can incorporate all these weights
into its cost model.
Once the inliner offers this functionality, we can change the hints
emitted here to a more precise per-callsite annotation. For now, this is
providing some measure of speedups with our internal benchmarks. I've
observed speedups of up to 23% (though the geo mean is about 3%). I expect
these numbers to improve as the inliner gets better annotations.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254212
91177308-0d34-0410-b5e6-
96231b3b80d8
Diego Novillo [Fri, 27 Nov 2015 23:14:49 +0000 (23:14 +0000)]
SamplePGO - Fix default threshold for hot callsites.
Based on testing of internal benchmarks, I'm lowering this threshold to
a value of 0.1%. This means that SamplePGO will respect 99.9% of the
original inline decisions when following a profile.
The performance difference is noticeable in some tests. With the
previous threshold, the speedups over baseline -O2 was about 0.63%. With
the new default, the speedups are around 3% on average.
The point of this threshold is not to do more aggressive inlining. When
an inlined callsite crosses this threshold, SamplePGO will redo the
inline decision so that it can better apply the input profile.
By respecting most original inline decisions, we can apply more of the
input profile because the shape of the code follows the profile more
closely.
In the next series, I'll be looking at adding some inline hints for the
cold callsites and for toplevel functions that are hot/cold as well.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254211
91177308-0d34-0410-b5e6-
96231b3b80d8
Rafael Espindola [Fri, 27 Nov 2015 23:13:17 +0000 (23:13 +0000)]
Modernize the test a bit
Remove out of date comment.
Pass .ll files to llvm-link.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254210
91177308-0d34-0410-b5e6-
96231b3b80d8
Rafael Espindola [Fri, 27 Nov 2015 20:28:19 +0000 (20:28 +0000)]
Simplify the linking of recursive data.
Now the ValueMapper has two callbacks. The first one maps the
declaration. The ValueMapper records the mapping and then materializes
the body/initializer.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254209
91177308-0d34-0410-b5e6-
96231b3b80d8
Artyom Skrobov [Fri, 27 Nov 2015 16:20:34 +0000 (16:20 +0000)]
Follow-up fix for r254201
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254202
91177308-0d34-0410-b5e6-
96231b3b80d8
Artyom Skrobov [Fri, 27 Nov 2015 15:30:51 +0000 (15:30 +0000)]
[ARM] Generate ABI_optimization_goals build attribute, as described in the ARM ARM.
Summary:
Since this build attribute corresponds to a whole module, and
different functions in a module may differ in the optimizations
enabled for them, this attribute is emitted after all functions,
and only in the case that the optimization goals for all
functions match.
Reviewers: logan, hans
Subscribers: aemerson, rengolin, llvm-commits
Differential Revision: http://reviews.llvm.org/D14934
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254201
91177308-0d34-0410-b5e6-
96231b3b80d8
Oliver Stannard [Fri, 27 Nov 2015 13:04:48 +0000 (13:04 +0000)]
[AArch64] Add ARMv8.2-A FP16 scalar instructions
ARMv8.2-A adds 16-bit floating point versions of all existing VFP
floating-point instructions. This is an optional extension, so all of
these instructions require the FeatureFullFP16 subtarget feature.
Most of these instructions are the same as the 32- and 64-bit versions,
but with the type field (bits 23-22) set to 0b11. Previously the top bit
of the size field was always 0, so the instruction classes only provided
a 1-bit size field, which I have widened to 2 bits.
Differential Revision: http://reviews.llvm.org/D15014
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254198
91177308-0d34-0410-b5e6-
96231b3b80d8
Adhemerval Zanella [Fri, 27 Nov 2015 12:42:39 +0000 (12:42 +0000)]
[sanitizer] [dfsan] Unify aarch64 mapping
This patch changes the DFSan instrumentation for aarch64 to instead
of using fixes application mask defined by SANITIZER_AARCH64_VMA
to read the application shadow mask value from compiler-rt. The value
is initialized based on runtime VAM detection.
