Daniel Dunbar [Tue, 8 May 2012 16:50:47 +0000 (16:50 +0000)]
[docs] Add support for building man pages using Sphinx.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@156386
91177308-0d34-0410-b5e6-
96231b3b80d8
Daniel Dunbar [Tue, 8 May 2012 16:50:43 +0000 (16:50 +0000)]
[docs] Integrate the command guide into the toctree.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@156385
91177308-0d34-0410-b5e6-
96231b3b80d8
Daniel Dunbar [Tue, 8 May 2012 16:50:35 +0000 (16:50 +0000)]
[docs] Add ReST version of all the man pages.
- The POD versions are slated for execution, but are still around until
llvm.org machinery is in place.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@156384
91177308-0d34-0410-b5e6-
96231b3b80d8
Nuno Lopes [Tue, 8 May 2012 16:16:20 +0000 (16:16 +0000)]
remove TYPE_CODE_FUNCTION_OLD type code. it is no longer in use and it was marked for removal in 3.0
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@156383
91177308-0d34-0410-b5e6-
96231b3b80d8
Jakob Stoklund Olesen [Tue, 8 May 2012 15:07:29 +0000 (15:07 +0000)]
s/CSR_Ghc/CSR_NoRegs/
Share the CalleeSavedRegs defs between all calling conventions having no
callee-saved registers.
Patch by Yiannis Tsiouris!
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@156382
91177308-0d34-0410-b5e6-
96231b3b80d8
NAKAMURA Takumi [Tue, 8 May 2012 14:31:52 +0000 (14:31 +0000)]
Lit: rewind WinWaitReleased() stuff in TestRunner.
r145222 "lit/TestRunner.py: [Win32] Introduce WinWaitReleased(f), to wait for file handles to be released by children."
r145223 "lit/TestRunner.py: Use RemoveForce()."
r145381 "lit/TestRunner.py: Try to catch ERROR_FILE_NOT_FOUND, too."
r152916 "lit/TestRunner.py: [Win32] Check all opened_files[] released, rather than (obsoleted) written_files[]."
r153172 "lit/TestRunner.py: [Win32] Rework WinWaitReleased() again! "win32file" from Python Win32 Extensions."
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@156381
91177308-0d34-0410-b5e6-
96231b3b80d8
NAKAMURA Takumi [Tue, 8 May 2012 14:31:46 +0000 (14:31 +0000)]
Windows/PathV2.inc: Retry rename() for (maximum) 2 seconds.
Files might be opend by system scanners (eg. file indexer, virus scanner, &c).
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@156380
91177308-0d34-0410-b5e6-
96231b3b80d8
Duncan Sands [Tue, 8 May 2012 12:16:05 +0000 (12:16 +0000)]
Calling ReassociateExpression recursively is extremely dangerous since it will
replace the operands of expressions with only one use with undef and generate
a new expression for the original without using RAUW to update the original.
Thus any copies of the original expression held in a vector may end up
referring to some bogus value - and using a ValueHandle won't help since there
is no RAUW. There is already a mechanism for getting the effect of recursion
non-recursively: adding the value to be recursed on to RedoInsts. But it wasn't
being used systematically. Have various places where recursion had snuck in at
some point use the RedoInsts mechanism instead. Fixes PR12169.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@156379
91177308-0d34-0410-b5e6-
96231b3b80d8
Stepan Dyatkovskiy [Tue, 8 May 2012 08:33:21 +0000 (08:33 +0000)]
Rejected r156374: Ordinary PR1255 patch. Due to clang-x86_64-debian-fnt buildbot failure.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@156377
91177308-0d34-0410-b5e6-
96231b3b80d8
Craig Topper [Tue, 8 May 2012 06:58:15 +0000 (06:58 +0000)]
Remove 256-bit AVX non-temporal store intrinsics. Similar was previously done for 128-bit.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@156375
91177308-0d34-0410-b5e6-
96231b3b80d8
Stepan Dyatkovskiy [Tue, 8 May 2012 06:36:08 +0000 (06:36 +0000)]
Ordinary patch for PR1255.
Added new case-ranges orientated methods for adding/removing cases in SwitchInst. After this patch cases will internally representated as ConstantArray-s instead of ConstantInt, externally cases wrapped within the ConstantRangesSet object.
