Benjamin Kramer [Thu, 31 Mar 2011 10:12:22 +0000 (10:12 +0000)]
InstCombine: fold fcmp (fneg x), (fneg y) -> fcmp x, y
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@128627
91177308-0d34-0410-b5e6-
96231b3b80d8
Benjamin Kramer [Thu, 31 Mar 2011 10:12:15 +0000 (10:12 +0000)]
InstCombine: fold fcmp pred (fneg x), C -> fcmp swap(pred) x, -C
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@128626
91177308-0d34-0410-b5e6-
96231b3b80d8
Benjamin Kramer [Thu, 31 Mar 2011 10:12:07 +0000 (10:12 +0000)]
InstCombine: Shrink "fcmp (fpext x), C" to "fcmp x, C" if C can be losslessly converted to the type of x.
Fixes PR9592.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@128625
91177308-0d34-0410-b5e6-
96231b3b80d8
Benjamin Kramer [Thu, 31 Mar 2011 10:11:58 +0000 (10:11 +0000)]
InstCombine: fold fcmp (fpext x), (fpext y) -> fcmp x, y.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@128624
91177308-0d34-0410-b5e6-
96231b3b80d8
Duncan Sands [Thu, 31 Mar 2011 10:03:32 +0000 (10:03 +0000)]
Will not compile without the spec!
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@128623
91177308-0d34-0410-b5e6-
96231b3b80d8
Duncan Sands [Thu, 31 Mar 2011 09:58:51 +0000 (09:58 +0000)]
Strip trailing whitespace.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@128622
91177308-0d34-0410-b5e6-
96231b3b80d8
Bill Wendling [Thu, 31 Mar 2011 08:13:57 +0000 (08:13 +0000)]
Testcase for r128619 (PR9571).
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@128620
91177308-0d34-0410-b5e6-
96231b3b80d8
Jakob Stoklund Olesen [Thu, 31 Mar 2011 03:54:44 +0000 (03:54 +0000)]
Pick a conservative register class when creating a small live range for remat.
The rematerialized instruction may require a more constrained register class
than the register being spilled. In the test case, the spilled register has been
inflated to the DPR register class, but we are rematerializing a load of the
ssub_0 sub-register which only exists for DPR_VFP2 registers.
The register class is reinflated after spilling, so the conservative choice is
only temporary.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@128610
91177308-0d34-0410-b5e6-
96231b3b80d8
Matt Beaumont-Gay [Thu, 31 Mar 2011 00:39:16 +0000 (00:39 +0000)]
Revert "- Implement asm parsing support for LDRT, LDRBT, STRT, STRBT and"
This revision introduced a dependency cycle, as nlewycky mentioned by email.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@128597
91177308-0d34-0410-b5e6-
96231b3b80d8
Nick Lewycky [Thu, 31 Mar 2011 00:23:57 +0000 (00:23 +0000)]
Fix typo in generated HTML.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@128594
91177308-0d34-0410-b5e6-
96231b3b80d8
Bob Wilson [Thu, 31 Mar 2011 00:09:35 +0000 (00:09 +0000)]
Use intrinsics for Neon vmull operations. Radar
9208957.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@128591
91177308-0d34-0410-b5e6-
96231b3b80d8
Owen Anderson [Wed, 30 Mar 2011 23:45:29 +0000 (23:45 +0000)]
Somehow we managed to forget to encode the lane index for a large swathe of NEON instructions. With this fix, the entire test-suite passes with the Thumb integrated assembler.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@128587
91177308-0d34-0410-b5e6-
96231b3b80d8
Evan Cheng [Wed, 30 Mar 2011 23:44:13 +0000 (23:44 +0000)]
Don't try to create zero-sized stack objects.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@128586
91177308-0d34-0410-b5e6-
96231b3b80d8
Bruno Cardoso Lopes [Wed, 30 Mar 2011 23:32:32 +0000 (23:32 +0000)]
- Implement asm parsing support for LDRT, LDRBT, STRT, STRBT and
{STR,LDC}{2}_PRE.
