Chris Bieneman [Tue, 26 Jan 2016 21:29:08 +0000 (21:29 +0000)]
Remove autoconf support
Summary:
This patch is provided in preparation for removing autoconf on 1/26. The proposal to remove autoconf on 1/26 was discussed on the llvm-dev thread here: http://lists.llvm.org/pipermail/llvm-dev/2016-January/093875.html
"I felt a great disturbance in the [build system], as if millions of [makefiles] suddenly cried out in terror and were suddenly silenced. I fear something [amazing] has happened."
- Obi Wan Kenobi
Reviewers: chandlerc, grosbach, bob.wilson, tstellarAMD, echristo, whitequark
Subscribers: chfast, simoncook, emaste, jholewinski, tberghammer, jfb, danalbert, srhines, arsenm, dschuff, jyknight, dsanders, joker.eph, llvm-commits
Differential Revision: http://reviews.llvm.org/D16471
llvm-svn: 258861
Derek Schuff [Tue, 26 Jan 2016 21:08:27 +0000 (21:08 +0000)]
[WebAssembly] Remove check for FrameIndex operands in WebAssemblyPeephole
This pass runs after FrameIndex elimination, so it should never see FI
operands. NFC
llvm-svn: 258860
Evgeniy Stepanov [Tue, 26 Jan 2016 21:06:26 +0000 (21:06 +0000)]
[cfi] Hide runtime implementation in a namespace.
Move all internal stuff into namespace __cfi.
Remove the double underscore prefix from anything that's now inside
the namespace.
llvm-svn: 258859
Sanjay Patel [Tue, 26 Jan 2016 21:05:00 +0000 (21:05 +0000)]
[x86] add materializeVectorConstant() helper function; NFC
LowerBUILD_VECTOR is still over 300 lines long, but it's a start...
llvm-svn: 258858
Evgeniy Stepanov [Tue, 26 Jan 2016 20:53:09 +0000 (20:53 +0000)]
[cfi] Support for dlopen and dlclose.
Add dlopen/dlclose interceptors to update CFI shadow for loaded/unloaded libraries.
llvm-svn: 258857
Hemant Kulkarni [Tue, 26 Jan 2016 20:38:15 +0000 (20:38 +0000)]
Fix comparison warning (r258845)
llvm-svn: 258856
Eric Fiselier [Tue, 26 Jan 2016 20:31:01 +0000 (20:31 +0000)]
Remove dead code missed in r258852.
llvm-svn: 258855
Hemant Kulkarni [Tue, 26 Jan 2016 20:28:15 +0000 (20:28 +0000)]
Fixes build break introduced by r258845
llvm-svn: 258854
JF Bastien [Tue, 26 Jan 2016 20:24:51 +0000 (20:24 +0000)]
WebAssembly NFC: update error message
I forgot to update this one in my previous patch.
llvm-svn: 258853
Eric Fiselier [Tue, 26 Jan 2016 20:24:30 +0000 (20:24 +0000)]
Fix PR26103 - Error calling is_convertible with incomplete type. Patch from Michael Daniels.
llvm-svn: 258852
JF Bastien [Tue, 26 Jan 2016 20:22:42 +0000 (20:22 +0000)]
WebAssembly: don't optimize memcpy/memmove/memcpy to frame index
r258781 optimized memcpy/memmove/memcpy so the intrinsic call can return its first argument, but missed the frame index case. Teach it to ignore that case so C code doesn't assert out in these cases.
llvm-svn: 258851
Yunzhong Gao [Tue, 26 Jan 2016 20:15:02 +0000 (20:15 +0000)]
Do not define GXX_RTTI macro for C.
This is same as GCC behavior (tested with GCC 4.8.2).
Differential Revision: http://reviews.llvm.org/D16365
llvm-svn: 258850
Mike Aizatsky [Tue, 26 Jan 2016 20:10:01 +0000 (20:10 +0000)]
[sanitizers] extracted process management functions
Differential Revision: http://reviews.llvm.org/D16546
llvm-svn: 258849
Cong Hou [Tue, 26 Jan 2016 20:09:38 +0000 (20:09 +0000)]
Add a missing test case for r258847.
llvm-svn: 258848
Cong Hou [Tue, 26 Jan 2016 20:08:01 +0000 (20:08 +0000)]
Allow X86::COND_NE_OR_P and X86::COND_NP_OR_E to be reversed.