Old methods of SwitchInst are also works well, but marked as deprecated. So on this stage we have no side effects except that I added support for case ranges in BitcodeReader/Writer, of course test for Bitcode is also added. Old "switch" format is also supported.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@156374
91177308-0d34-0410-b5e6-
96231b3b80d8
Andrew Trick [Tue, 8 May 2012 02:52:09 +0000 (02:52 +0000)]
Allow NULL LoopPassManager argument in UnrollLoop. PR12734.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@156358
91177308-0d34-0410-b5e6-
96231b3b80d8
Jakob Stoklund Olesen [Tue, 8 May 2012 00:08:35 +0000 (00:08 +0000)]
Extract methods for joining physregs.
No functional change.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@156345
91177308-0d34-0410-b5e6-
96231b3b80d8
Jakob Stoklund Olesen [Mon, 7 May 2012 23:46:16 +0000 (23:46 +0000)]
Naming convention and whitespace. No functional change.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@156342
91177308-0d34-0410-b5e6-
96231b3b80d8
Jakob Stoklund Olesen [Mon, 7 May 2012 22:57:55 +0000 (22:57 +0000)]
Coalesce subreg-subreg copies.
At least some of them:
%vreg1:sub_16bit = COPY %vreg2:sub_16bit; GR64:%vreg1, GR32: %vreg2
Previously, we couldn't figure out that the above copy could be
eliminated by coalescing %vreg2 with %vreg1:sub_32bit.
The new getCommonSuperRegClass() hook makes it possible.
This is not very useful yet since the unmodified part of the destination
register usually interferes with the source register. The coalescer
needs to understand sub-register interference checking first.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@156334
91177308-0d34-0410-b5e6-
96231b3b80d8
Pete Cooper [Mon, 7 May 2012 22:42:40 +0000 (22:42 +0000)]
Remove C Backend from the bugpoint docs
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@156333
91177308-0d34-0410-b5e6-
96231b3b80d8
Jakob Stoklund Olesen [Mon, 7 May 2012 22:10:26 +0000 (22:10 +0000)]
Add an MF argument to TRI::getPointerRegClass() and TII::getRegClass().
The getPointerRegClass() hook can return register classes that depend on
the calling convention of the current function (ptr_rc_tailcall).
So far, we have been able to infer the calling convention from the
subtarget alone, but as we add support for multiple calling conventions
per target, that no longer works.
Patch by Yiannis Tsiouris!
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@156328
91177308-0d34-0410-b5e6-
96231b3b80d8
Jakob Stoklund Olesen [Mon, 7 May 2012 21:59:31 +0000 (21:59 +0000)]
Fix bug in TRI::getCommonSuperRegClass().
Test cases for this code are coming. It is not used for anything yet.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@156327
91177308-0d34-0410-b5e6-
96231b3b80d8
Owen Anderson [Mon, 7 May 2012 20:51:25 +0000 (20:51 +0000)]
Teach DAG combine to fold x-x to 0.0 when unsafe FP math is enabled.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@156324
91177308-0d34-0410-b5e6-
96231b3b80d8
Owen Anderson [Mon, 7 May 2012 20:47:23 +0000 (20:47 +0000)]
Teach reassociate to commute FMul's and FAdd's in order to canonicalize the order of their operands across instructions. This allows for greater CSE opportunities.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@156323
91177308-0d34-0410-b5e6-
96231b3b80d8
Preston Gurd [Mon, 7 May 2012 19:38:40 +0000 (19:38 +0000)]
Make IntelJITEvents and OProfileJIT as optional libraries and add
optional library support to the llvm-build tool:
- Add new command line parameter to llvm-build: “--enable-optional-libraries”
- Add handing of new llvm-build library type “OptionalLibrary”
- Update Cmake and automake build systems to pass correct flags to llvm-build
based on configuration
Patch by Dan Malea!
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@156319
91177308-0d34-0410-b5e6-
96231b3b80d8
Jordy Rose [Mon, 7 May 2012 19:24:40 +0000 (19:24 +0000)]
Constify (trivially) ImmutableSet::iterator::getVisitState().
This was probably intended all along.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@156318
91177308-0d34-0410-b5e6-
96231b3b80d8
Jakob Stoklund Olesen [Mon, 7 May 2012 19:14:58 +0000 (19:14 +0000)]
Add TRI::getCommonSuperRegClass().