- Fixed the encoding in some places.
- Some of those instructions were using am2offset and now use addrmode2.
Codegen isn't affected, instructions which use SelectAddrMode2Offset were not
touched.
- Teach printAddrMode2Operand to check by the addressing mode which index
mode to print.
- This is a work in progress, more work to come. The idea is to change places
which use am2offset to use addrmode2 instead, as to unify assembly parser.
- Add testcases for assembly parser
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@128585
91177308-0d34-0410-b5e6-
96231b3b80d8
Cameron Zwarich [Wed, 30 Mar 2011 23:01:21 +0000 (23:01 +0000)]
Add a ARM-specific SD node for VBSL so that forms with a constant first operand
can be recognized. This fixes <rdar://problem/
9183078>.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@128584
91177308-0d34-0410-b5e6-
96231b3b80d8
Jim Grosbach [Wed, 30 Mar 2011 22:38:13 +0000 (22:38 +0000)]
Tidy up. Whitespace and 80-columns.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@128583
91177308-0d34-0410-b5e6-
96231b3b80d8
Bill Wendling [Wed, 30 Mar 2011 21:37:19 +0000 (21:37 +0000)]
* The DSE code that tested for overlapping needed to take into account the fact
that one of the numbers is signed while the other is unsigned. This could lead
to a wrong result when the signed was promoted to an unsigned int.
* Add the data layout line to the testcase so that it will test the appropriate
thing.
Patch by David Terei!
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@128577
91177308-0d34-0410-b5e6-
96231b3b80d8
Akira Hatanaka [Wed, 30 Mar 2011 21:15:35 +0000 (21:15 +0000)]
fixed typo
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@128574
91177308-0d34-0410-b5e6-
96231b3b80d8
Jakob Stoklund Olesen [Wed, 30 Mar 2011 18:32:53 +0000 (18:32 +0000)]
Don't add the same analysis implementation pair twice.
Prevent infinite growth of the list.
Patch by José Fonseca!
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@128568
91177308-0d34-0410-b5e6-
96231b3b80d8
Jakob Stoklund Olesen [Wed, 30 Mar 2011 18:32:51 +0000 (18:32 +0000)]
Reset StringMap's NumTombstones on clears and rehashes.
StringMap was not properly updating NumTombstones after a clear or rehash.
This was not fatal until now because the table was growing faster than
NumTombstones could, but with the previous change of preventing infinite
growth of the table the invariant (NumItems + NumTombstones <= NumBuckets)
stopped being observed, causing infinite loops in certain situations.
Patch by José Fonseca!
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@128567
91177308-0d34-0410-b5e6-
96231b3b80d8
Jakob Stoklund Olesen [Wed, 30 Mar 2011 18:32:48 +0000 (18:32 +0000)]
Prevent infinite growth of SmallPtrSet instances.
Rehash but don't grow when full of tombstones.
Patch by José Fonseca!
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@128566
91177308-0d34-0410-b5e6-
96231b3b80d8
Jakob Stoklund Olesen [Wed, 30 Mar 2011 18:32:44 +0000 (18:32 +0000)]
Prevent infinite growth of SmallMap instances.
Rehash but don't grow when full of tombstones.
Patch by José Fonseca!
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@128565
91177308-0d34-0410-b5e6-
96231b3b80d8
Jakob Stoklund Olesen [Wed, 30 Mar 2011 18:32:41 +0000 (18:32 +0000)]
Prevent infinite growth of the DenseMap.
When the hash function uses object pointers all free entries eventually
become tombstones as they are used at least once, regardless of the size.
DenseMap cannot function with zero empty keys, so it double size to get
get ridof the tombstones.
However DenseMap never shrinks automatically unless it is cleared, so
the net result is that certain tables grow infinitely.
The solution is to make a fresh copy of the table without tombstones
instead of doubling size, by simply calling grow with the current size.
Patch by José Fonseca!