Currently, AnalyzeBranch() fails non-equality comparison between floating points
on X86 (see https://llvm.org/bugs/show_bug.cgi?id=23875). This is because this
function can modify the branch by reversing the conditional jump and removing
unconditional jump if there is a proper fall-through. However, in the case of
non-equality comparison between floating points, this can turn the branch
"unanalyzable". Consider the following case:
jne.BB1
jp.BB1
jmp.BB2
.BB1:
...
.BB2:
...
AnalyzeBranch() will reverse "jp .BB1" to "jnp .BB2" and then "jmp .BB2" will be
removed:
jne.BB1
jnp.BB2
.BB1:
...
.BB2:
...
However, AnalyzeBranch() cannot analyze this branch anymore as there are two
conditional jumps with different targets. This may disable some optimizations
like block-placement: in this case the fall-through behavior is enforced even if
the fall-through block is very cold, which is suboptimal.
Actually this optimization is also done in block-placement pass, which means we
can remove this optimization from AnalyzeBranch(). However, currently
X86::COND_NE_OR_P and X86::COND_NP_OR_E are not reversible: there is no defined
negation conditions for them.
In order to reverse them, this patch defines two new CondCode X86::COND_E_AND_NP
and X86::COND_P_AND_NE. It also defines how to synthesize instructions for them.
Here only the second conditional jump is reversed. This is valid as we only need
them to do this "unconditional jump removal" optimization.
Differential Revision: http://reviews.llvm.org/D11393
llvm-svn: 258847
Davide Italiano [Tue, 26 Jan 2016 19:57:42 +0000 (19:57 +0000)]
[llvm-nm] Roll several conditions into a single if. NFCI.
llvm-svn: 258846
Hemant Kulkarni [Tue, 26 Jan 2016 19:46:39 +0000 (19:46 +0000)]
[llvm-readobj] Add -elf-section-groups option
Adds a way to inspect SHT_GROUP sections in ELF objects.
Displays signature, member sections of these sections.
Differential revision: http://reviews.llvm.org/D16555
llvm-svn: 258845
Jonathan Peyton [Tue, 26 Jan 2016 19:44:31 +0000 (19:44 +0000)]
Bypass Perl modules in build system
This change fixes the bug: https://llvm.org/bugs/show_bug.cgi?id=25975
by bypassing the perl module files which try to deduce system information.
These perl modules files don't offer useful information and are from the
original build system. They can be removed after this change.
llvm-svn: 258843
Chad Rosier [Tue, 26 Jan 2016 19:33:57 +0000 (19:33 +0000)]
[ScheduleDAGInstrs] Simplify logic to improve readability. NFC.
The call to isInvariantLoad() already returns false for non-load instructions.
llvm-svn: 258841
David Majnemer [Tue, 26 Jan 2016 19:30:26 +0000 (19:30 +0000)]
[MS ABI] Allow a member pointers' converted type to change
Member pointers in the MS ABI are tricky for a variety of reasons.
The size of a member pointer is indeterminate until the program reaches
a point where the representation is required to be known. However,
*pointers* to member pointers may exist without knowing the pointee
type's representation. In these cases, we synthesize an opaque LLVM
type for the pointee type.
However, we can be in a situation where the underlying member pointer's
representation became known mid-way through the program. To account for
this, we attempted to manicure CodeGen's type-cache so that we can
replace the opaque member pointer type with the real deal while leaving
the pointer types unperturbed. This, unfortunately, is a problematic
approach to take as we will violate CodeGen's invariants.
These violations are mostly harmless but let's do the right thing
instead: invalidate the type-cache if a member pointer's LLVM
representation changes.
This fixes PR26313.
llvm-svn: 258839
Sanjay Patel [Tue, 26 Jan 2016 19:30:14 +0000 (19:30 +0000)]
tidy up; NFC
llvm-svn: 258838
Davide Italiano [Tue, 26 Jan 2016 19:28:51 +0000 (19:28 +0000)]
[llvm-nm] Simplify. No functional changes intended.
llvm-svn: 258837
Eugene Zelenko [Tue, 26 Jan 2016 19:01:06 +0000 (19:01 +0000)]
Fix Clang-tidy modernize-use-nullptr warnings; other minor fixes.
Differential revision: http://reviews.llvm.org/D16567
llvm-svn: 258836
Jonathan Coe [Tue, 26 Jan 2016 18:55:55 +0000 (18:55 +0000)]
Test commit. Fix typo in comment.
llvm-svn: 258835
Manman Ren [Tue, 26 Jan 2016 18:52:43 +0000 (18:52 +0000)]
Class Property: parse property attribute (class).