This function is a generalization of getMatchingSuperRegClass() to the
symmetric case where both sides are using a sub-register index. It will
find a super-register class and sub-register indexes that make this
diagram commute:
PreA
SuperRC ----------> RCA
| |
| |
PreB | | SubA
| |
| |
V V
RCB ----------> SubRC
SubB
This can be used to coalesce copies like:
%vreg1:sub16 = COPY %vreg2:sub16; GR64:%vreg1, GR32: %vreg2
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@156317
91177308-0d34-0410-b5e6-
96231b3b80d8
Chad Rosier [Mon, 7 May 2012 18:47:44 +0000 (18:47 +0000)]
Fix a regression from r147481. This combine should only happen if there is a
single use.
rdar://
11360370
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@156316
91177308-0d34-0410-b5e6-
96231b3b80d8
Matt Beaumont-Gay [Mon, 7 May 2012 18:12:42 +0000 (18:12 +0000)]
Don't assume size_t is unsigned long long.
Fixes a -Woverflow warning from gcc when building for 32-bit platforms.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@156313
91177308-0d34-0410-b5e6-
96231b3b80d8
Manman Ren [Mon, 7 May 2012 18:06:23 +0000 (18:06 +0000)]
X86: optimization for -(x != 0)
This patch will optimize -(x != 0) on X86
FROM
cmpl $0x01,%edi
sbbl %eax,%eax
notl %eax
TO
negl %edi
sbbl %eax %eax
In order to generate negl, I added patterns in Target/X86/X86InstrCompiler.td:
def : Pat<(X86sub_flag 0, GR32:$src), (NEG32r GR32:$src)>;
rdar:
10961709
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@156312
91177308-0d34-0410-b5e6-
96231b3b80d8
Eric Christopher [Mon, 7 May 2012 06:25:19 +0000 (06:25 +0000)]
Add support for the 'x' constraint.
Patch by Jack Carter.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@156295
91177308-0d34-0410-b5e6-
96231b3b80d8
Eric Christopher [Mon, 7 May 2012 06:25:15 +0000 (06:25 +0000)]
Add support for the 'l' constraint.
Patch by Jack Carter.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@156294
91177308-0d34-0410-b5e6-
96231b3b80d8
Eric Christopher [Mon, 7 May 2012 06:25:10 +0000 (06:25 +0000)]
Add support for the 'c' constraint.
Patch by Jack Carter.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@156293
91177308-0d34-0410-b5e6-
96231b3b80d8
Eric Christopher [Mon, 7 May 2012 06:25:02 +0000 (06:25 +0000)]
Add support for the 'P' constraint.
Patch by Jack Carter.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@156292
91177308-0d34-0410-b5e6-
96231b3b80d8
John McCall [Mon, 7 May 2012 06:00:23 +0000 (06:00 +0000)]
Fix trivial typo in llvm_move.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@156288
91177308-0d34-0410-b5e6-
96231b3b80d8
Craig Topper [Mon, 7 May 2012 06:00:15 +0000 (06:00 +0000)]
Fix some issues in the f16c instructions.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@156287
91177308-0d34-0410-b5e6-
96231b3b80d8
Eric Christopher [Mon, 7 May 2012 05:46:48 +0000 (05:46 +0000)]
Add support for the 'O' constraint.
Patch by Jack Carter.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@156285
91177308-0d34-0410-b5e6-
96231b3b80d8
Eric Christopher [Mon, 7 May 2012 05:46:43 +0000 (05:46 +0000)]
Add support for the 'N' inline asm constraint.
Patch by Jack Carter.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@156284
91177308-0d34-0410-b5e6-
96231b3b80d8
Eric Christopher [Mon, 7 May 2012 05:46:37 +0000 (05:46 +0000)]
Add support for the 'L' inline asm constraint.
Patch by Jack Carter.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@156283
91177308-0d34-0410-b5e6-
96231b3b80d8
Eric Christopher [Mon, 7 May 2012 05:46:29 +0000 (05:46 +0000)]
Add support for the inline asm constraint 'K'.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@156282
91177308-0d34-0410-b5e6-
96231b3b80d8
Craig Topper [Mon, 7 May 2012 05:36:19 +0000 (05:36 +0000)]
Add SSE4A MOVNTSS/MOVNTSD instructions.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@156281
91177308-0d34-0410-b5e6-
96231b3b80d8
Eric Christopher [Mon, 7 May 2012 03:13:42 +0000 (03:13 +0000)]
Support the 'J' constraint.
Patch by Jack Carter.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@156280
91177308-0d34-0410-b5e6-
96231b3b80d8
Eric Christopher [Mon, 7 May 2012 03:13:32 +0000 (03:13 +0000)]
Add support for the 'I' inline asm constraint. Also add tests
from the previous 2 patches.
Patch by Jack Carter.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@156279
91177308-0d34-0410-b5e6-
96231b3b80d8
Eric Christopher [Mon, 7 May 2012 03:13:22 +0000 (03:13 +0000)]
Allow 64 bit integer values in gpu registers if arch and abi are 64 bit.
