Jakob Stoklund Olesen [Thu, 6 Sep 2012 18:15:23 +0000 (18:15 +0000)]
Allow overlaps between virtreg and physreg live ranges.
The RegisterCoalescer understands overlapping live ranges where one
register is defined as a copy of the other. With this change, register
allocators using LiveRegMatrix can do the same, at least for copies
between physical and virtual registers.
When a physreg is defined by a copy from a virtreg, allow those live
ranges to overlap:
%CL<def> = COPY %vreg11:sub_8bit; GR32_ABCD:%vreg11
%vreg13<def,tied1> = SAR32rCL %vreg13<tied0>, %CL<imp-use,kill>
We can assign %vreg11 to %ECX, overlapping the live range of %CL.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@163336
91177308-0d34-0410-b5e6-
96231b3b80d8
Jakob Stoklund Olesen [Thu, 6 Sep 2012 18:15:18 +0000 (18:15 +0000)]
Handle overlapping regunit intervals in LiveIntervals::addKillFlags().
We will soon allow virtual register live ranges to overlap regunit live
ranges when the physreg is defined as a copy of the virtreg:
%EAX = COPY %vreg5
FOO %vreg5
BAR %EAX<kill>
There is no real interference since %vreg5 and %EAX have the same value
where they overlap.
This patch prevents addKillFlags from adding virtreg kill flags to FOO
where the assigned physreg is overlapping the virtual register live
range.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@163335
91177308-0d34-0410-b5e6-
96231b3b80d8
Jakob Stoklund Olesen [Thu, 6 Sep 2012 18:15:15 +0000 (18:15 +0000)]
Clear kill flags while computing live ranges.
Kill flags are difficult to maintain, and liveness queries are better
handled by live intervals.
Kill flags are reinserted after register allocation by addKillFlags().
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@163334
91177308-0d34-0410-b5e6-
96231b3b80d8
Roman Divacky [Thu, 6 Sep 2012 15:42:13 +0000 (15:42 +0000)]
Dont cast away const needlessly. Found by gcc48 -Wcast-qual.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@163324
91177308-0d34-0410-b5e6-
96231b3b80d8
Tim Northover [Thu, 6 Sep 2012 15:27:12 +0000 (15:27 +0000)]
Diagnose invalid alignments on duplicating VLDn instructions.
Patch by Chris Lidbury.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@163323
91177308-0d34-0410-b5e6-
96231b3b80d8
Tim Northover [Thu, 6 Sep 2012 15:17:49 +0000 (15:17 +0000)]
Check for invalid alignment values when decoding VLDn/VSTn (single ln) instructions.
Patch by Chris Lidbury.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@163321
91177308-0d34-0410-b5e6-
96231b3b80d8
Arnold Schwaighofer [Thu, 6 Sep 2012 14:41:53 +0000 (14:41 +0000)]
BasicAA: Recognize cyclic NoAlias phis
Enhances basic alias analysis to recognize phis whose first incoming values are
NoAlias and whose other incoming values are just the phi node itself through
some amount of recursion.
Example: With this change basicaa reports that ptr_phi and ptr_phi2 do not alias
each other.
bb:
ptr = ptr2 + 1
loop:
ptr_phi = phi [bb, ptr], [loop, ptr_plus_one]
ptr2_phi = phi [bb, ptr2], [loop, ptr2_plus_one]
...
ptr_plus_one = gep ptr_phi, 1
ptr2_plus_one = gep ptr2_phi, 1
This enables the elimination of one load in code like the following:
extern int foo;
int test_noalias(int *ptr, int num, int* coeff) {
int *ptr2 = ptr;
int result = (*ptr++) * (*coeff--);
while (num--) {
*ptr2++ = *ptr;
result += (*coeff--) * (*ptr++);
}
*ptr = foo;
return result;
}
Part 2/2 of fix for PR13564.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@163319
91177308-0d34-0410-b5e6-
96231b3b80d8
Tim Northover [Thu, 6 Sep 2012 14:36:55 +0000 (14:36 +0000)]
Use correct part of complex operand to encode VST1 alignment.
Patch by Chris Lidbury.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@163318
91177308-0d34-0410-b5e6-
96231b3b80d8
Arnold Schwaighofer [Thu, 6 Sep 2012 14:31:51 +0000 (14:31 +0000)]
BasicAA: GEPs of NoAlias'ing base ptr with equivalent indices are NoAlias
If we can show that the base pointers of two GEPs don't alias each other using
precise analysis and the indices and base offset are equal then the two GEPs
also don't alias each other.
