Timur Iskhodzhanov [Thu, 25 Dec 2014 13:03:58 +0000 (13:03 +0000)]
Simplify allocator_returns_null.cc test
llvm-svn: 224837
Yaron Keren [Thu, 25 Dec 2014 12:21:56 +0000 (12:21 +0000)]
Initialize BackendConsumer::Context in constructor.
llvm-svn: 224836
Yaron Keren [Thu, 25 Dec 2014 11:38:15 +0000 (11:38 +0000)]
Initialize CodeGeneratorImpl::Ctx in constructor.
llvm-svn: 224835
Dmitry Vyukov [Thu, 25 Dec 2014 10:32:25 +0000 (10:32 +0000)]
tsan: fix trace initialization during thread id reuse
The current code leaves the first event in the trace part uninitialized
(from the previous thread). It can cause unpredictable behavior
during stack/mutexset restoration.
Initialize the first event to a fake harmless memory access.
llvm-svn: 224834
David Majnemer [Thu, 25 Dec 2014 10:03:23 +0000 (10:03 +0000)]
Silence GCC's -Wparentheses warning
No functionality change intended.
llvm-svn: 224833
Elena Demikhovsky [Thu, 25 Dec 2014 09:29:13 +0000 (09:29 +0000)]
Documentation for Masked Load and Store intrinsics.
llvm-svn: 224832
Simon Atanasyan [Thu, 25 Dec 2014 09:23:47 +0000 (09:23 +0000)]
[Mips] Replace stderr output by the `llvm_unreachable` call
If a regular symbol has microMIPS-bit we need to stop linking. Now the
LLD does not check the `applyRelocation` return value and continues
linking anyway. As a temporary workaround use the `llvm_unreachable`
call to stop the linker.
llvm-svn: 224831
Simon Atanasyan [Thu, 25 Dec 2014 09:23:37 +0000 (09:23 +0000)]
[Mips] Make the test more tolerant to the linker output order
The LLD output in the YAML mode depends on LLD_RUN_ROUNDTRIP_TEST
environment variable. Do not check unimportant YAML items like section-name.
llvm-svn: 224830
Elena Demikhovsky [Thu, 25 Dec 2014 07:49:20 +0000 (07:49 +0000)]
Masked Load/Store - Changed the order of parameters in intrinsics.
No functional changes.
The documentation is coming.
llvm-svn: 224829
David Majnemer [Wed, 24 Dec 2014 23:06:55 +0000 (23:06 +0000)]
CodeGen: Error on redefinitions instead of asserting
It's possible to have a prior definition of a symbol in module asm.
Raise an error instead of crashing.
llvm-svn: 224828
David Majnemer [Wed, 24 Dec 2014 22:44:29 +0000 (22:44 +0000)]
CodeGen: Allow aliases to be overridden by variables
llvm-svn: 224827
Simon Atanasyan [Wed, 24 Dec 2014 21:04:05 +0000 (21:04 +0000)]
[Mips] Support linking of microMIPS 32-bit code
The change is rather large but mainly it just adds handling of new relocations,
PLT entries etc.
llvm-svn: 224826
Saleem Abdulrasool [Wed, 24 Dec 2014 18:40:42 +0000 (18:40 +0000)]
MC: address some comments in deprecation checks
Bob Wilson pointed out the unnecessary checks that had been committed to the
instruction check predicates. The check was meant to ensure that the check was
not accidentally applied to non-ARM instructions. This is better served as an
assertion rather than a condition check.
llvm-svn: 224825
Evgeniy Stepanov [Wed, 24 Dec 2014 16:58:50 +0000 (16:58 +0000)]
[asan] Support ASAN_ACTIVATION_OPTIONS.
This is mostly useful for testing, as the only other way of specifying
activation options (Android system property) is system-wide and affects
concurrently running tests.
llvm-svn: 224824
Timur Iskhodzhanov [Wed, 24 Dec 2014 16:14:16 +0000 (16:14 +0000)]
[ASan/Win] Bandaid fix for PR22025 -- deadlocks when creating suspended threads
llvm-svn: 224823
Evgeniy Stepanov [Wed, 24 Dec 2014 13:59:35 +0000 (13:59 +0000)]
[sanitizer] Disable InternalMmapWithOffset test on OSX.
llvm-svn: 224822
Evgeniy Stepanov [Wed, 24 Dec 2014 13:57:11 +0000 (13:57 +0000)]
[sanitizer] Fix off-by-8x in direct coverage.
