OSDN Git Service
Amaury Sechet [Tue, 9 Feb 2016 23:15:02 +0000 (23:15 +0000)]
Add icmp and conditional branches in the C API echo test.
Summary:
Improving coverage.
Depends on D16912 .
Reviewers: bogner, chandlerc, echristo, dblaikie, joker.eph, Wallbraker
Subscribers: llvm-commits
Differential Revision: http://reviews.llvm.org/D16937
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@260321
91177308-0d34-0410-b5e6-
96231b3b80d8
Justin Lebar [Tue, 9 Feb 2016 23:03:22 +0000 (23:03 +0000)]
Add convergent-removing bits to FunctionAttrs pass.
Summary:
Remove the convergent attribute on any functions which provably do not
contain or invoke any convergent functions.
After this change, we'll be able to modify clang to conservatively add
'convergent' to all functions when compiling CUDA.
Reviewers: jingyue, joker.eph
Subscribers: llvm-commits, tra, jhen, hfinkel, resistor, chandlerc, arsenm
Differential Revision: http://reviews.llvm.org/D17013
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@260319
91177308-0d34-0410-b5e6-
96231b3b80d8
Justin Lebar [Tue, 9 Feb 2016 23:03:17 +0000 (23:03 +0000)]
Update documentation about convergent attribute.
Summary:
Be more explicit about what 'convergent' means, and indicate that the
compiler may remove the attribute from a function if it can prove that
the function doesn't in fact execute any convergent ops.
Reviewers: resistor, jingyue, joker.eph
Subscribers: hfinkel, chandlerc, arsenm, jhen, tra, llvm-commits
Differential Revision: http://reviews.llvm.org/D17012
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@260318
91177308-0d34-0410-b5e6-
96231b3b80d8
Peter Collingbourne [Tue, 9 Feb 2016 23:01:38 +0000 (23:01 +0000)]
Fix GCC build.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@260317
91177308-0d34-0410-b5e6-
96231b3b80d8
Ahmed Bougacha [Tue, 9 Feb 2016 22:54:12 +0000 (22:54 +0000)]
[CodeGen] Prefer "if (SDValue R = ...)" to "if (R.getNode())". NFCI.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@260316
91177308-0d34-0410-b5e6-
96231b3b80d8
Ahmed Bougacha [Tue, 9 Feb 2016 22:54:05 +0000 (22:54 +0000)]
[X86] Don't reuse an unrelated variable, create a new one. NFC.
Using Op makes it look like we're doing something with it.
We're really not.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@260315
91177308-0d34-0410-b5e6-
96231b3b80d8
Ahmed Bougacha [Tue, 9 Feb 2016 22:53:58 +0000 (22:53 +0000)]
[X86] Remove unnecessary assignment. NFC.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@260314
91177308-0d34-0410-b5e6-
96231b3b80d8
Amaury Sechet [Tue, 9 Feb 2016 22:50:53 +0000 (22:50 +0000)]
Add C binding for AllocaInst::getAllocatedType
Summary:
Comes with an awesome test.
Depends on D16912
Reviewers: bogner, chandlerc, echristo, dblaikie, joker.eph, Wallbraker
Subscribers: llvm-commits
Differential Revision: http://reviews.llvm.org/D16942
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@260313
91177308-0d34-0410-b5e6-
96231b3b80d8
Peter Collingbourne [Tue, 9 Feb 2016 22:50:34 +0000 (22:50 +0000)]
WholeProgramDevirt: introduce.
This pass implements whole program optimization of virtual calls in cases
where we know (via bitset information) that the list of callees is fixed. This
includes the following:
- Single implementation devirtualization: if a virtual call has a single
possible callee, replace all calls with a direct call to that callee.
- Virtual constant propagation: if the virtual function's return type is an
integer <=64 bits and all possible callees are readnone, for each class and
each list of constant arguments: evaluate the function, store the return
value alongside the virtual table, and rewrite each virtual call as a load
from the virtual table.
- Uniform return value optimization: if the conditions for virtual constant
propagation hold and each function returns the same constant value, replace
each virtual call with that constant.
- Unique return value optimization for i1 return values: if the conditions
for virtual constant propagation hold and a single vtable's function
returns 0, or a single vtable's function returns 1, replace each virtual
call with a comparison of the vptr against that vtable's address.
Differential Revision: http://reviews.llvm.org/D16795
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@260312
91177308-0d34-0410-b5e6-
96231b3b80d8
Amaury Sechet [Tue, 9 Feb 2016 22:36:41 +0000 (22:36 +0000)]
Improve the C API echo test tool to emit basic block is the right order.
Summary: As per title. Also add a facility method to get the name of a basic block from the C API.
Reviewers: bogner, chandlerc, echristo, dblaikie, joker.eph, Wallbraker
Subscribers: llvm-commits
Differential Revision: http://reviews.llvm.org/D16912
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@260309
91177308-0d34-0410-b5e6-
96231b3b80d8
Simon Atanasyan [Tue, 9 Feb 2016 22:31:49 +0000 (22:31 +0000)]
[mips] Extend MipsAsmParser class to handle %got(sym + const) expressions
Now the parser supports `%got(sym)` expressions only but `%got(sym + const)`
variant is also valid and accepted by GAS.
