OSDN Git Service
Rafael Espindola [Wed, 4 Sep 2013 19:19:32 +0000 (19:19 +0000)]
Merge these 2 tests in a single file.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@189975
91177308-0d34-0410-b5e6-
96231b3b80d8
Jim Grosbach [Wed, 4 Sep 2013 19:08:44 +0000 (19:08 +0000)]
ARM: Teach A15 SDOptimizer to properly handle D-reg by-lane.
These instructions, such as vmul.f32, require the second source operand to
be in D0-D15 rather than the full D0-D31. When optimizing, make sure to
account for that by constraining the register class of a replacement virtual
register to be compatible with the virtual register(s) it's replacing.
I've been unsuccessful in creating a non-fragile regression test. This issue
was detected by the LLVM nightly test suite running on an A15 (Bullet).
PR17093: http://llvm.org/bugs/show_bug.cgi?id=17093
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@189972
91177308-0d34-0410-b5e6-
96231b3b80d8
Rafael Espindola [Wed, 4 Sep 2013 18:53:21 +0000 (18:53 +0000)]
Small simplification given that insert of an empty range is a nop.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@189971
91177308-0d34-0410-b5e6-
96231b3b80d8
Bill Wendling [Wed, 4 Sep 2013 18:48:12 +0000 (18:48 +0000)]
Remove 'param' label from comments. They aren't used properly here.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@189970
91177308-0d34-0410-b5e6-
96231b3b80d8
Rafael Espindola [Wed, 4 Sep 2013 18:37:36 +0000 (18:37 +0000)]
Refactor duplicated logic to a helper function.
No functionality change.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@189969
91177308-0d34-0410-b5e6-
96231b3b80d8
Rafael Espindola [Wed, 4 Sep 2013 18:16:02 +0000 (18:16 +0000)]
Remove dead code.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@189967
91177308-0d34-0410-b5e6-
96231b3b80d8
Dmitri Gribenko [Wed, 4 Sep 2013 18:02:13 +0000 (18:02 +0000)]
MemoryBufer: add a test: check that a file with size that is a multiple of the
page size can be null terminated correctly by MemoryBuffer.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@189965
91177308-0d34-0410-b5e6-
96231b3b80d8
Rafael Espindola [Wed, 4 Sep 2013 17:44:24 +0000 (17:44 +0000)]
Rename variables to match the style guide and clang-format.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@189962
91177308-0d34-0410-b5e6-
96231b3b80d8
Arnold Schwaighofer [Wed, 4 Sep 2013 17:41:16 +0000 (17:41 +0000)]
Swift: Only build vldm/vstm with q register aligned register lists
Unaligned vldm/vstm need more uops and therefore are slower in general on swift.
radar://
14522102
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@189961
91177308-0d34-0410-b5e6-
96231b3b80d8
Silviu Baranga [Wed, 4 Sep 2013 17:05:18 +0000 (17:05 +0000)]
Fix scheduling for vldm/vstm instructions that load/store more than 32 bytes on Cortex-A9. This also makes the existing code more compact.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@189958
91177308-0d34-0410-b5e6-
96231b3b80d8
Rafael Espindola [Wed, 4 Sep 2013 16:09:01 +0000 (16:09 +0000)]
Revert "Add r159136 back now that pr13124 has been fixed."
This reverts commit r189886.
I found a corner case where this optimization is not valid:
Say we have a "linkonce_odr unnamed_addr" in two translation units:
* In TU 1 this optimization kicks in and makes it hidden.
* In TU 2 it gets const merged with a constant that is *not* unnamed_addr,
resulting in a non unnamed_addr constant with default visibility.
* The static linker rules for combining visibility them produce a hidden
symbol, which is incorrect from the point of view of the non unnamed_addr
constant.
The one place we can do this is when we know that the symbol is not used from
another TU in the same shared object, i.e., during LTO. I will move it there.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@189954
91177308-0d34-0410-b5e6-
96231b3b80d8
Alexander Kornienko [Wed, 4 Sep 2013 16:00:12 +0000 (16:00 +0000)]
Move generic isPrint and columnWidth implementations to a separate header/source to allow using both generic and system-dependent versions on win32.
Summary:
This is needed so we can use generic columnWidthUTF8 in clang-format on
win32 simultaneously with a separate system-dependent implementations of
isPrint/columnWidth in TextDiagnostic.cpp to avoid attempts to print Unicode
characters using narrow-character interfaces (which is not supported on Windows,
and we'll have to figure out how to handle this).
Reviewers: jordan_rose
Reviewed By: jordan_rose
CC: llvm-commits, klimek
Differential Revision: http://llvm-reviews.chandlerc.com/D1559
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@189952
91177308-0d34-0410-b5e6-
96231b3b80d8
Rafael Espindola [Wed, 4 Sep 2013 15:33:34 +0000 (15:33 +0000)]
Error on linking appending globals with different unnamed_addr.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@189950
91177308-0d34-0410-b5e6-
96231b3b80d8
Venkatraman Govindaraju [Wed, 4 Sep 2013 15:15:20 +0000 (15:15 +0000)]
[Sparc] Fix an assertion failure while lowering fcmp on long double.
