Zinovy Nis [Wed, 7 May 2014 09:51:22 +0000 (09:51 +0000)]
[BUG][REFACTOR]
1) Fix for printing debug locations for absolute paths.
2) Location printing is moved into public method DebugLoc::print() to avoid re-inventing the wheel.
Differential Revision: http://reviews.llvm.org/D3513
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@208177
91177308-0d34-0410-b5e6-
96231b3b80d8
Stepan Dyatkovskiy [Wed, 7 May 2014 09:05:10 +0000 (09:05 +0000)]
Second patch of patch series that improves MergeFunctions performance time from O(N*N) to
O(N*log(N)). The idea is to introduce total ordering among functions set.
It allows to build binary tree and perform function look-up procedure in O(log(N)) time.
This patch description:
Introduced total ordering among constants implemented in cmpConstants method.
Method performs lexicographical comparison between constants represented as
hypothetical numbers of next format:
<bitcastability-trait><raw-bit-contents>
Please, read cmpConstants declaration comments for more details.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@208173
91177308-0d34-0410-b5e6-
96231b3b80d8
Timur Iskhodzhanov [Wed, 7 May 2014 08:52:13 +0000 (08:52 +0000)]
Work-around MSVS build breakage due to r208148
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@208171
91177308-0d34-0410-b5e6-
96231b3b80d8
Evgeniy Stepanov [Wed, 7 May 2014 07:54:11 +0000 (07:54 +0000)]
[asan] Add a flag to control asm instrumentation.
With this change, asm instrumentation is disabled by default.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@208167
91177308-0d34-0410-b5e6-
96231b3b80d8
Joerg Sonnenberger [Wed, 7 May 2014 07:49:34 +0000 (07:49 +0000)]
Allow using normal .eh_frame based unwinding on ARM. Use the same
encodings as x86. Use this exception model for NetBSD.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@208166
91177308-0d34-0410-b5e6-
96231b3b80d8
David Blaikie [Wed, 7 May 2014 06:08:28 +0000 (06:08 +0000)]
PR19562: DebugInfo temporary MDNode leak: Don't include a temporary node to replace with a variable list for methods, since they're always declarations and thus never include variables
This field is used for a list of variables to ensure they are not lost
during optimization (they're only included when optimizations are
enabled).
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@208159
91177308-0d34-0410-b5e6-
96231b3b80d8
Tobias Grosser [Wed, 7 May 2014 06:07:47 +0000 (06:07 +0000)]
[C++11] Add NArySCEV->Operands iterator range
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@208158
91177308-0d34-0410-b5e6-
96231b3b80d8
Simon Atanasyan [Wed, 7 May 2014 05:18:51 +0000 (05:18 +0000)]
[obj2yaml] Recognize input file format using llvm::object::ObjectFile
interface methods isCOFF().
The '-coff' command line option has been removed. It was not used in any
test cases.
The patch reviewed by Michael Spencer.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@208157
91177308-0d34-0410-b5e6-
96231b3b80d8
Saleem Abdulrasool [Wed, 7 May 2014 03:03:31 +0000 (03:03 +0000)]
ARM: mark additional instructions as MachineFrameSetup
Mark up additional instructions which are part of the function prologue as
MachineFrameSetup. These instructions are part of the function prologue,
emitted by the PEI pass to setup the stack for use in the activating frame.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@208153
91177308-0d34-0410-b5e6-
96231b3b80d8
Saleem Abdulrasool [Wed, 7 May 2014 03:03:27 +0000 (03:03 +0000)]
ARM: fix WoA PEI instruction selection
The ARM::BLX instruction is an ARM mode instruction. The Windows on ARM target
is limited to Thumb instructions. Correctly use the thumb mode tBLXr
instruction. This would manifest as an errant write into the object file as the
instruction is 4-bytes in length rather than 2. The result would be a corrupted
object file that would eventually result in an executable that would crash at
runtime.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@208152
91177308-0d34-0410-b5e6-
96231b3b80d8
Justin Bogner [Wed, 7 May 2014 02:33:58 +0000 (02:33 +0000)]
llvm-cov: Document --no-output in the command guide
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@208151
91177308-0d34-0410-b5e6-
96231b3b80d8
Justin Bogner [Wed, 7 May 2014 02:11:23 +0000 (02:11 +0000)]
llvm-cov: Handle missing source files as GCOV does
If the source files referenced by a gcno file are missing, gcov
outputs a coverage file where every line is simply /*EOF*/. This also
occurs for lines in the coverage that are past the end of a file that
is found.