Differential Revision: http://reviews.llvm.org/D16885
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@260305
91177308-0d34-0410-b5e6-
96231b3b80d8
Sanjoy Das [Tue, 9 Feb 2016 21:54:18 +0000 (21:54 +0000)]
[MemoryBuiltins] Fix an issue with hasNoAliasAttr
Summary:
`hasNoAliasAttr` is buggy: it checks to see if the called function has
a `noalias` attribute, which is incorrect since functions are not even
allowed to have the `noalias` attribute. The comment on its only
caller, `llvm::isNoAliasFn`, makes it pretty clear that the intention
to do the `noalias` check on the return value, and not the callee.
Unfortunately I couldn't find a way to test this upstream -- fixing
this does not change the observable behavior of any of the passes that
use this. This is not very surprising, since `noalias` does not tell
anything about the contents of the allocated memory (so, e.g., you
still cannot fold loads). I'll be happy to be proven wrong though.
Reviewers: chandlerc, reames
Subscribers: mcrosier, llvm-commits
Differential Revision: http://reviews.llvm.org/D17037
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@260298
91177308-0d34-0410-b5e6-
96231b3b80d8
Philip Reames [Tue, 9 Feb 2016 21:43:12 +0000 (21:43 +0000)]
[Verifier] Add checks for masked.load and masked.store intrinsics
While trying to track down what appears to be a LoopVectorizer bug, I noticed that we had no validation of the correctness of calls emitted to @llvm.masked.load and @llvm.masked.store. This meant malformed IR was showing up much much later than it should. Hopefully, having Verifier rules in place will make this easier to isolate.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@260296
91177308-0d34-0410-b5e6-
96231b3b80d8
Sanjay Patel [Tue, 9 Feb 2016 21:42:04 +0000 (21:42 +0000)]
[SelectionDAG] make getMemBasePlusOffset() accessible; NFCI
I reinvented this functionality in http://reviews.llvm.org/D16828 because it was
hidden away as a static function. The changes in x86 are not based on a complete
audit. I suspect there are other possible uses there, and there are almost certainly
more potential users in other targets.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@260295
91177308-0d34-0410-b5e6-
96231b3b80d8
Rafael Espindola [Tue, 9 Feb 2016 21:39:49 +0000 (21:39 +0000)]
This brings back commit r259578.
But now it follows the llvm style, uses an early return and doesn't
include a file named 1.o.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@260293
91177308-0d34-0410-b5e6-
96231b3b80d8
Rafael Espindola [Tue, 9 Feb 2016 21:35:14 +0000 (21:35 +0000)]
Fix name style.
Functions start with a lower case, variables with an upper case.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@260287
91177308-0d34-0410-b5e6-
96231b3b80d8
Rafael Espindola [Tue, 9 Feb 2016 21:32:56 +0000 (21:32 +0000)]
Update comment style.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@260286
91177308-0d34-0410-b5e6-
96231b3b80d8
Rafael Espindola [Tue, 9 Feb 2016 21:30:10 +0000 (21:30 +0000)]
Revert "Correct size calculations for ELF files"
This reverts commit r259578.
There are enough issues with this small patch that it is better to
revert and then commit a fixed version (will be committed shortly).
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@260285
91177308-0d34-0410-b5e6-
96231b3b80d8
Chad Rosier [Tue, 9 Feb 2016 21:20:12 +0000 (21:20 +0000)]
[AArch64] This check is specific to merging instructions. NFC.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@260283
91177308-0d34-0410-b5e6-
96231b3b80d8
Ehsan Akhgari [Tue, 9 Feb 2016 21:17:58 +0000 (21:17 +0000)]
Remove system-libs.test for now
This test fails in the ninja-x64-msvc-RA-centos6 builder, so the
UNSUPPORTED: system-windows condition is insufficient. Removing it
for now; I will investigate how this can be fixed later.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@260281
91177308-0d34-0410-b5e6-
96231b3b80d8
Ehsan Akhgari [Tue, 9 Feb 2016 21:10:22 +0000 (21:10 +0000)]
Fix a test added in r260263
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@260280
91177308-0d34-0410-b5e6-
96231b3b80d8
Philip Reames [Tue, 9 Feb 2016 21:09:22 +0000 (21:09 +0000)]
[InstCombine][GC] Handle gc.relocations of vector type
We introduced gc.relocates of vector-of-pointer types a couple of weeks back. Somehow, I missed updating the InstCombine rule to account for this. If we hit this code path with a vector-of-pointers gc.relocate, we'd crash on a cast<PointerType>.
I also took the chance to do a bit of code style cleanup.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@260279
91177308-0d34-0410-b5e6-
96231b3b80d8
Geoff Berry [Tue, 9 Feb 2016 20:47:21 +0000 (20:47 +0000)]
[AArch64] AArch64LoadStoreOptimizer: fix bug in pre-inc check iterator
Summary:
Fix case where a pre-inc/dec load/store would not be formed if the
add/sub that forms the inc/dec part of the operation was the first
instruction in the block being examined.