This assertion is triggered because an integer constant is created with wrong
type.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@189948
91177308-0d34-0410-b5e6-
96231b3b80d8
Rafael Espindola [Wed, 4 Sep 2013 14:59:03 +0000 (14:59 +0000)]
Fix linking of unnamed_addr in functions.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@189945
91177308-0d34-0410-b5e6-
96231b3b80d8
Rafael Espindola [Wed, 4 Sep 2013 14:40:49 +0000 (14:40 +0000)]
Rename in preparation for adding tests for function linking.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@189944
91177308-0d34-0410-b5e6-
96231b3b80d8
Rafael Espindola [Wed, 4 Sep 2013 14:35:48 +0000 (14:35 +0000)]
Use CHECK-DAG instead of sort.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@189942
91177308-0d34-0410-b5e6-
96231b3b80d8
NAKAMURA Takumi [Wed, 4 Sep 2013 14:12:26 +0000 (14:12 +0000)]
Support/Process: Add comments about PageSize and AllocationGranularity on Cygwin and Win32.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@189940
91177308-0d34-0410-b5e6-
96231b3b80d8
NAKAMURA Takumi [Wed, 4 Sep 2013 14:12:19 +0000 (14:12 +0000)]
MemoryBuffer.cpp: Don't peek the next page if file is multiple of *physical* pagesize(4k) but is not multiple of AllocationGranularity(64k), when a null terminator is required, on cygwin and win32.
For example, r189780's SparcISelLowering.cpp has the size 98304. It crashed clang to touch a null terminator on cygwin.
FIXME: It's not good to hardcode 4096 here. dwPageSize shows 4096.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@189939
91177308-0d34-0410-b5e6-
96231b3b80d8
NAKAMURA Takumi [Wed, 4 Sep 2013 14:12:12 +0000 (14:12 +0000)]
Whitespace.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@189938
91177308-0d34-0410-b5e6-
96231b3b80d8
Rafael Espindola [Wed, 4 Sep 2013 14:05:09 +0000 (14:05 +0000)]
Fix linking of unnamed_addr.
This was regression from r134829. When linking we have to be conservative. If
one of the symbols has a significant address, then the result should have it
too.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@189935
91177308-0d34-0410-b5e6-
96231b3b80d8
Tim Northover [Wed, 4 Sep 2013 11:57:17 +0000 (11:57 +0000)]
InstCombine: allow unmasked icmps to be combined with logical ops
"(icmp op i8 A, B)" is equivalent to "(icmp op i8 (A & 0xff), B)" as a
degenerate case. Allowing this as a "masked" comparison when analysing "(icmp)
&/| (icmp)" allows us to combine them in more cases.
rdar://problem/
7625728
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@189931
91177308-0d34-0410-b5e6-
96231b3b80d8
Tim Northover [Wed, 4 Sep 2013 11:57:13 +0000 (11:57 +0000)]
InstCombine: look for masked compares with subset relation
Even in cases which aren't universally optimisable like "(A & B) != 0 && (A &
C) != 0", the masks can make one of the comparisons completely redundant. In
this case, since we've gone to the effort of spotting masked comparisons we
should combine them.
rdar://problem/
7625728
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@189930
91177308-0d34-0410-b5e6-
96231b3b80d8
Hao Liu [Wed, 4 Sep 2013 09:28:24 +0000 (09:28 +0000)]
Inplement aarch64 neon instructions in AdvSIMD(shift). About 24 shift instructions:
sshr,ushr,ssra,usra,srshr,urshr,srsra,ursra,sri,shl,sli,sqshlu,sqshl,uqshl,shrn,sqrshrun,sqshrn,uqshr,sqrshrn,uqrshrn,sshll,ushll
and 4 convert instructions:
scvtf,ucvtf,fcvtzs,fcvtzu
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@189925
91177308-0d34-0410-b5e6-
96231b3b80d8
Eric Christopher [Wed, 4 Sep 2013 05:08:32 +0000 (05:08 +0000)]
Use Intrinsic::ID for the pattern match templates, fixing a signed/unsigned
comparison warning.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@189921
91177308-0d34-0410-b5e6-
96231b3b80d8
Michael Gottesman [Wed, 4 Sep 2013 04:39:38 +0000 (04:39 +0000)]
Revert "Revert "Remove the darwin gdb option, that version of gdb is now dead and the rest of the compatibility should be done on a dwarf-N level.""
This reverts commit r189913.
Talked with Eric on IRC. I am going to XFAIL the failing test since it
is using what Eric described as "the member hack" which was needed on
that old GDB.