This is the third patch in a series of patches to support class properties
in addition to instance properties in objective-c.
rdar://
23891898
llvm-svn: 258834
Sanjay Patel [Tue, 26 Jan 2016 18:49:36 +0000 (18:49 +0000)]
[x86] simplify getOnesVector() ; NFCI
Let DAG.getConstant() handle the splatting; there's no need
to repeat that logic here.
llvm-svn: 258833
Arpith Chacko Jacob [Tue, 26 Jan 2016 18:48:41 +0000 (18:48 +0000)]
[OpenMP] Parsing + sema for target parallel directive.
Summary:
This patch adds parsing + sema for the target parallel directive and its clauses along with testcases.
Reviewers: ABataev
Differential Revision: http://reviews.llvm.org/D16553
Rebased to current trunk and updated test cases.
llvm-svn: 258832
Eugene Zelenko [Tue, 26 Jan 2016 18:48:36 +0000 (18:48 +0000)]
Fix Clang-tidy modernize-use-nullptr and modernize-use-override warnings; other minor fixes.
Differential revision: reviews.llvm.org/D16568
llvm-svn: 258831
Aditya Nandakumar [Tue, 26 Jan 2016 18:42:36 +0000 (18:42 +0000)]
Reassociate: Reprocess RedoInsts after each inst
Previously the RedoInsts was processed at the end of the block.
However it was possible that it left behind some instructions that
were not canonicalized.
This should guarantee that any previous instruction in the basic
block is canonicalized before we process a new instruction.
llvm-svn: 258830
Eugene Zelenko [Tue, 26 Jan 2016 18:27:37 +0000 (18:27 +0000)]
Fix Clang-tidy modernize-use-nullptr warnings in include/lld/Core/range.h; other minor fixes.
Differential revision: http://reviews.llvm.org/D16565
llvm-svn: 258829
Sanjay Patel [Tue, 26 Jan 2016 18:22:50 +0000 (18:22 +0000)]
[x86, AVX] tighten checks
llvm-svn: 258828
Benjamin Kramer [Tue, 26 Jan 2016 18:21:38 +0000 (18:21 +0000)]
Update wasm target for r258819.
llvm-svn: 258827
Kevin Enderby [Tue, 26 Jan 2016 18:20:49 +0000 (18:20 +0000)]
Update the comments for the macho-invalid-zero-ncmds test and fix
llvm-objdump when printing the Mach Header to print the unknown
cputype and cpusubtype fields as decimal instead of not printing
them at all. And change the test to check for that.
llvm-svn: 258826
Sanjay Patel [Tue, 26 Jan 2016 18:14:37 +0000 (18:14 +0000)]
fix formatting; NFC
llvm-svn: 258825
Manman Ren [Tue, 26 Jan 2016 18:05:23 +0000 (18:05 +0000)]
Use instance_properties instead of properties. NFC.
All current properties are instance properties.
This is the second patch in a series of patches to support class properties
in addition to instance properties in objective-c.
rdar://
23891898
llvm-svn: 258824
Justin Lebar [Tue, 26 Jan 2016 17:47:20 +0000 (17:47 +0000)]
[CUDA] Add -fcuda-allow-variadic-functions.
Summary:
Turns out the variadic function checking added in r258643 was too strict
for some existing users; give them an escape valve. When
-fcuda-allow-variadic-functions is passed, the front-end makes no
attempt to disallow C-style variadic functions. Calls to va_arg are
still not allowed.
Reviewers: tra
Subscribers: cfe-commits, jhen, echristo, bkramer
Differential Revision: http://reviews.llvm.org/D16559
llvm-svn: 258822
Saleem Abdulrasool [Tue, 26 Jan 2016 17:43:48 +0000 (17:43 +0000)]
Revert r258546.
Seems that the patch was rebased on top of another change which obsoleted the
change but wasnt caught.
Thanks to nbjoerg for pointing this out!
llvm-svn: 258821
Sanjay Patel [Tue, 26 Jan 2016 17:06:13 +0000 (17:06 +0000)]
don't repeat names in documentation comments; NFC
llvm-svn: 258820
Benjamin Kramer [Tue, 26 Jan 2016 16:45:00 +0000 (16:45 +0000)]
Update for LLVM change
llvm-svn: 258819
Benjamin Kramer [Tue, 26 Jan 2016 16:44:37 +0000 (16:44 +0000)]
Reflect the MC/MCDisassembler split on the include/ level.