This change mimics gcov.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@208149
91177308-0d34-0410-b5e6-
96231b3b80d8
Justin Bogner [Wed, 7 May 2014 02:11:18 +0000 (02:11 +0000)]
llvm-cov: Implement --no-output
In gcov, there's a -n/--no-output option, which disables the writing
of any .gcov files, so that it emits only the summary info on stdout.
This implements the same behaviour in llvm-cov.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@208148
91177308-0d34-0410-b5e6-
96231b3b80d8
Argyrios Kyrtzidis [Tue, 6 May 2014 23:30:56 +0000 (23:30 +0000)]
[Support/MemoryBuffer] Remove the assertion that the file size did not shrink.
This can happen in practice with the user changing files and we can recover from it.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@208143
91177308-0d34-0410-b5e6-
96231b3b80d8
Nico Weber [Tue, 6 May 2014 23:17:26 +0000 (23:17 +0000)]
Fix ASan init function detection after clang r208128.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@208141
91177308-0d34-0410-b5e6-
96231b3b80d8
Rafael Espindola [Tue, 6 May 2014 22:44:30 +0000 (22:44 +0000)]
Special case aliases in GlobalValue::getSection.
This is similar to the getAlignment patch, but is done just for
completeness. It looks like we never call getSection on an alias. All the
tests still pass if the if is replaced with an assert.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@208139
91177308-0d34-0410-b5e6-
96231b3b80d8
Andrew Trick [Tue, 6 May 2014 22:18:43 +0000 (22:18 +0000)]
Update an embarassing out-of-date comment.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@208137
91177308-0d34-0410-b5e6-
96231b3b80d8
Eric Christopher [Tue, 6 May 2014 21:20:29 +0000 (21:20 +0000)]
Use a range based for loop for the SubtargetFeatures print function.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@208132
91177308-0d34-0410-b5e6-
96231b3b80d8
David Blaikie [Tue, 6 May 2014 21:07:17 +0000 (21:07 +0000)]
Revert "Try simplifying LexicalScopes ownership again."
Speculatively reverting due to a suspicious failure on a Windows
buildbot.
This reverts commit
10c37a012ea11596d44cd9059fe09c959caf30c8.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@208131
91177308-0d34-0410-b5e6-
96231b3b80d8
Eric Christopher [Tue, 6 May 2014 21:04:27 +0000 (21:04 +0000)]
Fix odd formatting that snuck into last patch.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@208130
91177308-0d34-0410-b5e6-
96231b3b80d8
Joerg Sonnenberger [Tue, 6 May 2014 20:43:01 +0000 (20:43 +0000)]
If a function needs a frame pointer, but r11 (aka fp) has not been used,
remove it from the list of unspilled registers. Otherwise the following
attempt to keep the stack aligned by picking an extra GPR register to
spill will not work as it picks up r11.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@208129
91177308-0d34-0410-b5e6-
96231b3b80d8
Eric Christopher [Tue, 6 May 2014 20:23:04 +0000 (20:23 +0000)]
ArrayRef-ize the Feature and Processor tables for SubtargetFeatures.
This removes arguments passed everywhere and allows the use of
standard iteration over lists.
Should be no functional change.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@208127
91177308-0d34-0410-b5e6-
96231b3b80d8
Reid Kleckner [Tue, 6 May 2014 20:08:20 +0000 (20:08 +0000)]
Copy the full TailCallKind in CallInst::clone_impl
Split from the musttail inliner change. This will be covered by an opt
test when the inliner change lands.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@208126
91177308-0d34-0410-b5e6-
96231b3b80d8
Diego Novillo [Tue, 6 May 2014 19:14:00 +0000 (19:14 +0000)]
Do not make -pass-remarks additive.
