Evan Cheng [Wed, 15 Aug 2012 17:44:53 +0000 (17:44 +0000)]
Use vld1/vst1 to load/store f64 if alignment is < 4 and the target allows unaligned access. rdar://
12091029
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161962
91177308-0d34-0410-b5e6-
96231b3b80d8
Owen Anderson [Wed, 15 Aug 2012 16:42:53 +0000 (16:42 +0000)]
Fix typo in comment.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161956
91177308-0d34-0410-b5e6-
96231b3b80d8
Jakob Stoklund Olesen [Wed, 15 Aug 2012 16:17:24 +0000 (16:17 +0000)]
Add missing Rfalse operand to the predicated pseudo-instructions.
When predicating this instruction:
Rd = ADD Rn, Rm
We need an extra operand to represent the value given to Rd when the
predicate is false:
Rd = ADDCC Rfalse, Rn, Rm, pred
The Rd and Rfalse operands are different registers while in SSA form.
Rfalse is tied to Rd to make sure they get the same register during
register allocation.
Previously, Rd and Rn were tied, but that is not required.
Compare to MOVCC:
Rd = MOVCC Rfalse, Rtrue, pred
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161955
91177308-0d34-0410-b5e6-
96231b3b80d8
Bill Wendling [Wed, 15 Aug 2012 12:22:35 +0000 (12:22 +0000)]
Set the branch probability of branching to the 'normal' destination of an invoke
instruction to something absurdly high, while setting the probability of
branching to the 'unwind' destination to the bare minimum. This should set cause
the normal destination's invoke blocks to be moved closer to the invoke.
PR13612
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161944
91177308-0d34-0410-b5e6-
96231b3b80d8
Benjamin Kramer [Wed, 15 Aug 2012 10:26:44 +0000 (10:26 +0000)]
Fix a const violation in the generated disassembler.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161940
91177308-0d34-0410-b5e6-
96231b3b80d8
Kostya Serebryany [Wed, 15 Aug 2012 08:58:58 +0000 (08:58 +0000)]
[asan] implement --asan-always-slow-path, which is a part of the improvement to handle unaligned partially OOB accesses. See code.google.com/p/address-sanitizer/issues/detail?id=100
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161937
91177308-0d34-0410-b5e6-
96231b3b80d8
Owen Anderson [Wed, 15 Aug 2012 05:39:46 +0000 (05:39 +0000)]
Fix a problem with APFloat::roundToIntegral where it would return incorrect results for negative inputs to trunc. Add unit tests to verify this behavior.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161929
91177308-0d34-0410-b5e6-
96231b3b80d8
Michael Liao [Wed, 15 Aug 2012 03:49:59 +0000 (03:49 +0000)]
fix infinite loop in instcombine with more than 4GB memcpy
- memcpy size is wrongly truncated into 32-bit and treat 8GB memcpy is
0-sized memcpy
- as 0-sized memcpy/memset is already removed before SimplifyMemTransfer
and SimplifyMemSet in visitCallInst, replace 0 checking with
assertions.
- replace getZExtValue() with getLimitedValue() according to
Eli Friedman
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161923
91177308-0d34-0410-b5e6-
96231b3b80d8
Nick Lewycky [Wed, 15 Aug 2012 02:37:07 +0000 (02:37 +0000)]
Fix a typo that led to a failure to correctly verify bitcast instructions.
Patch by Stephen Hines!
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161921
91177308-0d34-0410-b5e6-
96231b3b80d8
Richard Smith [Wed, 15 Aug 2012 01:39:31 +0000 (01:39 +0000)]
Fix undefined behavior: don't perform array indexing through a potentially null
pointer.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161919
91177308-0d34-0410-b5e6-
96231b3b80d8
Anton Korobeynikov [Tue, 14 Aug 2012 23:36:01 +0000 (23:36 +0000)]
The names of VFP variants of half-to-float conversion instructions were
reversed. This leads to wrong codegen for float-to-half conversion
intrinsics which are used to support storage-only fp16 type.