No functional change, just moving code around.
llvm-svn: 258818
Arpith Chacko Jacob [Tue, 26 Jan 2016 16:37:23 +0000 (16:37 +0000)]
[OpenMP] Parsing + sema for defaultmap clause.
Summary:
This patch adds parsing + sema for the defaultmap clause associated with the target directive (among others).
Reviewers: ABataev
Differential Revision: http://reviews.llvm.org/D16527
llvm-svn: 258817
Sanjay Patel [Tue, 26 Jan 2016 16:17:24 +0000 (16:17 +0000)]
[LibCallSimplifier] fold memset(malloc(x), 0, x) --> calloc(1, x)
This is a step towards solving PR25892:
https://llvm.org/bugs/show_bug.cgi?id=25892
It won't handle the reported case. As noted by the 'TODO' comments in the patch,
we need to relax the hasOneUse() constraint and also match patterns that include
memset_chk() and the llvm.memset() intrinsic in addition to memset().
Differential Revision: http://reviews.llvm.org/D16337
llvm-svn: 258816
Chad Rosier [Tue, 26 Jan 2016 16:16:53 +0000 (16:16 +0000)]
Revert "[Driver] Make sure -fno-math-builtin option is being passed by the driver."
This reverts commit r258814.
llvm-svn: 258815
Chad Rosier [Tue, 26 Jan 2016 15:52:05 +0000 (15:52 +0000)]
[Driver] Make sure -fno-math-builtin option is being passed by the driver.
Support for the -fno-math-builtin option was added in r186899. The codegen side
is being tested in test/CodeGen/nomathbuiltin.c. The missing part was just
passing the option through the driver.
PR26317
llvm-svn: 258814
Chad Rosier [Tue, 26 Jan 2016 15:46:29 +0000 (15:46 +0000)]
[Driver] Update FIXME comment now that PR4941 has been addressed.
The actual fix should be addressed by someone who can test on Darwin.
llvm-svn: 258813
Matthew Simpson [Tue, 26 Jan 2016 15:45:49 +0000 (15:45 +0000)]
Revert "Reapply commit r258404 with fix"
This commit exposes a crash in computeKnownBits on the Chromium buildbots.
Reverting to investigate.
Reference: https://llvm.org/bugs/show_bug.cgi?id=26307
llvm-svn: 258812
Igor Laevsky [Tue, 26 Jan 2016 15:09:42 +0000 (15:09 +0000)]
Re-submit r256008 "Improve DWARFDebugFrame::parse to also handle __eh_frame."
Originally this change was causing failures on windows buildbots.
But those problems were fixed in r258806.
llvm-svn: 258811
Dan Gohman [Tue, 26 Jan 2016 14:55:17 +0000 (14:55 +0000)]
[WebAssembly] Fix a typo in a comment.
llvm-svn: 258810
Michael Kruse [Tue, 26 Jan 2016 13:33:27 +0000 (13:33 +0000)]
Unique phi write accesses
Ensure that there is at most one phi write access per PHINode and
ScopStmt. In particular, this would be possible for non-affine
subregions with multiple exiting blocks. We replace multiple MAY_WRITE
accesses by one MUST_WRITE access. The written value is constructed
using a PHINode of all exiting blocks. The interpretation of the PHI
WRITE's "accessed value" changed from the incoming value to the PHI like
for PHI READs since there is no unique incoming value.
Because region simplification shuffles around PHI nodes -- particularly
with exit node PHIs -- the PHINodes at analysis time does not always
exist anymore in the code generation pass. We instead remember the
incoming block/value pair in the MemoryAccess.
Differential Revision: http://reviews.llvm.org/D15681
llvm-svn: 258809
Michael Kruse [Tue, 26 Jan 2016 13:33:15 +0000 (13:33 +0000)]
Unique value read accesses
Keep at most one value read MemoryAccess per value and statement;
multiple generated loads do not have any additional effect. As one such
MemoryAccess can cater multiple uses within the statement, the
AccessInstruction property is not unique any more and set to nullptr.
Differential Revision: http://reviews.llvm.org/D15510
llvm-svn: 258808
Michael Kruse [Tue, 26 Jan 2016 13:33:10 +0000 (13:33 +0000)]
Unique value write accesses
Ensure there is at most one write access per definition of an
llvm::Value. Keep track of already created value write access by using
a (dense) map.