Summary:
When I initially introduced -pass-remarks, I thought it would be a
neat idea to make it additive. So, if one used it as:
$ llc -pass-remarks=inliner --pass-remarks=loop.*
the compiler would build the regular expression '(inliner)|(loop.*)'.
The more I think about it, the more I regret it. This is not how
other flags work. The standard semantics are right-to-left overrides.
This is how clang interprets -Rpass. And I think the two should be
compatible in this respect.
Reviewers: qcolombet
Subscribers: llvm-commits
Differential Revision: http://reviews.llvm.org/D3614
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@208122
91177308-0d34-0410-b5e6-
96231b3b80d8
Benjamin Kramer [Tue, 6 May 2014 18:36:23 +0000 (18:36 +0000)]
TTI: Estimate @llvm.fmuladd cost as fmul + fadd when FMA's aren't legal on the target.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@208115
91177308-0d34-0410-b5e6-
96231b3b80d8
Andrea Di Biagio [Tue, 6 May 2014 17:09:03 +0000 (17:09 +0000)]
[X86] Improve the lowering of BITCAST dag nodes from type f64 to type v2i32 (and vice versa).
Before this patch, the backend always emitted a store+load sequence to
bitconvert from f64 to i64 the input operand of a ISD::BITCAST dag node that
performed a bitconvert from type MVT::f64 to type MVT::v2i32. The resulting
i64 node was then used to build a v2i32 vector.
With this patch, the backend now produces a cheaper SCALAR_TO_VECTOR from
MVT::f64 to MVT::v2f64. That SCALAR_TO_VECTOR is then followed by a "free"
bitcast to type MVT::v4i32. The elements of the resulting
v4i32 are then extracted to build a v2i32 vector (which is illegal and
therefore promoted to MVT::v2i64).
This is in general cheaper than emitting a stack store+load sequence
to bitconvert the operand from type f64 to type i64.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@208107
91177308-0d34-0410-b5e6-
96231b3b80d8
Renato Golin [Tue, 6 May 2014 16:51:25 +0000 (16:51 +0000)]
Implememting named register intrinsics
This patch implements the infrastructure to use named register constructs in
programs that need access to specific registers (bare metal, kernels, etc).
So far, only the stack pointer is supported as a technology preview, but as it
is, the intrinsic can already support all non-allocatable registers from any
architecture.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@208104
91177308-0d34-0410-b5e6-
96231b3b80d8
Rafael Espindola [Tue, 6 May 2014 16:48:58 +0000 (16:48 +0000)]
Special case aliases in GlobalValue::getAlignment.
An alias has the address of what it points to, so it also has the same
alignment.
This allows a few optimizations to see past aliases for free.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@208103
91177308-0d34-0410-b5e6-
96231b3b80d8
Eric Christopher [Tue, 6 May 2014 16:29:50 +0000 (16:29 +0000)]
Have the SubtargetFeature help routine just not return a number and
fall back to the normal path without a cpu. While doing this fix
llc to just exit when we don't have a module to process instead of
asserting.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@208102
91177308-0d34-0410-b5e6-
96231b3b80d8
Bob Wilson [Tue, 6 May 2014 15:58:06 +0000 (15:58 +0000)]
Add some details to the llvm-cov documentation. <rdar://problem/
15819496>
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@208098
91177308-0d34-0410-b5e6-
96231b3b80d8
Rafael Espindola [Tue, 6 May 2014 14:59:14 +0000 (14:59 +0000)]
Be more strict about not allowing setSection on aliases.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@208095
91177308-0d34-0410-b5e6-
96231b3b80d8
Rafael Espindola [Tue, 6 May 2014 14:51:36 +0000 (14:51 +0000)]
Be more strict about not calling setAlignment on global aliases.
The fact that GlobalAlias::setAlignment exists at all is a side effect of
how the classes are organized, it should never be used.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@208094
91177308-0d34-0410-b5e6-
96231b3b80d8
Tim Northover [Tue, 6 May 2014 14:15:14 +0000 (14:15 +0000)]
AArch64/ARM64: implement diagnosis of unpredictable loads & stores
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@208091
91177308-0d34-0410-b5e6-
96231b3b80d8
Tim Northover [Tue, 6 May 2014 12:50:58 +0000 (12:50 +0000)]
AArch64/ARM64: add two more MC tests to ARM64 set.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@208085
91177308-0d34-0410-b5e6-
96231b3b80d8
Tim Northover [Tue, 6 May 2014 12:50:55 +0000 (12:50 +0000)]
AArch64/ARM64: enable MC-level diagnostic tests for NEON insts.