NEON variants of same instructions are fine.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161907
91177308-0d34-0410-b5e6-
96231b3b80d8
Eric Christopher [Tue, 14 Aug 2012 23:32:15 +0000 (23:32 +0000)]
This needs braces. Spotted by Bill.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161906
91177308-0d34-0410-b5e6-
96231b3b80d8
Michael Liao [Tue, 14 Aug 2012 22:53:17 +0000 (22:53 +0000)]
minor fix of X86ISD::VSEXT_MOVL dump
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161902
91177308-0d34-0410-b5e6-
96231b3b80d8
Michael J. Spencer [Tue, 14 Aug 2012 22:42:31 +0000 (22:42 +0000)]
[yaml2obj] Fix incorrect use of signed values.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161900
91177308-0d34-0410-b5e6-
96231b3b80d8
Michael Liao [Tue, 14 Aug 2012 21:24:47 +0000 (21:24 +0000)]
fix PR11334
- FP_EXTEND only support extending from vectors with matching elements.
This results in the scalarization of extending to v2f64 from v2f32,
which will be legalized to v4f32 not matching with v2f64.
- add X86-specific VFPEXT supproting extending from v4f32 to v2f64.
- add BUILD_VECTOR lowering helper to recover back the original
extending from v4f32 to v2f64.
- test case is enhanced to include different vector width.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161894
91177308-0d34-0410-b5e6-
96231b3b80d8
Jim Grosbach [Tue, 14 Aug 2012 19:06:05 +0000 (19:06 +0000)]
Switch the fixed-length disassembler to be table-driven.
Refactor the TableGen'erated fixed length disassemblmer to use a
table-driven state machine rather than a massive set of nested
switch() statements.
As a result, the ARM Disassembler (ARMDisassembler.cpp) builds much more
quickly and generates a smaller end result. For a Release+Asserts build on
a 16GB 3.4GHz i7 iMac w/ SSD:
Time to compile at -O2 (averaged w/ hot caches):
Previous: 35.5s
New: 8.9s
TEXT size:
Previous: 447,251
New: 297,661
Builds in 25% of the time previously required and generates code 66% of
the size.
Execution time of the disassembler is only slightly slower (7% disassembling
10 million ARM instructions, 19.6s vs 21.0s). The new implementation has
not yet been tuned, however, so the performance should almost certainly
be recoverable should it become a concern.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161888
91177308-0d34-0410-b5e6-
96231b3b80d8
Owen Anderson [Tue, 14 Aug 2012 18:51:15 +0000 (18:51 +0000)]
Fix the construction of the magic constant for roundToIntegral to be 64-bit safe. Fixes c-torture/execute/990826-0.c
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161885
91177308-0d34-0410-b5e6-
96231b3b80d8
David Blaikie [Tue, 14 Aug 2012 18:33:50 +0000 (18:33 +0000)]
Fix the make update target to work even when sub repo repositories are ignored.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161883
91177308-0d34-0410-b5e6-
96231b3b80d8
Daniel Dunbar [Tue, 14 Aug 2012 18:14:22 +0000 (18:14 +0000)]
Update configure.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161881
91177308-0d34-0410-b5e6-
96231b3b80d8
Daniel Dunbar [Tue, 14 Aug 2012 18:14:20 +0000 (18:14 +0000)]
[configure] Add a --enable-keep-symbols configure flag.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161880
91177308-0d34-0410-b5e6-
96231b3b80d8
Kostya Serebryany [Tue, 14 Aug 2012 14:05:50 +0000 (14:05 +0000)]
[asan] insert crash basic blocks inline as opposed to inserting them at the end of the function. This doesn't seem to fix or break anything, but is considered to be more friendly to downstream passes (test change)
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161871
91177308-0d34-0410-b5e6-
96231b3b80d8
Kostya Serebryany [Tue, 14 Aug 2012 14:04:51 +0000 (14:04 +0000)]
[asan] insert crash basic blocks inline as opposed to inserting them at the end of the function. This doesn't seem to fix or break anything, but is considered to be more friendly to downstream passes
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161870
91177308-0d34-0410-b5e6-
96231b3b80d8
Craig Topper [Tue, 14 Aug 2012 08:18:43 +0000 (08:18 +0000)]
Factor duplicate calls to getUNDEF in several functions.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161860
91177308-0d34-0410-b5e6-
96231b3b80d8
Craig Topper [Tue, 14 Aug 2012 07:43:25 +0000 (07:43 +0000)]
Re-factor intrinsic lowering to combine common parts of similar intrinsics. Reduces compiled code size a little bit.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161859
91177308-0d34-0410-b5e6-
96231b3b80d8
Craig Topper [Tue, 14 Aug 2012 07:32:05 +0000 (07:32 +0000)]
Change greater than to greater than or equal so that an identical sized store to the same offset is treated as completing overwriting.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161857
91177308-0d34-0410-b5e6-
96231b3b80d8
Richard Smith [Tue, 14 Aug 2012 05:31:26 +0000 (05:31 +0000)]
Fix undefined behavior: binding null pointer to reference. No functionality change.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161853
91177308-0d34-0410-b5e6-
96231b3b80d8
Nadav Rotem [Tue, 14 Aug 2012 05:19:07 +0000 (05:19 +0000)]
During the CodeGenPrepare we often lower intrinsics (such as objsize)
and allow some optimizations to turn conditional branches into unconditional.