Reviewers: mcrosier, jmolloy, t.p.northover, junbuml
Subscribers: aemerson, rengolin, mcrosier, llvm-commits
Differential Revision: http://reviews.llvm.org/D16785
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@260275
91177308-0d34-0410-b5e6-
96231b3b80d8
Chad Rosier [Tue, 9 Feb 2016 20:44:41 +0000 (20:44 +0000)]
[AArch64] Bail even earlier if the instructions modifieds the base register. NFC.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@260274
91177308-0d34-0410-b5e6-
96231b3b80d8
Chad Rosier [Tue, 9 Feb 2016 20:27:45 +0000 (20:27 +0000)]
[AArch64] Simplify. NFC.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@260273
91177308-0d34-0410-b5e6-
96231b3b80d8
Chad Rosier [Tue, 9 Feb 2016 20:18:07 +0000 (20:18 +0000)]
[AArch64] Add an assert to ensure we don't scale an offset that can't be scaled.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@260272
91177308-0d34-0410-b5e6-
96231b3b80d8
Ehsan Akhgari [Tue, 9 Feb 2016 20:02:35 +0000 (20:02 +0000)]
Fix a typo in r260263
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@260269
91177308-0d34-0410-b5e6-
96231b3b80d8
Chad Rosier [Tue, 9 Feb 2016 19:42:19 +0000 (19:42 +0000)]
[AArch64] Add a FIXME about invalid KILL markers after the ld/st opt pass.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@260264
91177308-0d34-0410-b5e6-
96231b3b80d8
Ehsan Akhgari [Tue, 9 Feb 2016 19:41:14 +0000 (19:41 +0000)]
llvm-config: Add preliminary Windows support
Summary:
This patch adds Windows support for a few of the llvm-config commands,
including cflags, ldflags, libs, and system-libs.
Currently llvm-config is untested, so this patch adds tests for the
commands that it fixes as well.
Reviewers: rnk
Subscribers: llvm-commits
Differential Revision: http://reviews.llvm.org/D16762
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@260263
91177308-0d34-0410-b5e6-
96231b3b80d8
Ehsan Akhgari [Tue, 9 Feb 2016 19:37:27 +0000 (19:37 +0000)]
Turn on LLVM_ENABLE_PLUGINS by default on Windows
Reviewers: rnk
Subscribers: llvm-commits
Differential Revision: http://reviews.llvm.org/D16760
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@260262
91177308-0d34-0410-b5e6-
96231b3b80d8
Ehsan Akhgari [Tue, 9 Feb 2016 19:37:25 +0000 (19:37 +0000)]
Add support for importing and exporting Registry objects on Windows
On Windows, the DLL containing the registry will get its own global head
and tail variables, so the entries registered in the DLL will be
invisible to the consumer.
In order to solve this, we need to export a getter function from the
plugin DLL per registry and copy over the data inside it. This patch
adds support for this. This will be used to support clang plugins on
Windows.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@260261
91177308-0d34-0410-b5e6-
96231b3b80d8
Chad Rosier [Tue, 9 Feb 2016 19:33:42 +0000 (19:33 +0000)]
[AArch64] Remove redundant calls and clang format. NFC.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@260260
91177308-0d34-0410-b5e6-
96231b3b80d8
Colin LeMahieu [Tue, 9 Feb 2016 19:18:02 +0000 (19:18 +0000)]
[Hexagon] Fixing relocation generation and adding tests.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@260259
91177308-0d34-0410-b5e6-
96231b3b80d8
Colin LeMahieu [Tue, 9 Feb 2016 19:17:34 +0000 (19:17 +0000)]
[MC] Adding GOTREL expression variant.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@260258
91177308-0d34-0410-b5e6-
96231b3b80d8
Chad Rosier [Tue, 9 Feb 2016 19:17:18 +0000 (19:17 +0000)]
[AArch64] Hoist now common logic. NFC.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@260257
91177308-0d34-0410-b5e6-
96231b3b80d8
Chad Rosier [Tue, 9 Feb 2016 19:09:22 +0000 (19:09 +0000)]
[AArch64] Rename variable to make it clear we're merging here, not pairing.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@260256
91177308-0d34-0410-b5e6-
96231b3b80d8
Chad Rosier [Tue, 9 Feb 2016 19:02:12 +0000 (19:02 +0000)]
[AArch64] Separage the codegen logic for widening vs. pairing. NFC.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@260249
91177308-0d34-0410-b5e6-
96231b3b80d8
Simon Atanasyan [Tue, 9 Feb 2016 18:45:35 +0000 (18:45 +0000)]
[llvm-readobj][ELF] Show MIPS GOT content when there is another zero-sized section at the same address
It is possible to have .got section and one or more zero-sized section
at the same address. This patch first checks that GOT (or GOT PLT)
section should have non-zero size using corresponding dynamic tags. Then
it looks up not empty section at the specified address.
Differential Revision: http://reviews.llvm.org/D16968
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@260245
91177308-0d34-0410-b5e6-
96231b3b80d8
Sanjoy Das [Tue, 9 Feb 2016 18:40:40 +0000 (18:40 +0000)]
[FunctionAttrs] Fix SCC logic around operand bundles
FunctionAttrs does an "optimistic" analysis of SCCs as a unit, which
means normally it is able to disregard calls from an SCC into itself.
However, calls and invokes with operand bundles are allowed to have
memory effects not fully described by the memory effects on the call
target, so we can't be optimistic around operand-bundled calls from an
SCC into itself.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@260244
91177308-0d34-0410-b5e6-
96231b3b80d8
Kevin Enderby [Tue, 9 Feb 2016 18:33:15 +0000 (18:33 +0000)]
Fix a formatting problems with llvm-size and the -m option.
It was using format() with a string for 64-bit types but was
passed a 32-bit type in places when printing values for
32-bit Mach-O files.
rdar://
24542509
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@260243
91177308-0d34-0410-b5e6-
96231b3b80d8
Chad Rosier [Tue, 9 Feb 2016 18:10:20 +0000 (18:10 +0000)]
[AArch64] Cleanup to simplify logic when widening vs. pairing loads/stores. NFC.