Patch by Jack Carter.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@156278
91177308-0d34-0410-b5e6-
96231b3b80d8
Eric Christopher [Mon, 7 May 2012 03:13:16 +0000 (03:13 +0000)]
When using inline asm constraints representing
non-floating point general registers allow 8 and 16-bit
elements.
Patch by Jack Carter.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@156277
91177308-0d34-0410-b5e6-
96231b3b80d8
Jim Grosbach [Mon, 7 May 2012 02:25:53 +0000 (02:25 +0000)]
Tidy up. Whitespace.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@156276
91177308-0d34-0410-b5e6-
96231b3b80d8
Craig Topper [Sun, 6 May 2012 19:46:21 +0000 (19:46 +0000)]
Use MVT instead of EVT as the argument to all the shuffle decode functions. Simplify some of the decode functions.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@156268
91177308-0d34-0410-b5e6-
96231b3b80d8
Craig Topper [Sun, 6 May 2012 18:54:26 +0000 (18:54 +0000)]
Add VPERMQ/VPERMPD to the list of target specific shuffles that can be looked through for DAG combine purposes.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@156266
91177308-0d34-0410-b5e6-
96231b3b80d8
Craig Topper [Sun, 6 May 2012 18:44:02 +0000 (18:44 +0000)]
Add shuffle decode support for VPERMQ/VPERMPD.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@156265
91177308-0d34-0410-b5e6-
96231b3b80d8
Jim Grosbach [Sun, 6 May 2012 17:33:14 +0000 (17:33 +0000)]
TableGen: AsmMatcher diagnostic when missing instruction mnemonic.
Previously, if an instruction definition was missing the mnemonic,
the next line would just assert(). Issue a real diagnostic instead.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@156263
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Sun, 6 May 2012 16:20:49 +0000 (16:20 +0000)]
make SourceMgr tolerate empty SMLoc()'s better.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@156260
91177308-0d34-0410-b5e6-
96231b3b80d8
Benjamin Kramer [Sun, 6 May 2012 14:25:16 +0000 (14:25 +0000)]
Switch the select to branch transformation on by default.
The primitive conservative heuristic seems to give a slight overall
improvement while not regressing stuff. Make it available to wider
testing. If you notice any speed regressions (or significant code
size regressions) let me know!
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@156258
91177308-0d34-0410-b5e6-
96231b3b80d8
Jakub Staszak [Sun, 6 May 2012 13:52:31 +0000 (13:52 +0000)]
Remove trailing spaces.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@156257
91177308-0d34-0410-b5e6-
96231b3b80d8
NAKAMURA Takumi [Sun, 6 May 2012 08:24:24 +0000 (08:24 +0000)]
Unix/Process.inc: Give more useful random seed to srand. Workaround for PR12743.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@156252
91177308-0d34-0410-b5e6-
96231b3b80d8
NAKAMURA Takumi [Sun, 6 May 2012 08:24:18 +0000 (08:24 +0000)]
Support/Process: Move llvm::sys::Process::GetRandomNumber() from Process.cpp to Unix/Process.inc.
FIXME: GetRandomNumber() is not implemented in Win32.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@156251
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Sat, 5 May 2012 22:17:32 +0000 (22:17 +0000)]
reapply my patch, with a fix for an off-by-one error. Turned out to be a lot
of work for a drive-by fix :)
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@156246
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Sat, 5 May 2012 22:11:04 +0000 (22:11 +0000)]
revert my patches, which are causing problems.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@156245
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Sat, 5 May 2012 22:04:11 +0000 (22:04 +0000)]
add missing header <shame>
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@156244
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Sat, 5 May 2012 21:39:51 +0000 (21:39 +0000)]
refactor some code to expose column numbers more and make diagnostic printing slightly more efficient.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@156243
91177308-0d34-0410-b5e6-
96231b3b80d8
Jim Grosbach [Sat, 5 May 2012 17:45:12 +0000 (17:45 +0000)]
Nuke a few dead remnants of the CBE.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@156241
91177308-0d34-0410-b5e6-
96231b3b80d8
Daniel Dunbar [Sat, 5 May 2012 16:49:11 +0000 (16:49 +0000)]
[Support] Add missing include.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@156240
91177308-0d34-0410-b5e6-
96231b3b80d8
Daniel Dunbar [Sat, 5 May 2012 16:39:22 +0000 (16:39 +0000)]
[Support] Fix up comments.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@156239
91177308-0d34-0410-b5e6-
96231b3b80d8
Daniel Dunbar [Sat, 5 May 2012 16:36:24 +0000 (16:36 +0000)]
[Support] Rewrite sys::fs::unique_file to not be stupid with /dev/urandom.