File mapping offset was calculated by offsetting (uptr *) instead of (char *).
llvm-svn: 224821
Evgeniy Stepanov [Wed, 24 Dec 2014 13:05:21 +0000 (13:05 +0000)]
[sanitizer] Disable InternalMmapWithOffset test on Windows.
llvm-svn: 224820
Evgeniy Stepanov [Wed, 24 Dec 2014 12:58:09 +0000 (12:58 +0000)]
[sanitizer] mmap2 syscall works with 4096-byte units instead of bytes.
Ouch.
llvm-svn: 224819
Evgeniy Stepanov [Wed, 24 Dec 2014 12:23:15 +0000 (12:23 +0000)]
[asan] Disable one test on Android.
It used to pass by chance and started failing on recent Android builds.
llvm-svn: 224818
Simon Atanasyan [Wed, 24 Dec 2014 12:19:42 +0000 (12:19 +0000)]
[Mips] Join two if statements
No functional changes.
llvm-svn: 224817
Simon Atanasyan [Wed, 24 Dec 2014 12:19:36 +0000 (12:19 +0000)]
[Mips] Make the comment more descriptive
No functional changes.
llvm-svn: 224816
Simon Atanasyan [Wed, 24 Dec 2014 12:19:30 +0000 (12:19 +0000)]
[Mips] Factor out the code checks a symbol's binding
No functional changes.
llvm-svn: 224815
Simon Atanasyan [Wed, 24 Dec 2014 12:19:24 +0000 (12:19 +0000)]
[Mips] Rename the function s/readAddend/getAddend/
No functional changes.
llvm-svn: 224814
Simon Atanasyan [Wed, 24 Dec 2014 12:19:18 +0000 (12:19 +0000)]
[Mips] Use OR operation to set the microMIPS bit
llvm-svn: 224813
Jay Foad [Wed, 24 Dec 2014 11:48:40 +0000 (11:48 +0000)]
[ASan] Fix asan_preload_test-2 on PowerPC64 Linux
Summary:
This test failed because clang compiled the call to memset() into a
single sth instruction, instead of a call. Fix it by using write() instead
of memset().
Reviewers: kcc, samsonov, garious, eugenis
Reviewed By: eugenis
Subscribers: llvm-commits
Differential Revision: http://reviews.llvm.org/D6776
llvm-svn: 224812
David Majnemer [Wed, 24 Dec 2014 10:27:50 +0000 (10:27 +0000)]
MC: Label definitions are permitted after .set directives
.set directives may be overridden by other .set directives as well as
label definitions.
This fixes PR22019.
llvm-svn: 224811
Saleem Abdulrasool [Wed, 24 Dec 2014 06:32:43 +0000 (06:32 +0000)]
IAS: correct debug line info for asm macros
Correct the line information generation for preprocessed assembly. Although we
tracked the source information for the macro instantiation, we failed to account
for the fact that we were instantiating a macro, which is populated into a new
buffer and that the line information would be relative to the definition rather
than the actual instantiation location. This could cause the line number
associated with the statement to be very high due to wrapping of the difference
calculated for the preprocessor line information emitted into the stream.
Properly calculate the line for the macro instantiation, referencing the line
where the macro is actually used as GCC/gas do.
The test case uses x86, though the same problem exists on any other target using
the LLVM IAS.
llvm-svn: 224810
Craig Topper [Wed, 24 Dec 2014 06:05:22 +0000 (06:05 +0000)]
[X86] Remove the single AdSize indicator and replace it with separate AdSize16/32/64 flags.
This removes a hardcoded list of instructions in the CodeEmitter. Eventually I intend to remove the predicates on the affected instructions since in any given mode two of them are valid if we supported addr32/addr16 prefixes in the assembler.
llvm-svn: 224809
David Majnemer [Wed, 24 Dec 2014 04:11:42 +0000 (04:11 +0000)]
MC: Don't emit .no_dead_strip on targets which don't support it
llvm-svn: 224808
Matthias Braun [Wed, 24 Dec 2014 02:35:07 +0000 (02:35 +0000)]
LiveInterval: Remove accidentally committed debug code.
llvm-svn: 224807
Matthias Braun [Wed, 24 Dec 2014 02:11:51 +0000 (02:11 +0000)]
LiveInterval: Introduce createMainRangeFromSubranges().
This function constructs the main liverange by merging all subranges if
subregister liveness tracking is available. This should be slightly
faster to compute instead of performing the liveness calculation again
for the main range. More importantly it avoids cases where the main
liverange would cover positions where no subrange was live. These cases
happened for partial definitions where the actual defined part was dead
and only the undefined parts used later.