This commit adds a simple control-flow optimization which merges two consecutive
basic blocks which are connected by a single edge. This allows the codegen to
operate on larger basic blocks.
rdar://
11973998
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161852
91177308-0d34-0410-b5e6-
96231b3b80d8
Eric Christopher [Tue, 14 Aug 2012 05:13:29 +0000 (05:13 +0000)]
Grammar.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161851
91177308-0d34-0410-b5e6-
96231b3b80d8
Richard Smith [Tue, 14 Aug 2012 02:56:51 +0000 (02:56 +0000)]
Avoid undefined behavior in DenseMap::shrink_and_clear(). Log2_32_Ceil(0)
returns 32. This change mirrors the corresponding code in
SmallDenseMap::shrink_and_clear().
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161829
91177308-0d34-0410-b5e6-
96231b3b80d8
Eric Christopher [Tue, 14 Aug 2012 01:09:10 +0000 (01:09 +0000)]
Typo.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161826
91177308-0d34-0410-b5e6-
96231b3b80d8
NAKAMURA Takumi [Tue, 14 Aug 2012 00:56:06 +0000 (00:56 +0000)]
llvm/test/CodeGen/ARM/floorf.ll: Add explicit -mtriple=arm-unknown-unknown, or it fails on msvc.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161825
91177308-0d34-0410-b5e6-
96231b3b80d8
Owen Anderson [Mon, 13 Aug 2012 23:32:49 +0000 (23:32 +0000)]
Add a roundToIntegral method to APFloat, which can be parameterized over various rounding modes. Use this to implement SelectionDAG constant folding of FFLOOR, FCEIL, and FTRUNC.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161807
91177308-0d34-0410-b5e6-
96231b3b80d8
Jakob Stoklund Olesen [Mon, 13 Aug 2012 23:13:25 +0000 (23:13 +0000)]
Transfer weights in transferSuccessorsAndUpdatePHIs().
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161805
91177308-0d34-0410-b5e6-
96231b3b80d8
Jakob Stoklund Olesen [Mon, 13 Aug 2012 23:13:23 +0000 (23:13 +0000)]
Print out MachineBasicBlock successor weights when available.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161804
91177308-0d34-0410-b5e6-
96231b3b80d8
Nadav Rotem [Mon, 13 Aug 2012 23:06:54 +0000 (23:06 +0000)]
LICM uses AliasSet information to hoist and sink instructions. However, other passes, such as LoopRotate
may invalidate its AliasSet because SSAUpdater does not update the AliasSet properly.
This patch teaches SSAUpdater to notify AliasSet that it made changes.
The testcase in PR12901 is too big to be useful and I could not reduce it to a normal size.
rdar://
11872059 PR12901
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161803
91177308-0d34-0410-b5e6-
96231b3b80d8
Nadav Rotem [Mon, 13 Aug 2012 23:03:43 +0000 (23:03 +0000)]
MemoryDependenceAnalysis attempts to find the first memory dependency for function calls.