- Just use sys::Process::GetRandomNumber instead of having two poor
implementations.
- This is ~70 times (!) faster on my OS X machine.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@156238
91177308-0d34-0410-b5e6-
96231b3b80d8
Daniel Dunbar [Sat, 5 May 2012 16:36:20 +0000 (16:36 +0000)]
[Support] Add sys::Process::GetRandomNumber().
- Primitive API, but we rarely have need for random numbers.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@156237
91177308-0d34-0410-b5e6-
96231b3b80d8
Daniel Dunbar [Sat, 5 May 2012 16:36:16 +0000 (16:36 +0000)]
[build] Add build check for ::arc4random().
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@156236
91177308-0d34-0410-b5e6-
96231b3b80d8
Benjamin Kramer [Sat, 5 May 2012 15:02:39 +0000 (15:02 +0000)]
Update all outdated autoconf files in the sample project.
We might just use symlinks here, but I'm afraid of possible portability issues.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@156235
91177308-0d34-0410-b5e6-
96231b3b80d8
Benjamin Kramer [Sat, 5 May 2012 12:49:22 +0000 (12:49 +0000)]
CodeGenPrepare: Add a transform to turn selects into branches in some cases.
This came up when a change in block placement formed a cmov and slowed down a
hot loop by 50%:
ucomisd (%rdi), %xmm0
cmovbel %edx, %esi
cmov is a really bad choice in this context because it doesn't get branch
prediction. If we emit it as a branch, an out-of-order CPU can do a better job
(if the branch is predicted right) and avoid waiting for the slow load+compare
instruction to finish. Of course it won't help if the branch is unpredictable,
but those are really rare in practice.
This patch uses a dumb conservative heuristic, it turns all cmovs that have one
use and a direct memory operand into branches. cmovs usually save some code
size, so we disable the transform in -Os mode. In-Order architectures are
unlikely to benefit as well, those are included in the
"predictableSelectIsExpensive" flag.
It would be better to reuse branch probability info here, but BPI doesn't
support select instructions currently. It would make sense to use the same
heuristics as the if-converter pass, which does the opposite direction of this
transform.
Test suite shows a small improvement here and there on corei7-level machines,
but the actual results depend a lot on the used microarchitecture. The
transformation is currently disabled by default and available by passing the
-enable-cgp-select2branch flag to the code generator.
Thanks to Chandler for the initial test case to him and Evan Cheng for providing
me with comments and test-suite numbers that were more stable than mine :)
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@156234
91177308-0d34-0410-b5e6-
96231b3b80d8
Benjamin Kramer [Sat, 5 May 2012 12:49:14 +0000 (12:49 +0000)]
Add a new target hook "predictableSelectIsExpensive".
This will be used to determine whether it's profitable to turn a select into a
branch when the branch is likely to be predicted.
Currently enabled for everything but Atom on X86 and Cortex-A9 devices on ARM.
I'm not entirely happy with the name of this flag, suggestions welcome ;)
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@156233
91177308-0d34-0410-b5e6-
96231b3b80d8
Benjamin Kramer [Sat, 5 May 2012 11:22:02 +0000 (11:22 +0000)]
NVPTX: Initialize the UseF32FTZ flag.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@156232
91177308-0d34-0410-b5e6-
96231b3b80d8
Stepan Dyatkovskiy [Sat, 5 May 2012 07:09:40 +0000 (07:09 +0000)]
Small fix in InstCombineCasts.cpp. Restored "alloca + bitcast" reducing for case when alloca's size is calculated within the "add/sub/... nsw".
Also added fix to 2011-06-13-nsw-alloca.ll test.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@156231
91177308-0d34-0410-b5e6-
96231b3b80d8
Eric Christopher [Sat, 5 May 2012 01:16:06 +0000 (01:16 +0000)]
Typo.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@156226
91177308-0d34-0410-b5e6-
96231b3b80d8
Jakob Stoklund Olesen [Fri, 4 May 2012 23:12:22 +0000 (23:12 +0000)]
Order register classes by spill size first, members last.
This is still a topological ordering such that every register class gets
a smaller enum value than its sub-classes.