Obviously we can't expect the two backends to produce identical diagnostics,
since what's possible depends quite a bit on how the .td files are structured.
I think the ARM64 diagnostics are basically of the same quality in all the
changed cases, so I've split the CHECK lines.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@208084
91177308-0d34-0410-b5e6-
96231b3b80d8
Tim Northover [Tue, 6 May 2014 12:50:51 +0000 (12:50 +0000)]
AArch64/ARM64: make NEON vector list parsing a bit more robust
It doesn't change the results, but it seems silly not to diagnose obvious
problems early on.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@208083
91177308-0d34-0410-b5e6-
96231b3b80d8
Tim Northover [Tue, 6 May 2014 12:50:47 +0000 (12:50 +0000)]
AArch64/ARM64: add more specific diagnostic for floating imm 0.0.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@208082
91177308-0d34-0410-b5e6-
96231b3b80d8
Tim Northover [Tue, 6 May 2014 12:50:44 +0000 (12:50 +0000)]
AArch64/ARM64: add more specific diagnostic for invalid vector lanes
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@208081
91177308-0d34-0410-b5e6-
96231b3b80d8
Tim Northover [Tue, 6 May 2014 11:18:53 +0000 (11:18 +0000)]
AArch64/ARM64: produce more informative diagnostic assembling some immediates
No tests here, they'll be added when the entire neon-diagnostics.s test from
AArch64 is enabled.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@208079
91177308-0d34-0410-b5e6-
96231b3b80d8
Christian Pirker [Tue, 6 May 2014 10:05:11 +0000 (10:05 +0000)]
ARM: For thumb fixups store halfwords high first and low second
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@208076
91177308-0d34-0410-b5e6-
96231b3b80d8
Kevin Qin [Tue, 6 May 2014 09:48:52 +0000 (09:48 +0000)]
[ARM64] Enable alignment control option in front-end for ARM64.
This is the modification in llvm part.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@208074
91177308-0d34-0410-b5e6-
96231b3b80d8
Evgeniy Stepanov [Tue, 6 May 2014 09:46:06 +0000 (09:46 +0000)]
Disable -Wcomment when building with GCC.
GCC version of -Wcomment is not compatible with ascii art graph diagrams.
Reverts r207629.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@208073
91177308-0d34-0410-b5e6-
96231b3b80d8
Richard Smith [Tue, 6 May 2014 07:45:39 +0000 (07:45 +0000)]
Update programmers manual to cover llvm::function_ref, and add a note to the
coding standard suggesting using it instead of the (unavailable) std::function.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@208067
91177308-0d34-0410-b5e6-
96231b3b80d8
Craig Topper [Tue, 6 May 2014 07:04:32 +0000 (07:04 +0000)]
Use X86 memory operand enums instead of hardcoding.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@208064
91177308-0d34-0410-b5e6-
96231b3b80d8
Owen Anderson [Tue, 6 May 2014 05:05:59 +0000 (05:05 +0000)]
Fix some obvious Doxygen comment bugs.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@208059
91177308-0d34-0410-b5e6-
96231b3b80d8
David Blaikie [Tue, 6 May 2014 03:53:10 +0000 (03:53 +0000)]
Update comment from a recent commit.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@208057
91177308-0d34-0410-b5e6-
96231b3b80d8
Adam Nemet [Tue, 6 May 2014 03:49:45 +0000 (03:49 +0000)]
[Build] Add rule to generate fully-expanded .td file
I found it useful in the past and now again to have a version of the .td file
where all the records are expanded. This adds a makefile rule to generate
this on demand.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@208056
91177308-0d34-0410-b5e6-
96231b3b80d8
David Blaikie [Tue, 6 May 2014 03:41:57 +0000 (03:41 +0000)]
PR19598: Provide the ability to RAUW a declaration with itself, creating a non-temporary copy and using that to RAUW.