Currently, if GetLocation reports that it did not find a valid pointer (this is the case for volatile load/stores),
we ignore the result. This patch adds code to handle the cases where we did not obtain a valid pointer.
rdar://
11872864 PR12899
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161802
91177308-0d34-0410-b5e6-
96231b3b80d8
Jim Grosbach [Mon, 13 Aug 2012 22:25:44 +0000 (22:25 +0000)]
ARM: Move Thumb2 tests to Thumb2 test file and fix CHECK lines.
These tests weren't actually being run before (missing ':' after CHECK).
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161800
91177308-0d34-0410-b5e6-
96231b3b80d8
Jakob Stoklund Olesen [Mon, 13 Aug 2012 21:52:57 +0000 (21:52 +0000)]
Remove the TII::scheduleTwoAddrSource() hook.
It never does anything when running 'make check', and it get's in the
way of updating live intervals in 2-addr.
The hook was originally added to help form IT blocks in Thumb2 code
before register allocation, but the pass ordering has changed since
then, and we run if-conversion after register allocation now.
When the MI scheduler is enabled, there will be no less than two
schedulers between 2-addr and Thumb2ITBlockPass, so this hook is
unlikely to help anything.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161794
91177308-0d34-0410-b5e6-
96231b3b80d8
Bill Wendling [Mon, 13 Aug 2012 21:32:42 +0000 (21:32 +0000)]
Rename test since it's not linux-specific.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161792
91177308-0d34-0410-b5e6-
96231b3b80d8
Manman Ren [Mon, 13 Aug 2012 21:22:50 +0000 (21:22 +0000)]
ARM: enable struct byval for AAPCS-VFP.
This change is to be enabled in clang.
rdar://
9877866
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161789
91177308-0d34-0410-b5e6-
96231b3b80d8
Bill Wendling [Mon, 13 Aug 2012 21:20:43 +0000 (21:20 +0000)]
Whitespace cleanup.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161788
91177308-0d34-0410-b5e6-
96231b3b80d8
Jakob Stoklund Olesen [Mon, 13 Aug 2012 21:03:27 +0000 (21:03 +0000)]
Count triangles and diamonds in early if-conversion.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161783
91177308-0d34-0410-b5e6-
96231b3b80d8
Jakob Stoklund Olesen [Mon, 13 Aug 2012 21:03:25 +0000 (21:03 +0000)]
Delete dead typedef.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161782
91177308-0d34-0410-b5e6-
96231b3b80d8
Jakob Stoklund Olesen [Mon, 13 Aug 2012 20:49:04 +0000 (20:49 +0000)]
Handle extra Tail predecessors in if-conversion.
It is still possible to if-convert if the tail block has extra
predecessors, but the tail phis must be rewritten instead of being
removed.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161781
91177308-0d34-0410-b5e6-
96231b3b80d8
Arnold Schwaighofer [Mon, 13 Aug 2012 19:54:01 +0000 (19:54 +0000)]
[Hexagon] Don't mark callee saved registers as clobbered by a tail call
This was causing unnecessary spills/restores of callee saved registers.
Fixes PR13572.
Patch by Pranav Bhandarkar!
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161778
91177308-0d34-0410-b5e6-
96231b3b80d8
Manman Ren [Mon, 13 Aug 2012 19:34:29 +0000 (19:34 +0000)]
Fix failure on Atom bot due to r161769
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161777
91177308-0d34-0410-b5e6-
96231b3b80d8
Nadav Rotem [Mon, 13 Aug 2012 18:52:44 +0000 (18:52 +0000)]
Do not optimize (or (and X,Y), Z) into BFI and other sequences if the AND ISDNode has more than one user.
rdar://
11876519
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161775
91177308-0d34-0410-b5e6-
96231b3b80d8
Manman Ren [Mon, 13 Aug 2012 18:29:41 +0000 (18:29 +0000)]
X86: move Int_CVTSD2SSrr, Int_CVTSI2SSrr, Int_CVTSI2SDrr, Int_CVTSS2SDrr from
OpTbl1 to OpTbl2 since they have 3 operands and the last operand can be changed
to a memory operand.