The logic to pair instructions and merge narrow instructions has become cloogy
and error prone. This patch beings to unravel these two similar, but distinct
optimizations.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@260242
91177308-0d34-0410-b5e6-
96231b3b80d8
Sanjay Patel [Tue, 9 Feb 2016 17:39:58 +0000 (17:39 +0000)]
[x86] make getOneTrueElt() a helper function ; NFC
As mentioned in http://reviews.llvm.org/D16828 , the related masked load transform
will need this logic, so I'm moving it out to make that patch smaller.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@260240
91177308-0d34-0410-b5e6-
96231b3b80d8
Chad Rosier [Tue, 9 Feb 2016 15:59:57 +0000 (15:59 +0000)]
[AArch64] Rename variable to improve readability. NFC.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@260228
91177308-0d34-0410-b5e6-
96231b3b80d8
Chad Rosier [Tue, 9 Feb 2016 15:51:33 +0000 (15:51 +0000)]
[AArch64] Remove stale comment.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@260226
91177308-0d34-0410-b5e6-
96231b3b80d8
Teresa Johnson [Tue, 9 Feb 2016 13:28:44 +0000 (13:28 +0000)]
Fix comment for MD5Hash.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@260221
91177308-0d34-0410-b5e6-
96231b3b80d8
Simon Pilgrim [Tue, 9 Feb 2016 08:19:19 +0000 (08:19 +0000)]
[X86][AVX2] Fix SIGN_EXTEND vector handling on AVX2 targets.
On AVX2 target we are poorly legalizing SIGN_EXTEND ops for which the input's legalized type doesn't have the same number of elements as the destination, resulting in an ANY_EXTEND followed by a SIGN_EXTEND_INREG.
This patch uses the existing SIGN_EXTEND -> SIGN_EXTEND_VECTOR_INREG combine to extend the input to the size of the result and using SIGN_EXTEND_VECTOR_INREG instead.
Differential Revision: http://reviews.llvm.org/D16994
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@260210
91177308-0d34-0410-b5e6-
96231b3b80d8
Xinliang David Li [Tue, 9 Feb 2016 05:47:08 +0000 (05:47 +0000)]
Add comments to some tests
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@260200
91177308-0d34-0410-b5e6-
96231b3b80d8
Xinliang David Li [Tue, 9 Feb 2016 05:36:57 +0000 (05:36 +0000)]
Further reduce test overhead
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@260198
91177308-0d34-0410-b5e6-
96231b3b80d8
Teresa Johnson [Tue, 9 Feb 2016 05:12:44 +0000 (05:12 +0000)]
Refactor PGO function naming and MD5 hashing support out of ProfileData
Summary:
Move the function renaming logic into the Function class, and the
MD5Hash routine into the MD5 header.
This will enable these routines to be shared with ThinLTO, which
will be changed to store the MD5 hash instead of full function name
in the combined index for significant size reductions. And using the same
function naming for locals in the function index facilitates future
integration with indirect call value profiles.
Reviewers: davidxl
Subscribers: llvm-commits
Differential Revision: http://reviews.llvm.org/D17006
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@260197
91177308-0d34-0410-b5e6-
96231b3b80d8
Nick Lewycky [Tue, 9 Feb 2016 04:47:58 +0000 (04:47 +0000)]
Use std::forward to make ErrorOr<T> constructible from a value that has a user-defined conversion to T. No functionality change intended.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@260196
91177308-0d34-0410-b5e6-
96231b3b80d8
Sanjoy Das [Tue, 9 Feb 2016 02:31:47 +0000 (02:31 +0000)]
[GMR/OperandBundles] Teach getModRefBehavior about operand bundles
In general, memory restrictions on a called function (e.g. readnone)
cannot be transferred to a CallSite that has operand bundles. It is
possible to make this inference smarter, but lets fix the behavior to be
correct first.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@260193
91177308-0d34-0410-b5e6-
96231b3b80d8
Richard Smith [Tue, 9 Feb 2016 02:09:16 +0000 (02:09 +0000)]
Remove TrailingObjects::operator delete. It's still suffering from
compiler-specific issues. Instead, repeat an 'operator delete' definition in
each derived class that is actually deleted, and give up on the static type
safety of an error when sized delete is accidentally used on a type derived
from TrailingObjects.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@260190
91177308-0d34-0410-b5e6-
96231b3b80d8
David L Kreitzer [Tue, 9 Feb 2016 01:35:45 +0000 (01:35 +0000)]
Fix the LLVM_ENABLE_MODULES build after adding TargetOpcodes.def in r259726.
Differential Revision: http://reviews.llvm.org/D17005
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@260186
91177308-0d34-0410-b5e6-
96231b3b80d8
Sanjoy Das [Tue, 9 Feb 2016 01:21:57 +0000 (01:21 +0000)]
Add an "addUsedAAAnalyses" helper function
Summary:
Passes that call `getAnalysisIfAvailable<T>` also need to call
`addUsedIfAvailable<T>` in `getAnalysisUsage` to indicate to the
legacy pass manager that it uses `T`. This contract was being
violated by passes that used `createLegacyPMAAResults`. This change
fixes this by exposing a helper in AliasAnalysis.h,
`addUsedAAAnalyses`, that is complementary to createLegacyPMAAResults
and does the right thing when called from `getAnalysisUsage`.