Sorry for the noise!
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@189914
91177308-0d34-0410-b5e6-
96231b3b80d8
Michael Gottesman [Wed, 4 Sep 2013 04:31:56 +0000 (04:31 +0000)]
Revert "Remove the darwin gdb option, that version of gdb is now dead and the rest of the compatibility should be done on a dwarf-N level."
This reverts commit r189903.
This commit broke the phase 1 buildbot for a while.
http://lab.llvm.org:8013/builders/clang-x86_64-darwin11-nobootstrap-RAincremental/builds/6684
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@189913
91177308-0d34-0410-b5e6-
96231b3b80d8
Michael Gottesman [Wed, 4 Sep 2013 04:26:09 +0000 (04:26 +0000)]
Add llvm namespace to llvm::next.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@189912
91177308-0d34-0410-b5e6-
96231b3b80d8
Michael Gottesman [Wed, 4 Sep 2013 04:19:01 +0000 (04:19 +0000)]
Use llvm::next() instead of incrementing begin iterators of std::vector.
Iterator of std::vector may be implemented as a raw pointer. In
this case begin iterators are rvalues and cannot be incremented.
For example, this is the case with STDCXX implementation of vector.
Patch by Konstantin Tokarev <annulen@yandex.ru>.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@189911
91177308-0d34-0410-b5e6-
96231b3b80d8
NAKAMURA Takumi [Wed, 4 Sep 2013 02:10:32 +0000 (02:10 +0000)]
MCDwarf.h: Prune a few stray \param(s). [-Wdocumentation]
\param should be used to describe individual parameters. Use a command like \a or \c for visual enhancements.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@189905
91177308-0d34-0410-b5e6-
96231b3b80d8
Eric Christopher [Wed, 4 Sep 2013 02:02:10 +0000 (02:02 +0000)]
Remove the darwin gdb option, that version of gdb is now dead and
the rest of the compatibility should be done on a dwarf-N level.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@189903
91177308-0d34-0410-b5e6-
96231b3b80d8
Eric Christopher [Wed, 4 Sep 2013 01:38:30 +0000 (01:38 +0000)]
Make the default dwarf version 3 for darwin when we can't find one
in the module. Add a FIXME with a comment about darwin's ld.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@189902
91177308-0d34-0410-b5e6-
96231b3b80d8
Eric Christopher [Wed, 4 Sep 2013 00:58:10 +0000 (00:58 +0000)]
It's a very large constant. Say so.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@189899
91177308-0d34-0410-b5e6-
96231b3b80d8
Rafael Espindola [Tue, 3 Sep 2013 23:34:36 +0000 (23:34 +0000)]
Add r159136 back now that pr13124 has been fixed.
Original message:
If a constant or a function has linkonce_odr linkage and unnamed_addr, mark
hidden. Being linkonce_odr guarantees that it is available in every dso that
needs it. Being a constant/function with unnamed_addr guarantees that the
copies don't have to be merged.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@189886
91177308-0d34-0410-b5e6-
96231b3b80d8
Daniel Dunbar [Tue, 3 Sep 2013 23:32:55 +0000 (23:32 +0000)]
[lit] Allow config files to pass arbitrary values to child configs.
- This aligns with how existing test suites end up wanting to use the local
config files, conceptually it makes sense to consider them to be inherited.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@189885
91177308-0d34-0410-b5e6-
96231b3b80d8
Jim Grosbach [Tue, 3 Sep 2013 23:02:00 +0000 (23:02 +0000)]
X86: Mark non-crashing report_fatal_errors() as such.
Previously, the clang crash handling code would kick in and give a crash
report for these, even though they're not that sort of error.
rdar://
14882264
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@189878
91177308-0d34-0410-b5e6-
96231b3b80d8
Michael Gottesman [Tue, 3 Sep 2013 22:40:56 +0000 (22:40 +0000)]
[objc-arc] Remove dead code from previous commit.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@189870
91177308-0d34-0410-b5e6-
96231b3b80d8
Michael Gottesman [Tue, 3 Sep 2013 22:40:54 +0000 (22:40 +0000)]
[objc-arc] Turn off the objc_retainBlock -> objc_retain optimization.
The reason that I am turning off this optimization is that there is an
additional case where a block can escape that has come up. Specifically, this
occurs when a block is used in a scope outside of its current scope.
This can cause a captured retainable object pointer whose life is preserved by
the objc_retainBlock to be deallocated before the block is invoked.
An example of the code needed to trigger the bug is:
----
\#import <Foundation/Foundation.h>
int main(int argc, const char * argv[]) {
void (^somethingToDoLater)();
{
NSObject *obj = [NSObject new];
somethingToDoLater = ^{
[obj self]; // Crashes here
};
}
NSLog(@"test.");
somethingToDoLater();
return 0;
}
----
In the next commit, I remove all the dead code that results from this.