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@128564
91177308-0d34-0410-b5e6-
96231b3b80d8
Jakob Stoklund Olesen [Wed, 30 Mar 2011 18:14:07 +0000 (18:14 +0000)]
Fix evil VirtRegRewriter bug.
The rewriter can keep track of multiple stack slots in the same register if they
happen to have the same value. When an instruction modifies a stack slot by
defining a register that is mapped to a stack slot, other stack slots in that
register are no longer valid.
This is a very rare problem, and I don't have a simple test case. I get the
impression that VirtRegRewriter knows it is about to be deleted, inventing a
last opaque problem.
<rdar://problem/
9204040>
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@128562
91177308-0d34-0410-b5e6-
96231b3b80d8
Jakob Stoklund Olesen [Wed, 30 Mar 2011 18:14:04 +0000 (18:14 +0000)]
Teach VirtRegRewriter about the new virtual register numbers. No functional change.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@128561
91177308-0d34-0410-b5e6-
96231b3b80d8
Benjamin Kramer [Wed, 30 Mar 2011 17:02:54 +0000 (17:02 +0000)]
Avoid turning a floating point division with a constant power of two into a denormal multiplication.
Some platforms may treat denormals as zero, on other platforms multiplication
with a subnormal is slower than dividing by a normal.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@128555
91177308-0d34-0410-b5e6-
96231b3b80d8
Benjamin Kramer [Wed, 30 Mar 2011 15:42:35 +0000 (15:42 +0000)]
InstCombine: If the divisor of an fdiv has an exact inverse, turn it into an fmul.
Fixes PR9587.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@128546
91177308-0d34-0410-b5e6-
96231b3b80d8
Benjamin Kramer [Wed, 30 Mar 2011 15:42:27 +0000 (15:42 +0000)]
Add APFloat::getExactInverse.
The idea is, that if an ieee 754 float is divided by a power of two, we can
turn the division into a cheaper multiplication. This function sees if we can
get an exact multiplicative inverse for a divisor and returns it if possible.
This is the hard part of PR9587.
I tested many inputs against llvm-gcc's frotend implementation of this
optimization and didn't find any difference. However, floating point is the
land of weird edge cases, so any review would be appreciated.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@128545
91177308-0d34-0410-b5e6-
96231b3b80d8
Jay Foad [Wed, 30 Mar 2011 15:31:02 +0000 (15:31 +0000)]
Fix more zero length memset warnings.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@128543
91177308-0d34-0410-b5e6-
96231b3b80d8
Jay Foad [Wed, 30 Mar 2011 13:29:06 +0000 (13:29 +0000)]
Add a comment on PHINode::Create().
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@128540
91177308-0d34-0410-b5e6-
96231b3b80d8
Jay Foad [Wed, 30 Mar 2011 11:28:46 +0000 (11:28 +0000)]
Remove PHINode::reserveOperandSpace(). Instead, add a parameter to
PHINode::Create() giving the (known or expected) number of operands.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@128537
91177308-0d34-0410-b5e6-
96231b3b80d8
Jay Foad [Wed, 30 Mar 2011 11:19:20 +0000 (11:19 +0000)]
(Almost) always call reserveOperandSpace() on newly created PHINodes.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@128535
91177308-0d34-0410-b5e6-
96231b3b80d8
Bob Wilson [Wed, 30 Mar 2011 05:26:51 +0000 (05:26 +0000)]
Revert a minor comment change inadvertently included with r128502.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@128526
91177308-0d34-0410-b5e6-
96231b3b80d8
Jakob Stoklund Olesen [Wed, 30 Mar 2011 02:52:39 +0000 (02:52 +0000)]
Treat clones the same as their origin.
When DCE clones a live range because it separates into connected components,
make sure that the clones enter the same register allocator stage as the
register they were cloned from.
For instance, clones may be split even when they where created during spilling.