This is primarily needed for the follow up patch that analyses NoAlias'ing PHI
nodes.
Part 1/2 of fix for PR13564.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@163317
91177308-0d34-0410-b5e6-
96231b3b80d8
Nadav Rotem [Thu, 6 Sep 2012 14:27:06 +0000 (14:27 +0000)]
Disable stack coloring by default in order to resolve the i386 failures.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@163316
91177308-0d34-0410-b5e6-
96231b3b80d8
Tom Stellard [Thu, 6 Sep 2012 14:15:52 +0000 (14:15 +0000)]
Tablegen: Add OperandWithDefaultOps Operand type
This Operand type takes a default argument, and is initialized to
this value if it does not appear in a patter.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@163315
91177308-0d34-0410-b5e6-
96231b3b80d8
Elena Demikhovsky [Thu, 6 Sep 2012 12:42:01 +0000 (12:42 +0000)]
AVX2 optimization.
Added generation of VPSHUB instruction for <32 x i8> vector shuffle when possible.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@163312
91177308-0d34-0410-b5e6-
96231b3b80d8
Nadav Rotem [Thu, 6 Sep 2012 11:13:55 +0000 (11:13 +0000)]
Fix a few old-GCC warnings. No functional change.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@163309
91177308-0d34-0410-b5e6-
96231b3b80d8
Nadav Rotem [Thu, 6 Sep 2012 10:33:33 +0000 (10:33 +0000)]
Fix the test by specifying an exact cpu model.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@163307
91177308-0d34-0410-b5e6-
96231b3b80d8
James Molloy [Thu, 6 Sep 2012 10:32:08 +0000 (10:32 +0000)]
Fix self-host; ensure signedness is consistent.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@163306
91177308-0d34-0410-b5e6-
96231b3b80d8
Hans Wennborg [Thu, 6 Sep 2012 10:10:35 +0000 (10:10 +0000)]
Fix switch_to_lookup_table.ll test from r163302.
The lookup tables did not get built in a deterministic order.
This makes them get built in the order that the corresponding phi nodes
were found.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@163305
91177308-0d34-0410-b5e6-
96231b3b80d8
James Molloy [Thu, 6 Sep 2012 09:55:02 +0000 (09:55 +0000)]
Improve codegen for BUILD_VECTORs on ARM.
If we have a BUILD_VECTOR that is mostly a constant splat, it is often better to splat that constant then insertelement the non-constant lanes instead of insertelementing every lane from an undef base.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@163304
91177308-0d34-0410-b5e6-
96231b3b80d8
Hans Wennborg [Thu, 6 Sep 2012 09:43:28 +0000 (09:43 +0000)]
Build lookup tables for switches (PR884)
This adds a transformation to SimplifyCFG that attemps to turn switch
instructions into loads from lookup tables. It works on switches that
are only used to initialize one or more phi nodes in a common successor
basic block, for example:
int f(int x) {
switch (x) {
case 0: return 5;
case 1: return 4;
case 2: return -2;
case 5: return 7;
case 6: return 9;
default: return 42;
}
This speeds up the code by removing the hard-to-predict jump, and
reduces code size by removing the code for the jump targets.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@163302
91177308-0d34-0410-b5e6-
96231b3b80d8
Nadav Rotem [Thu, 6 Sep 2012 09:17:37 +0000 (09:17 +0000)]
Add a new optimization pass: Stack Coloring, that merges disjoint static allocations (allocas). Allocas are known to be
disjoint if they are marked by disjoint lifetime markers (@llvm.lifetime.XXX intrinsics).
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@163299
91177308-0d34-0410-b5e6-
96231b3b80d8
James Molloy [Thu, 6 Sep 2012 09:16:01 +0000 (09:16 +0000)]
Optimize codegen for VSETLNi{8,16,32} operating on Q registers. Degenerate to a VSETLN on D registers, instead of an (INSERT_SUBREG (VSETLN (EXTRACT_SUBREG ))) sequence to help the register coalescer.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@163298
91177308-0d34-0410-b5e6-
96231b3b80d8
Michael Liao [Thu, 6 Sep 2012 07:11:22 +0000 (07:11 +0000)]
Remove duplicated helper function
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@163295
91177308-0d34-0410-b5e6-
96231b3b80d8
Craig Topper [Thu, 6 Sep 2012 06:09:01 +0000 (06:09 +0000)]
Use iPTR instead of i32 for extract_subvector/insert_subvector index in lowering and patterns. This makes it consistent with the incoming DAG nodes from the DAG builder.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@163293
91177308-0d34-0410-b5e6-
96231b3b80d8
Craig Topper [Thu, 6 Sep 2012 05:15:01 +0000 (05:15 +0000)]
Add patterns for converting stores of subvector_extracts of lower 128-bits of a 256-bit vector to VMOVAPSmr/VMOVUPSmr.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@163292
91177308-0d34-0410-b5e6-
96231b3b80d8
Jim Grosbach [Thu, 6 Sep 2012 03:24:09 +0000 (03:24 +0000)]
Revert "Enable MCJIT tests on Darwin."