PR13576
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161769
91177308-0d34-0410-b5e6-
96231b3b80d8
Eric Christopher [Mon, 13 Aug 2012 18:18:52 +0000 (18:18 +0000)]
Add support for the %H output modifier.
Patch by Weiming Zhao.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161768
91177308-0d34-0410-b5e6-
96231b3b80d8
David Blaikie [Mon, 13 Aug 2012 17:50:44 +0000 (17:50 +0000)]
Ignore known externally-homed tools.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161765
91177308-0d34-0410-b5e6-
96231b3b80d8
Manman Ren [Mon, 13 Aug 2012 17:26:46 +0000 (17:26 +0000)]
X86: when auto-detecting the subtarget features, make sure use IsIntel to detect
Nehalem, Westmere and Sandy Bridge. AMD also has processor family 6.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161763
91177308-0d34-0410-b5e6-
96231b3b80d8
Nadav Rotem [Mon, 13 Aug 2012 15:29:53 +0000 (15:29 +0000)]
Fix a documentation typo.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161758
91177308-0d34-0410-b5e6-
96231b3b80d8
Kostya Serebryany [Mon, 13 Aug 2012 14:08:46 +0000 (14:08 +0000)]
[asan] remove the code for --asan-merge-callbacks as it appears to be a bad idea. (partly related to Bug 13225)
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161757
91177308-0d34-0410-b5e6-
96231b3b80d8
Alexander Kornienko [Mon, 13 Aug 2012 10:43:36 +0000 (10:43 +0000)]
Added test for non-static use of cl::opt (fixed in r160170)
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161751
91177308-0d34-0410-b5e6-
96231b3b80d8
Tim Northover [Mon, 13 Aug 2012 10:38:45 +0000 (10:38 +0000)]
Add test for previous commit correcting NEON load patterns.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161750
91177308-0d34-0410-b5e6-
96231b3b80d8
Tim Northover [Mon, 13 Aug 2012 09:06:31 +0000 (09:06 +0000)]
Use correct loads for vector types during extending-load operations.
Previously, we used VLD1.32 in all cases, however there are both 16 and 64-bit
accesses being selected, so we need to use an appropriate width load in those
cases.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161748
91177308-0d34-0410-b5e6-
96231b3b80d8
Craig Topper [Mon, 13 Aug 2012 03:42:38 +0000 (03:42 +0000)]
Tidy up VSETCC lowering code a bit more by adding an llvm_unreachable and putting an a couple if conditions in a better order.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161746
91177308-0d34-0410-b5e6-
96231b3b80d8
Craig Topper [Mon, 13 Aug 2012 02:34:03 +0000 (02:34 +0000)]
Refactor code a bit to share commonalities. No functional change intended.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161745
91177308-0d34-0410-b5e6-
96231b3b80d8
Craig Topper [Mon, 13 Aug 2012 01:26:45 +0000 (01:26 +0000)]
Fix an unused variable warning from r161742.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161743
91177308-0d34-0410-b5e6-
96231b3b80d8
Craig Topper [Mon, 13 Aug 2012 01:23:55 +0000 (01:23 +0000)]
Remove the LowerMMXCONCAT_VECTORS function. It could never execute because there are no legal 64-bit vector types that could be used as inputs to a 128-bit concat_vectors. Remove a target specific SDNode and its patterns that become unused as a result.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161742
91177308-0d34-0410-b5e6-
96231b3b80d8
Nick Lewycky [Sun, 12 Aug 2012 08:21:27 +0000 (08:21 +0000)]
Give this test an explicit triple.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161740
91177308-0d34-0410-b5e6-
96231b3b80d8
Nick Lewycky [Sun, 12 Aug 2012 08:09:45 +0000 (08:09 +0000)]
When emitting the PC range in an FDE, use the same data encoding for both ends
of the range. Fixes PR13581!