Along with this patch a compiler-rt one will also be added to export
the shadow mask variable.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254196
91177308-0d34-0410-b5e6-
96231b3b80d8
Davide Italiano [Fri, 27 Nov 2015 08:05:40 +0000 (08:05 +0000)]
[SimplifyLibCalls] Use range-based loop. NFC.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254193
91177308-0d34-0410-b5e6-
96231b3b80d8
Craig Topper [Fri, 27 Nov 2015 05:44:04 +0000 (05:44 +0000)]
[TableGen] Sort pattern predicates before concatenating into a string so that different orders of the same set will produce the same string. This can reduce the number of unique predicates in the isel tables. NFC
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254192
91177308-0d34-0410-b5e6-
96231b3b80d8
Craig Topper [Fri, 27 Nov 2015 05:44:02 +0000 (05:44 +0000)]
[X86] Pair a NoVLX with HasAVX512 to match the others and remove a unique predicate check in the isel tables. NFC
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254191
91177308-0d34-0410-b5e6-
96231b3b80d8
Andrew Wilkins [Fri, 27 Nov 2015 05:07:26 +0000 (05:07 +0000)]
test: bail early if tool_path is None
tool_path will be None for llvm-go if Go cannot be found
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254190
91177308-0d34-0410-b5e6-
96231b3b80d8
Andrew Wilkins [Fri, 27 Nov 2015 04:51:13 +0000 (04:51 +0000)]
test: check if go_executable is set
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254189
91177308-0d34-0410-b5e6-
96231b3b80d8
Andrew Wilkins [Fri, 27 Nov 2015 04:44:51 +0000 (04:44 +0000)]
Use $GO_EXECUTABLE in Go-based lit tests
Summary:
When running tests, pass the GO_EXECUTABLE CMake
cache variable to llvm-go. The "go" binary may
not be in $PATH, or may be different to the one
passed to CMake.
Reviewers: pcc
Subscribers: llvm-commits
Differential Revision: http://reviews.llvm.org/D14041
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254187
91177308-0d34-0410-b5e6-
96231b3b80d8
Rafael Espindola [Fri, 27 Nov 2015 03:50:34 +0000 (03:50 +0000)]
Test both input file orders.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254186
91177308-0d34-0410-b5e6-
96231b3b80d8
Rafael Espindola [Fri, 27 Nov 2015 03:47:29 +0000 (03:47 +0000)]
Add missing file.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254185
91177308-0d34-0410-b5e6-
96231b3b80d8
Rafael Espindola [Fri, 27 Nov 2015 02:07:37 +0000 (02:07 +0000)]
Make the test a bit more interesting.
It now covers a regular function replacing an available_externally one.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254184
91177308-0d34-0410-b5e6-
96231b3b80d8
Peter Collingbourne [Thu, 26 Nov 2015 23:29:27 +0000 (23:29 +0000)]
MC: Simplify handling of temporary symbols in COFF writer.
The COFF object writer was previously adding unnecessary symbols to its
temporary data structures and cleaning them up later. This made the code
harder to understand and caused a bug (aliases classed as temporary symbols
would cause an assertion failure). A much simpler way of handling such
symbols is to ask the layout for their section-relative position when needed.
Tested with a bootstrap on Windows and by building Chrome.
Differential Revision: http://reviews.llvm.org/D14975
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254183
91177308-0d34-0410-b5e6-
96231b3b80d8
Simon Pilgrim [Thu, 26 Nov 2015 20:53:28 +0000 (20:53 +0000)]
[X86][FMA] Begun adding AVX512 FMA tests
As discussed on D14909
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254180
91177308-0d34-0410-b5e6-
96231b3b80d8
Charlie Turner [Thu, 26 Nov 2015 20:39:51 +0000 (20:39 +0000)]
[LoopVectorize] Use MapVector rather than DenseMap for MinBWs.
The order in which instructions are truncated in truncateToMinimalBitwidths
effects code generation. Switch to a map with a determinisic order, since the
iteration order over a DenseMap is not defined.
This code is not hot, so the difference in container performance isn't
interesting.
Many thanks to David Blaikie for making me aware of MapVector!
Fixes PR25490.