Once I put in the fixing commit I will bring back the tests that I deleted in
this commit.
rdar://
14802782.
rdar://
14868830.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@189869
91177308-0d34-0410-b5e6-
96231b3b80d8
Michael Gottesman [Tue, 3 Sep 2013 22:40:50 +0000 (22:40 +0000)]
[objc-arc] Move some block tests from basic.ll -> retain-block.ll and add some missing CHECK-LABELS.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@189868
91177308-0d34-0410-b5e6-
96231b3b80d8
Bill Wendling [Tue, 3 Sep 2013 22:16:52 +0000 (22:16 +0000)]
Add comments to the CFI instructions and reformat with clang-format. No functionality change.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@189864
91177308-0d34-0410-b5e6-
96231b3b80d8
Eric Christopher [Tue, 3 Sep 2013 22:08:28 +0000 (22:08 +0000)]
Fix copy and pasto with CMake files for unittest.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@189863
91177308-0d34-0410-b5e6-
96231b3b80d8
Eric Christopher [Tue, 3 Sep 2013 21:57:57 +0000 (21:57 +0000)]
Add a hashing routine that handles hashing types. Add a test for
hashing the contents of DW_FORM_data1 on top of a type with attributes.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@189862
91177308-0d34-0410-b5e6-
96231b3b80d8
Eric Christopher [Tue, 3 Sep 2013 21:57:50 +0000 (21:57 +0000)]
Sentences end with periods.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@189861
91177308-0d34-0410-b5e6-
96231b3b80d8
Nadav Rotem [Tue, 3 Sep 2013 21:33:17 +0000 (21:33 +0000)]
Enable late-vectorization by default.
This patch changes the default setting for the LateVectorization flag that controls where the loop-vectorizer is ran.
Perf gains:
SingleSource/Benchmarks/Shootout/matrix -37.33%
MultiSource/Benchmarks/PAQ8p/paq8p -22.83%
SingleSource/Benchmarks/Linpack/linpack-pc -16.22%
SingleSource/Benchmarks/Shootout-C++/ary3 -15.16%
MultiSource/Benchmarks/TSVC/NodeSplitting-flt/NodeSplitting-flt -10.34%
MultiSource/Benchmarks/TSVC/NodeSplitting-dbl/NodeSplitting-dbl -7.12%
Regressions:
SingleSource/Benchmarks/Misc/lowercase 15.10%
MultiSource/Benchmarks/TSVC/Equivalencing-flt/Equivalencing-flt 13.18%
SingleSource/Benchmarks/Shootout-C++/matrix 8.27%
SingleSource/Benchmarks/CoyoteBench/lpbench 7.30%
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@189858
91177308-0d34-0410-b5e6-
96231b3b80d8
Matt Arsenault [Tue, 3 Sep 2013 21:05:51 +0000 (21:05 +0000)]
Fix grammar
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@189855
91177308-0d34-0410-b5e6-
96231b3b80d8
Matt Arsenault [Tue, 3 Sep 2013 21:05:48 +0000 (21:05 +0000)]
Teach InstCombineLoadCast about address spaces.
This is another one that doesn't matter much,
but uses the right GEP index types in the first
place.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@189854
91177308-0d34-0410-b5e6-
96231b3b80d8
Matt Arsenault [Tue, 3 Sep 2013 21:05:15 +0000 (21:05 +0000)]
Use type form of getIntPtrType in alloca visitor.
This doesn't actually matter, since alloca is always
0 address space, but this is more consistent.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@189853
91177308-0d34-0410-b5e6-
96231b3b80d8
Bill Wendling [Tue, 3 Sep 2013 20:59:07 +0000 (20:59 +0000)]
WIP: Refactor some code so that it can be called by more than just one method. No functionality change.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@189849
91177308-0d34-0410-b5e6-
96231b3b80d8
Eric Christopher [Tue, 3 Sep 2013 20:51:14 +0000 (20:51 +0000)]
Alphabetize.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@189848
91177308-0d34-0410-b5e6-
96231b3b80d8
Joerg Sonnenberger [Tue, 3 Sep 2013 20:43:54 +0000 (20:43 +0000)]
Add a Python-like join function to merge a list of strings with a
separator between each two elements.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@189846
91177308-0d34-0410-b5e6-
96231b3b80d8
Jim Grosbach [Tue, 3 Sep 2013 20:08:17 +0000 (20:08 +0000)]
Revert "Revert "ARM: Improve pattern for isel mul of vector by scalar.""
This reverts commit r189648.
Fixes for the previously failing clang-side arm_neon_intrinsics test
cases will be checked in separately.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@189841
91177308-0d34-0410-b5e6-
96231b3b80d8
Eric Christopher [Tue, 3 Sep 2013 20:00:20 +0000 (20:00 +0000)]
Add the rest of the stock attributes to the attribute table.