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161739
91177308-0d34-0410-b5e6-
96231b3b80d8
Craig Topper [Sun, 12 Aug 2012 05:31:32 +0000 (05:31 +0000)]
Remove call to setOperationAction for SETCC of v4f32. SETCC returns an integer type not an FP type.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161738
91177308-0d34-0410-b5e6-
96231b3b80d8
Craig Topper [Sun, 12 Aug 2012 05:15:16 +0000 (05:15 +0000)]
Remove unnecessary call to setOperationAction for SETCC of v2i64 under SSE42. It was already called for the same under SSE2.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161737
91177308-0d34-0410-b5e6-
96231b3b80d8
Arnold Schwaighofer [Sun, 12 Aug 2012 05:11:56 +0000 (05:11 +0000)]
Revert 161581: Patch to implement UMLAL/SMLAL instructions for the ARM
architecture
It broke MultiSource/Applications/JM/ldecod/ldecod on armv7 thumb O0 g and armv7
thumb O3.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161736
91177308-0d34-0410-b5e6-
96231b3b80d8
Craig Topper [Sun, 12 Aug 2012 03:16:37 +0000 (03:16 +0000)]
Change addTypeForNeon to use MVT instead of EVT so all the calls to getSimpleVT can be removed.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161735
91177308-0d34-0410-b5e6-
96231b3b80d8
Craig Topper [Sun, 12 Aug 2012 02:23:29 +0000 (02:23 +0000)]
Make replace many calls to getSizeInBits() with is128BitVector/is256BitVector
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161734
91177308-0d34-0410-b5e6-
96231b3b80d8
Craig Topper [Sun, 12 Aug 2012 00:34:56 +0000 (00:34 +0000)]
Use MVT.isXBitVector instead of EVT.isXBitVector when setting up operation actions. Compiles to smaller code.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161733
91177308-0d34-0410-b5e6-
96231b3b80d8
Michael Liao [Sat, 11 Aug 2012 23:47:06 +0000 (23:47 +0000)]
fix PR13577, an issue introduced by r161687
- FCMOV only supports a subset of X86 conditions. Skip boolean
simplification if X86 condition is not valid for FCMOV.
- add a minimal test case for PR13577.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161732
91177308-0d34-0410-b5e6-
96231b3b80d8
Craig Topper [Sat, 11 Aug 2012 23:31:18 +0000 (23:31 +0000)]
Create isXBitVector methods in MVT and call them from EVT. This allows targets to call them with an MVT without needing to convert to EVT.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161731
91177308-0d34-0410-b5e6-
96231b3b80d8
Craig Topper [Sat, 11 Aug 2012 22:34:26 +0000 (22:34 +0000)]
Move setOperationAction for CONCAT_VECTORS for 256-bit vectors into loop since all 256-bit types are supported.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161730
91177308-0d34-0410-b5e6-
96231b3b80d8
Benjamin Kramer [Sat, 11 Aug 2012 20:42:59 +0000 (20:42 +0000)]
MachineCSE: Hoist isConstantPhysReg out of the loop, it checks for overlaps already.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161729
91177308-0d34-0410-b5e6-
96231b3b80d8
Benjamin Kramer [Sat, 11 Aug 2012 19:05:13 +0000 (19:05 +0000)]
PR13578: Teach MachineCSE that instructions that use a constant register can be CSE'd safely.
This is common e.g. when doing rip-relative addressing on x86_64.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161728
91177308-0d34-0410-b5e6-
96231b3b80d8
Craig Topper [Sat, 11 Aug 2012 17:53:00 +0000 (17:53 +0000)]
Tidy up indentation. No functional change.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161727
91177308-0d34-0410-b5e6-
96231b3b80d8
Craig Topper [Sat, 11 Aug 2012 17:46:16 +0000 (17:46 +0000)]
Fix a cast that was casting away 'const' unnecessarily
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161726
91177308-0d34-0410-b5e6-
96231b3b80d8
Craig Topper [Sat, 11 Aug 2012 17:44:14 +0000 (17:44 +0000)]
Add a couple default: llvm_unreachable() to some switch statements. Fix a bad message in an existing llvm_unreachable.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161725
91177308-0d34-0410-b5e6-
96231b3b80d8
Manman Ren [Fri, 10 Aug 2012 23:43:32 +0000 (23:43 +0000)]
X86: when we are auto-detecting the subtarget features, make sure we turn on
FeatureFastUAMem for Nehalem, Westmere and Sandy Bridge.