Replace addValueWriteAccess by ensureValueStore which can be uses more
liberally without worrying to add redundant accesses. It will be used,
e.g. in a logical correspondant for value reads -- ensureValueReload --
to ensure that the expected definition has been written when loading it.
Differential Revision: http://reviews.llvm.org/D15483
llvm-svn: 258807
Igor Laevsky [Tue, 26 Jan 2016 13:31:11 +0000 (13:31 +0000)]
[DebugInfo] Fix DWARFDebugFrame instruction operand ordering
We can't rely on the evalution order of function arguments.
Differential Revision: http://reviews.llvm.org/D16509
llvm-svn: 258806
Alexey Bataev [Tue, 26 Jan 2016 12:20:39 +0000 (12:20 +0000)]
[OPENMP 4.5] Allow arrays in 'reduction' clause.
OpenMP 4.5, alogn with array sections, allows to use variables of array type in reductions.
llvm-svn: 258804
Johannes Doerfert [Tue, 26 Jan 2016 11:03:25 +0000 (11:03 +0000)]
[FIX] Domain generation error due to loops in non-affine regions
llvm-svn: 258803
Johannes Doerfert [Tue, 26 Jan 2016 11:01:41 +0000 (11:01 +0000)]
[FIX] Build correct domain for non-affine region SCoPs
llvm-svn: 258802
Alexander Kornienko [Tue, 26 Jan 2016 10:56:27 +0000 (10:56 +0000)]
Fix crashing on user-defined conversion.
Summary: Fix the assertion failure for the user-defined conversion method. e.g.: operator bool()
Reviewers: alexfh, aaron.ballman
Subscribers: aaron.ballman, cfe-commits
Patch by Cong Liu!
Differential Revision: http://reviews.llvm.org/D16536
llvm-svn: 258801
Ewan Crawford [Tue, 26 Jan 2016 10:41:08 +0000 (10:41 +0000)]
[RenderScript] Provide option to specify a single allocation to print
Patch replaces the 'renderscript allocation list' command flag --refresh, with a new option --id <ID>.
This new option only prints the details of a single allocation with a given id, rather than printing all the allocations.
Functionality from the removed '--refresh' flag will be moved into its own command in a subsequent commit.
llvm-svn: 258800
Tobias Grosser [Tue, 26 Jan 2016 10:01:35 +0000 (10:01 +0000)]
BlockGenerators: Replace getNewScalarValue with getNewValue
Both functions implement the same functionality, with the difference that
getNewScalarValue assumes that globals and out-of-scop scalars can be directly
reused without loading them from their corresponding stack slot. This is correct
for sequential code generation, but causes issues with outlining code e.g. for
OpenMP code generation. getNewValue handles such cases correctly.
Hence, we can replace getNewScalarValue with getNewValue. This is not only more
future proof, but also eliminates a bunch of code.
The only functionality that was available in getNewScalarValue that is lost
is the on-demand creation of scalar values. However, this is not necessary any
more as scalars are always loaded at the beginning of each basic block and will
consequently always be available when scalar stores are generated. As this was
not the case in older versions of Polly, it seems the on-demand loading is just
some older code that has not yet been removed.
Finally, generateScalarLoads also generated loads for values that are loop
invariant, available in GlobalMap and which are preferred over the ones loaded
in generateScalarLoads. Hence, we can just skip the code generation of such
scalar values, avoiding the generation of dead code.
Differential Revision: http://reviews.llvm.org/D16522
llvm-svn: 258799
Simon Pilgrim [Tue, 26 Jan 2016 09:30:08 +0000 (09:30 +0000)]
[X86][SSE] Add zero element and general 64-bit VZEXT_LOAD support to EltsFromConsecutiveLoads
This patch adds support for trailing zero elements to VZEXT_LOAD loads (and checks that no zero elts occur within the consecutive load).
It also generalizes the 64-bit VZEXT_LOAD load matching to work for loads other than 2x32-bit loads.
After this patch it will also be easier to add support for other basic load patterns like 32-bit VZEXT_LOAD loads, PMOVZX and subvector load insertion.
Differential Revision: http://reviews.llvm.org/D16217
llvm-svn: 258798
Ismail Donmez [Tue, 26 Jan 2016 08:24:57 +0000 (08:24 +0000)]
Fix compilations with msvc's /Zc:strictStrings
llvm-svn: 258797
Rui Ueyama [Tue, 26 Jan 2016 07:17:29 +0000 (07:17 +0000)]
Simplify. NFC.
llvm-svn: 258796
Rui Ueyama [Tue, 26 Jan 2016 07:17:27 +0000 (07:17 +0000)]
Simplify. NFC.
llvm-svn: 258795
Matt Arsenault [Tue, 26 Jan 2016 06:37:54 +0000 (06:37 +0000)]
AMDGPU: Add amdgcn cube builtins
llvm-svn: 258794
Craig Topper [Tue, 26 Jan 2016 06:10:15 +0000 (06:10 +0000)]
[X86] Mark LDS/LES as not being allowed in 64-bit mode.