This reverts commit 163278.
Works OK on x86_64, but not i386. Will re-enable when that's cleared up.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@163290
91177308-0d34-0410-b5e6-
96231b3b80d8
NAKAMURA Takumi [Thu, 6 Sep 2012 03:02:56 +0000 (03:02 +0000)]
Whitespace.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@163289
91177308-0d34-0410-b5e6-
96231b3b80d8
NAKAMURA Takumi [Thu, 6 Sep 2012 03:01:43 +0000 (03:01 +0000)]
Unix/Signals.inc: Fix a typo. Thanks to Dani Berg!
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@163288
91177308-0d34-0410-b5e6-
96231b3b80d8
Jack Carter [Thu, 6 Sep 2012 02:31:34 +0000 (02:31 +0000)]
There are some Mips instructions that are lowered by the
assembler such as shifts greater than 32. In the case
of direct object, the code gen needs to do this lowering
since the assembler is not involved.
With the advent of the llvm-mc assembler, it also needs
to do the same lowering.
This patch makes that specific lowering code accessible
to both the direct object output and the assembler.
This patch does not affect generated output.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@163287
91177308-0d34-0410-b5e6-
96231b3b80d8
Jim Grosbach [Thu, 6 Sep 2012 00:59:08 +0000 (00:59 +0000)]
Update function names to conform to guidelines.
No functional change.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@163279
91177308-0d34-0410-b5e6-
96231b3b80d8
Jim Grosbach [Thu, 6 Sep 2012 00:59:06 +0000 (00:59 +0000)]
Enable MCJIT tests on Darwin.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@163278
91177308-0d34-0410-b5e6-
96231b3b80d8
Jack Carter [Thu, 6 Sep 2012 00:43:26 +0000 (00:43 +0000)]
Mips specific llvm assembler support for branch and jump instructions.
Test case included.
Contributer: Vladimir Medic
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@163277
91177308-0d34-0410-b5e6-
96231b3b80d8
Eli Friedman [Thu, 6 Sep 2012 00:12:55 +0000 (00:12 +0000)]
Don't include stdint.h directly.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@163276
91177308-0d34-0410-b5e6-
96231b3b80d8
Jakob Stoklund Olesen [Wed, 5 Sep 2012 23:58:04 +0000 (23:58 +0000)]
Remove predicated pseudo-instructions.
These pseudos are no longer needed now that it is possible to represent
predicated instructions in SSA form.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@163275
91177308-0d34-0410-b5e6-
96231b3b80d8
Jakob Stoklund Olesen [Wed, 5 Sep 2012 23:58:02 +0000 (23:58 +0000)]
Use predication instead of pseudo-opcodes when folding into MOVCC.
Now that it is possible to dynamically tie MachineInstr operands,
predicated instructions are possible in SSA form:
%vreg3<def> = SUBri %vreg1, -
2147483647, pred:14, pred:%noreg, %opt:%noreg
%vreg4<def,tied1> = MOVCCr %vreg3<tied0>, %vreg1, %pred:12, pred:%CPSR
Becomes a predicated SUBri with a tied imp-use:
SUBri %vreg1, -
2147483647, pred:13, pred:%CPSR, opt:%noreg, %vreg1<imp-use,tied0>
This means that any instruction that is safe to move can be folded into
a MOVCC, and the *CC pseudo-instructions are no longer needed.
The test case changes reflect that Thumb2SizeReduce recognizes the
predicated instructions. It didn't understand the pseudos.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@163274
91177308-0d34-0410-b5e6-
96231b3b80d8
Chad Rosier [Wed, 5 Sep 2012 23:57:37 +0000 (23:57 +0000)]
[ms-inline asm] Use the asm dialect from the MI to set the parser dialect.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@163273
91177308-0d34-0410-b5e6-
96231b3b80d8
Nick Lewycky [Wed, 5 Sep 2012 23:52:20 +0000 (23:52 +0000)]
Add missing file for test.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@163272
91177308-0d34-0410-b5e6-
96231b3b80d8
Nick Lewycky [Wed, 5 Sep 2012 23:48:54 +0000 (23:48 +0000)]
Teach libObject about some more ELF relocations. llvm-objdump -r now knows
every relocation in C++ hello world built with debug info.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@163271
91177308-0d34-0410-b5e6-
96231b3b80d8
Manman Ren [Wed, 5 Sep 2012 23:45:58 +0000 (23:45 +0000)]
JumpThreading: when default destination is the destination of some cases in a
switch, make sure we include the value for the cases when calculating edge
value from switch to the default destination.