The register coalescing requires that every part covered by the main
live range has at least one subrange live.
I also expect this function to become usefull later for places where the
subranges are modified in a way that it is hard to correctly fix the
main liverange in the machine scheduler, we can simply reconstruct it
from subranges then.
llvm-svn: 224806
Matthias Braun [Wed, 24 Dec 2014 02:11:48 +0000 (02:11 +0000)]
RegisterCoalescer: With subrange liveness there may be no RedefVNI for unused lanes.
llvm-svn: 224805
Matthias Braun [Wed, 24 Dec 2014 02:11:46 +0000 (02:11 +0000)]
LiveRangeEdit: Check for completely empy subranges after removing ValNos.
Completely empty subranges are not allowed and must be removed when
subreg liveness is enabled.
llvm-svn: 224804
Matthias Braun [Wed, 24 Dec 2014 02:11:43 +0000 (02:11 +0000)]
LiveIntervalAnalysis: Fix performance bug that I introduced in r224663.
Without a reference the code did not remember when moving the iterators
of the subranges/registerunit ranges forward and instead would scan from
the beginning again at the next position.
llvm-svn: 224803
Peter Zotov [Wed, 24 Dec 2014 01:58:45 +0000 (01:58 +0000)]
[OCaml] PR21901: Update tests.
This finishes the fix partially applied by r224782.
llvm-svn: 224802
Peter Zotov [Wed, 24 Dec 2014 01:52:51 +0000 (01:52 +0000)]
[OCaml] Expose Llvm_executionengine.get_{global_value,function}_address.
Patch by Ramkumar Ramachandra <artagnon@gmail.com>.
Also remove Llvm_executionengine.get_pointer_to_global, as it
is actually deprecated and didn't appear in a stable release.
llvm-svn: 224801
Chandler Carruth [Wed, 24 Dec 2014 01:48:09 +0000 (01:48 +0000)]
[SROA] Update the documentation and names for accessing the slices
within a partition of an alloca in SROA.
This reflects the fact that the organization of the slices isn't really
ideal for analysis, but is the naive way in which the slices are
available while we're processing them in the core partitioning
algorithm.
It is possible we could improve matters, and I've left a FIXME with
one of my ideas for how to do this, but it is a lot of work, the benefit
is somewhat minor, and it isn't clear that it would be strictly better.
=/ Not really satisfying, but I'm out of really good ideas.
This also improves one place where the debug logging failed to mark some
split partitions. Now we log in one place, slightly later, and with
accurate information about whether the slice is split by the partition
being rewritten.
llvm-svn: 224800
Adrian Prantl [Wed, 24 Dec 2014 01:17:51 +0000 (01:17 +0000)]
Debug Info: In symmetry to DW_TAG_pointer_type, do not emit the byte size
of a DW_TAG_ptr_to_member_type.
This restores the behavior from before r224780-r224781.
llvm-svn: 224799
Chandler Carruth [Wed, 24 Dec 2014 01:05:14 +0000 (01:05 +0000)]
[SROA] Refactor the integer and vector promotion testing logic to
operate in terms of the new Partition class, and generally have a more
clear set of arguments. No functionality changed.
The most notable improvements here are consistently using the
terminology of 'partition' for a collection of slices that will be
rewritten together and 'slice' for a region of an alloca that is used by
a particular instruction.
This also makes it more clear that the split things are actually slices
as well, just ones that will be split by the proposed partition.
This doesn't yet address the confusing aspects of the partition's
interface where slices that will be split by the partition and start
prior to the partition are accesssed via Partition::splitSlices() while
the core range of slices exposed by a Partition includes both unsplit
slices and slices which will be split by the end, but started within the
offset range of the partition. This is particularly hard to address
because the algorithm which computes partitions quite literally doesn't
know which slices these will end up being until too late. I'm looking at
whether I can fix that or not, but I'm not optimistic. I'll update the
comments and/or names to further explain this either way. I've also
added one FIXME in this patch relating to this confusion so that I don't
forget about it.
llvm-svn: 224798
Anton Yartsev [Wed, 24 Dec 2014 01:04:38 +0000 (01:04 +0000)]
[analyzer] Updated progress for memory.ZeroAlloc
llvm-svn: 224797
Anton Yartsev [Wed, 24 Dec 2014 00:46:56 +0000 (00:46 +0000)]
[analyzer] +memory.ZeroAlloc
llvm-svn: 224796
Colin LeMahieu [Wed, 24 Dec 2014 00:43:00 +0000 (00:43 +0000)]
[Hexagon] Removing old classes.