Also, provide the ability to create temporary and non-temporary
declarations, as not all declarations may be replaced by definitions
later on.
This provides the necessary infrastructure for Clang to fix PR19598,
leaking temporary MDNodes in Clang's debug info generation.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@208054
91177308-0d34-0410-b5e6-
96231b3b80d8
Eric Christopher [Tue, 6 May 2014 02:37:26 +0000 (02:37 +0000)]
Revert "Walk back commits for unused function parameters - they're still being"
this reapplies 208012 and 208002.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@208037
91177308-0d34-0410-b5e6-
96231b3b80d8
Duncan P. N. Exon Smith [Tue, 6 May 2014 01:57:42 +0000 (01:57 +0000)]
blockfreq: Move include to .cpp
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@208035
91177308-0d34-0410-b5e6-
96231b3b80d8
Richard Smith [Tue, 6 May 2014 01:46:26 +0000 (01:46 +0000)]
Revert accidentally-committed files.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@208034
91177308-0d34-0410-b5e6-
96231b3b80d8
Richard Smith [Tue, 6 May 2014 01:44:26 +0000 (01:44 +0000)]
Re-commit r208025, reverted in r208030, with a fix for a conformance issue
which GCC detects and Clang does not!
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@208033
91177308-0d34-0410-b5e6-
96231b3b80d8
Richard Smith [Tue, 6 May 2014 01:26:00 +0000 (01:26 +0000)]
Revert r208025, which made buildbots unhappy for unknown reasons.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@208030
91177308-0d34-0410-b5e6-
96231b3b80d8
Reid Kleckner [Tue, 6 May 2014 01:20:42 +0000 (01:20 +0000)]
Fix i128 div/mod on mingw64
The Win64 docs are very clear that anything larger than 8 bytes is
passed by reference, and GCC MinGW64 honors that for __modti3 and
friends.
Patch by Jameson Nash!
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@208029
91177308-0d34-0410-b5e6-
96231b3b80d8
Argyrios Kyrtzidis [Tue, 6 May 2014 01:03:52 +0000 (01:03 +0000)]
[Support/MemoryBuffer] Rename IsVolatile -> IsVolatileSize and add a comment about the use case for the new parameter.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@208026
91177308-0d34-0410-b5e6-
96231b3b80d8
Richard Smith [Tue, 6 May 2014 01:01:29 +0000 (01:01 +0000)]
Add llvm::function_ref (and a couple of uses of it), representing a type-erased reference to a callable object.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@208025
91177308-0d34-0410-b5e6-
96231b3b80d8
Reid Kleckner [Tue, 6 May 2014 00:57:33 +0000 (00:57 +0000)]
Include intrin.h before windows.h as a workaround for the x64 self-host
On x64, windows.h doesn't include intrin.h for intrinsics. It just
declares them in the global namespace and uses them, expecting the
compiler to lower it as a builtin. We basically need to do this in
clang, eventually.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@208023
91177308-0d34-0410-b5e6-
96231b3b80d8
Argyrios Kyrtzidis [Tue, 6 May 2014 00:51:45 +0000 (00:51 +0000)]
[Support/MemoryBuffer] Move the IsVolatile check inside shouldUseMmap() and make sure to zero-initialize the rest
of the buffer if we unexpectedly reach end-of-file while reading.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@208021
91177308-0d34-0410-b5e6-
96231b3b80d8
Nick Lewycky [Tue, 6 May 2014 00:46:20 +0000 (00:46 +0000)]
Detabify.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@208019
91177308-0d34-0410-b5e6-
96231b3b80d8
Nick Lewycky [Mon, 5 May 2014 23:59:03 +0000 (23:59 +0000)]
Improve 'tail' call marking in TRE. A bootstrap of clang goes from 375k calls marked tail in the IR to 470k, however this improvement does not carry into an improvement of the call/jmp ratio on x86. The most common pattern is a tail call + br to a block with nothing but a 'ret'.
The number of tail call to loop conversions remains the same (1618 by my count).