rdar://
12241132
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@163270
91177308-0d34-0410-b5e6-
96231b3b80d8
Jack Carter [Wed, 5 Sep 2012 23:34:03 +0000 (23:34 +0000)]
Mips specific llvm assembler support for ALU instructions. This includes
register support. Test case included.
Contributer: Vladimir Medic
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@163268
91177308-0d34-0410-b5e6-
96231b3b80d8
Chad Rosier [Wed, 5 Sep 2012 22:40:13 +0000 (22:40 +0000)]
Cleanup a few magic numbers.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@163263
91177308-0d34-0410-b5e6-
96231b3b80d8
Roman Divacky [Wed, 5 Sep 2012 22:26:57 +0000 (22:26 +0000)]
Stop casting away const qualifier needlessly.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@163258
91177308-0d34-0410-b5e6-
96231b3b80d8
Chad Rosier [Wed, 5 Sep 2012 22:17:43 +0000 (22:17 +0000)]
[ms-inline asm] We only need one bit to represent the AsmDialect in the
MachineInstr.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@163257
91177308-0d34-0410-b5e6-
96231b3b80d8
Roman Divacky [Wed, 5 Sep 2012 22:15:49 +0000 (22:15 +0000)]
Constify this properly. Found by gcc48 -Wcast-qual.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@163256
91177308-0d34-0410-b5e6-
96231b3b80d8
Roman Divacky [Wed, 5 Sep 2012 22:09:23 +0000 (22:09 +0000)]
Mark checkSignature const, and in turn stop casting away const from
ArchiveMemberHeader. Found by gcc48 -Wcast-qual.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@163255
91177308-0d34-0410-b5e6-
96231b3b80d8
Roman Divacky [Wed, 5 Sep 2012 22:03:34 +0000 (22:03 +0000)]
Constify SDNodeIterator an stop its only non-const user being cast stripped
of its constness. Found by gcc48 -Wcast-qual.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@163254
91177308-0d34-0410-b5e6-
96231b3b80d8
Roman Divacky [Wed, 5 Sep 2012 21:43:57 +0000 (21:43 +0000)]
Constify subtarget info properly so that we dont cast away the const in
the SubtargetInfoKV tables. Found by gcc48 -Wcast-qual.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@163251
91177308-0d34-0410-b5e6-
96231b3b80d8
Roman Divacky [Wed, 5 Sep 2012 21:17:34 +0000 (21:17 +0000)]
Use const properly so that we dont remove const qualifier from region and MII
by casting. Found with gcc48.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@163247
91177308-0d34-0410-b5e6-
96231b3b80d8
Chad Rosier [Wed, 5 Sep 2012 21:00:58 +0000 (21:00 +0000)]
[ms-inline asm] Propagate the asm dialect into the MachineInstr representation.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@163243
91177308-0d34-0410-b5e6-
96231b3b80d8
Jan Wen Voung [Wed, 5 Sep 2012 20:56:00 +0000 (20:56 +0000)]
Fix a bug.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@163242
91177308-0d34-0410-b5e6-
96231b3b80d8
Jan Wen Voung [Wed, 5 Sep 2012 20:55:57 +0000 (20:55 +0000)]
revert the additional stuff.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@163241
91177308-0d34-0410-b5e6-
96231b3b80d8
Jan Wen Voung [Wed, 5 Sep 2012 20:55:54 +0000 (20:55 +0000)]
Clean up llvm-bcanalyzer to print to consistent streams.
Avoid interleaving fprintf(stderr,...) and outs() << ...;
Also add a column to show "bytes-per" for each record.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@163240
91177308-0d34-0410-b5e6-
96231b3b80d8
Michael J. Spencer [Wed, 5 Sep 2012 19:44:47 +0000 (19:44 +0000)]
[Docs] Fix Sphinx incremental build. Patch by Sean Silva!
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@163235
91177308-0d34-0410-b5e6-
96231b3b80d8
Hal Finkel [Wed, 5 Sep 2012 19:22:27 +0000 (19:22 +0000)]
Move the PPC TOC defs into the PPC64 InstrInfo file.
