RPCS3/llvm-mirror
1
0
Fork 0
mirror of https://github.com/RPCS3/llvm-mirror.git synced 2025-01-13 15:11:39 +00:00
Code Issues Actions Packages Projects Releases Wiki Activity
89,365 Commits 4 Branches 5 Tags
Commit Graph

89365 Commits

This Branch
This Branch
All Branches
Author SHA1 Message Date
Pekka Jaaskelainen
7e2908d0f3 Metadata for annotating loops as parallel. The first consumer for this 
metadata is the loop vectorizer.

See the documentation update for more info.

llvm-svn: 175060
 2013-02-13 18:08:57 +00:00
Krzysztof Parzyszek
b6d2a1c1ee Add registration for PPC-specific passes to allow the IR to be dumped 
via -print-after-all.

llvm-svn: 175058
 2013-02-13 17:40:07 +00:00
Benjamin Kramer
34ab81b7fa X86: Disable generation of rep;movsl when %esi is used as a base pointer. 
This happens when there is both stack realignment and a dynamic alloca in the
function. If we overwrite %esi (rep;movsl uses fixed registers) we'll lose the
base pointer and the next register spill will write into oblivion.

Fixes PR15249 and unbreaks firefox on i386/freebsd. Mozilla uses dynamic allocas
and freebsd a 4 byte stack alignment.

llvm-svn: 175057
 2013-02-13 13:40:35 +00:00
Tim Northover
6f37a3070c Mention AArch64 in release notes for 3.3 
llvm-svn: 175055
 2013-02-13 12:46:32 +00:00
Evgeniy Stepanov
be7d518073 Fix MSan annotations inclusion guard. 
#cmakedefine does not behave the way I though it was.

llvm-svn: 175050
 2013-02-13 10:40:01 +00:00
Bill Wendling
8d3d898234 Use array_pod_sort. 
llvm-svn: 175048
 2013-02-13 09:26:26 +00:00
Manuel Klimek
1cd92e6900 Update phab docs to clarify how to accept a change. 
llvm-svn: 175047
 2013-02-13 09:07:18 +00:00
Bill Wendling
45be8c5d89 Add some accessor and query methods for retrieving Attribute objects and such. 
llvm-svn: 175046
 2013-02-13 08:42:21 +00:00
Reed Kotler
49229780c8 Make jumptables work for -static 
llvm-svn: 175044
 2013-02-13 08:32:14 +00:00
Elena Demikhovsky
a4a4bded4d Prevent insertion of "vzeroupper" before call that preserves YMM registers, since a caller uses preserved registers across the call. 
llvm-svn: 175043
 2013-02-13 08:02:04 +00:00
Craig Topper
dc6619d1a6 Remove unnecessary condtional assignment. The next line ignores the result of the assignment with the same condition. 
llvm-svn: 175042
 2013-02-13 07:44:17 +00:00
Eric Christopher
5821e56ae7 This is actually located at the end, not the middle. 
llvm-svn: 175041
 2013-02-13 07:22:25 +00:00
Eric Christopher
a2c85e433f Check i1 as well as i8 variables for 8 bit registers for x86 inline 
assembly.

llvm-svn: 175036
 2013-02-13 06:01:05 +00:00
Eric Christopher
25984ae5af Finish obviously broken thought. 
llvm-svn: 175035
 2013-02-13 06:01:00 +00:00
Kostya Serebryany
2d4f4b284c [tsan] disable load widening in ThreadSanitizer mode 
llvm-svn: 175034
 2013-02-13 05:59:45 +00:00
Kostya Serebryany
c6f889301b [asan] fix confusing indentation 
llvm-svn: 175033
 2013-02-13 05:14:12 +00:00
Chris Lattner
4eca27141d use memcpy instead of dubious union to type pun two values, 
thanks to David Blaike for pointing this out.

llvm-svn: 175032
 2013-02-13 04:53:40 +00:00
Eric Christopher
49b50f690c Fix comment. 
llvm-svn: 175024
 2013-02-13 02:29:18 +00:00
Manman Ren
7c3626e9c0 Debug Info: LiveDebugVarible can remove DBG_VALUEs, make sure we emit them back. 
RegisterCoalescer used to depend on LiveDebugVariable. LDV removes DBG_VALUEs
without emitting them at the end.