Other registers created during spilling are not candidates for splitting or even
(re-)spilling.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@128524
91177308-0d34-0410-b5e6-
96231b3b80d8
Bill Wendling [Wed, 30 Mar 2011 01:03:48 +0000 (01:03 +0000)]
Remove dead code.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@128519
91177308-0d34-0410-b5e6-
96231b3b80d8
Johnny Chen [Wed, 30 Mar 2011 01:02:06 +0000 (01:02 +0000)]
Add a test case for thumb stc2 instruction.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@128517
91177308-0d34-0410-b5e6-
96231b3b80d8
Argyrios Kyrtzidis [Wed, 30 Mar 2011 00:22:00 +0000 (00:22 +0000)]
ClangSAEmClangSACheckersEmitter, emit info about groups.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@128515
91177308-0d34-0410-b5e6-
96231b3b80d8
Bill Wendling [Tue, 29 Mar 2011 23:31:06 +0000 (23:31 +0000)]
Set the unnamed_addr only when we're creating a new GV in the dest module.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@128507
91177308-0d34-0410-b5e6-
96231b3b80d8
Bill Wendling [Tue, 29 Mar 2011 23:28:02 +0000 (23:28 +0000)]
Revert r128501. It caused test failures.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@128506
91177308-0d34-0410-b5e6-
96231b3b80d8
Jim Grosbach [Tue, 29 Mar 2011 23:20:22 +0000 (23:20 +0000)]
Tidy up. 80 columns and trailing whitespace.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@128504
91177308-0d34-0410-b5e6-
96231b3b80d8
Francois Pichet [Tue, 29 Mar 2011 23:18:51 +0000 (23:18 +0000)]
Update CMake link dependency.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@128503
91177308-0d34-0410-b5e6-
96231b3b80d8
Evan Cheng [Tue, 29 Mar 2011 23:06:19 +0000 (23:06 +0000)]
Add intrinsics @llvm.arm.neon.vmulls and @llvm.arm.neon.vmullu.* back. Frontends
was lowering them to sext / uxt + mul instructions. Unfortunately the
optimization passes may hoist the extensions out of the loop and separate them.
When that happens, the long multiplication instructions can be broken into
several scalar instructions, causing significant performance issue.
Note the vmla and vmls intrinsics are not added back. Frontend will codegen them
as intrinsics vmull* + add / sub. Also note the isel optimizations for catching
mul + sext / zext are not changed either.
First part of rdar://
8832507, rdar://
9203134
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@128502
91177308-0d34-0410-b5e6-
96231b3b80d8
Bill Wendling [Tue, 29 Mar 2011 23:05:41 +0000 (23:05 +0000)]
We need to copy over the unnamed_addr attribute.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@128501
91177308-0d34-0410-b5e6-
96231b3b80d8
Matt Beaumont-Gay [Tue, 29 Mar 2011 22:42:41 +0000 (22:42 +0000)]
Revert accidental change to LLVMLibDeps.cmake
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@128499
91177308-0d34-0410-b5e6-
96231b3b80d8
Matt Beaumont-Gay [Tue, 29 Mar 2011 22:25:36 +0000 (22:25 +0000)]
Quiet a gcc warning about changed name lookup rules
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@128497
91177308-0d34-0410-b5e6-
96231b3b80d8
Benjamin Kramer [Tue, 29 Mar 2011 22:06:41 +0000 (22:06 +0000)]
InstCombine: Add a few missing combines for ANDs and ORs of sign bit tests.
On x86 we now compile "if (a < 0 && b < 0)" into
testl %edi, %esi
js IF.THEN
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@128496
91177308-0d34-0410-b5e6-
96231b3b80d8
Kevin Enderby [Tue, 29 Mar 2011 21:54:10 +0000 (21:54 +0000)]
Adding a test for "-inf" as well.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@128495
91177308-0d34-0410-b5e6-
96231b3b80d8
Johnny Chen [Tue, 29 Mar 2011 21:52:02 +0000 (21:52 +0000)]
Add a test case for MSRi.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@128494
91177308-0d34-0410-b5e6-
96231b3b80d8
Cameron Zwarich [Tue, 29 Mar 2011 21:41:55 +0000 (21:41 +0000)]
Add Neon SINT_TO_FP and UINT_TO_FP lowering from v4i16 to v4f32. Fixes
<rdar://problem/
8875309> and <rdar://problem/
9057191>.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@128492
91177308-0d34-0410-b5e6-
96231b3b80d8
Jakob Stoklund Olesen [Tue, 29 Mar 2011 21:20:19 +0000 (21:20 +0000)]
Recompute register class and hint for registers created during spilling.