The new algorithm does a local scan over the use-def chains to identify local "alloca-derived" values, as well as points where the alloca could escape. Then, a visit over the CFG marks blocks as being before or after the allocas have escaped, and annotates the calls accordingly.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@208017
91177308-0d34-0410-b5e6-
96231b3b80d8
Eric Christopher [Mon, 5 May 2014 23:26:59 +0000 (23:26 +0000)]
Walk back commits for unused function parameters - they're still being
used via dragonegg for now.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@208016
91177308-0d34-0410-b5e6-
96231b3b80d8
Yi Jiang [Mon, 5 May 2014 23:14:46 +0000 (23:14 +0000)]
Reapply: Add slp vectorization to LTO passes. The bug it exposed has been fixed by r207983. <radar://
16641956>
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@208013
91177308-0d34-0410-b5e6-
96231b3b80d8
Eric Christopher [Mon, 5 May 2014 22:36:07 +0000 (22:36 +0000)]
Remove a now unnecessary function since all calls have one version
and inline it into its caller.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@208012
91177308-0d34-0410-b5e6-
96231b3b80d8
Eric Christopher [Mon, 5 May 2014 22:01:47 +0000 (22:01 +0000)]
Remove a call to std::exit in a library. Make "Help" return
a 0 as a default answer.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@208009
91177308-0d34-0410-b5e6-
96231b3b80d8
Argyrios Kyrtzidis [Mon, 5 May 2014 21:55:51 +0000 (21:55 +0000)]
[Support/MemoryBuffer] Introduce a boolean parameter (false by default) 'IsVolatile' for the open file functions.
This provides a hint that the file may be changing often so mmap is avoided.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@208007
91177308-0d34-0410-b5e6-
96231b3b80d8
Eric Christopher [Mon, 5 May 2014 21:50:57 +0000 (21:50 +0000)]
Fix typo.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@208006
91177308-0d34-0410-b5e6-
96231b3b80d8
Tom Stellard [Mon, 5 May 2014 21:47:15 +0000 (21:47 +0000)]
R600: Expand i64 ISD:SUB
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@208005
91177308-0d34-0410-b5e6-
96231b3b80d8
Eric Christopher [Mon, 5 May 2014 21:40:44 +0000 (21:40 +0000)]
Remove unused argument from AddFeature.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@208002
91177308-0d34-0410-b5e6-
96231b3b80d8
Eric Christopher [Mon, 5 May 2014 21:40:41 +0000 (21:40 +0000)]
Fix typo (also tab character).
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@208001
91177308-0d34-0410-b5e6-
96231b3b80d8
Rafael Espindola [Mon, 5 May 2014 20:06:41 +0000 (20:06 +0000)]
Use a range loop.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207996
91177308-0d34-0410-b5e6-
96231b3b80d8
Filipe Cabecinhas [Mon, 5 May 2014 19:40:36 +0000 (19:40 +0000)]
Revert "Optimize shufflevector that copies an i64/f64 and zeros the rest."
This reverts commit 207992. I misread the phab number on the LGTM.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207993
91177308-0d34-0410-b5e6-
96231b3b80d8
Filipe Cabecinhas [Mon, 5 May 2014 19:36:28 +0000 (19:36 +0000)]
Optimize shufflevector that copies an i64/f64 and zeros the rest.
Summary:
Also ran clang-format on the function. The code added is the last else
if block.
Reviewers: nadav, craig.topper
Subscribers: llvm-commits
Differential Revision: http://reviews.llvm.org/D3518
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207992
91177308-0d34-0410-b5e6-
96231b3b80d8
Marek Olsak [Mon, 5 May 2014 19:30:54 +0000 (19:30 +0000)]
R600/SI: allow 5 more input SGPRs to a shader
Our OpenGL driver needs 22 SGPRs (16 user SGPRs + 6 streamout non-user SGPRs).