We fix this by removing LDV from RegisterCoalescer. Also add an assertion to
make sure we call emitDebugValues if DBG_VALUEs are removed at
runOnMachineFunction.

rdar://problem/13183203
Reviewed by Andy & Jakob

llvm-svn: 175023
 2013-02-13 01:14:49 +00:00
Chad Rosier
271b9e136a [ms-inline-asm] Make sure the AsmRewrite list is sorted in lexical order. 
rdar://13202662

llvm-svn: 175021
 2013-02-13 01:03:13 +00:00
David Peixotto
84c964ec93 Test commit. Fixed typo. 
llvm-svn: 175020
 2013-02-13 00:36:35 +00:00
Arnaud A. de Grandmaison
2c730cd330 Fix comment 
visitSExt is an adapted copy of the related visitZExt method, so adapt the comment accordingly.

llvm-svn: 175019
 2013-02-13 00:19:19 +00:00
Michael Gottesman
da8220bd2f Changed isStoredObjCPointer => IsStoredObjCPointer. No functionality change. 
llvm-svn: 175017
 2013-02-12 23:35:08 +00:00
Bill Wendling
42ec85cc86 Add a default empty string to the 'value' of a string attribute. 
llvm-svn: 175016
 2013-02-12 23:14:31 +00:00
Dan Gohman
1ea13bd49f Actually delete this code, since it's really not clear what it's 
trying to do.

llvm-svn: 175014
 2013-02-12 22:26:41 +00:00
Chad Rosier
fccc6d00f0 [ms-inline-asm] Implement align directive (which is roughly equivalent to .align). 
Also, allow _EMIT and __EMIT for the emit directive.  We already do the same
for TYPE, SIZE, and LENGTH.
rdar://13200215

llvm-svn: 175008
 2013-02-12 21:33:51 +00:00
Jack Carter
78a165943b This patch just fixes up various llvm formatting 
violations such as tabs, blanks at eol and long 
lines.


 

llvm-svn: 175007
 2013-02-12 21:29:39 +00:00
Guy Benyei
92dac48079 Add static cast to unsigned char whenever a character classification function is called with a signed char argument, in order to avoid assertions in Windows Debug configuration. 
llvm-svn: 175006
 2013-02-12 21:21:59 +00:00
Dan Gohman
a04e614c0f Record PRE predecessors with a SmallVector instead of a DenseMap, and 
avoid a second pred_iterator traversal.

llvm-svn: 175001
 2013-02-12 19:49:10 +00:00
Chad Rosier
3eb029b4c7 [ms-inline asm] Pass the length of the IDVal, so we can do a proper AsmRewrite. 
llvm-svn: 174999
 2013-02-12 19:42:32 +00:00
Chad Rosier
24b5149dc9 [ms-inline asm] Accept the emit directive as either _emit or __emit. 
llvm-svn: 174998
 2013-02-12 19:31:23 +00:00
Dan Gohman
79869cd16c When disabling PRE for a value is directly redundant with itself 
(through a loop), don't continue to iterate through the reamining
predecessors.

llvm-svn: 174994
 2013-02-12 19:05:10 +00:00
Krzysztof Parzyszek
9d40323823 Remove target-specific info from the testcase for DWARF/pubnames. 
llvm-svn: 174992
 2013-02-12 18:53:21 +00:00
Dan Gohman
6815bf69b1 Check that pointers are removed from maps before calling delete on the pointers, 
for tidiness' sake.

llvm-svn: 174988
 2013-02-12 18:44:43 +00:00
Dan Gohman
69f4c8b640 Minor code simplification. 
llvm-svn: 174985
 2013-02-12 18:38:36 +00:00
Chad Rosier
5fc603a34e [ms-inline asm] Add support for lexing binary integers with a [bB] suffix. 
This is complicated by backward labels (e.g., 0b can be both a backward label
and a binary zero).  The current implementation assumes [0-9]b is always a
label and thus it's possible for 0b and 1b to not be interpreted correctly for
ms-style inline assembly.  However, this is relatively simple to fix in the
inline assembly (i.e., drop the [bB]).