The spill weight is not recomputed for an unspillable register - it stays infinite.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@128490
91177308-0d34-0410-b5e6-
96231b3b80d8
Argyrios Kyrtzidis [Tue, 29 Mar 2011 21:16:19 +0000 (21:16 +0000)]
In ClangSACheckersEmitter:
- Also emit a list of packages and groups sorted by name
- Avoid iterating over DenseSet so that the output of the arrays is deterministic.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@128489
91177308-0d34-0410-b5e6-
96231b3b80d8
Kevin Enderby [Tue, 29 Mar 2011 21:11:52 +0000 (21:11 +0000)]
Added support symbolic floating point constants in the MC assembler for Infinity
and Nans with the same strings as GAS supports. rdar://
8673024
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@128488
91177308-0d34-0410-b5e6-
96231b3b80d8
Johnny Chen [Tue, 29 Mar 2011 21:09:30 +0000 (21:09 +0000)]
Add a thumb test file for printf (iOS 4.3).
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@128487
91177308-0d34-0410-b5e6-
96231b3b80d8
Jim Grosbach [Tue, 29 Mar 2011 21:03:05 +0000 (21:03 +0000)]
Instantiate a JITMemoryManager for MCJIT Dyld
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@128485
91177308-0d34-0410-b5e6-
96231b3b80d8
Oscar Fuentes [Tue, 29 Mar 2011 20:51:08 +0000 (20:51 +0000)]
Fixed the build of Clang's unit tests on MinGW. Also removed some
unnecesary conditionals and introduced a new convenience function.
The problem was that the list of libraries for Clang's unit tests was
<clang libraries> <system libraries> <llvm libraries>. As the llvm
libraries references symbols defined on the system libraries, those
were reported as undefined.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@128484
91177308-0d34-0410-b5e6-
96231b3b80d8
Benjamin Kramer [Tue, 29 Mar 2011 20:28:57 +0000 (20:28 +0000)]
DSE: Remove an early exit optimization that depended on the ordering of a SmallPtrSet.
Fixes PR9569 and will hopefully make selfhost on ASLR-enabled systems more deterministic.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@128482
91177308-0d34-0410-b5e6-
96231b3b80d8
Owen Anderson [Tue, 29 Mar 2011 20:27:38 +0000 (20:27 +0000)]
Check early if this is an unsupported opcode, so that we can avoid needlessly instantiating the base register in some cases.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@128481
91177308-0d34-0410-b5e6-
96231b3b80d8
Johnny Chen [Tue, 29 Mar 2011 19:49:38 +0000 (19:49 +0000)]
A8.6.188 STC, STC2
The STC_OPTION and STC2_OPTION instructions should have their coprocessor option enclosed in {}.
rdar://problem/
9200661
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@128478
91177308-0d34-0410-b5e6-
96231b3b80d8
Johnny Chen [Tue, 29 Mar 2011 19:10:06 +0000 (19:10 +0000)]
Rename invalid-VLDMSDB-arm.txt to be invalid-VLDMSDB_UPD-arm.txt.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@128477
91177308-0d34-0410-b5e6-
96231b3b80d8
Johnny Chen [Tue, 29 Mar 2011 19:08:52 +0000 (19:08 +0000)]
Add and modify some tests.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@128476
91177308-0d34-0410-b5e6-
96231b3b80d8
Argyrios Kyrtzidis [Tue, 29 Mar 2011 18:53:00 +0000 (18:53 +0000)]
For ClangSACheckersEmitter, allow a package to belong to checker group, in which all its checkers will go into the group.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@128474
91177308-0d34-0410-b5e6-
96231b3b80d8
Jakob Stoklund Olesen [Tue, 29 Mar 2011 17:47:02 +0000 (17:47 +0000)]
Remember to use the correct register when rematerializing for snippets.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@128469
91177308-0d34-0410-b5e6-
96231b3b80d8
Jakob Stoklund Olesen [Tue, 29 Mar 2011 17:47:00 +0000 (17:47 +0000)]
Run dead code elimination immediately after rematerialization.