Signed-off-by: Marek Olšák <marek.olsak@amd.com>
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207990
91177308-0d34-0410-b5e6-
96231b3b80d8
Michael Zolotukhin [Mon, 5 May 2014 18:10:15 +0000 (18:10 +0000)]
Move test from r207969 to another folder and rename it.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207984
91177308-0d34-0410-b5e6-
96231b3b80d8
Yi Jiang [Mon, 5 May 2014 17:59:14 +0000 (17:59 +0000)]
Always set alignment of vectorized LD/ST in SLP-Vectorizer. <rdar://problem/
16812145>
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207983
91177308-0d34-0410-b5e6-
96231b3b80d8
Joerg Sonnenberger [Mon, 5 May 2014 17:58:46 +0000 (17:58 +0000)]
Fix spelling.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207982
91177308-0d34-0410-b5e6-
96231b3b80d8
Rafael Espindola [Mon, 5 May 2014 17:53:29 +0000 (17:53 +0000)]
Be a bit more specific in the release notes.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207981
91177308-0d34-0410-b5e6-
96231b3b80d8
Duncan P. N. Exon Smith [Mon, 5 May 2014 17:40:44 +0000 (17:40 +0000)]
LTO: -internalize sets visibility to default
Visibility is meaningless when the linkage is local. Change
`-internalize` to reset the visibility to `default`.
<rdar://problem/
16141113>
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207979
91177308-0d34-0410-b5e6-
96231b3b80d8
Rafael Espindola [Mon, 5 May 2014 17:33:26 +0000 (17:33 +0000)]
Remove the -disable-cfi option.
This also add a release note about it. If this stays I will cleanup MC
next week.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207977
91177308-0d34-0410-b5e6-
96231b3b80d8
Adam Nemet [Mon, 5 May 2014 17:17:27 +0000 (17:17 +0000)]
[Test] Remove substitution for clang
clang should not be used in the llvm tests. The topic was discussed in this
thread:
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-
20140428/214905.html
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207976
91177308-0d34-0410-b5e6-
96231b3b80d8
Rafael Espindola [Mon, 5 May 2014 16:47:07 +0000 (16:47 +0000)]
Modify test to not use -disable-cfi.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207974
91177308-0d34-0410-b5e6-
96231b3b80d8
Kaelyn Takata [Mon, 5 May 2014 16:32:10 +0000 (16:32 +0000)]
Select bdver2 instead of bdver1 if TBM support is present on models < 0x10.
Tested that the right -target-cpu is set in the clang -cc1 command line
when running "clang -march=native -E -v - </dev/null" on both an FX-8150
and an FX-8350. Both are family 15h; the FX-8150 (Bulldozer processor)
reports a model number of 1, and the FX-8350 (Piledriver processor)
reports a model number of 2.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207973
91177308-0d34-0410-b5e6-
96231b3b80d8
Rafael Espindola [Mon, 5 May 2014 16:14:37 +0000 (16:14 +0000)]
Move test to the ARM64 directory.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207972
91177308-0d34-0410-b5e6-
96231b3b80d8
Rafael Espindola [Mon, 5 May 2014 15:34:13 +0000 (15:34 +0000)]
Convert a CodeGen test into a MC test.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207971
91177308-0d34-0410-b5e6-
96231b3b80d8
Michael Zolotukhin [Mon, 5 May 2014 14:46:53 +0000 (14:46 +0000)]
Fix test from r207966 and add a comment there.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207969
91177308-0d34-0410-b5e6-
96231b3b80d8
Timur Iskhodzhanov [Mon, 5 May 2014 14:28:38 +0000 (14:28 +0000)]
[ASan/Win] Fix issue 305 -- don't instrument .CRT initializer/terminator callbacks
See https://code.google.com/p/address-sanitizer/issues/detail?id=305
Reviewed at http://reviews.llvm.org/D3607
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207968
91177308-0d34-0410-b5e6-
96231b3b80d8
Rafael Espindola [Mon, 5 May 2014 14:18:16 +0000 (14:18 +0000)]
Trivial simplification. No functionality change.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207967
91177308-0d34-0410-b5e6-
96231b3b80d8
Michael Zolotukhin [Mon, 5 May 2014 14:05:25 +0000 (14:05 +0000)]
Add regression test for r207692.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207966
91177308-0d34-0410-b5e6-
96231b3b80d8
Patrik Hagglund [Mon, 5 May 2014 12:24:08 +0000 (12:24 +0000)]
Fix gcc -pedantic warning in lto.cpp.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207959