Differential Revision: http://reviews.llvm.org/D14981
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254179
91177308-0d34-0410-b5e6-
96231b3b80d8
Craig Topper [Thu, 26 Nov 2015 20:21:29 +0000 (20:21 +0000)]
[X86] Now that X86VPermt2 is used in all the avx512_perm_t_sizes just hardcode it into the patterns instead of passing as an argument. NFC
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254177
91177308-0d34-0410-b5e6-
96231b3b80d8
Craig Topper [Thu, 26 Nov 2015 20:02:01 +0000 (20:02 +0000)]
[X86] Merge X86VPermt2Fp and X86VPermt2Int back together by weakening them just enough. The SDTCisSameSizeAs introduced in r254138 helps here.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254176
91177308-0d34-0410-b5e6-
96231b3b80d8
Rafael Espindola [Thu, 26 Nov 2015 19:53:12 +0000 (19:53 +0000)]
Add a few passing lto tests.
I found these while trying to get a prototype to bootstrap.
They cover things like
* Handling of non linker visible stuff (append, available_externally)
* Type merging
* Alias to dropped globals
* Dropping linkage when converting to a declaration.
These should hopefully be generally useful for anyone refactoring the
plugin.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254174
91177308-0d34-0410-b5e6-
96231b3b80d8
Craig Topper [Thu, 26 Nov 2015 19:41:34 +0000 (19:41 +0000)]
[X86] Split ISD node for Vfpclass and Vfpclasss so that we can write strong type constraints for each that don't cause ambiguous isel.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254172
91177308-0d34-0410-b5e6-
96231b3b80d8
Hal Finkel [Thu, 26 Nov 2015 19:23:49 +0000 (19:23 +0000)]
[bugpoint] Fix "Alias must point to a definition" problems
GlobalAliases may reference function definitions, but not function declarations.
bugpoint would sometimes create invalid IR by deleting a function's body (thus
mutating a function definition into a declaration) without first 'fixing' any
GlobalAliases that reference that function definition.
This change iteratively prevents that issue. Before deleting a function's body,
it scans the module for GlobalAliases which reference that function. When
found, it eliminates them using replaceAllUsesWith.
Fixes PR20788.
Patch by Nick Johnson!
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254171
91177308-0d34-0410-b5e6-
96231b3b80d8
Rafael Espindola [Thu, 26 Nov 2015 19:22:59 +0000 (19:22 +0000)]
Disallow aliases to available_externally.
They are as much trouble as aliases to declarations. They are requiring
the code generator to define a symbol with the same value as another
symbol, but the second symbol is undefined.
If representing this is important for some optimization, we could add
support for available_externally aliases. They would be *required* to
point to a declaration (or available_externally definition).
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254170
91177308-0d34-0410-b5e6-
96231b3b80d8
Craig Topper [Thu, 26 Nov 2015 19:13:05 +0000 (19:13 +0000)]
[X86] Revert part of r254167 to recover bots.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254169
91177308-0d34-0410-b5e6-
96231b3b80d8
Krzysztof Parzyszek [Thu, 26 Nov 2015 18:38:27 +0000 (18:38 +0000)]
[Hexagon] Lowering of V60/HVX vector types
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254168
91177308-0d34-0410-b5e6-
96231b3b80d8
Craig Topper [Thu, 26 Nov 2015 18:31:19 +0000 (18:31 +0000)]
[X86] Strengthen more type constraints to reduce isel table size.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254167
91177308-0d34-0410-b5e6-
96231b3b80d8
Krzysztof Parzyszek [Thu, 26 Nov 2015 16:54:33 +0000 (16:54 +0000)]
[Hexagon] Hexagon V60 HVX intrinsic defintions
Author: Ron Lieberman <ronl@codeaurora.org>
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254165
91177308-0d34-0410-b5e6-
96231b3b80d8
Daniel Sanders [Thu, 26 Nov 2015 16:35:41 +0000 (16:35 +0000)]
[mips][ias] Range check uimm5 operands and fix several bugs this revealed.
Summary:
The bugs were:
* append, prepend, and balign were not tested
* balign takes a uimm2 not a uimm5.
* drotr32 was correctly implemented with a uimm5 but the tests expected
'52' to be valid.
* li/la were implemented with a uimm5 instead of simm32. simm32 isn't
completely correct either but I'll fix that when I get to simm32.
A notable omission are some of the shift instructions. Several of these
have been implemented using a single uimm6 instruction (rather than two
uimm5 instructions and a CodeGen-only uimm6 pseudo). These will be updated
in the uimm6 patch.