Reviewers: chandlerc
Subscribers: mcrosier, llvm-commits
Differential Revision: http://reviews.llvm.org/D17010
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@260183
91177308-0d34-0410-b5e6-
96231b3b80d8
Sanjoy Das [Tue, 9 Feb 2016 01:21:50 +0000 (01:21 +0000)]
Remove SCEVAAWrapperPass from createLegacyPMAAResults; NFC
Summary:
createLegacyPMAAResults is only called by CGSCC and Module passes, so
the call to getAnalysisIfAvailable<SCEVAAWrapperPass>() never
succeeds (SCEVAAWrapperPass is a function pass).
Reviewers: chandlerc
Subscribers: mcrosier, llvm-commits
Differential Revision: http://reviews.llvm.org/D17009
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@260182
91177308-0d34-0410-b5e6-
96231b3b80d8
Richard Smith [Tue, 9 Feb 2016 01:03:42 +0000 (01:03 +0000)]
Re-commit r259942 (reverted in r260053) with a different workaround for the MSVC bug.
This fixes undefined behavior in C++14 due to the size of the object being
deleted being different from sizeof(dynamic type) when it is allocated with
trailing objects.
MSVC seems to have several bugs around using-declarations changing the access
of a member inherited from a base class, so use forwarding functions instead of
using-declarations to make TrailingObjects::operator delete accessible where
desired.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@260180
91177308-0d34-0410-b5e6-
96231b3b80d8
David Blaikie [Tue, 9 Feb 2016 01:02:24 +0000 (01:02 +0000)]
Simplify some expressions involving unique_ptr and ErrorOr
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@260179
91177308-0d34-0410-b5e6-
96231b3b80d8
Wei Mi [Tue, 9 Feb 2016 00:07:08 +0000 (00:07 +0000)]
This patch is to fix PR26529 caused by r259736.
IndVarSimplify assumes scAddRecExpr to be expanded in literal form instead of
canonical form by calling disableCanonicalMode after it creates SCEVExpander.
When CanonicalMode is disabled, SCEVExpander::expand should always return PHI
node for scAddRecExpr. r259736 broke the assumption.
The fix is to let SCEVExpander::expand skip the reuse Value logic if
CanonicalMode is false.
In addition, Besides IndVarSimplify, LSR pass also calls disableCanonicalMode
before doing rewrite. We can remove the original check of LSRMode in reuse
Value logic and use CanonicalMode instead.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@260174
91177308-0d34-0410-b5e6-
96231b3b80d8
Davide Italiano [Mon, 8 Feb 2016 23:50:23 +0000 (23:50 +0000)]
[llvm-nm] Remove excessive parenthesis, noticed by David Blaikie.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@260173
91177308-0d34-0410-b5e6-
96231b3b80d8
Rong Xu [Mon, 8 Feb 2016 23:11:16 +0000 (23:11 +0000)]
[PGO] Revert r260146 as it breaks Darwin platforms.
r260146 | xur | 2016-02-08 13:07:46 -0800 (Mon, 08 Feb 2016) | 13 lines
[PGO] Differentiate Clang instrumentation and IR level instrumentation profiles
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@260170
91177308-0d34-0410-b5e6-
96231b3b80d8
Michael Zolotukhin [Mon, 8 Feb 2016 23:03:59 +0000 (23:03 +0000)]
Factor out UnrollAnalyzer to Analysis, and add unit tests for it.
Summary:
Unrolling Analyzer is already pretty complicated, and it becomes harder and harder to exercise it with usual IR tests, as with them we can only check the final decision: whether the loop is unrolled or not. This change factors this framework out from LoopUnrollPass to analyses, which allows to use unit tests.
The change itself is supposed to be NFC, except adding a couple of tests.
I plan to add more tests as I add new functionality and find/fix bugs.
Reviewers: chandlerc, hfinkel, sanjoy
Subscribers: zzheng, sanjoy, llvm-commits
Differential Revision: http://reviews.llvm.org/D16623
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@260169
91177308-0d34-0410-b5e6-
96231b3b80d8
Simon Pilgrim [Mon, 8 Feb 2016 23:03:46 +0000 (23:03 +0000)]
[X86][SSE1] Add MOVLHPS/MOVHLPS lowering and memory folding support
As discussed on PR26491, this patch adds support for lowering v4f32 shuffles to the MOVLHPS/MOVHLPS instructions. It also adds support for memory folding with their MOVLPS/MOVHPS load equivalents.
This first patch only really helps SSE1 targets as SSE2+ targets will widen the shuffle mask and use v2f64 equivalents (although they still combine to MOVLHPS/MOVHLPS for v2f64 splats). This will have to be addressed in a future patch, most likely when we add support for binary target shuffle combines.
Differential Revision: http://reviews.llvm.org/D16956
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@260168
91177308-0d34-0410-b5e6-
96231b3b80d8
Davide Italiano [Mon, 8 Feb 2016 22:58:26 +0000 (22:58 +0000)]
[llvm-nm] Yet another attempt of simplifying code.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@260166
91177308-0d34-0410-b5e6-
96231b3b80d8
Andrew Kaylor [Mon, 8 Feb 2016 22:52:51 +0000 (22:52 +0000)]
[regalloc][WinEH] Do not mark intervals as not spillable if they contain a regmask
Differential Revision: http://reviews.llvm.org/D16831
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@260164
91177308-0d34-0410-b5e6-
96231b3b80d8
Justin Bogner [Mon, 8 Feb 2016 22:49:40 +0000 (22:49 +0000)]
llvm-cov: Fix reading gcov data that does not have function names
In order for recent gcov versions to read the coverage data, you have
to use UseCfgChecksum=true and FunctionNamesInData=false options for
coverage profiling pass. This is because gcov is expecting the
function section in .gcda to be exactly 3 words in size, containing
ident and two checksums.