This may eliminate some uses of the spilled registers, and we don't want to
insert reloads for that.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@128468
91177308-0d34-0410-b5e6-
96231b3b80d8
Owen Anderson [Tue, 29 Mar 2011 17:42:25 +0000 (17:42 +0000)]
Add safety check that didn't show up in testing.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@128467
91177308-0d34-0410-b5e6-
96231b3b80d8
Devang Patel [Tue, 29 Mar 2011 17:27:08 +0000 (17:27 +0000)]
Document llvm.dbg.sp, llvm.dbg.gv and llvm.dbg.enum
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@128466
91177308-0d34-0410-b5e6-
96231b3b80d8
Bill Wendling [Tue, 29 Mar 2011 17:12:55 +0000 (17:12 +0000)]
Inline check that's used only once.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@128465
91177308-0d34-0410-b5e6-
96231b3b80d8
Owen Anderson [Tue, 29 Mar 2011 16:45:53 +0000 (16:45 +0000)]
Get rid of the non-writeback versions VLDMDB and VSTMDB, which don't actually exist.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@128461
91177308-0d34-0410-b5e6-
96231b3b80d8
Bill Wendling [Tue, 29 Mar 2011 07:28:52 +0000 (07:28 +0000)]
Rework the logic (and removing the bad check for an unreachable block) so that
the FailBB dominator is correctly calculated. Believe it or not, there isn't a
functionality change here.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@128455
91177308-0d34-0410-b5e6-
96231b3b80d8
Cameron Zwarich [Tue, 29 Mar 2011 05:19:52 +0000 (05:19 +0000)]
Do some simple copy propagation through integer loads and stores when promoting
vector types. This helps a lot with inlined functions when using the ARM soft
float ABI. Fixes <rdar://problem/
9184212>.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@128453
91177308-0d34-0410-b5e6-
96231b3b80d8
Bill Wendling [Tue, 29 Mar 2011 05:15:48 +0000 (05:15 +0000)]
Don't try to add stack protector logic to a dead basic block. It messes up
dominator information.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@128452
91177308-0d34-0410-b5e6-
96231b3b80d8
Bill Wendling [Tue, 29 Mar 2011 04:28:26 +0000 (04:28 +0000)]
Spruce up the error output.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@128451
91177308-0d34-0410-b5e6-
96231b3b80d8
Jakob Stoklund Olesen [Tue, 29 Mar 2011 03:12:04 +0000 (03:12 +0000)]
Handle the special case when all uses follow the last split point.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@128450
91177308-0d34-0410-b5e6-
96231b3b80d8
Jakob Stoklund Olesen [Tue, 29 Mar 2011 03:12:02 +0000 (03:12 +0000)]
Properly enable rematerialization when spilling after live range splitting.