llvm-svn: 224795
Kevin Enderby [Wed, 24 Dec 2014 00:16:51 +0000 (00:16 +0000)]
Another attempt to fix the LLVM Windows build bot lld-x86_64-win7, one last place to fix I think.
llvm-svn: 224794
Kevin Enderby [Tue, 23 Dec 2014 23:43:59 +0000 (23:43 +0000)]
Attempt to fix the LLVM Windows build bot lld-x86_64-win7.
llvm-svn: 224793
Kevin Enderby [Tue, 23 Dec 2014 22:56:39 +0000 (22:56 +0000)]
Add printing the LC_THREAD load commands with llvm-objdump’s -private-headers.
llvm-svn: 224792
Nico Weber [Tue, 23 Dec 2014 22:55:34 +0000 (22:55 +0000)]
clang-cl: Various changes to /Zc: handling.
* /Zc:trigraphs and /Zc:trigraphs- are now honored
* /Zc:strictStrings is now honored
* /Zc:auto is now honored/ignored (clang does the Right Thing for this already)
Also add a dedicated test for the various /Zc: flags.
clang-cl doesn't always agree with cl.exe on the default values for /Zc flags.
For example, I think clang always behaves as if /Zc:inline is passed, and
warns if the user explicitly passes /Zc:inline-
Fixes PR21974.
llvm-svn: 224791
Nico Weber [Tue, 23 Dec 2014 22:32:37 +0000 (22:32 +0000)]
Add driver flags -ftrigraphs, -fno-trigraphs.
-trigraphs is now an alias for -ftrigraphs. -fno-trigraphs makes it possible
to explicitly disable trigraphs, which couldn't be done before.
clang -std=c++11 -fno-trigraphs
now builds without GNU extensions, but with trigraphs disabled. Previously,
trigraphs were only disabled in GNU modes or with -std=c++1z.
Make the new -f flags the cc1 interface too. This requires changing -trigraphs
to -ftrigraphs in a few cc1 tests.
Related to PR21974.
llvm-svn: 224790
Kostya Serebryany [Tue, 23 Dec 2014 22:32:17 +0000 (22:32 +0000)]
[asan] change the coverage collection scheme so that we can easily emit coverage for the entire process as a single bit set, and if coverage_bitset=1 actually emit that bitset
llvm-svn: 224789
Hal Finkel [Tue, 23 Dec 2014 22:29:40 +0000 (22:29 +0000)]
[PowerPC] Ensure that the TOC reload directly follows bctrl on PPC64
On non-Darwin PPC64, the TOC reload needs to come directly after the bctrl
instruction (for indirect calls) because the 'bctrl/ld 2, 40(1)' instruction
sequence is interpreted by the unwinding code in libgcc. To make sure these
occur as a pair, as with other pairings interpreted by the linker, fuse the two
instructions into one instruction (for code generation only).
In the future, we might wish to do this by emitting CFI directives instead,
but this solution is simpler, and mirrors what GCC does. Additional discussion
on this point is contained in the PR.
Fixes PR22015.
llvm-svn: 224788
Colin LeMahieu [Tue, 23 Dec 2014 20:44:59 +0000 (20:44 +0000)]
[Hexagon] Adding doubleword load.
llvm-svn: 224787
Colin LeMahieu [Tue, 23 Dec 2014 20:02:16 +0000 (20:02 +0000)]
[Hexagon] Reapplying 224775 load words.
llvm-svn: 224786
Jozef Kolek [Tue, 23 Dec 2014 19:55:34 +0000 (19:55 +0000)]
[mips][microMIPS] Implement CACHE, PREF, SSNOP, EHB and PAUSE instructions
Differential Revision: http://reviews.llvm.org/D5204
llvm-svn: 224785
Nico Weber [Tue, 23 Dec 2014 19:30:13 +0000 (19:30 +0000)]
Remove unused test input.
llvm-svn: 224784
Colin LeMahieu [Tue, 23 Dec 2014 19:22:59 +0000 (19:22 +0000)]
Reverting 224775 until mayLoad flag is addressed.
llvm-svn: 224783
Rafael Espindola [Tue, 23 Dec 2014 19:16:45 +0000 (19:16 +0000)]
Finish removing DestroySource.
Fixes pr21901.
llvm-svn: 224782
Adrian Prantl [Tue, 23 Dec 2014 19:11:54 +0000 (19:11 +0000)]
Debug Info: Pass the pointer size into createMemberPointerType().