Reviewers: vkalintiris
Subscribers: llvm-commits, dsanders
Differential Revision: http://reviews.llvm.org/D14712
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254164
91177308-0d34-0410-b5e6-
96231b3b80d8
Oliver Stannard [Thu, 26 Nov 2015 15:34:44 +0000 (15:34 +0000)]
[AArch64] Add ARMv8.2-A new AT instruction variants
ARMv8.2-A adds new variants of the "at" (address translate) system
instruction, which take the PSTATE.PAN bit (added in ARMv8.1-A). These
are a required part of ARMv8.2-A, so no additional subtarget features
are required.
Differential Revision: http://reviews.llvm.org/D15018
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254159
91177308-0d34-0410-b5e6-
96231b3b80d8
Martell Malone [Thu, 26 Nov 2015 15:34:03 +0000 (15:34 +0000)]
ARM: address WOA unsigned division overflow crash
Building on r253865 the crash is not limited to signed overflows.
Disable custom handling of unsigned 32-bit and 64-bit integer divide.
Add test cases for both 32-bit and 64-bit unsigned integer overflow.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254158
91177308-0d34-0410-b5e6-
96231b3b80d8
Oliver Stannard [Thu, 26 Nov 2015 15:32:30 +0000 (15:32 +0000)]
[AArch64] Add ARMv8.2-A UAO PSTATE bit
ARMv8.2-A adds a new PSTATE bit, PSTATE.UAO, which allows the LDTR/STTR
instructions to behave the same as LDR/STR with respect to execute-only
pages at higher privilege levels. New variants of the MSR/MRS
instructions are added to allow reading and writing this bit. It is a
required part of ARMv8.2-A, so no additional subtarget features are
required.
Differential Revision: http://reviews.llvm.org/D15020
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254157
91177308-0d34-0410-b5e6-
96231b3b80d8
Oliver Stannard [Thu, 26 Nov 2015 15:28:47 +0000 (15:28 +0000)]
[AArch64] Add ARMv8.2-A persistent memory instruction
ARMv8.2-A adds the "dc cvap" instruction, which is a system instruction
that cleans caches to the point of persistence (for systems that have
persistent memory). It is a required part of ARMv8.2-A, so no additional
subtarget features are required.
Differential Revision: http://reviews.llvm.org/D15016
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254156
91177308-0d34-0410-b5e6-
96231b3b80d8
Oliver Stannard [Thu, 26 Nov 2015 15:26:10 +0000 (15:26 +0000)]
[AArch64] Add ARMv8.2-A ID_A64MMFR2_EL1 register
ARMv8.2-A adds a new ID register, ID_A64MMFR2_EL1, which behaves in the
same way as ID_A64MMFR0_EL1 and ID_A64MMFR1_EL1. It is a required part
of ARMv8.2-A, so no additional subtarget features are required.
Differential Revision: http://reviews.llvm.org/D15017
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254155
91177308-0d34-0410-b5e6-
96231b3b80d8
Oliver Stannard [Thu, 26 Nov 2015 15:23:32 +0000 (15:23 +0000)]
[AArch64] Add subtarget features for ARMv8.2-A
This adds subtarget features for ARMv8.2-A, which builds on (and
requires the features from) ARMv8.1-A. Most assembler-visible features
of ARMv8.2-A are system instructions, and are all required parts of the
architecture, so just depend on the HasV8_2aOps subtarget feature. There
is also one large, optional feature, which adds 16-bit floating point
versions of all existing floating-point instructions (VFP and SIMD),
this is represented by the FeatureFullFP16 subtarget feature.
Differential Revision: http://reviews.llvm.org/D15013
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254154
91177308-0d34-0410-b5e6-
96231b3b80d8
Daniel Sanders [Thu, 26 Nov 2015 11:23:03 +0000 (11:23 +0000)]
[mips][ias] Explicitly disable IAS on tests that depend on not assembling.