This won't affect the kinds of hashes we test for as we actually
do hashing based on form and attribute. Change the fission-hash
testcase one last time to handle DW_AT_comp_dir.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@189840
91177308-0d34-0410-b5e6-
96231b3b80d8
Vincent Lejeune [Tue, 3 Sep 2013 19:43:28 +0000 (19:43 +0000)]
TableGen: Enumerate Schedule Model too.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@189839
91177308-0d34-0410-b5e6-
96231b3b80d8
Yi Jiang [Tue, 3 Sep 2013 17:26:04 +0000 (17:26 +0000)]
In this patch we are trying to do two things:
1) If the width of vectorization list candidate is bigger than vector reg width, we will break it down to fit the vector reg.
2) We do not vectorize the width which is not power of two.
The performance result shows it will help some spec benchmarks. mesa improved 6.97% and ammp improved 1.54%.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@189830
91177308-0d34-0410-b5e6-
96231b3b80d8
Joerg Sonnenberger [Tue, 3 Sep 2013 17:13:53 +0000 (17:13 +0000)]
Print string value for DT_RPATH and DT_RUNPATH.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@189829
91177308-0d34-0410-b5e6-
96231b3b80d8
Joey Gouly [Tue, 3 Sep 2013 16:18:19 +0000 (16:18 +0000)]
[MC] Revert part of my previous change, I was a bit overzealous.
A change to test the previous commit will be coming soon.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@189825
91177308-0d34-0410-b5e6-
96231b3b80d8
Eric Christopher [Tue, 3 Sep 2013 16:11:13 +0000 (16:11 +0000)]
Migrate the DWARF_VERSION constant to 4. It largely needs to go away.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@189824
91177308-0d34-0410-b5e6-
96231b3b80d8
Richard Sandiford [Tue, 3 Sep 2013 15:38:35 +0000 (15:38 +0000)]
[SystemZ] Add support for TMHH, TMHL, TMLH and TMLL
For now this just handles simple comparisons of an ANDed value with zero.
The CC value provides enough information to do any comparison for a
2-bit mask, and some nonzero comparisons with more populated masks,
but that's all future work.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@189819
91177308-0d34-0410-b5e6-
96231b3b80d8
Joey Gouly [Tue, 3 Sep 2013 15:03:36 +0000 (15:03 +0000)]
[MC] AvailableFeatures needs to be a uint64_t to match FeatureBits in MCSubtargetInfo.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@189815
91177308-0d34-0410-b5e6-
96231b3b80d8
Evgeniy Stepanov [Tue, 3 Sep 2013 13:05:29 +0000 (13:05 +0000)]
[msan] Fix handling of select with struct arguments.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@189796
91177308-0d34-0410-b5e6-
96231b3b80d8
Evgeniy Stepanov [Tue, 3 Sep 2013 10:04:11 +0000 (10:04 +0000)]
[msan] Fix select instrumentation.
Select condition shadow was being ignored resulting in false negatives.
This change OR-s sign-extended condition shadow into the result shadow.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@189785
91177308-0d34-0410-b5e6-
96231b3b80d8
Daniel Sanders [Tue, 3 Sep 2013 09:45:20 +0000 (09:45 +0000)]
[mips][msa] Added IntrNoMem and removed Commutative from sub intrinsics.
This changes the SelectionDAG nodes from ISD::INTRINSIC_W_CHAIN to
ISD::INTRINSIC_WO_CHAIN which enables easy lowering to equivalent SelectionDAG
nodes (e.g. __builtin_msa_sub_w -> ISD::SUB) in future patches since nodes
such as ISD::SUB do not have a chain.
It also corrects an obvious mistake, namely that the subtract intrinsics were
marked as being commutative.
As per a similar change in r189106
(http://llvm.org/viewvc/llvm-project?rev=189106&view=rev) there isn’t a new
testcase in this patch since the existing tests should test the intrinsics to
the same standard and the best I can do for a testcase would be a fragile
pass/maybe test of whether memory operations can (and do) cross the intrinsic.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@189784
91177308-0d34-0410-b5e6-
96231b3b80d8
Daniel Sanders [Tue, 3 Sep 2013 09:35:20 +0000 (09:35 +0000)]
[mips][msa] Added IntrNoMem to the floating-point intrinsics.
This changes the SelectionDAG nodes from ISD::INTRINSIC_W_CHAIN to
ISD::INTRINSIC_WO_CHAIN which enables easy lowering to equivalent SelectionDAG
nodes (e.g. __builtin_msa_fadd_w -> ISD::FADD) in future patches since nodes
such as ISD::FADD do not have a chain.
As per a similar change in r189106
(http://llvm.org/viewvc/llvm-project?rev=189106&view=rev) there isn’t a new
testcase in this patch since the existing tests should test the intrinsics to
the same standard and the best I can do for a testcase would be a fragile
pass/maybe test of whether memory operations can (and do) cross the intrinsic.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@189782
91177308-0d34-0410-b5e6-
96231b3b80d8
Venkatraman Govindaraju [Tue, 3 Sep 2013 04:11:59 +0000 (04:11 +0000)]
[Sparc] Add support for soft long double (fp128).