Paired commit with LLVM.
llvm-svn: 224781
Adrian Prantl [Tue, 23 Dec 2014 19:11:47 +0000 (19:11 +0000)]
DIBuilder: Similar to createPointerType, make createMemberPointerType take
a size and alignment. Several assertions in DwarfDebug rely on all variable
types to report back a size, or to be derived from a type with a size.
Tested in CFE.
llvm-svn: 224780
Mehdi Amini [Tue, 23 Dec 2014 18:59:02 +0000 (18:59 +0000)]
Always assert in DAGCombine and not only when -debug is enabled
Right now in DAG Combine check the validity of the returned type
only when -debug is given on the command line. However usually
the test cases in the validation does not use -debug.
An Assert build should always check this.
llvm-svn: 224779
Nico Weber [Tue, 23 Dec 2014 18:52:41 +0000 (18:52 +0000)]
Tweak lld's checkout instructions.
llvm-svn: 224778
Rafael Espindola [Tue, 23 Dec 2014 18:39:02 +0000 (18:39 +0000)]
Pass LSAN_OPTIONS down so that it is possible to add suppressions.
llvm-svn: 224777
Rafael Espindola [Tue, 23 Dec 2014 18:18:37 +0000 (18:18 +0000)]
Fix a leak found by asan.
llvm-svn: 224776
Colin LeMahieu [Tue, 23 Dec 2014 18:06:56 +0000 (18:06 +0000)]
[Hexagon] Adding word loads.
llvm-svn: 224775
Colin LeMahieu [Tue, 23 Dec 2014 17:25:57 +0000 (17:25 +0000)]
[Hexagon] Adding signed halfword loads.
llvm-svn: 224774
Rafael Espindola [Tue, 23 Dec 2014 17:20:23 +0000 (17:20 +0000)]
Fix a leak found by asan.
llvm-svn: 224773
Colin LeMahieu [Tue, 23 Dec 2014 16:42:57 +0000 (16:42 +0000)]
[Hexagon] Adding unsigned halfword load.
llvm-svn: 224772
Jozef Kolek [Tue, 23 Dec 2014 16:16:33 +0000 (16:16 +0000)]
[mips][microMIPS] Implement LWSP and SWSP instructions
Differential Revision: http://reviews.llvm.org/D6416
llvm-svn: 224771
Rafael Espindola [Tue, 23 Dec 2014 15:57:12 +0000 (15:57 +0000)]
This reverts commit r224668 and r224667.
r224667 broke bootstrap on Fedora 20 X86_64 (at least).
See pr22006 for the details.
r224668 depends on r224667.
llvm-svn: 224770
Peter Zotov [Tue, 23 Dec 2014 13:09:59 +0000 (13:09 +0000)]
[OCaml] PR22014: OCaml bindings didn't link to libLLVM-*.so with -Wl,--as-needed
Patch by Evangelos Foutras <evangelos@foutrelis.com>.
llvm-svn: 224766
Michael Kuperstein [Tue, 23 Dec 2014 11:33:41 +0000 (11:33 +0000)]
[ValueTracking] Move GlobalAlias handling to be after the max depth check in computeKnownBits()
GlobalAlias handling used to be after GlobalValue handling, which meant it was, in practice, dead code. r220165 moved GlobalAlias handling to be before GlobalValue handling, but also moved it to be before the max depth check, causing an assert due to a recursion depth limit violation.
This moves GlobalAlias handling forward to where it's safe, and changes the GlobalValue handling to only look at GlobalObjects.
Differential Revision: http://reviews.llvm.org/D6758
llvm-svn: 224765
Elena Demikhovsky [Tue, 23 Dec 2014 10:30:39 +0000 (10:30 +0000)]
AVX-512: Added FMA instructions, intrinsics an tests for KNL and SKX targets
by Asaf Badouh
http://reviews.llvm.org/D6456
llvm-svn: 224764
Evgeniy Stepanov [Tue, 23 Dec 2014 10:18:47 +0000 (10:18 +0000)]
[asan] Fix line >80 chars.
llvm-svn: 224763
Evgeniy Stepanov [Tue, 23 Dec 2014 10:12:51 +0000 (10:12 +0000)]
[asan] Better error message in coverage.
llvm-svn: 224762
Hal Finkel [Tue, 23 Dec 2014 09:45:06 +0000 (09:45 +0000)]
[PowerPC] Don't mark the return-address slot as immutable
It is tempting to mark the fixed stack slot used to store the return address as
immutable when lowering @llvm.returnaddress(i32 0). Unfortunately, within the
function, it is not completely immutable: it is written during the function
prologue. When using post-RA instruction scheduling, the prologue instructions
are available for scheduling, and we're not free to interchange the order of a
particular store in the prologue with loads from that stack location.