Their opcodes are used as part of the VEX prefix in 64-bit mode. Clearly the disassembler implicitly decoded them as AVX instructions in 64-bit mode, but I think the AsmParser would have encoded them.
llvm-svn: 258793
Rui Ueyama [Tue, 26 Jan 2016 04:58:58 +0000 (04:58 +0000)]
Simplify. NFC.
This new code should be logically equivalent to the previous code.
llvm-svn: 258792
Enrico Granata [Tue, 26 Jan 2016 04:53:10 +0000 (04:53 +0000)]
Reverting r258759 as it is breaking the OSX build
llvm-svn: 258791
Matt Arsenault [Tue, 26 Jan 2016 04:49:24 +0000 (04:49 +0000)]
AMDGPU: Move AMDGPU intrinsics only used by R600
llvm-svn: 258790
Matt Arsenault [Tue, 26 Jan 2016 04:49:22 +0000 (04:49 +0000)]
AMDGPU: Tidy minor td file issues
Make comments and indentation more consistent.
Rearrange a few things to be in a more consistent order,
such as organizing subtarget features from those describing
an actual device property, and those used as options.
llvm-svn: 258789
Matt Arsenault [Tue, 26 Jan 2016 04:43:48 +0000 (04:43 +0000)]
AMDGPU: Make v32i8/v64i8 illegal types
Old intrinsics were forcing these, but they have now all
been removed. This fixes large i8 vector operations generally
being broken.
llvm-svn: 258788
Matt Arsenault [Tue, 26 Jan 2016 04:38:08 +0000 (04:38 +0000)]
AMDGPU: Remove old sample intrinsics
I did my best to try to update all the uses in tests that
just happened to use the old ones to the newer intrinsics.
I'm not sure I got all of the immediate operand conversions
correct, since the value seems to have been ignored by the
old pattern but I don't think it really matters.
llvm-svn: 258787
Matt Arsenault [Tue, 26 Jan 2016 04:29:56 +0000 (04:29 +0000)]
AMDGPU: Add new amdgcn intrinsics for cube instructions
More cleanup to try to get all intrinsics using the correct
amdgcn prefix that are as close to the instruction as possible.
llvm-svn: 258786
Matt Arsenault [Tue, 26 Jan 2016 04:29:24 +0000 (04:29 +0000)]
AMDGPU: Implement read_register and write_register intrinsics
Some of the special intrinsics now that now correspond to a instruction
also have special setting of some registers, e.g. llvm.SI.sendmsg sets
m0 as well as use s_sendmsg. Using these explicit register intrinsics
may be a better option.
Reading the exec mask and others may be useful for debugging. For this
I'm not sure this is entirely correct because we would want this to
be convergent, although it's possible this is already treated
sufficently conservatively.
llvm-svn: 258785
Matt Arsenault [Tue, 26 Jan 2016 04:29:15 +0000 (04:29 +0000)]
AMDGPU: Note mesa version in release notes
llvm-svn: 258784
Matt Arsenault [Tue, 26 Jan 2016 04:14:16 +0000 (04:14 +0000)]
AMDGPU: Restore AMDGPU prefixed rsq intrinsic for now
Also move into backend intrinsics to discourage use of the old name.
llvm-svn: 258783
Xiuli Pan [Tue, 26 Jan 2016 04:03:48 +0000 (04:03 +0000)]
Recommit: R258773 [OpenCL] Pipe builtin functions
Fix arc patch fuzz error.
Summary:
Support for the pipe built-in functions for OpenCL 2.0.
The pipe builtin functions may have infinite kinds of element types, one approach
would be to just generate calls that would always use generic types such as void*.
This patch is based on bader's opencl support patch on SPIR-V branch.
Reviewers: Anastasia, pekka.jaaskelainen
Subscribers: keryell, bader, cfe-commits
Differential Revision: http://reviews.llvm.org/D15914
llvm-svn: 258782
Dan Gohman [Tue, 26 Jan 2016 04:01:11 +0000 (04:01 +0000)]
[WebAssembly] Optimize memcpy/memmove/memcpy calls.