FeatureFastUAMem is already on if we pass in nehalem or westmere as a command
argument.
rdar:
7252306
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161717
91177308-0d34-0410-b5e6-
96231b3b80d8
Jakob Stoklund Olesen [Fri, 10 Aug 2012 22:27:31 +0000 (22:27 +0000)]
Add a proper if-conversion cost model.
Detect when there is not enough available ILP, so if-conversion can't
speculate instructions for free.
Compute the lengthening of the critical path when inserting a select
instruction that depends on the condition as well as both sides of the
if.
Reject conversions that would stretch the critical path by more than
half a mispredict penalty.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161713
91177308-0d34-0410-b5e6-
96231b3b80d8
Jakob Stoklund Olesen [Fri, 10 Aug 2012 22:27:29 +0000 (22:27 +0000)]
Give MachineTraceMetrics its own debug tag.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161712
91177308-0d34-0410-b5e6-
96231b3b80d8
Jakob Stoklund Olesen [Fri, 10 Aug 2012 22:27:27 +0000 (22:27 +0000)]
Add more trace query functions.
Trace::getResourceLength() computes the number of cycles required to
execute the trace when ignoring data dependencies. The number can be
compared to the critical path to estimate the trace ILP.
Trace::getPHIDepth() computes the data dependency depth of a PHI in a
trace successor that isn't necessarily part of the trace.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161711
91177308-0d34-0410-b5e6-
96231b3b80d8
Eli Friedman [Fri, 10 Aug 2012 20:55:20 +0000 (20:55 +0000)]
The normal edge of an invoke is not allowed to branch to a block with a
landingpad. Enforce it in the verifier, and fix the regression tests to match.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161697
91177308-0d34-0410-b5e6-
96231b3b80d8
Manman Ren [Fri, 10 Aug 2012 20:39:38 +0000 (20:39 +0000)]
ARM: enable struct byval for AAPCS.
This change is to be enabled in clang.
rdar://
9877866
PR://13350
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161693
91177308-0d34-0410-b5e6-
96231b3b80d8
Jakob Stoklund Olesen [Fri, 10 Aug 2012 20:19:17 +0000 (20:19 +0000)]
Add getTPred() and getFPred() functions.
They identify the PHI predecessors in both diamonds and triangles.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161689
91177308-0d34-0410-b5e6-
96231b3b80d8
Jakob Stoklund Olesen [Fri, 10 Aug 2012 20:11:38 +0000 (20:11 +0000)]
Include loop-carried dependencies when computing instr heights.
When a trace ends with a back-edge, include PHIs in the loop header in
the height computations. This makes the critical path through a loop
more accurate by including the latencies of the last instructions in the
loop.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161688
91177308-0d34-0410-b5e6-
96231b3b80d8
Michael Liao [Fri, 10 Aug 2012 19:58:13 +0000 (19:58 +0000)]
add X86-specific DAG optimization to simplify boolean test
- if a boolean test (X86ISD::CMP or X86ISD:SUB) checks a boolean value
generated from X86ISD::SETCC, try to simplify the boolean value
generation and checking by reusing the original EFLAGS with proper
condition code
- add hooks to X86 specific SETCC/BRCOND/CMOV, the major 3 places
consuming EFLAGS
part of patches fixing PR12312
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161687
91177308-0d34-0410-b5e6-
96231b3b80d8
Eli Bendersky [Fri, 10 Aug 2012 18:30:44 +0000 (18:30 +0000)]
A couple of addition comment fixes
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161678
91177308-0d34-0410-b5e6-
96231b3b80d8
Eli Bendersky [Fri, 10 Aug 2012 18:26:20 +0000 (18:26 +0000)]
Fix a couple of typos in comments
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161677
91177308-0d34-0410-b5e6-
96231b3b80d8
Rafael Espindola [Fri, 10 Aug 2012 15:55:25 +0000 (15:55 +0000)]
Constify some basic blocks, no functionality change.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161668
91177308-0d34-0410-b5e6-
96231b3b80d8
Michael Liao [Fri, 10 Aug 2012 14:39:24 +0000 (14:39 +0000)]