Since TOC is just defined for PPC64, move its definition to PPC64 td file.
Patch by Adhemerval Zanella.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@163234
91177308-0d34-0410-b5e6-
96231b3b80d8
Chad Rosier [Wed, 5 Sep 2012 19:16:22 +0000 (19:16 +0000)]
Clean up.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@163233
91177308-0d34-0410-b5e6-
96231b3b80d8
Chad Rosier [Wed, 5 Sep 2012 19:00:49 +0000 (19:00 +0000)]
[ms-inline asm] Enumerate the InlineAsm dialects and rename the nsdialect to
inteldialect.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@163231
91177308-0d34-0410-b5e6-
96231b3b80d8
Tim Northover [Wed, 5 Sep 2012 18:37:53 +0000 (18:37 +0000)]
Strip old MachineInstrs *after* we know we can put them back.
Previous patch accidentally decided it couldn't convert a VFP to a
NEON instruction after it had already destroyed the old one. Not a
good move.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@163230
91177308-0d34-0410-b5e6-
96231b3b80d8
Benjamin Kramer [Wed, 5 Sep 2012 18:19:08 +0000 (18:19 +0000)]
Clean up includes.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@163229
91177308-0d34-0410-b5e6-
96231b3b80d8
Jim Grosbach [Wed, 5 Sep 2012 18:15:08 +0000 (18:15 +0000)]
Update CMakeList.txt for new lli sources.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@163228
91177308-0d34-0410-b5e6-
96231b3b80d8
Roman Divacky [Wed, 5 Sep 2012 17:55:46 +0000 (17:55 +0000)]
Remove unused typedefs gcc4.8 warns about.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@163225
91177308-0d34-0410-b5e6-
96231b3b80d8
Jim Grosbach [Wed, 5 Sep 2012 16:50:40 +0000 (16:50 +0000)]
MCJIT: getPointerToFunction() references target address space.
Make sure to return a pointer into the target memory, not the local memory.
Often they are the same, but we can't assume that.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@163217
91177308-0d34-0410-b5e6-
96231b3b80d8
Jim Grosbach [Wed, 5 Sep 2012 16:50:34 +0000 (16:50 +0000)]
MCJIT: Add faux remote target execution to lli for the MCJIT.
Simulate a remote target address space by allocating a seperate chunk of
memory for the target and re-mapping section addresses to that prior to
execution. Later we'll want to have a truly remote process, but for now
this gets us closer to being able to test the remote target
functionality outside LLDB.
rdar://
12157052
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@163216
91177308-0d34-0410-b5e6-
96231b3b80d8
Benjamin Kramer [Wed, 5 Sep 2012 16:49:37 +0000 (16:49 +0000)]
Switch BasicAliasAnalysis' cache to SmallDenseMap.
It relies on clear() being fast and the cache rarely has more than 1 or 2
elements, so give it an inline capacity and always shrink it back down in case
it grows. DenseMap will grow to 64 buckets which makes clear() a lot slower.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@163215
91177308-0d34-0410-b5e6-
96231b3b80d8
Pranav Bhandarkar [Wed, 5 Sep 2012 16:01:40 +0000 (16:01 +0000)]
LLVM Bug Fix 13709: Remove needless lsr(Rp, #32) instruction access the
subreg_hireg of register pair Rp.
* lib/Target/Hexagon/HexagonPeephole.cpp(PeepholeDoubleRegsMap): New
DenseMap similar to PeepholeMap that additionally records subreg info
too.
(runOnMachineFunction): Record information in PeepholeDoubleRegsMap
and copy propagate the high sub-reg of Rp0 in Rp1 = lsr(Rp0, #32) to
the instruction Rx = COPY Rp1:logreg_subreg.