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@189780
91177308-0d34-0410-b5e6-
96231b3b80d8
Craig Topper [Tue, 3 Sep 2013 03:56:17 +0000 (03:56 +0000)]
Add hadSideEffects=0 to some instructions.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@189779
91177308-0d34-0410-b5e6-
96231b3b80d8
Venkatraman Govindaraju [Mon, 2 Sep 2013 18:32:45 +0000 (18:32 +0000)]
[Sparc] Implement spill and load for long double(f128) registers.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@189768
91177308-0d34-0410-b5e6-
96231b3b80d8
Tilmann Scheller [Mon, 2 Sep 2013 17:09:01 +0000 (17:09 +0000)]
ARM: Default to the Swift CPU when targeting armv7s/thumbv7s.
Test cases adjusted accordingly.
This fixes rdar://
14871821.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@189766
91177308-0d34-0410-b5e6-
96231b3b80d8
Tilmann Scheller [Mon, 2 Sep 2013 15:48:17 +0000 (15:48 +0000)]
Revert 189756 for now, it doesn't match what rdar://
14871821 really wants.
What we really want is to enable Swift by default for *v7s triples (and there already seems to be some logic which attempts to do that). In that case the iOS version doesn't matter.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@189763
91177308-0d34-0410-b5e6-
96231b3b80d8
Tilmann Scheller [Mon, 2 Sep 2013 12:01:58 +0000 (12:01 +0000)]
ARM: Default to Swift when compiling for iOS 6 or later.
Test cases adjusted accordingly.
This fixes rdar://
14871821.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@189756
91177308-0d34-0410-b5e6-
96231b3b80d8
NAKAMURA Takumi [Mon, 2 Sep 2013 12:00:53 +0000 (12:00 +0000)]
FileCheck-ize three tests of llvm/test/CodeGen/X86/h-register(s).
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@189755
91177308-0d34-0410-b5e6-
96231b3b80d8
NAKAMURA Takumi [Mon, 2 Sep 2013 12:00:46 +0000 (12:00 +0000)]
llvm/test/CodeGen/X86: Update tests with -mattr=-bmi not to take BMI, corresponding to Craig's r189742.
AMD Piledriver builder detected failures.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@189754
91177308-0d34-0410-b5e6-
96231b3b80d8
Alexey Samsonov [Mon, 2 Sep 2013 09:15:01 +0000 (09:15 +0000)]
[CMake] Add -O1 in debug builds with LLVM_USE_SANITIZER
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@189747
91177308-0d34-0410-b5e6-
96231b3b80d8
Craig Topper [Mon, 2 Sep 2013 07:53:17 +0000 (07:53 +0000)]
Create BEXTR instructions for (and ((sra or srl) x, imm), (2**size - 1)). Fixes PR17028.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@189742
91177308-0d34-0410-b5e6-
96231b3b80d8
Elena Demikhovsky [Mon, 2 Sep 2013 07:41:01 +0000 (07:41 +0000)]
AVX-512: updated the list of high-latency instructions.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@189740
91177308-0d34-0410-b5e6-
96231b3b80d8
Chandler Carruth [Mon, 2 Sep 2013 07:19:04 +0000 (07:19 +0000)]
Add a clang-format file so that the tool can automatically detect the
correct style for the LLVM tree as well as projects checked out under
projects and tools.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@189737
91177308-0d34-0410-b5e6-
96231b3b80d8
Elena Demikhovsky [Mon, 2 Sep 2013 07:12:29 +0000 (07:12 +0000)]
AVX-512: gather-scatter tests; added foldable instructions;
Specify GATHER/SCATTER as heavy instructions.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@189736
91177308-0d34-0410-b5e6-
96231b3b80d8
Elena Demikhovsky [Mon, 2 Sep 2013 06:40:09 +0000 (06:40 +0000)]
llvm interpreter: select, shuffle and insertelement instructions.
This patch implements vector support for select instruction and adds specific vector instructions : shuffle and insertelement. (tests are also included)
and functions lle_X_memset, lle_X_memcpy added.
Done by Veselov, Yuri (mailto:Yuri.Veselov@intel.com)
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@189735
91177308-0d34-0410-b5e6-
96231b3b80d8
Chandler Carruth [Mon, 2 Sep 2013 05:55:10 +0000 (05:55 +0000)]
Fix some rather confusing indentation and control flow in the errno
printing routine. This is made harder to see due to the surprising
formatting, inconsistent brace usage, and repeated conditions that all
test the same thing.