While llvm-cov is compatible with UseCfgChecksum=true, it always
expects a function name in .gcda function sections (it's not
compatible with FunctionNamesInData=false). Thus it's currently
impossible to generate one set of coverage files that works with both
gcov and llvm-cov.
This change fixes the reading of coverage information to only read the
function name if it's present.
Patch by Arseny Kapoulkine. Thanks!
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@260162
91177308-0d34-0410-b5e6-
96231b3b80d8
Justin Bogner [Mon, 8 Feb 2016 21:55:19 +0000 (21:55 +0000)]
cmake: Use "set" instead of "option" for LLVM_ENABLE_LTO
Apparently option is for bools and cmake-gui will display this
strangely with option.
Pointed out by edward-san - thanks!
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@260154
91177308-0d34-0410-b5e6-
96231b3b80d8
Dan Gohman [Mon, 8 Feb 2016 21:50:13 +0000 (21:50 +0000)]
[WebAssembly] Update the br_if instructions' operand orders to match the spec.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@260152
91177308-0d34-0410-b5e6-
96231b3b80d8
Sanjay Patel [Mon, 8 Feb 2016 21:32:43 +0000 (21:32 +0000)]
rangify; NFC
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@260151
91177308-0d34-0410-b5e6-
96231b3b80d8
Rong Xu [Mon, 8 Feb 2016 21:07:46 +0000 (21:07 +0000)]
[PGO] Differentiate Clang instrumentation and IR level instrumentation profiles
This patch uses one bit in profile version to differentiate Clang
instrumentation and IR level instrumentation profiles.
PGOInstrumenation generates a COMDAT variable __llvm_profile_raw_version so
that the compiler runtime can set the right profile kind.
PGOInstrumenation now checks this bit to make sure it's an IR level
instrumentation profile.
Differential Revision: http://reviews.llvm.org/D15540
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@260146
91177308-0d34-0410-b5e6-
96231b3b80d8
Sanjay Patel [Mon, 8 Feb 2016 21:05:08 +0000 (21:05 +0000)]
[x86] convert masked store of one element to scalar store
Another opportunity to reduce masked stores: in D16691, we decided not to attempt the 'one mask element is set'
transform in InstCombine, but this should be a win for any AVX machine.
Code comments note that this transform could be extended for other targets / cases.
Differential Revision: http://reviews.llvm.org/D16828
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@260145
91177308-0d34-0410-b5e6-
96231b3b80d8
Justin Bogner [Mon, 8 Feb 2016 21:01:24 +0000 (21:01 +0000)]
cmake: Accept "thin" or "full" as arguments to -DLLVM_ENABLE_LTO
Mehdi suggested in a review of r259766 that it's also useful to easily
set the type of LTO. Augment the cmake variable to support that.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@260143
91177308-0d34-0410-b5e6-
96231b3b80d8
Xinliang David Li [Mon, 8 Feb 2016 20:08:21 +0000 (20:08 +0000)]
Fix build bot failure
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@260138
91177308-0d34-0410-b5e6-
96231b3b80d8
Tom Stellard [Mon, 8 Feb 2016 19:49:20 +0000 (19:49 +0000)]
AMDGPU/SI: Implement a work-around for smrd corrupting vccz bit
Summary:
We will hit this once we have enabled uniform branches. The
smrd-vccz-bug.ll test will be added with the uniform branch commit.
Reviewers: mareko, arsenm
Subscribers: arsenm, llvm-commits
Differential Revision: http://reviews.llvm.org/D16725
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@260137
91177308-0d34-0410-b5e6-
96231b3b80d8
Hans Wennborg [Mon, 8 Feb 2016 19:45:24 +0000 (19:45 +0000)]
Add triple to h-registers-3.ll to make bots happy after r260133
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@260136
91177308-0d34-0410-b5e6-
96231b3b80d8
Hans Wennborg [Mon, 8 Feb 2016 19:34:30 +0000 (19:34 +0000)]
[X86] Don't zero/sign-extend i1, i8, or i16 return values to 32 bits (PR22532)
This matches GCC and MSVC's behaviour, and saves on code size.
We were already not extending i1 return values on x86_64 after r127766. This
takes that patch further by applying it to x86 target as well, and also for i8
and i16.
The ABI docs have been unclear about the required behaviour here. The new i386
psABI [1] clearly states (Table 2.4, page 14) that i1, i8, and i16 return
vales do not need to be extended beyond 8 bits. The x86_64 ABI doc is being
updated to say the same [2].
Differential Revision: http://reviews.llvm.org/D16907
[1]. https://01.org/sites/default/files/file_attach/intel386-psabi-1.0.pdf
[2]. https://groups.google.com/d/msg/x86-64-abi/E8O33onbnGQ/_RFWw_ixDQAJ
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@260133
91177308-0d34-0410-b5e6-
96231b3b80d8
Tim Northover [Mon, 8 Feb 2016 19:33:18 +0000 (19:33 +0000)]
AArch64: match correct order in subtraction pattern.