Placing the smaller spill sizes first makes a difference for the
super-register class bit masks. When looking for a super-register class,
we usually want the smallest possible kind of super-register. That is
now available as the first bit set in the bit mask.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@156222
91177308-0d34-0410-b5e6-
96231b3b80d8
Jakob Stoklund Olesen [Fri, 4 May 2012 22:53:28 +0000 (22:53 +0000)]
Make sure findRepresentativeClass picks the widest super-register.
We want the representative register class to contain the largest
super-registers available. This makes the function less sensitive to the
register class numbering.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@156220
91177308-0d34-0410-b5e6-
96231b3b80d8
Jakob Stoklund Olesen [Fri, 4 May 2012 22:53:26 +0000 (22:53 +0000)]
Remove extra comma in debug output.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@156219
91177308-0d34-0410-b5e6-
96231b3b80d8
David Blaikie [Fri, 4 May 2012 22:34:16 +0000 (22:34 +0000)]
Fix warnings in release build.
This fixes a couple of Clang warnings in release builds of LLVM:
* Missing return in ISelLowering
* Unused variable in NVPTXutil.cpp
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@156216
91177308-0d34-0410-b5e6-
96231b3b80d8
Kevin Enderby [Fri, 4 May 2012 22:09:52 +0000 (22:09 +0000)]
Tweak to the fix in r156212, as with the change in removing the shift the
SignExtend32<22>(Val<<1) also needs to change to SignExtend32<21>(Val) .
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@156213
91177308-0d34-0410-b5e6-
96231b3b80d8
Kevin Enderby [Fri, 4 May 2012 22:02:27 +0000 (22:02 +0000)]
Fix a bug in the ARM disassembler for wide branch conditional instructions
where the symbolic operand's displacement was incorrectly shifted left by 1.
rdar://
11387046
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@156212
91177308-0d34-0410-b5e6-
96231b3b80d8
Chandler Carruth [Fri, 4 May 2012 21:35:49 +0000 (21:35 +0000)]
Fix a Clang warning in the new NVPTX backend:
In file included from ../lib/Target/NVPTX/VectorElementize.cpp:53:
../lib/Target/NVPTX/NVPTX.h:44:3: warning: default label in switch which covers all enumeration values [-Wcovered-switch-default]
default: assert(0 && "Unknown condition code");
^
1 warning generated.
The prevailing pattern in LLVM is to not use a default label, and instead to
use llvm_unreachable to denote that the switch in fact covers all return paths
from the function.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@156209
91177308-0d34-0410-b5e6-
96231b3b80d8
Chandler Carruth [Fri, 4 May 2012 21:33:30 +0000 (21:33 +0000)]
Teach the code extractor how to extract a sequence of blocks from
RegionInfo's RegionNode. This mirrors the logic for automating the
extraction from a Loop.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@156208
91177308-0d34-0410-b5e6-
96231b3b80d8
Chandler Carruth [Fri, 4 May 2012 20:55:23 +0000 (20:55 +0000)]
Rename the Region::block_iterator to Region::block_node_iterator, and
add a new Region::block_iterator which actually iterates over the basic
blocks of the region.
The old iterator, now call 'block_node_iterator' iterates over
RegionNodes which contain a single basic block. This works well with the
GraphTraits-based iterator design, however most users actually want an
iterator over the BasicBlocks inside these RegionNodes. Now the
'block_iterator' is a wrapper which exposes exactly this interface.
Internally it uses the block_node_iterator to walk all nodes which are
single basic blocks, but transparently unwraps the basic block to make
user code simpler.
While this patch is a bit of a wash, most of the updates are to internal
users, not external users of the RegionInfo. I have an accompanying
patch to Polly that is a strict simplification of every user of this
interface, and I'm working on a pass that also wants the same simplified
interface.
This patch alone should have no functional impact.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@156202
91177308-0d34-0410-b5e6-
96231b3b80d8
Justin Holewinski [Fri, 4 May 2012 20:18:50 +0000 (20:18 +0000)]
This patch adds a new NVPTX back-end to LLVM which supports code generation for NVIDIA PTX 3.0. This back-end will (eventually) replace the current PTX back-end, while maintaining compatibility with it.
The new target machines are:
nvptx (old ptx32) => 32-bit PTX
nvptx64 (old ptx64) => 64-bit PTX
The sources are based on the internal NVIDIA NVPTX back-end, and
contain more functionality than the current PTX back-end currently
provides.