Summary:
no-odd-spreg-msa.ll: This test deliberately uses an odd-numbered register
in inline assembly and expects the compiler to insert a move to an
even-numbered register.
inlineasm-operand-code.ll and inlineasm_constraint.ll:
Checks for IAS's output will be added once a matcher bug is resolved. This bug
causes the canonical output emitted by IAS to be incorrect for uimm16 constants
with the MSB set. We will still need the non-IAS checks at this point since
these tests primarily test formatting of operands.
Reviewers: vkalintiris
Subscribers: dsanders, llvm-commits
Differential Revision: http://reviews.llvm.org/D14705
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254148
91177308-0d34-0410-b5e6-
96231b3b80d8
Daniel Sanders [Thu, 26 Nov 2015 10:26:18 +0000 (10:26 +0000)]
[mips][ias] Replace anchor comments with anchor instructions in tests.
Summary:
This is because IAS will delete the comments. NFC at the moment but it will
prevent a failure once IAS is the default.
Reviewers: vkalintiris
Subscribers: llvm-commits, dsanders
Differential Revision: http://reviews.llvm.org/D14704
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254147
91177308-0d34-0410-b5e6-
96231b3b80d8
Benjamin Kramer [Thu, 26 Nov 2015 09:51:17 +0000 (09:51 +0000)]
[SimplifyLibCalls] Don't depend on a called function having a name, it might be an indirect call.
Fixes the crasher in PR25651 and related crashers using the same pattern.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254145
91177308-0d34-0410-b5e6-
96231b3b80d8
Craig Topper [Thu, 26 Nov 2015 07:58:20 +0000 (07:58 +0000)]
[X86] Strengthen more type constraints to reduce isel table size.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254142
91177308-0d34-0410-b5e6-
96231b3b80d8
Vyacheslav Klochkov [Thu, 26 Nov 2015 07:45:30 +0000 (07:45 +0000)]
X86-FMA3: Improved/enabled the memory folding optimization for scalar loads
generated for _mm_losd_s{s,d}() intrinsics and used in scalar FMAs generated
for FMA intrinsics _mm_f{madd,msub,nmadd,nmsub}_s{s,d}().
Reviewer: David Kreitzer
Differential Revision: http://reviews.llvm.org/D14762
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254140
91177308-0d34-0410-b5e6-
96231b3b80d8
Craig Topper [Thu, 26 Nov 2015 07:02:21 +0000 (07:02 +0000)]
[X86] Strengthen the type constraints on X86psadbw and X86dbpsadbw to reduce some of the type checks in the isel matching tables.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254139
91177308-0d34-0410-b5e6-
96231b3b80d8
Craig Topper [Thu, 26 Nov 2015 07:02:18 +0000 (07:02 +0000)]
[SelectionDAG] Add a SDTCisSameSizeAs type constraint that can be used to ensure vector widths match even if the element size and count don't.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254138
91177308-0d34-0410-b5e6-
96231b3b80d8
Craig Topper [Thu, 26 Nov 2015 06:30:42 +0000 (06:30 +0000)]
Add type constraints to masked_load/masked_store to ensure the mask vector has the same number of elements as the other vectors.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254137
91177308-0d34-0410-b5e6-
96231b3b80d8
Craig Topper [Thu, 26 Nov 2015 06:30:40 +0000 (06:30 +0000)]
[TableGen] Flip reversed comments.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254136
91177308-0d34-0410-b5e6-
96231b3b80d8
Krzysztof Parzyszek [Thu, 26 Nov 2015 04:33:11 +0000 (04:33 +0000)]
[Hexagon] HVX vector register classes and more isel patterns
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254132
91177308-0d34-0410-b5e6-
96231b3b80d8
Tom Stellard [Thu, 26 Nov 2015 02:04:11 +0000 (02:04 +0000)]
AMDGPU: Fix typo
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254120
91177308-0d34-0410-b5e6-
96231b3b80d8
Sanjoy Das [Thu, 26 Nov 2015 01:16:05 +0000 (01:16 +0000)]
[OperandBundles] Treat "deopt" operand bundles specially
Teach LLVM optimize to more precisely in the presence of "deopt" operand
bundles. "deopt" operand bundles imply that the call they're attached
to is at least `readonly` (i.e. they don't imply clobber semantics), and
they don't capture their bundle operands.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254118
91177308-0d34-0410-b5e6-
96231b3b80d8