* test/CodeGen/Hexagon/remove_lsr.ll: New test.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@163214
91177308-0d34-0410-b5e6-
96231b3b80d8
Kostya Serebryany [Wed, 5 Sep 2012 09:00:18 +0000 (09:00 +0000)]
[asan] fix lint
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@163205
91177308-0d34-0410-b5e6-
96231b3b80d8
Silviu Baranga [Wed, 5 Sep 2012 08:57:21 +0000 (08:57 +0000)]
Fixed the DAG combiner to better handle the folding of AND nodes for vector types. The previous code was making the assumption that the length of the bitmask returned by isConstantSplat was equal to the size of the vector type. Now we first make sure that the splat value has at least the length of the vector lane type, then we only use as many fields as we have available in the splat value.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@163203
91177308-0d34-0410-b5e6-
96231b3b80d8
Kostya Serebryany [Wed, 5 Sep 2012 07:29:56 +0000 (07:29 +0000)]
[asan] extend the blacklist functionality to handle global-init. Patch by Reid Watson
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@163199
91177308-0d34-0410-b5e6-
96231b3b80d8
Craig Topper [Wed, 5 Sep 2012 07:26:35 +0000 (07:26 +0000)]
Remove some of the patterns added in r163196. Increasing the complexity on insert_subvector into undef accomplishes the same thing.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@163198
91177308-0d34-0410-b5e6-
96231b3b80d8
Craig Topper [Wed, 5 Sep 2012 06:58:39 +0000 (06:58 +0000)]
Add patterns for integer forms of VINSERTF128/VINSERTI128 folded with loads. Also add patterns to turn subvector inserts with loads to index 0 of an undef into VMOVAPS.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@163196
91177308-0d34-0410-b5e6-
96231b3b80d8
Chad Rosier [Wed, 5 Sep 2012 06:28:52 +0000 (06:28 +0000)]
Add a FIXME that assumes we maintain backward compatibility until the next major release.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@163195
91177308-0d34-0410-b5e6-
96231b3b80d8
Logan Chien [Wed, 5 Sep 2012 06:28:26 +0000 (06:28 +0000)]
Reorder the comments of EmitExceptionTable.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@163194
91177308-0d34-0410-b5e6-
96231b3b80d8
Logan Chien [Wed, 5 Sep 2012 06:17:17 +0000 (06:17 +0000)]
Fix UseInitArray option for MIPS target.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@163193
91177308-0d34-0410-b5e6-
96231b3b80d8
Craig Topper [Wed, 5 Sep 2012 05:48:09 +0000 (05:48 +0000)]
Convert vextracti128/vextractf128 intrinsics to extract_subvector at DAG build time. Similar was previously done for vinserti128/vinsertf128. Add patterns for folding these extract_subvectors with stores.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@163192
91177308-0d34-0410-b5e6-
96231b3b80d8
Marshall Clow [Wed, 5 Sep 2012 03:18:55 +0000 (03:18 +0000)]
Removed Trie.h; unused in a long time
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@163191
91177308-0d34-0410-b5e6-
96231b3b80d8
Richard Smith [Wed, 5 Sep 2012 01:41:37 +0000 (01:41 +0000)]
Remove redundant semicolons to fix -pedantic-errors build.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@163190
91177308-0d34-0410-b5e6-
96231b3b80d8
Chad Rosier [Wed, 5 Sep 2012 01:15:43 +0000 (01:15 +0000)]
Fix function name per coding standard.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@163187
91177308-0d34-0410-b5e6-
96231b3b80d8
Chad Rosier [Wed, 5 Sep 2012 01:02:38 +0000 (01:02 +0000)]
Fix function name per coding standard.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@163186
91177308-0d34-0410-b5e6-
96231b3b80d8
Chad Rosier [Wed, 5 Sep 2012 00:56:20 +0000 (00:56 +0000)]
[ms-inline asm] Add support for the nsdialect keyword in the Bitcode
Reader/Writer.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@163185
91177308-0d34-0410-b5e6-
96231b3b80d8
Chad Rosier [Wed, 5 Sep 2012 00:51:02 +0000 (00:51 +0000)]
[ms-inline asm] Add the nsdialect keyword to the lexer.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@163184
91177308-0d34-0410-b5e6-
96231b3b80d8
Chad Rosier [Wed, 5 Sep 2012 00:08:17 +0000 (00:08 +0000)]
[ms-inline asm] Emit the (new) inline asm Non-Standard Dialect attribute.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@163181
91177308-0d34-0410-b5e6-
96231b3b80d8
Dan Gohman [Tue, 4 Sep 2012 23:16:20 +0000 (23:16 +0000)]
Make provenance checking conservative in cases when
pointers-to-strong-pointers may be in play. These can lead to retains and
releases happening in unstructured ways, foiling the optimizer. This fixes
rdar://
12150909.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@163180
91177308-0d34-0410-b5e6-
96231b3b80d8
Jakub Staszak [Tue, 4 Sep 2012 23:11:11 +0000 (23:11 +0000)]
BypassSlowDivision: Assign to reference, don't copy the object.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@163179
91177308-0d34-0410-b5e6-
96231b3b80d8
Jakob Stoklund Olesen [Tue, 4 Sep 2012 22:59:30 +0000 (22:59 +0000)]
Search the whole instruction for tied operands.