NV_CONTRIB
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@156196
91177308-0d34-0410-b5e6-
96231b3b80d8
Sebastian Pop [Fri, 4 May 2012 19:53:56 +0000 (19:53 +0000)]
Added missing CMN case in Thumb2SizeReduction pass so that LLVM emits 16-bits encoding of CMN instructions.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@156195
91177308-0d34-0410-b5e6-
96231b3b80d8
Preston Gurd [Fri, 4 May 2012 19:26:37 +0000 (19:26 +0000)]
Adds Intel Atom scheduling latencies to X86InstrSystem.td.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@156194
91177308-0d34-0410-b5e6-
96231b3b80d8
Matt Beaumont-Gay [Fri, 4 May 2012 18:34:27 +0000 (18:34 +0000)]
Pacify GCC's -Wreturn-type
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@156189
91177308-0d34-0410-b5e6-
96231b3b80d8
Chandler Carruth [Fri, 4 May 2012 11:20:27 +0000 (11:20 +0000)]
Factor the computation of input and output sets into a public interface
of the CodeExtractor utility. This allows speculatively computing input
and output sets to measure the likely size impact of the code
extraction.
These sets cannot be reused sadly -- we mutate the function prior to
forming the final sets used by the actual extraction.
The interface has been revamped slightly to make it easier to use
correctly by making the interface const and sinking the computation of
the number of exit blocks into the full extraction function and away
from the rest of this logic which just computed two output parameters.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@156168
91177308-0d34-0410-b5e6-
96231b3b80d8
Chandler Carruth [Fri, 4 May 2012 11:17:06 +0000 (11:17 +0000)]
Rather than trying to gracefully handle input sequences with repeated
blocks, assert that this doesn't happen. We don't want to bother trying
to support this call pattern as it isn't necessary.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@156167
91177308-0d34-0410-b5e6-
96231b3b80d8
Chandler Carruth [Fri, 4 May 2012 11:14:19 +0000 (11:14 +0000)]
Fix a goof with my previous commit by completely returning when we
detect an in-eligible block rather than just breaking out of the loop.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@156166
91177308-0d34-0410-b5e6-
96231b3b80d8
Chandler Carruth [Fri, 4 May 2012 10:26:45 +0000 (10:26 +0000)]
Hoist a safety assert from the extraction method into the construction
of the extractor itself.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@156164
91177308-0d34-0410-b5e6-
96231b3b80d8
Chandler Carruth [Fri, 4 May 2012 10:18:49 +0000 (10:18 +0000)]
Move the CodeExtractor utility to a dedicated header file / source file,
and expose it as a utility class rather than as free function wrappers.
The simple free-function interface works well for the bugpoint-specific
pass's uses of code extraction, but in an upcoming patch for more
advanced code extraction, they simply don't expose a rich enough
interface. I need to expose various stages of the process of doing the
code extraction and query information to decide whether or not to
actually complete the extraction or give up.
Rather than build up a new predicate model and pass that into these
functions, just take the class that was actually implementing the
functions and lift it up into a proper interface that can be used to
perform code extraction. The interface is cleaned up and re-documented
to work better in a header. It also is now setup to accept the blocks to
be extracted in the constructor rather than in a method.
In passing this essentially reverts my previous commit here exposing
a block-level query for eligibility of extraction. That is no longer
necessary with the more rich interface as clients can query the
extraction object for eligibility directly. This will reduce the number
of walks of the input basic block sequence by quite a bit which is
useful if this enters the normal optimization pipeline.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@156163
91177308-0d34-0410-b5e6-
96231b3b80d8
Hans Wennborg [Fri, 4 May 2012 09:40:39 +0000 (09:40 +0000)]
Make ARM and Mips use TargetMachine::getTLSModel()
This moves the logic for selecting a TLS model to a single place,
instead of the previous three (ARM, Mips, and X86 which already
uses this function).
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@156162
91177308-0d34-0410-b5e6-
96231b3b80d8
Craig Topper [Fri, 4 May 2012 06:39:13 +0000 (06:39 +0000)]
Fix some loops to match coding standards. No functional change intended.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@156159
91177308-0d34-0410-b5e6-
96231b3b80d8
Craig Topper [Fri, 4 May 2012 06:18:33 +0000 (06:18 +0000)]
Fix up some spacing. No functional change.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@156158
91177308-0d34-0410-b5e6-
96231b3b80d8
Craig Topper [Fri, 4 May 2012 05:49:51 +0000 (05:49 +0000)]
Simplify broadcast lowering code. No functional change intended.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@156157
91177308-0d34-0410-b5e6-
96231b3b80d8
Craig Topper [Fri, 4 May 2012 04:44:49 +0000 (04:44 +0000)]
Allow v16i16 and v32i8 shuffles to be rewritten as narrower shuffles.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@156156
91177308-0d34-0410-b5e6-
96231b3b80d8
Bill Wendling [Fri, 4 May 2012 04:22:32 +0000 (04:22 +0000)]
Add 'landingpad' instructions to the list of instructions to ignore.