These calls return their first argument, but because LLVM uses an intrinsic
with a void return type, they can't use the returned attribute. Generalize
the store results pass to optimize these calls too.
llvm-svn: 258781
Dan Gohman [Tue, 26 Jan 2016 03:43:48 +0000 (03:43 +0000)]
[WebAssembly] Remove a completed entry from the README.txt.
llvm-svn: 258780
Dan Gohman [Tue, 26 Jan 2016 03:39:31 +0000 (03:39 +0000)]
[WebAssembly] Implement unaligned loads and stores.
Differential Revision: http://reviews.llvm.org/D16534
llvm-svn: 258779
Richard Trieu [Tue, 26 Jan 2016 02:51:55 +0000 (02:51 +0000)]
Fix -Wnull-conversion for long macros.
Move the function to get a macro name from DiagnosticRenderer.cpp to Lexer.cpp
so that other files can use it. Lexer now has two functions to get the
immediate macro name, the newly added one is better for diagnostic purposes.
Make -Wnull-conversion use this function for better NULL macro detection.
llvm-svn: 258778
Haicheng Wu [Tue, 26 Jan 2016 02:27:47 +0000 (02:27 +0000)]
[LIR] Add support for structs and hand unrolled loops
This is a recommit of r258620 which causes PR26293.
The original message:
Now LIR can turn following codes into memset:
typedef struct foo {
int a;
int b;
} foo_t;
void bar(foo_t *f, unsigned n) {
for (unsigned i = 0; i < n; ++i) {
f[i].a = 0;
f[i].b = 0;
}
}
void test(foo_t *f, unsigned n) {
for (unsigned i = 0; i < n; i += 2) {
f[i] = 0;
f[i+1] = 0;
}
}
llvm-svn: 258777
Ehsan Akhgari [Tue, 26 Jan 2016 02:23:05 +0000 (02:23 +0000)]
Recommit the test for r258720 using -###
llvm-svn: 258776
David Majnemer [Tue, 26 Jan 2016 02:22:31 +0000 (02:22 +0000)]
Revert "[OpenCL] Pipe builtin functions"
This reverts commit r258773, it broke the build bots:
http://bb.pgr.jp/builders/cmake-clang-x86_64-linux/builds/43853
llvm-svn: 258775
Reid Kleckner [Tue, 26 Jan 2016 02:06:41 +0000 (02:06 +0000)]
Use binary search for intrinsic ID lookups
This improves compile time of Function.cpp from 57s to 37s for me
locally. Intrinsic IDs are cached on the Function object, so this
shouldn't regress performance.
llvm-svn: 258774
Xiuli Pan [Tue, 26 Jan 2016 02:06:04 +0000 (02:06 +0000)]
[OpenCL] Pipe builtin functions
Summary:
Support for the pipe built-in functions for OpenCL 2.0.
The pipe builtin functions may have infinite kinds of element types, one approach
would be to just generate calls that would always use generic types such as void*.
This patch is based on bader's opencl support patch on SPIR-V branch.
Reviewers: Anastasia, pekka.jaaskelainen
Subscribers: keryell, bader, cfe-commits
Differential Revision: http://reviews.llvm.org/D15914
llvm-svn: 258773
Ehsan Akhgari [Tue, 26 Jan 2016 01:51:47 +0000 (01:51 +0000)]
Revert the test for r258720 temporarily
This test is failing on a bot for reasons that are unclear to me.
Reverting for now...
llvm-svn: 258772
Matthias Braun [Tue, 26 Jan 2016 01:40:48 +0000 (01:40 +0000)]
LiveIntervalAnalysis: Improve some comments
As recommended by Justin.
llvm-svn: 258771
David Majnemer [Tue, 26 Jan 2016 01:39:17 +0000 (01:39 +0000)]
[Sema] Remove stray semicolons.
No functional change is intended.
llvm-svn: 258769
David Majnemer [Tue, 26 Jan 2016 01:37:01 +0000 (01:37 +0000)]
[Sema] Incomplete types are OK for covariant returns
Per C++14 [class.virtual]p8, it is OK for the return type's class type
to be incomplete so long as the return type is the same between the base
and complete classes.