The accumulator in multiply-and-subtract instructions is actually subtracted
*from* so these patterns were computing the wrong value.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@260131
91177308-0d34-0410-b5e6-
96231b3b80d8
Sanjay Patel [Mon, 8 Feb 2016 19:27:33 +0000 (19:27 +0000)]
fix typos; NFC
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@260130
91177308-0d34-0410-b5e6-
96231b3b80d8
Adrian Prantl [Mon, 8 Feb 2016 19:13:15 +0000 (19:13 +0000)]
Simplify this unittest.
Thanks to dblaikie for the suggestion!
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@260125
91177308-0d34-0410-b5e6-
96231b3b80d8
Matt Arsenault [Mon, 8 Feb 2016 19:06:01 +0000 (19:06 +0000)]
AMDGPU: Remove bfi and bfm intrinsics
Nothing is using them.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@260123
91177308-0d34-0410-b5e6-
96231b3b80d8
Teresa Johnson [Mon, 8 Feb 2016 18:47:20 +0000 (18:47 +0000)]
[ThinLTO] Remove imported available externally defs from comdats.
Summary:
Available externally definitions are considered declarations for the
linker and eventually dropped. As such they are not allowed to be
in comdats. Remove any such imported functions from comdats.
Reviewers: rafael
Subscribers: davidxl, llvm-commits, joker.eph
Differential Revision: http://reviews.llvm.org/D16120
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@260122
91177308-0d34-0410-b5e6-
96231b3b80d8
Xinliang David Li [Mon, 8 Feb 2016 18:13:49 +0000 (18:13 +0000)]
[PGO] Enable compression in pgo instrumentation
This reduces sizes of instrumented object files, final binaries,
process images, and raw profile data.
The format of the indexed profile data remain the same.
Differential Revision: http://reviews.llvm.org/D16388
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@260117
91177308-0d34-0410-b5e6-
96231b3b80d8
Silviu Baranga [Mon, 8 Feb 2016 17:02:45 +0000 (17:02 +0000)]
[SCEV][LAA] Re-commit r260085 and r260086, this time with a fix for the memory
sanitizer issue. The PredicatedScalarEvolution's copy constructor
wasn't copying the Generation value, and was leaving it un-initialized.
Original commit message:
[SCEV][LAA] Add no wrap SCEV predicates and use use them to improve strided pointer detection
Summary:
This change adds no wrap SCEV predicates with:
- support for runtime checking
- support for expression rewriting:
(sext ({x,+,y}) -> {sext(x),+,sext(y)}
(zext ({x,+,y}) -> {zext(x),+,sext(y)}
Note that we are sign extending the increment of the SCEV, even for
the zext case. This is needed to cover the fairly common case where y would
be a (small) negative integer. In order to do this, this change adds two new
flags: nusw and nssw that are applicable to AddRecExprs and permit the
transformations above.
We also change isStridedPtr in LAA to be able to make use of
these predicates. With this feature we should now always be able to
work around overflow issues in the dependence analysis.
Reviewers: mzolotukhin, sanjoy, anemet
Subscribers: mzolotukhin, sanjoy, llvm-commits, rengolin, jmolloy, hfinkel
Differential Revision: http://reviews.llvm.org/D15412
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@260112
91177308-0d34-0410-b5e6-
96231b3b80d8
Adrian Prantl [Mon, 8 Feb 2016 17:02:34 +0000 (17:02 +0000)]
Add a unit test for r259973.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@260111
91177308-0d34-0410-b5e6-
96231b3b80d8
Haicheng Wu [Mon, 8 Feb 2016 17:00:39 +0000 (17:00 +0000)]
[JumpThreading] Change a return of ComputeValueKnownInPredecessors()
Change a return statement of ComputeValueKnownInPredecessors() to be the same as
the rest return statements of the function. Otherwise, it might return true with
an empty Result when the current basic block has no predecessors and trigger the
first assert of JumpThreading::ProcessThreadableEdges().
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@260110
91177308-0d34-0410-b5e6-
96231b3b80d8
Matt Arsenault [Mon, 8 Feb 2016 16:28:19 +0000 (16:28 +0000)]
SelectionDAG: Lower some range metadata to AssertZext
If a range has a lower bound of 0, add an AssertZext from the
nearest floor power of two.
This allows operations with some workitem intrinsics with known
maximum ranges to use fast 24-bit multiplies.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@260109
91177308-0d34-0410-b5e6-
96231b3b80d8
Michael Zuckerman [Mon, 8 Feb 2016 15:13:32 +0000 (15:13 +0000)]
[AVX512][PROLQ][PROLD] Change imm8 to int
Differential Revision: http://reviews.llvm.org/D16983
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@260101
91177308-0d34-0410-b5e6-
96231b3b80d8
Igor Breger [Mon, 8 Feb 2016 14:11:39 +0000 (14:11 +0000)]
[SLP] Fix placement of debug statement (NFC)
By Ayal Zaks (ayal.zaks@intel.com)
Differential Revision: http://reviews.llvm.org/D16976
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@260094
91177308-0d34-0410-b5e6-
96231b3b80d8
Igor Breger [Mon, 8 Feb 2016 12:38:03 +0000 (12:38 +0000)]
AVX512: Change builtin function name for scalar intrinsics. Add "mask" to function name to reflect the function behavior.