91177308-0d34-0410-b5e6-
96231b3b80d8
Simon Atanasyan [Mon, 5 May 2014 06:48:34 +0000 (06:48 +0000)]
Add range access to ELFFile's sections collection.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207952
91177308-0d34-0410-b5e6-
96231b3b80d8
Saleem Abdulrasool [Sun, 4 May 2014 23:13:21 +0000 (23:13 +0000)]
CodeGen: correct memset emittance for WoA
Windows on ARM does not conform to AEABI. However, memset would be emitted
using the AEABI signature, resulting in inverted parameters. Handle this
special case appropriately.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207943
91177308-0d34-0410-b5e6-
96231b3b80d8
Saleem Abdulrasool [Sun, 4 May 2014 23:13:18 +0000 (23:13 +0000)]
CodeGen: strengthen WoA AEABI avoidance tests
Add additional test cases for WoA AEABI avoidance checking.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207942
91177308-0d34-0410-b5e6-
96231b3b80d8
Saleem Abdulrasool [Sun, 4 May 2014 23:13:15 +0000 (23:13 +0000)]
MC: support FK_SecRel_4 for Windows on ARM
Add handling for FK_SecRel_4 (4-byte section relative relocations). These are
used by the generation of DWARF debug information (the abbrevations use section
relative relocations). This will also be used in generation of CodeView line
tables.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207941
91177308-0d34-0410-b5e6-
96231b3b80d8
Benjamin Kramer [Sun, 4 May 2014 19:12:38 +0000 (19:12 +0000)]
LoopUnroll: If we're doing partial unrolling, use the PartialThreshold to limit unrolling.
Otherwise we use the same threshold as for complete unrolling, which is
way too high. This made us unroll any loop smaller than 150 instructions
by 8 times, but only if someone specified -march=core2 or better,
which happens to be the default on darwin.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207940
91177308-0d34-0410-b5e6-
96231b3b80d8
Arnold Schwaighofer [Sun, 4 May 2014 17:10:15 +0000 (17:10 +0000)]
SLPVectorizer: Bring back the insertelement patch (r205965) with fixes
When can't assume a vectorized tree is rooted in an instruction. The IRBuilder
could have constant folded it. When we rebuild the build_vector (the series of
InsertElement instructions) use the last original InsertElement instruction. The
vectorized tree root is guaranteed to be before it.
Also, we can't assume that the n-th InsertElement inserts the n-th element into
a vector.
This reverts r207746 which reverted the revert of the revert of r205018 or so.
Fixes the test case in PR19621.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207939
91177308-0d34-0410-b5e6-
96231b3b80d8
Elena Demikhovsky [Sun, 4 May 2014 13:35:37 +0000 (13:35 +0000)]
AVX-512: minor change in rndscale intrinsic
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207937
91177308-0d34-0410-b5e6-
96231b3b80d8
Chandler Carruth [Sun, 4 May 2014 09:38:32 +0000 (09:38 +0000)]
[LCG] Add the last (and most complex) of the edge insertion mutation
operations on the call graph. This one forms a cycle, and while not as
complex as removing an internal edge from an SCC, it involves
a reasonable amount of work to find all of the nodes newly connected in
a cycle.
Also somewhat alarming is the worst case complexity here: it might have
to walk roughly the entire SCC inverse DAG to insert a single edge. This
is carefully documented in the API (I hope).
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207935
91177308-0d34-0410-b5e6-
96231b3b80d8
Chandler Carruth [Sun, 4 May 2014 09:38:23 +0000 (09:38 +0000)]
[LCG] Reorder the tests to be a bit more logical: inter-SCC mutation
before intra-SCC mutation, insertion before removal.
No functionality changed.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207934
91177308-0d34-0410-b5e6-
96231b3b80d8
Saleem Abdulrasool [Sun, 4 May 2014 01:54:17 +0000 (01:54 +0000)]
X86: further range-loopify AsmPrinter
Use more range loops in the X86AsmPrinter. NFC.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207928
91177308-0d34-0410-b5e6-
96231b3b80d8
Saleem Abdulrasool [Sun, 4 May 2014 01:54:12 +0000 (01:54 +0000)]
X86: remove X86COFFMachineModuleInfo
Remove dead code. This is vestigial after r98384.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207927
91177308-0d34-0410-b5e6-
96231b3b80d8