Fixes PR21976.
llvm-svn: 224761
Elena Demikhovsky [Tue, 23 Dec 2014 09:36:28 +0000 (09:36 +0000)]
AVX-512: BLENDM - fixed encoding of the broadcast version
Added more intrinsics and encoding tests.
llvm-svn: 224760
Michael Kuperstein [Tue, 23 Dec 2014 08:59:45 +0000 (08:59 +0000)]
[DagCombine] Improve DAGCombiner BUILD_VECTOR when it has two sources of elements
This partially fixes PR21943.
For AVX, we go from:
vmovq (%rsi), %xmm0
vmovq (%rdi), %xmm1
vpermilps $-27, %xmm1, %xmm2 ## xmm2 = xmm1[1,1,2,3]
vinsertps $16, %xmm2, %xmm1, %xmm1 ## xmm1 = xmm1[0],xmm2[0],xmm1[2,3]
vinsertps $32, %xmm0, %xmm1, %xmm1 ## xmm1 = xmm1[0,1],xmm0[0],xmm1[3]
vpermilps $-27, %xmm0, %xmm0 ## xmm0 = xmm0[1,1,2,3]
vinsertps $48, %xmm0, %xmm1, %xmm0 ## xmm0 = xmm1[0,1,2],xmm0[0]
To the expected:
vmovq (%rdi), %xmm0
vmovhpd (%rsi), %xmm0, %xmm0
retq
Fixing this for AVX2 is still open.
Differential Revision: http://reviews.llvm.org/D6749
llvm-svn: 224759
Hal Finkel [Tue, 23 Dec 2014 08:38:50 +0000 (08:38 +0000)]
[PowerPC] Don't attempt a 64-bit pow2 division on PPC32
In r224033, in moving the signed power-of-2 division expansion into
BuildSDIVPow2, I accidentally made it possible to attempt the lowering for a
64-bit division on PPC32. This later asserts.
Fixes PR21928.
llvm-svn: 224758
Michael Liao [Tue, 23 Dec 2014 08:26:55 +0000 (08:26 +0000)]
[SimplifyCFG] Revise common code sinking
- Fix the case where more than 1 common instructions derived from the same
operand cannot be sunk. When a pair of value has more than 1 derived values
in both branches, only 1 derived value could be sunk.
- Replace BB1 -> (BB2, PN) map with joint value map, i.e.
map of (BB1, BB2) -> PN, which is more accurate to track common ops.
llvm-svn: 224757
Michael Kuperstein [Tue, 23 Dec 2014 08:23:45 +0000 (08:23 +0000)]
Remove a bad cast in CloneModule()
A cast that was introduced in r209007 was accidentally left in after the changes made to GlobalAlias rules in r210062. This crashes if the aliasee is a now-leggal ConstantExpr.
llvm-svn: 224756
Dmitry Vyukov [Tue, 23 Dec 2014 07:21:23 +0000 (07:21 +0000)]
tsan: fix style
Usually we roll the variable declaration into the condition in cases like this.
llvm-svn: 224755
Ahmed Bougacha [Tue, 23 Dec 2014 06:07:31 +0000 (06:07 +0000)]
[ARM] Don't break alignment when combining base updates into load/stores.
r223862/r224203 tried to also combine base-updating load/stores.
There was a mistake there: the alignment was added as is as an operand to
the ARMISD::VLD/VST node. However, the VLD/VST selection logic doesn't care
about less-than-standard alignment attributes.
For example, no matter the alignment of a v2i64 load (say 1), SelectVLD picks
VLD1q64 (because of the memory type). But VLD1q64 ("vld1.64 {dXX, dYY}") is
8-aligned, per ARMARMv7a 3.2.1.
For the 1-aligned load, what we really want is VLD1q8.
This commit introduces bitcasts if necessary, and changes the vld/vst type to
one whose standard alignment matches the original load/store alignment.
Differential Revision: http://reviews.llvm.org/D6759
llvm-svn: 224754
Eric Fiselier [Tue, 23 Dec 2014 05:54:34 +0000 (05:54 +0000)]
Prevent ill-formed instantiation of __invoke_of<...> during the evaluation of a bind expression. Fixes PR22003.
The SFINAE on the function __mu(Fn, Args...) that evaluates nested bind
expressions always tries to deduce the return type for Fn(Args...) even when Fn
is not a nested bind expression. This can cause hard compile errors when the
instantation of Fn(Args...) is ill-formed. This patch prevents the instantation
of __invoke_of<Fn, Args...> unless Fn is actually a bind expression.