The only "consequence" of this bug is re-assigning 'str' to an empty
string when computing the error string for an error number of 0 in the
event of a non-GNU strerror_r routine. So, nothing to see here other
than cleanup. It did help me find PR17055 in clang-format though.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@189734
91177308-0d34-0410-b5e6-
96231b3b80d8
Dmitri Gribenko [Mon, 2 Sep 2013 01:18:56 +0000 (01:18 +0000)]
Added std:: qualifier to find() invocation
Iterator of std::vector may be implemented as a raw pointer. In
this case ADL does not find the find() function in the std namespace.
For example, this is the case with STDCXX implementation of vector.
Patch by Konstantin Tokarev.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@189733
91177308-0d34-0410-b5e6-
96231b3b80d8
Chandler Carruth [Sun, 1 Sep 2013 23:42:27 +0000 (23:42 +0000)]
Nuke the hilariously out of date suggestion to unpack llvm-gcc 4.2 as
part of getting started with LLVM.
The LLVM getting started document is in woeful need of attention. I may
get to some of this, but some random notes for folks interested:
1) We need to separate the getting started steps for folks who are
interested in the core LLVM libs and nothing else, folks interested
in a nifty C++ toolchain and nothing else, and folks interested in
both.
2) We should include documentation for both release archives, svn, and
git in equal portion, and we should document all of the various
repositories of interest: llvm, clang, clang-tools-extra,
compiler-rt, lld, libcxx, test-suite.
3) We should document the CMake build. We should probably document the
CMake build first, and give a fall-back set of docs for the Makefile
build for the use cases where that is still the preferred solution.
This would more closely match the use cases that folks in the open
source community are likely to have, and would remove a point of
discrepancy between Linux, Windows, and Mac instructions.
4) Probably a ton of other modernization stuff that I've not thought of
here.
Anyways, if anyone at all is interested, please help clean up this
document. It is much needed.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@189732
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Sun, 1 Sep 2013 15:48:08 +0000 (15:48 +0000)]
Revert r189704, which removed the guidance about not duplicating doc comments.
This is under active discussion.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@189730
91177308-0d34-0410-b5e6-
96231b3b80d8
Elena Demikhovsky [Sun, 1 Sep 2013 14:24:41 +0000 (14:24 +0000)]
AVX-512: Added GATHER and SCATTER instructions.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@189729
91177308-0d34-0410-b5e6-
96231b3b80d8
Charles Davis [Sun, 1 Sep 2013 04:28:48 +0000 (04:28 +0000)]
Move everything depending on Object/MachOFormat.h over to Support/MachO.h.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@189728
91177308-0d34-0410-b5e6-
96231b3b80d8
Reed Kotler [Sun, 1 Sep 2013 04:12:59 +0000 (04:12 +0000)]
Make sure we don't generate stubs for any of these functions because they
don't exist in libc. This is really not the right way to solve this problem;
but it's not clear to me at this time exactly what is the right way.
If we create stubs here, they will cause link errors because these functions
do not exist in libc.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@189727
91177308-0d34-0410-b5e6-
96231b3b80d8
Benjamin Kramer [Sat, 31 Aug 2013 21:20:04 +0000 (21:20 +0000)]
Mark an unreachable code path with llvm_unreachable. Pacifies GCC.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@189726
91177308-0d34-0410-b5e6-
96231b3b80d8
Benjamin Kramer [Sat, 31 Aug 2013 20:10:01 +0000 (20:10 +0000)]
Free PressureDiffs instead of leaking.
Found by valgrind.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@189725
91177308-0d34-0410-b5e6-
96231b3b80d8
Benjamin Kramer [Sat, 31 Aug 2013 18:19:35 +0000 (18:19 +0000)]
SimplifyLibCalls: When emitting an overloaded fp function check that it's available.
The existing code missed some edge cases when e.g. we're going to emit sqrtf but
only the availability of sqrt was checked. This happens on odd platforms like
windows.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@189724
91177308-0d34-0410-b5e6-
96231b3b80d8
Andrew Trick [Sat, 31 Aug 2013 05:17:58 +0000 (05:17 +0000)]
Fix my previous checkin to updatePressureDiffs.
There was one case that we could hit a DebugValue where I didn't think
to check. DebugValues are evil. No checkinable test case, sorry. It's
an obvious fix.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@189717
91177308-0d34-0410-b5e6-
96231b3b80d8
Bill Schmidt [Sat, 31 Aug 2013 02:33:40 +0000 (02:33 +0000)]
[PowerPC] Fast-isel cleanup patch.
Here are a few miscellaneous things to tidy up the PPC64 fast-isel
implementation. I corrected a couple of commentary lapses, and added
documentation of future opportunities. I also implemented
TargetMaterializeAlloca, which I somehow forgot when I split up the
original huge patch.
Finally, I decided to delete SelectCmp. I hadn't previously hooked it
in to TargetSelectInstruction(), and when I did I realized it wasn't
serving any useful purpose. This is only useful for compares that
don't feed a branch in the same block, and to handle that we would
have to have logic to interpret i1 as a condition register. This
could probably be done, but would require Unseemly Hackery, and
honestly does not seem worth the hassle.