remove tailing whitespaces and test commit
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161664
91177308-0d34-0410-b5e6-
96231b3b80d8
Rafael Espindola [Fri, 10 Aug 2012 14:05:55 +0000 (14:05 +0000)]
Move BasicBlockEdge to the cpp file. No functionality change.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161663
91177308-0d34-0410-b5e6-
96231b3b80d8
Joerg Sonnenberger [Fri, 10 Aug 2012 10:56:09 +0000 (10:56 +0000)]
stdcxx's cstdio doesn't include stdio.h, but the code using PathV2.inc
includes both. Deal with feof and ferror potentially being macros.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161658
91177308-0d34-0410-b5e6-
96231b3b80d8
Joerg Sonnenberger [Fri, 10 Aug 2012 10:53:56 +0000 (10:53 +0000)]
Add some missing includes for the build against stdcxx.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161657
91177308-0d34-0410-b5e6-
96231b3b80d8
Pete Cooper [Fri, 10 Aug 2012 03:26:36 +0000 (03:26 +0000)]
Fix crash when when do lto on Bullet. Dynamic GEPs in SROA were incorrectly being applied to all accesses to an alloca, not just the ones which read from the GEP. Thanks to Evan for reducing the test. rdar://
11861001
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161654
91177308-0d34-0410-b5e6-
96231b3b80d8
Jakob Stoklund Olesen [Fri, 10 Aug 2012 03:23:27 +0000 (03:23 +0000)]
Update edge weights correctly in replaceSuccessor().
When replacing Old with New, it can happen that New is already a
successor. Add the old and new edge weights instead of creating a
duplicate edge.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161653
91177308-0d34-0410-b5e6-
96231b3b80d8
Rafael Espindola [Fri, 10 Aug 2012 01:57:52 +0000 (01:57 +0000)]
Remove references to compression in llvm-ar. It has been a long time since we
switched from a bytecode+bzip2 to the current bitcode.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161651
91177308-0d34-0410-b5e6-
96231b3b80d8
Bob Wilson [Fri, 10 Aug 2012 00:25:30 +0000 (00:25 +0000)]
Use the final .version number for LLVM_MINOR_VERSION in Apple llvmCore builds.
We've switched to a 3-component version numbering scheme for Apple releases,
and with this scheme, the final number is the one most relevant for setting
LLVM_MINOR_VERSION. <rdar://problem/
12071459>
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161645
91177308-0d34-0410-b5e6-
96231b3b80d8
Jakob Stoklund Olesen [Fri, 10 Aug 2012 00:21:30 +0000 (00:21 +0000)]
Reapply r161633-161634 "Partition use lists so defs always come before uses.""
No changes to these patches, MRI needed to be notified when changing
uses into defs and vice versa.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161644
91177308-0d34-0410-b5e6-
96231b3b80d8
Jakob Stoklund Olesen [Fri, 10 Aug 2012 00:21:26 +0000 (00:21 +0000)]
Also update MRI use lists when changing a use to a def and vice versa.
This was the cause of the buildbot failures.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161643
91177308-0d34-0410-b5e6-
96231b3b80d8
Chad Rosier [Fri, 10 Aug 2012 00:00:22 +0000 (00:00 +0000)]
[ms-inline asm] Add a new Inline Asm Non-Standard Dialect attribute.
This new attribute is intended to be used by the backend to determine how
the inline asm string should be parsed/printed. This patch adds the
ia_nsdialect attribute and also adds a test case to ensure the IR is
correctly parsed, but there is no functional change at this time.
The standard dialect is assumed to be AT&T. Therefore, this attribute
should only be added to MS-style inline assembly statements, which use
the Intel dialect. If we ever support more dialects we'll need to
add additional state to the attribute.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161641
91177308-0d34-0410-b5e6-
96231b3b80d8
Jakob Stoklund Olesen [Thu, 9 Aug 2012 23:31:36 +0000 (23:31 +0000)]
Revert r161633-161634 "Partition use lists so defs always come before uses."
These commits broke a number of buildbots.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161640
91177308-0d34-0410-b5e6-
96231b3b80d8