Implicit uses can be dynamically tied to defs. This will soon be used
for predicated instructions on ARM.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@163177
91177308-0d34-0410-b5e6-
96231b3b80d8
Chad Rosier [Tue, 4 Sep 2012 22:46:24 +0000 (22:46 +0000)]
[ms-inline asm] Add the inline assembly dialect, AsmDialect, to the InlineAsm
class.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@163175
91177308-0d34-0410-b5e6-
96231b3b80d8
Chad Rosier [Tue, 4 Sep 2012 22:29:45 +0000 (22:29 +0000)]
[ms-inline asm] Remove the Inline Asm Non-Standard Dialect attribute. This
implementation does not co-exist well with how the sideeffect and alignstack
attributes are handled. The reverts r161641.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@163174
91177308-0d34-0410-b5e6-
96231b3b80d8
David Blaikie [Tue, 4 Sep 2012 22:09:04 +0000 (22:09 +0000)]
[LIT] Add a clang_tools_extra_site_cfg to match the various other site_cfg.
This doesn't seem ideal, perhaps we could just keep the llvm_site_cfg and have
other config (clang and clang-tools-extra) derive their site_cfg from that.
Suggestions/complaints/ideas welcome.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@163171
91177308-0d34-0410-b5e6-
96231b3b80d8
Jakub Staszak [Tue, 4 Sep 2012 21:16:59 +0000 (21:16 +0000)]
Fix my previous patch (r163164). It does now what it is supposed to do:
Doesn't set MadeChange to TRUE if BypassSlowDivision doesn't change anything.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@163165
91177308-0d34-0410-b5e6-
96231b3b80d8
Jakub Staszak [Tue, 4 Sep 2012 20:48:24 +0000 (20:48 +0000)]
Return false if BypassSlowDivision doesn't change anything.
Also a few minor changes:
- use pre-inc instead of post-inc
- use isa instead of dyn_cast
- 80 col
- trailing spaces
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@163164
91177308-0d34-0410-b5e6-
96231b3b80d8
Jakub Staszak [Tue, 4 Sep 2012 19:49:17 +0000 (19:49 +0000)]
Remove unneeded code.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@163160
91177308-0d34-0410-b5e6-
96231b3b80d8
Jakob Stoklund Olesen [Tue, 4 Sep 2012 18:44:43 +0000 (18:44 +0000)]
Typo.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@163154
91177308-0d34-0410-b5e6-
96231b3b80d8
Jakob Stoklund Olesen [Tue, 4 Sep 2012 18:43:25 +0000 (18:43 +0000)]
Actually use the MachineOperand field for isRegTiedToDefOperand().
The MachineOperand::TiedTo field was maintained, but not used.
This patch enables it in isRegTiedToDefOperand() and
isRegTiedToUseOperand() which are the actual functions use by the
register allocator.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@163153
91177308-0d34-0410-b5e6-
96231b3b80d8
Jakob Stoklund Olesen [Tue, 4 Sep 2012 18:38:28 +0000 (18:38 +0000)]
Move tie checks into MachineVerifier::visitMachineOperand.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@163152
91177308-0d34-0410-b5e6-
96231b3b80d8
Jakob Stoklund Olesen [Tue, 4 Sep 2012 18:36:28 +0000 (18:36 +0000)]
Allow tied uses and defs in different orders.
After much agonizing, use a full 4 bits of precious MachineOperand space
to encode this. This uses existing padding, and doesn't grow
MachineOperand beyond its current 32 bytes.
This allows tied defs among the first 15 operands on a normal
instruction, just like the current MCInstrDesc constraint encoding.
Inline assembly needs to be able to tie more than the first 15 operands,
and gets special treatment.
Tied uses can appear beyond 15 operands, as long as they are tied to a
def that's in range.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@163151
91177308-0d34-0410-b5e6-
96231b3b80d8
Preston Gurd [Tue, 4 Sep 2012 18:22:17 +0000 (18:22 +0000)]
Generic Bypass Slow Div
- CodeGenPrepare pass for identifying div/rem ops
- Backend specifies the type mapping using addBypassSlowDivType
- Enabled only for Intel Atom with O2 32-bit -> 8-bit
- Replace IDIV with instructions which test its value and use DIVB if the value
is positive and less than 256.
- In the case when the quotient and remainder of a divide are used a DIV
and a REM instruction will be present in the IR. In the non-Atom case
they are both lowered to IDIVs and CSE removes the redundant IDIV instruction,
using the quotient and remainder from the first IDIV. However,
due to this optimization CSE is not able to eliminate redundant
IDIV instructions because they are located in different basic blocks.