The instruction to be rematerialized may not be the one defining the register
that is being spilled. The traceSiblingValue() function sees through sibling
copies to find the remat candidate.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@128449
91177308-0d34-0410-b5e6-
96231b3b80d8
Daniel Dunbar [Tue, 29 Mar 2011 02:30:34 +0000 (02:30 +0000)]
C-API: Include DataTypes.h instead of stdint.h.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@128446
91177308-0d34-0410-b5e6-
96231b3b80d8
Rafael Espindola [Tue, 29 Mar 2011 02:18:54 +0000 (02:18 +0000)]
Reduce test case.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@128445
91177308-0d34-0410-b5e6-
96231b3b80d8
Evan Cheng [Tue, 29 Mar 2011 01:56:09 +0000 (01:56 +0000)]
Optimizing (zext A + zext B) * C, to (VMULL A, C) + (VMULL B, C) during
isel lowering to fold the zero-extend's and take advantage of no-stall
back to back vmul + vmla:
vmull q0, d4, d6
vmlal q0, d5, d6
is faster than
vaddl q0, d4, d5
vmovl q1, d6
vmul q0, q0, q1
This allows us to vmull + vmlal for:
f = vmull_u8( vget_high_u8(s), c);
f = vmlal_u8(f, vget_low_u8(s), c);
rdar://
9197392
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@128444
91177308-0d34-0410-b5e6-
96231b3b80d8
Francois Pichet [Tue, 29 Mar 2011 00:30:01 +0000 (00:30 +0000)]
Fix the MSVC build.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@128441
91177308-0d34-0410-b5e6-
96231b3b80d8
Devang Patel [Tue, 29 Mar 2011 00:01:39 +0000 (00:01 +0000)]
Expoert c interface for disassembler.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@128440
91177308-0d34-0410-b5e6-
96231b3b80d8
Bill Wendling [Mon, 28 Mar 2011 23:02:18 +0000 (23:02 +0000)]
In some cases, the "fail BB dominator" may be null after the BB was split (and
becomes reachable when before it wasn't). Check to make sure that it's not null
before trying to use it.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@128434
91177308-0d34-0410-b5e6-
96231b3b80d8
Daniel Dunbar [Mon, 28 Mar 2011 22:49:19 +0000 (22:49 +0000)]
Integrated-As: Add support for setting the AllowTemporaryLabels flag via
integrated-as.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@128431
91177308-0d34-0410-b5e6-
96231b3b80d8
Daniel Dunbar [Mon, 28 Mar 2011 22:49:15 +0000 (22:49 +0000)]
MC: Add support for disabling "temporary label" behavior. Useful for debugging
on Darwin.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@128430
91177308-0d34-0410-b5e6-
96231b3b80d8
Ted Kremenek [Mon, 28 Mar 2011 20:43:53 +0000 (20:43 +0000)]
Unbreak CMake build.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@128426
91177308-0d34-0410-b5e6-
96231b3b80d8
Devang Patel [Mon, 28 Mar 2011 20:28:30 +0000 (20:28 +0000)]
Remove scripts used by TEST=dbg from here. They now live inside llvm test suite.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@128425
91177308-0d34-0410-b5e6-
96231b3b80d8
Johnny Chen [Mon, 28 Mar 2011 18:41:58 +0000 (18:41 +0000)]
Fix ARM disassembly for PLD/PLDW/PLI which suffers from code rot and add some test cases.
Add comments to ThumbDisassemblerCore.h for recent change made for t2PLD disassembly.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@128417
91177308-0d34-0410-b5e6-
96231b3b80d8
Kevin Enderby [Mon, 28 Mar 2011 18:25:07 +0000 (18:25 +0000)]
Again adding a C API to the disassembler for use by such tools as Darwin's
otool(1), this time with the needed fix for case sensitive file systems :) .
This is a work in progress as the interface for producing symbolic operands is
not done. But a hacked prototype using information from the object file's
relocation entiries and replacing immediate operands with MCExpr's has been
shown to work with no changes to the instrucion printer. These APIs will be
moved into a dynamic library at some point.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@128415
91177308-0d34-0410-b5e6-
96231b3b80d8
Nick Lewycky [Mon, 28 Mar 2011 17:48:26 +0000 (17:48 +0000)]
Remove tabs I accidentally added.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@128413
91177308-0d34-0410-b5e6-
96231b3b80d8
Jay Foad [Mon, 28 Mar 2011 13:03:10 +0000 (13:03 +0000)]
Make more use of PHINode::getNumIncomingValues().