Bug reportand patch from Michel Morin.
http://llvm.org/bugs/show_bug.cgi?id=22003
llvm-svn: 224753
Alexey Samsonov [Tue, 23 Dec 2014 04:15:53 +0000 (04:15 +0000)]
Fix UBSan bootstrap: replace shift of negative value with multiplication.
llvm-svn: 224752
Alexey Samsonov [Tue, 23 Dec 2014 04:15:47 +0000 (04:15 +0000)]
Fix UBSan bootstrap: don't bind reference to nullptr.
llvm-svn: 224751
Chandler Carruth [Tue, 23 Dec 2014 02:58:14 +0000 (02:58 +0000)]
Revert r224739: Debug info: Teach SROA how to update debug info for
fragmented variables.
This caused codegen to start crashing when we built somewhat large
programs with debug info and optimizations. 'check-msan' hit in, and
I suspect a bootstrap would as well. I mailed a test case to the
review thread.
llvm-svn: 224750
Kuba Brecka [Tue, 23 Dec 2014 01:52:53 +0000 (01:52 +0000)]
Change the CMake build to generate full debugging info when COMPILER_RT_DEBUG=ON
The compiler-rt CMake build currently uses -gline-tables-only for all builds. While this makes the debugger correctly map source files, debugging of the runtime libraries in general is pretty tough, because you don't see any local variables, parameters, etc. Let's allow emitting full debugging info by setting COMPILER_RT_DEBUG=ON.
Reviewed at http://reviews.llvm.org/D6764
llvm-svn: 224749
Marshall Clow [Tue, 23 Dec 2014 01:30:39 +0000 (01:30 +0000)]
Remove non-const test to get test passing. Will come back later and (correctly) add non-const tests
llvm-svn: 224748
Nico Weber [Tue, 23 Dec 2014 01:07:10 +0000 (01:07 +0000)]
Add a triple to try and get this test passing on the ARM bots.
llvm-svn: 224747
Jim Grosbach [Tue, 23 Dec 2014 00:35:23 +0000 (00:35 +0000)]
X86: Don't over-align combined loads.
When combining consecutive loads+inserts into a single vector load,
we should keep the alignment of the base load. Doing otherwise can, and does,
lead to using overly aligned instructions. In the included test case, for
example, using a 32-byte vmovaps on a 16-byte aligned value. Oops.
rdar://
19190968
llvm-svn: 224746
Reid Kleckner [Mon, 22 Dec 2014 23:58:37 +0000 (23:58 +0000)]
Make musttail more robust for vector types on x86
Previously I tried to plug musttail into the existing vararg lowering
code. That turned out to be a mistake, because non-vararg calls use
significantly different register lowering, even on x86. For example, AVX
vectors are usually passed in registers to normal functions and memory
to vararg functions. Now musttail uses a completely separate lowering.
Hopefully this can be used as the basis for non-x86 perfect forwarding.
Reviewers: majnemer
Differential Revision: http://reviews.llvm.org/D6156
llvm-svn: 224745
David Blaikie [Mon, 22 Dec 2014 23:12:42 +0000 (23:12 +0000)]
Remove dynamic allocation/indirection from GCOVBlocks owned by GCOVFunction
Since these are all created in the DenseMap before they are referenced,
there's no problem with pointer validity by the time it's required. This
removes another use of DeleteContainerSeconds/manual memory management
which I'm cleaning up from time to time.
llvm-svn: 224744
Adrian Prantl [Mon, 22 Dec 2014 23:09:14 +0000 (23:09 +0000)]
Thumb1 frame lowering: Mark CFI instructions with the FrameSetup flag.
Followup to r224294:
ARM/AArch64: Attach the FrameSetup MIFlag to CFI instructions.
Debug info marks the first instruction without the FrameSetup flag
as being the end of the function prologue. Any CFI instructions in the
middle of the function prologue would cause debug info to end the prologue
too early and worse, attach the line number of the CFI instruction, which
incidentally is often 0.
llvm-svn: 224743
Chandler Carruth [Mon, 22 Dec 2014 22:46:00 +0000 (22:46 +0000)]
[SROA] Lift the logic for traversing the alloca slices one partition at
a time into a partition iterator and a Partition class.
There is a lot of knock-on simplification that this enables, largely
stemming from having a Partition object to refer to in lots of helpers.
I've only done a minimal amount of that because enoguh stuff is changing
as-is in this commit.