This is overcome by calculating both the quotient (DIV) and remainder (REM)
in each basic block that is inserted by the optimization and reusing the result
values when a subsequent DIV or REM instruction uses the same operands.
- Test cases check for the presents of the optimization when calculating
either the quotient, remainder, or both.
Patch by Tyler Nowicki!
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@163150
91177308-0d34-0410-b5e6-
96231b3b80d8
Bob Wilson [Tue, 4 Sep 2012 17:42:53 +0000 (17:42 +0000)]
Make sure macros in the include subdirectory are not used without being defined.
Rationale: For each preprocessor macro, either the definedness is what's
meaningful, or the value is what's meaningful, or both. If definedness is
meaningful, we should use #ifdef. If the value is meaningful, we should use
and #ifdef interchangeably for the same macro, seems ugly to me, even if
undefined macros are zero if used.
This also has the benefit that including an LLVM header doesn't prevent
you from compiling with -Wundef -Werror.
Patch by John Garvin!
<rdar://problem/
12189979>
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@163148
91177308-0d34-0410-b5e6-
96231b3b80d8
Sergei Larin [Tue, 4 Sep 2012 14:49:56 +0000 (14:49 +0000)]
Porting Hexagon MI Scheduler to the new API.
Change current Hexagon MI scheduler to use new converging
scheduler. Integrates DFA resource model into it.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@163137
91177308-0d34-0410-b5e6-
96231b3b80d8
Arnold Schwaighofer [Tue, 4 Sep 2012 14:37:49 +0000 (14:37 +0000)]
Patch to implement UMLAL/SMLAL instructions for the ARM architecture
This patch corrects the definition of umlal/smlal instructions and adds support
for matching them to the ARM dag combiner.
Bug 12213
Patch by Yin Ma!
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@163136
91177308-0d34-0410-b5e6-
96231b3b80d8
Elena Demikhovsky [Tue, 4 Sep 2012 12:49:02 +0000 (12:49 +0000)]
This patch optimizes shuffle instruction - generates 2 instructions instead of 4.
Since this specific shuffle is widely used in many workloads we have ~10% performance on them.
shufflevector <8 x float> %A, <8 x float> %B, <8 x i32> <i32 0, i32 8, i32 2, i32 10, i32 4, i32 12, i32 6, i32 14>
vmovaps (%rdx), %ymm0
vshufps $8, %ymm0, %ymm0, %ymm0
vmovaps (%rcx), %ymm1
vshufps $8, %ymm0, %ymm1, %ymm1
vunpcklps %ymm0, %ymm1, %ymm0
vmovaps (%rcx), %ymm0
vmovsldup (%rdx), %ymm1
vblendps $85, %ymm0, %ymm1, %ymm0
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@163134
91177308-0d34-0410-b5e6-
96231b3b80d8
Nadav Rotem [Tue, 4 Sep 2012 10:25:04 +0000 (10:25 +0000)]
LICM may hoist an instruction with undefined behavior above a trap.
Scan the body of the loop and find instructions that may trap.
Use this information when deciding if it is safe to hoist or sink instructions.
Notice that we can optimize the search of instructions that may throw in the case of nested loops.
rdar://
11518836
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@163132
91177308-0d34-0410-b5e6-
96231b3b80d8
Evgeniy Stepanov [Tue, 4 Sep 2012 09:14:45 +0000 (09:14 +0000)]
Fix Android build of gtest and lib/Support.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@163131
91177308-0d34-0410-b5e6-
96231b3b80d8
Alexey Samsonov [Tue, 4 Sep 2012 08:12:33 +0000 (08:12 +0000)]
Add support for fetching inlining context (stack of source code locations)
by instruction address from DWARF.
Add --inlining flag to llvm-dwarfdump to demonstrate and test this functionality,
so that "llvm-dwarfdump --inlining --address=0x..." now works much like
"addr2line -i 0x...", provided that the binary has debug info
(Clang's -gline-tables-only *is* enough).
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@163128
91177308-0d34-0410-b5e6-
96231b3b80d8
Bob Wilson [Tue, 4 Sep 2012 03:30:13 +0000 (03:30 +0000)]
Be conservative about allocations that may alias the accessed pointer.
If an allocation has a must-alias relation to the access pointer, we treat it
as a Def. Otherwise, without this check, the code here was just skipping over
the allocation call and ignoring it. I noticed this by inspection and don't
have a specific testcase that it breaks, but it seems like we need to treat
a may-alias allocation as a Clobber.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@163127
91177308-0d34-0410-b5e6-
96231b3b80d8