Also combine the code in the 'assert' statement.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@156155
91177308-0d34-0410-b5e6-
96231b3b80d8
Craig Topper [Fri, 4 May 2012 04:08:44 +0000 (04:08 +0000)]
Simplify shuffle narrowing code a bit. No functional change intended.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@156154
91177308-0d34-0410-b5e6-
96231b3b80d8
Jakob Stoklund Olesen [Fri, 4 May 2012 03:30:34 +0000 (03:30 +0000)]
Remove the SubRegClasses field from RegisterClass descriptions.
This information in now computed by TableGen.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@156152
91177308-0d34-0410-b5e6-
96231b3b80d8
Jakob Stoklund Olesen [Fri, 4 May 2012 03:30:28 +0000 (03:30 +0000)]
Remove TargetRegisterClass::SuperRegClasses.
This manually enumerated list of super-register classes has been
superceeded by the automatically computed super-register class masks
available through SuperRegClassIterator.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@156151
91177308-0d34-0410-b5e6-
96231b3b80d8
Rafael Espindola [Fri, 4 May 2012 03:23:36 +0000 (03:23 +0000)]
Pass -fcolor-diagnostics when it is supported. This makes a difference when
using cmake+ninja, since ninja buffers the compiler output.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@156150
91177308-0d34-0410-b5e6-
96231b3b80d8
Jakob Stoklund Olesen [Fri, 4 May 2012 02:19:22 +0000 (02:19 +0000)]
Use SuperRegClassIterator for findRepresentativeClass().
The masks returned by SuperRegClassIterator are computed automatically
by TableGen. This is better than depending on the manually specified
SuperRegClasses.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@156147
91177308-0d34-0410-b5e6-
96231b3b80d8
Jakob Stoklund Olesen [Fri, 4 May 2012 02:16:39 +0000 (02:16 +0000)]
Initialize SparcInstrInfo before SparcTargetLowering.
The TargetLowering construction needs to use a valid TargetRegisterInfo
instance.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@156146
91177308-0d34-0410-b5e6-
96231b3b80d8
Jakob Stoklund Olesen [Fri, 4 May 2012 01:48:29 +0000 (01:48 +0000)]
Add a SuperRegClassIterator class.
This iterator class provides a more abstract interface to the (Idx,
Mask) lists of super-registers for a register class. The layout of the
tables shouldn't be exposed to clients.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@156144
91177308-0d34-0410-b5e6-
96231b3b80d8
Chandler Carruth [Fri, 4 May 2012 00:58:03 +0000 (00:58 +0000)]
A pile of long over-due refactorings here. There are some very, *very*
minor behavior changes with this, but nothing I have seen evidence of in
the wild or expect to be meaningful. The real goal is unifying our logic
and simplifying the interfaces. A summary of the changes follows:
- Make 'callIsSmall' actually accept a callsite so it can handle
intrinsics, and simplify callers appropriately.
- Nuke a completely bogus declaration of 'callIsSmall' that was still
lurking in InlineCost.h... No idea how this got missed.
- Teach the 'isInstructionFree' about the various more intelligent
'free' heuristics that got added to the inline cost analysis during
review and testing. This mostly surrounds int->ptr and ptr->int casts.
- Switch most of the interesting parts of the inline cost analysis that
were essentially computing 'is this instruction free?' to use the code
metrics routine instead. This way we won't keep duplicating logic.
All of this is motivated by the desire to allow other passes to compute
a roughly equivalent 'cost' metric for a particular basic block as the
inline cost analysis. Sadly, re-using the same analysis for both is
really messy because only the actual inline cost analysis is ever going
to go to the contortions required for simplification, SROA analysis,
etc.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@156140
91177308-0d34-0410-b5e6-
96231b3b80d8
Chandler Carruth [Thu, 3 May 2012 23:38:34 +0000 (23:38 +0000)]
Add a FoldingSetVector datastructure which is analogous to a SetVector,
but using a FoldingSet underneath and with a largely compatible
interface to that of FoldingSet. This can be used anywhere a FoldingSet
would be natural, but iteration order is significant. The initial
intended use case is in Clang's template specialization lists to
preserve instantiation order iteration.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@156131
91177308-0d34-0410-b5e6-
96231b3b80d8