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@128406
91177308-0d34-0410-b5e6-
96231b3b80d8
Che-Liang Chiou [Mon, 28 Mar 2011 10:23:13 +0000 (10:23 +0000)]
ptx: clean up branch code a bit
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@128405
91177308-0d34-0410-b5e6-
96231b3b80d8
NAKAMURA Takumi [Mon, 28 Mar 2011 06:27:12 +0000 (06:27 +0000)]
docs/GettingStarted.html: [PR8850] Add a note for x86_64-w64-mingw32.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@128404
91177308-0d34-0410-b5e6-
96231b3b80d8
NAKAMURA Takumi [Mon, 28 Mar 2011 06:27:06 +0000 (06:27 +0000)]
docs/GettingStarted.html: Add blurb "--enable-shared" on cygming.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@128403
91177308-0d34-0410-b5e6-
96231b3b80d8
Frits van Bommel [Sun, 27 Mar 2011 23:32:31 +0000 (23:32 +0000)]
Add some debug output when -instcombine uses RAUW. This can make debug output for those cases much clearer since without this it only showed that the original instruction was removed, not what it was replaced with.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@128399
91177308-0d34-0410-b5e6-
96231b3b80d8
Jakob Stoklund Olesen [Sun, 27 Mar 2011 22:49:23 +0000 (22:49 +0000)]
Amend debug output.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@128398
91177308-0d34-0410-b5e6-
96231b3b80d8
Jakob Stoklund Olesen [Sun, 27 Mar 2011 22:49:21 +0000 (22:49 +0000)]
Drop interference reassignment in favor of eviction.
The reassignment phase was able to move interference with a higher spill weight,
but it didn't happen very often and it was fairly expensive.
The existing interference eviction picks up the slack.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@128397
91177308-0d34-0410-b5e6-
96231b3b80d8
Nick Lewycky [Sun, 27 Mar 2011 19:51:23 +0000 (19:51 +0000)]
Teach the transformation that moves binary operators around selects to preserve
the subclass optional data.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@128388
91177308-0d34-0410-b5e6-
96231b3b80d8
Benjamin Kramer [Sun, 27 Mar 2011 15:04:38 +0000 (15:04 +0000)]
Use APInt's umul_ov instead of rolling our own overflow detection.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@128380
91177308-0d34-0410-b5e6-
96231b3b80d8
Frits van Bommel [Sun, 27 Mar 2011 14:26:13 +0000 (14:26 +0000)]
Constant folding support for calls to umul.with.overflow(), basically identical to the smul.with.overflow() code.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@128379
91177308-0d34-0410-b5e6-
96231b3b80d8
Duncan Sands [Sun, 27 Mar 2011 13:52:32 +0000 (13:52 +0000)]
Partially revert commit 127155: I think it is much more convenient
to have structured log files rather than one big file produced by
piping output.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@128378
91177308-0d34-0410-b5e6-
96231b3b80d8
Nick Lewycky [Sun, 27 Mar 2011 07:30:57 +0000 (07:30 +0000)]
Add a small missed optimization: turn X == C ? X : Y into X == C ? C : Y. This
removes one use of X which helps it pass the many hasOneUse() checks.
In my analysis, this turns up very often where X = A >>exact B and that can't be
simplified unless X has one use (except by increasing the lifetime of A which is
generally a performance loss).
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@128373
91177308-0d34-0410-b5e6-
96231b3b80d8
NAKAMURA Takumi [Sun, 27 Mar 2011 01:44:40 +0000 (01:44 +0000)]
Fix whitespace.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@128370
91177308-0d34-0410-b5e6-
96231b3b80d8
Jakob Stoklund Olesen [Sat, 26 Mar 2011 22:16:41 +0000 (22:16 +0000)]
Use individual register classes when spilling snippets.
The main register class may have been inflated by live range splitting, so that
register class is not necessarily valid for the snippet instructions.
Use the original register class for the stack slot interval.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@128351
91177308-0d34-0410-b5e6-
96231b3b80d8