This fixes PR26297.
llvm-svn: 258768
Rui Ueyama [Tue, 26 Jan 2016 01:32:00 +0000 (01:32 +0000)]
Add missing template instantiations.
llvm-svn: 258767
Rafael Espindola [Tue, 26 Jan 2016 01:30:07 +0000 (01:30 +0000)]
Fix MSVC build.
llvm-svn: 258766
Zachary Turner [Tue, 26 Jan 2016 01:19:50 +0000 (01:19 +0000)]
Fix TestRerun.py on Windows.
This is another example of a test that was looking for the thread
at index 0 instead of requesting the thread that was stopped at
the created breakpoint. This assumption isn't true on Windows 10.
llvm-svn: 258764
Sean Callanan [Tue, 26 Jan 2016 01:15:57 +0000 (01:15 +0000)]
Fix the lldbinline tests so they make well-formed Makefiles.
lldbinline tests previously did not run correctly unless there was already a
Makefile for them. This was because the syntax of the emitted Makefile made the
default make rule be the "cleanup" rule, which is pretty unhelpful. Now the
default rule is the one included from Makefile.rules, which is much better.
llvm-svn: 258763
David Majnemer [Tue, 26 Jan 2016 01:12:17 +0000 (01:12 +0000)]
[MS Compat] Don't crash if __GetExceptionInfo is in global scope
__GetExceptionInfo triggered Sema::LazilyCreateBuiltin which tries to
create a non-templated function decl. This is unnecessary and
ill-advised, there is no need for us to create a declaration for such a
builtin.
This fixes PR26298.
llvm-svn: 258762
Zachary Turner [Tue, 26 Jan 2016 01:09:38 +0000 (01:09 +0000)]
Remove XFAIL Windows from a test that was fixed by r258758.
llvm-svn: 258761
Rui Ueyama [Tue, 26 Jan 2016 01:03:21 +0000 (01:03 +0000)]
Move code to create RELATIVE reloc for TLS_IE to one place.
llvm-svn: 258760
Zachary Turner [Tue, 26 Jan 2016 00:59:42 +0000 (00:59 +0000)]
Write the session log file in UTF-8.
Previously we were writing in the default encoding, which depends
on the operating system and is not guaranteed to be unicode aware.
On Python 3, this would lead to a situation where writing unicode
text to the log file generates an exception. The fix here is to
write session logs using the proper encoding, which incidentally
fixes another test, so xfail is removed from that.
llvm-svn: 258759
Adrian McCarthy [Tue, 26 Jan 2016 00:58:09 +0000 (00:58 +0000)]
Set symbol types for function symbols loaded from PE/COFF
This fixes the regression of several tests on Windows after rL258621.
The root problem is that ObjectFilePECOFF was not setting type information for the symbols, and the new CL rejects symbols without type information, breaking functionality like thread step-over.
The fix sets the type information for functions (and creates a TODO for other types).
Along the way, I fixed some typos and formatting that made the code I was debugging harder to understand.
In the long run, we should consider replacing most of ObjectFilePECOFF with the COFF parsing code from LLVM.
Differential Revision: http://reviews.llvm.org/D16563
llvm-svn: 258758
Reid Kleckner [Tue, 26 Jan 2016 00:55:00 +0000 (00:55 +0000)]
Sort intrinsics by LLVM intrinsic name, rather than tablegen def name
Step one towards using a simple binary search to lookup intrinsic IDs
instead of our crazy table generated switch+memcmp+startswith code that
makes Function.cpp take about a minute to compile. See PR24785 and
PR11951 for why we should do this.
The X86 backend contains tables that need to be sorted on intrinsic ID,
so reorder those.
llvm-svn: 258757
Matthias Braun [Tue, 26 Jan 2016 00:43:50 +0000 (00:43 +0000)]
LiveIntervalAnalysis: Cleanup handleMove{Down|Up}() functions, NFC
These two functions are hard to reason about. This commit makes the code
more comprehensible:
- Use four distinct variables (OldIdxIn, OldIdxOut, NewIdxIn, NewIdxOut)
with a fixed value instead of a changing iterator I that points to
different things during the function.
- Remove the early explanation before the function in favor of more
detailed comments inside the function. Should have more/clearer comments now
stating which conditions are tested and which invariants hold at
different points in the functions.
The behaviour of the code was not changed.
I hope that this will make it easier to review the changes in
http://reviews.llvm.org/D9067 which I will adapt next.
Differential Revision: http://reviews.llvm.org/D16379
llvm-svn: 258756
Dan Gohman [Tue, 26 Jan 2016 00:35:07 +0000 (00:35 +0000)]
Followup to 258750; update more tests to use .p2align .
llvm-svn: 258755