This shouldn't change any observable behavior. I've worked hard to
preserve the *exact* traversal semantics which were originally present
even though some of them make no sense. I'll be changing some of this in
subsequent commits now that the logic is carefully factored into
a reusable place.
The primary motivation for this change is to break the rewriting into
phases in order to support more intelligent rewriting. For example, I'm
planning to change how split loads and stores are rewritten to remove
the significant overuse of integer bit packing in the resulting code and
allow more effective secondary splitting of aggregates. For any of this
to work, they have to share the exact traversal logic.
llvm-svn: 224742
Eric Fiselier [Mon, 22 Dec 2014 22:38:59 +0000 (22:38 +0000)]
[libcxx] Consolidate new/delete replacement in tests and disable it when using sanitizers.
Summary:
MSAN and ASAN also replace new/delete which leads to a link error in these tests. Currently they are unsupported but I think it would be useful if these tests could run with sanitizers.
This patch creates a support header that consolidates the new/delete replacement functionality and checking.
When we are using sanitizers new and delete are no longer replaced and the checks always return true.
Reviewers: mclow.lists, danalbert, jroelofs, EricWF
Reviewed By: EricWF
Subscribers: cfe-commits
Differential Revision: http://reviews.llvm.org/D6562
llvm-svn: 224741
Bruno Cardoso Lopes [Mon, 22 Dec 2014 22:35:46 +0000 (22:35 +0000)]
[LCSSA] Handle PHI insertion in disjoint loops
Take two disjoint Loops L1 and L2.
LoopSimplify fails to simplify some loops (e.g. when indirect branches
are involved). In such situations, it can happen that an exit for L1 is
the header of L2. Thus, when we create PHIs in one of such exits we are
also inserting PHIs in L2 header.
This could break LCSSA form for L2 because these inserted PHIs can also
have uses in L2 exits, which are never handled in the current
implementation. Provide a fix for this corner case and test that we
don't assert/crash on that.
Differential Revision: http://reviews.llvm.org/D6624
rdar://problem/
19166231
llvm-svn: 224740
Adrian Prantl [Mon, 22 Dec 2014 22:26:00 +0000 (22:26 +0000)]
Debug info: Teach SROA how to update debug info for fragmented variables.
This allows us to generate debug info for extremely advanced code such as
typedef struct { long int a; int b;} S;
int foo(S s) {
return s.b;
}
which at -O1 on x86_64 is codegen'd into
define i32 @foo(i64 %s.coerce0, i32 %s.coerce1) #0 {
ret i32 %s.coerce1, !dbg !24
}
with this patch we emit the following debug info for this
TAG_formal_parameter [3]
AT_location( 0x00000000
0x0000000000000000 - 0x0000000000000006: rdi, piece 0x00000008, rsi, piece 0x00000004
0x0000000000000006 - 0x0000000000000008: rdi, piece 0x00000008, rax, piece 0x00000004 )
AT_name( "s" )
AT_decl_file( "/Volumes/Data/llvm/_build.ninja.release/test.c" )
Thanks to chandlerc, dblaikie, and echristo for their feedback on all
previous iterations of this patch!
llvm-svn: 224739
Reid Kleckner [Mon, 22 Dec 2014 22:10:08 +0000 (22:10 +0000)]
Fix Windows unwind info for functions in sections other than .text
Previously we assumed the section name had the form .text$foo, which is
what we used to do for inline functions. If the dollar wasn't present,
we'd put unwind data in the .pdata and .xdata sections for the main
.text section, which is incorrect.
Fixes PR22001.
llvm-svn: 224738
Alexey Samsonov [Mon, 22 Dec 2014 21:46:10 +0000 (21:46 +0000)]
[Sanitizer] Make CommonFlags immutable after initialization.
Summary:
Protect CommonFlags singleton by adding const qualifier to
common_flags() accessor. The only ways to modify the flags are
SetCommonFlagsDefaults(), ParseCommonFlagsFromString() and
OverrideCommonFlags() functions, which are only supposed to be
called during initialization.
Test Plan: regression test suite
Reviewers: kcc, eugenis, glider
Subscribers: llvm-commits
Differential Revision: http://reviews.llvm.org/D6741
llvm-svn: 224736
Colin LeMahieu [Mon, 22 Dec 2014 21:40:43 +0000 (21:40 +0000)]
[Hexagon] Adding memb instruction. Fixing whitespace in test from 224730.
llvm-svn: 224735
David Blaikie [Mon, 22 Dec 2014 21:26:38 +0000 (21:26 +0000)]
Use iterators rather than indices to make this forwards-compatible with a change to the underlying container (to std::list)
llvm-svn: 224734