This ends the current patch series.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@189715
91177308-0d34-0410-b5e6-
96231b3b80d8
Bill Schmidt [Fri, 30 Aug 2013 23:31:33 +0000 (23:31 +0000)]
[PowerPC] Add integer truncation support to fast-isel.
This is the last substantive patch I'm planning for fast-isel in the
near future, adding fast selection of integer truncates. There are
certainly more things that can be improved (many of which are called
out in FIXMEs), but for now we are catching most of the important
cases.
I'll document some of the remaining work in a cleanup patch shortly.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@189706
91177308-0d34-0410-b5e6-
96231b3b80d8
Bill Schmidt [Fri, 30 Aug 2013 23:25:30 +0000 (23:25 +0000)]
Correct partially defined variable
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@189705
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Fri, 30 Aug 2013 23:06:20 +0000 (23:06 +0000)]
Remove the suggestion to not duplicate comments in header and
implementation files. While doc generation systems don't need this,
humans do benefit from it. Not everyone reads all code through doxygen.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@189704
91177308-0d34-0410-b5e6-
96231b3b80d8
Bill Schmidt [Fri, 30 Aug 2013 22:18:55 +0000 (22:18 +0000)]
[PowerPC] Call support for fast-isel.
This patch adds fast-isel support for calls (but not intrinsic calls
or varargs calls). It also removes a badly-formed assert. There are
some new tests just for calls, and also for folding loads into
arguments on calls to avoid extra extends.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@189701
91177308-0d34-0410-b5e6-
96231b3b80d8
Richard Mitton [Fri, 30 Aug 2013 21:32:42 +0000 (21:32 +0000)]
Build fix
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@189699
91177308-0d34-0410-b5e6-
96231b3b80d8
Richard Mitton [Fri, 30 Aug 2013 21:19:48 +0000 (21:19 +0000)]
Fixed a bug where diassembling an instruction that had a prefix would cause LLVM to identify a 1-byte instruction, but then upon querying it for that 1-byte instruction would cause an undefined opcode.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@189698
91177308-0d34-0410-b5e6-
96231b3b80d8
Bill Wendling [Fri, 30 Aug 2013 21:07:33 +0000 (21:07 +0000)]
Compulsive reformatting.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@189697
91177308-0d34-0410-b5e6-
96231b3b80d8
Daniel Dunbar [Fri, 30 Aug 2013 20:39:21 +0000 (20:39 +0000)]
[conf] Add config variable to disable crash related overrides.
- We do some nasty things w.r.t. installing or overriding signal handlers in
order to improve our crash recovery support or interaction with crash
reporting software, and those things are not necessarily appropriate when
LLVM is being linked into a client application that has its own ideas about
how to do things. This gives those clients a way to disable that handling at
build time.
- Currently, the code this guards is all Apple specific, but other platforms
might have the same concerns so I went for a more generic configure
name. Someone who is more familiar with library embedding on Windows can
handle choosing which of the Windows/Signals.inc behaviors might make sense
to go under this flag.
- This also fixes the proper autoconf'ing of ENABLE_BACKTRACES. The code
expects it to be undefined when disabled, but the autoconf check was just
defining it to 0.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@189694
91177308-0d34-0410-b5e6-
96231b3b80d8
Reed Kotler [Fri, 30 Aug 2013 19:40:56 +0000 (19:40 +0000)]
Fix a problem with dual mips16/mips32 mode. When the underlying processor
has hard float, when you compile the mips32 code you have to make sure
that it knows to compile any mips32 routines as hard float. I need to clean
up the way mips16 hard float is specified but I need to first think through
all the details. Mips16 always has a form of soft float, the difference being
whether the underlying hardware has floating point. So it's not really
necessary to pass the -soft-float to llvm since soft-float is always true
for mips16 by virtue of the fact that it will not register floating point
registers. By using this fact, I can simplify the way this is all handled.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@189690
91177308-0d34-0410-b5e6-
96231b3b80d8
Andrew Trick [Fri, 30 Aug 2013 17:58:49 +0000 (17:58 +0000)]
Use LiveRangeQuery for instruction-level liveness queries.
Remove redundant or bug-prone LiveInterval APIs.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@189685
91177308-0d34-0410-b5e6-
96231b3b80d8
Bill Schmidt [Fri, 30 Aug 2013 15:18:11 +0000 (15:18 +0000)]
[PowerPC] Add handling for conversions to fast-isel.
Yet another chunk of fast-isel code. This one handles various
conversions involving floating-point. (It also includes some
miscellaneous handling throughout the back end for LWA_32 and LWAX_32
that should have been part of the load-store patch.)
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@189677
91177308-0d34-0410-b5e6-
96231b3b80d8