Differential Revision: http://reviews.llvm.org/D16958
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@260089
91177308-0d34-0410-b5e6-
96231b3b80d8
Silviu Baranga [Mon, 8 Feb 2016 11:56:15 +0000 (11:56 +0000)]
Revert r260086 and r260085. They have broken the memory
sanitizer bots.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@260087
91177308-0d34-0410-b5e6-
96231b3b80d8
Silviu Baranga [Mon, 8 Feb 2016 11:15:29 +0000 (11:15 +0000)]
[LoopVersioning] Don't assert when there are no memchecks
We shouldn't assert when there are no memchecks, since we
can have SCEV checks. There is already an assert covering
the case where there are no SCEV checks or memchecks.
This also changes the LAA pointer wrapping versioning test
to use the loop versioning pass (this was how I managed to
trigger the assert in the loop versioning pass).
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@260086
91177308-0d34-0410-b5e6-
96231b3b80d8
Silviu Baranga [Mon, 8 Feb 2016 10:45:50 +0000 (10:45 +0000)]
[SCEV][LAA] Add no wrap SCEV predicates and use use them to improve strided pointer detection
Summary:
This change adds no wrap SCEV predicates with:
- support for runtime checking
- support for expression rewriting:
(sext ({x,+,y}) -> {sext(x),+,sext(y)}
(zext ({x,+,y}) -> {zext(x),+,sext(y)}
Note that we are sign extending the increment of the SCEV, even for
the zext case. This is needed to cover the fairly common case where y would
be a (small) negative integer. In order to do this, this change adds two new
flags: nusw and nssw that are applicable to AddRecExprs and permit the
transformations above.
We also change isStridedPtr in LAA to be able to make use of
these predicates. With this feature we should now always be able to
work around overflow issues in the dependence analysis.
Reviewers: mzolotukhin, sanjoy, anemet
Subscribers: mzolotukhin, sanjoy, llvm-commits, rengolin, jmolloy, hfinkel
Differential Revision: http://reviews.llvm.org/D15412
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@260085
91177308-0d34-0410-b5e6-
96231b3b80d8
Maxim Ostapenko [Mon, 8 Feb 2016 08:30:57 +0000 (08:30 +0000)]
[asan] Introduce new hidden -asan-use-private-alias option.
As discussed in https://github.com/google/sanitizers/issues/398, with current
implementation of poisoning globals we can have some CHECK failures or false
positives in case of mixing instrumented and non-instrumented code due to ASan
poisons innocent globals from non-sanitized binary/library. We can use private
aliases to avoid such errors. In addition, to preserve ODR violation detection,
we introduce new __odr_asan_gen_XXX symbol for each instrumented global that
indicates if this global was already registered. To detect ODR violation in
runtime, we should only check the value of indicator and report an error if it
isn't equal to zero.
Differential Revision: http://reviews.llvm.org/D15642
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@260075
91177308-0d34-0410-b5e6-
96231b3b80d8
Dan Gohman [Mon, 8 Feb 2016 03:42:36 +0000 (03:42 +0000)]
[WebAssembly] Add another optimization idea to README.txt.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@260070
91177308-0d34-0410-b5e6-
96231b3b80d8
Craig Topper [Mon, 8 Feb 2016 01:23:15 +0000 (01:23 +0000)]
[X86] Change FeatureIFMA string to 'avx512ifma'. Matches gcc and fixes PR26461.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@260069
91177308-0d34-0410-b5e6-
96231b3b80d8
Craig Topper [Mon, 8 Feb 2016 01:03:01 +0000 (01:03 +0000)]
[Support] Use hexdigit. NFC
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@260068
91177308-0d34-0410-b5e6-
96231b3b80d8
Craig Topper [Mon, 8 Feb 2016 01:02:55 +0000 (01:02 +0000)]
[Support] Fix the examples and assertion for format_hex_no_prefix to take into account that there are no prefix characters to include in Width.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@260067
91177308-0d34-0410-b5e6-
96231b3b80d8
NAKAMURA Takumi [Sun, 7 Feb 2016 23:03:38 +0000 (23:03 +0000)]
Disable llvm/test/tools/llvm-profdata/value-prof.proftext on win32 for now. Investigating.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@260064
91177308-0d34-0410-b5e6-
96231b3b80d8
Simon Pilgrim [Sun, 7 Feb 2016 22:51:06 +0000 (22:51 +0000)]
[X86][SSE] Resolve target shuffle inputs to sentinels to permit more combines
The combineX86ShufflesRecursively only supports unary shuffles, but was missing the opportunity to combine binary shuffles with a zero / undef second input.
This patch resolves target shuffle inputs, converting the shuffle mask elements to SM_SentinelUndef/SM_SentinelZero where possible. It then resolves the updated mask to check if we have created a faux unary shuffle.
Additionally, we now attempt to recursively call combineX86ShufflesRecursively for all input operands (we used to just recurse for unary integer shuffles and unary unpacks) - it safely returns early if its not a target shuffle.
Differential Revision: http://reviews.llvm.org/D16683
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@260063
91177308-0d34-0410-b5e6-
96231b3b80d8
Simon Pilgrim [Sun, 7 Feb 2016 22:22:09 +0000 (22:22 +0000)]
[X86][SSE] Regenerate PSHUFB shuffle mask comments tests
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@260061
91177308-0d34-0410-b5e6-
96231b3b80d8
Daniel Berlin [Sun, 7 Feb 2016 20:57:46 +0000 (20:57 +0000)]
Make check line consistent
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@260055
91177308-0d34-0410-b5e6-
96231b3b80d8