This patch also limits backward labels to [0-9]b, so that only 0b and 1b are
ambiguous.
Part of rdar://12470373

llvm-svn: 174983
 2013-02-12 18:29:02 +00:00
Dmitri Gribenko
e7a571489a Documentation: HowToUseAttributes: formatting (use monospaced font) 
llvm-svn: 174982
 2013-02-12 18:26:08 +00:00
Krzysztof Parzyszek
bb3116f965 Allow optionally generating pubnames section in DWARF info. Introduce 
option "generate-dwarf-pubnames" to control it, set to "false" by default.

llvm-svn: 174981
 2013-02-12 18:00:14 +00:00
Kay Tiong Khoo
63416d2726 added test cases for r174920 (prefetch disassembly) 
llvm-svn: 174979
 2013-02-12 17:07:44 +00:00
Sergei Larin
e598e9d599 Equal treatment of labels and other terminators in MI DAG construction. 
MI sched DAG construction allows targets to include terminators into scheduling DAG.
Extend this functionality to labels as well.

llvm-svn: 174977
 2013-02-12 16:36:03 +00:00
Krzysztof Parzyszek
de39fad3d2 Add support for the pubnames section to llvm-dwarfdump. 
llvm-svn: 174976
 2013-02-12 16:20:28 +00:00
Paul Redmond
7d75c870f1 Fix the lit test added in r174972 
Patch by: Kevin Schoedel

llvm-svn: 174974
 2013-02-12 16:07:27 +00:00
Jyotsna Verma
abd979fd30 Hexagon: Add support to generate predicated absolute addressing mode 
instructions.

llvm-svn: 174973
 2013-02-12 16:06:23 +00:00
Paul Redmond
add1aa5259 PR14562 - Truncation of left shift became undef 
DAGCombiner::ReduceLoadWidth was converting (trunc i32 (shl i64 v, 32))
into (shl i32 v, 32) into undef. To prevent this, check the shift count
against the final result size.

Patch by: Kevin Schoedel
Reviewed by: Nadav Rotem

llvm-svn: 174972
 2013-02-12 15:21:21 +00:00
Justin Holewinski
9a248309f0 [NVPTX] Disable vector registers 
Vectors were being manually scalarized by the backend.  Instead,
let the target-independent code do all of the work.  The manual
scalarization was from a time before good target-independent support
for scalarization in LLVM. However, this forces us to specially-handle
vector loads and stores, which we can turn into PTX instructions that
produce/consume multiple operands.

llvm-svn: 174968
 2013-02-12 14:18:49 +00:00
Alexander Potapenko
aa2cd19415 [ASan] Do not use kDefaultShort64bitShadowOffset on Mac, where the binaries may get mapped at 0x100000000+ and thus may interleave with the shadow. 
llvm-svn: 174964
 2013-02-12 12:41:12 +00:00
Michel Danzer
6e93c3c0af R600: Fix regression with shadow array sampler on pre-SI GPUs. 
'R600/SI: Use proper instructions for array/shadow samplers.' removed two
cases from TEX_SHADOW. Vincent Lejeune reported on IRC that this broke some
shadow array piglit tests with the r600g driver. Reinstating the removed
cases should fix this, and still works with radeonsi as well.

I will follow up with some lit tests which would have caught the regression.

NOTE: This is a candidate for the Mesa stable branch.

Tested-by: Vincent Lejeune <vljn@ovi.com>
Reviewed-by: Tom Stellard <thomas.stellard@amd.com>
llvm-svn: 174963
 2013-02-12 12:11:23 +00:00
Joe Abbey
2a0a9b8280 Adding a HowTo for Attributes. 
This is based on Bill Wendling's email.  No additional content has been added,
but now there's a place for Attributes to capture future information.

llvm-svn: 174961
 2013-02-12 11:45:22 +00:00
Kostya Serebryany
4709ce39c2 [asan] fix tests for the new ABI 
llvm-svn: 174959
 2013-02-12 11:14:24 +00:00
Kostya Serebryany
b7c0f19f8a [asan] change the default mapping offset on x86_64 to 0x7fff8000. This gives roughly 5% speedup. Since this is an ABI change, bump the asan ABI version by renaming __asan_init to __asan_init_v1. llvm part, compiler-rt part will follow 
llvm-svn: 174957
 2013-02-12 11:11:02 +00:00