oota-llvm.git
11 years agodocs: Convert ProgrammersManual to reST.
Sean Silva [Tue, 4 Dec 2012 03:20:08 +0000 (03:20 +0000)]
docs: Convert ProgrammersManual to reST.

Patch by Alexander Zinenko!

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@169208 91177308-0d34-0410-b5e6-96231b3b80d8

11 years agoASan: add initial support for handling llvm.lifetime intrinsics in ASan - emit calls...
Alexey Samsonov [Tue, 4 Dec 2012 01:34:23 +0000 (01:34 +0000)]
ASan: add initial support for handling llvm.lifetime intrinsics in ASan - emit calls into runtime library that poison memory for local variables when their lifetime is over and unpoison memory when their lifetime begins.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@169200 91177308-0d34-0410-b5e6-96231b3b80d8

11 years agoSimplify code. No functionality change.
Jakub Staszak [Tue, 4 Dec 2012 01:00:52 +0000 (01:00 +0000)]
Simplify code. No functionality change.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@169198 91177308-0d34-0410-b5e6-96231b3b80d8

11 years agoStack Alignment: when creating stack objects in MachineFrameInfo, make sure
Manman Ren [Tue, 4 Dec 2012 00:52:33 +0000 (00:52 +0000)]
Stack Alignment: when creating stack objects in MachineFrameInfo, make sure
the alignment is clamped to TargetFrameLowering.getStackAlignment if the target
does not support stack realignment or the option "realign-stack" is off.

This will cause miscompile if the address is treated as aligned and add is
replaced with or in DAGCombine.

Added a bool StackRealignable to TargetFrameLowering to check whether stack
realignment is implemented for the target. Also added a bool RealignOption
to MachineFrameInfo to check whether the option "realign-stack" is on.

rdar://12713765

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@169197 91177308-0d34-0410-b5e6-96231b3b80d8

11 years agoUse dyn_cast instead of isa and cast. No functionality change.
Jakub Staszak [Tue, 4 Dec 2012 00:50:06 +0000 (00:50 +0000)]
Use dyn_cast instead of isa and cast. No functionality change.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@169196 91177308-0d34-0410-b5e6-96231b3b80d8

11 years agoLoopVectorize.cpp: Suppress a warning. [-Wunused-variable]
NAKAMURA Takumi [Tue, 4 Dec 2012 00:49:34 +0000 (00:49 +0000)]
LoopVectorize.cpp: Suppress a warning. [-Wunused-variable]

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@169195 91177308-0d34-0410-b5e6-96231b3b80d8

11 years agoFix whitespace.
NAKAMURA Takumi [Tue, 4 Dec 2012 00:49:28 +0000 (00:49 +0000)]
Fix whitespace.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@169194 91177308-0d34-0410-b5e6-96231b3b80d8

11 years agoRemove the old TRI::ResolveRegAllocHint() and getRawAllocationOrder() hooks.
Jakob Stoklund Olesen [Tue, 4 Dec 2012 00:46:13 +0000 (00:46 +0000)]
Remove the old TRI::ResolveRegAllocHint() and getRawAllocationOrder() hooks.

These functions have been replaced by TRI::getRegAllocationHints() which
provides the same capabilities.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@169192 91177308-0d34-0410-b5e6-96231b3b80d8

11 years agoRemove VirtRegMap::getRegAllocPref().
Jakob Stoklund Olesen [Tue, 4 Dec 2012 00:35:59 +0000 (00:35 +0000)]
Remove VirtRegMap::getRegAllocPref().

Now that there can be multiple hint registers from targets, it doesn't
make sense to have a function that returns 'the' preferred register.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@169190 91177308-0d34-0410-b5e6-96231b3b80d8

11 years agoUse MRI::getSimpleHint() instead of getRegAllocPref() in remaining cases.
Jakob Stoklund Olesen [Tue, 4 Dec 2012 00:30:22 +0000 (00:30 +0000)]
Use MRI::getSimpleHint() instead of getRegAllocPref() in remaining cases.

Targets can provide multiple hints now, so getRegAllocPref() doesn't
make sense any longer because it only returns one preferred register.
Replace it with getSimpleHint() in the remaining heuristics. This
function only

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@169188 91177308-0d34-0410-b5e6-96231b3b80d8

11 years agoStack Alignment: move functions from header file MachineFrameInfo.h.
Manman Ren [Tue, 4 Dec 2012 00:26:44 +0000 (00:26 +0000)]
Stack Alignment: move functions from header file MachineFrameInfo.h.

No functional change for this commit. The follow-up patch will add more stuff to
these functions.

rdar://12713765

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@169186 91177308-0d34-0410-b5e6-96231b3b80d8

11 years agoRuntimeDyld: Fix up r169178. MSVC doesn't like "or".
NAKAMURA Takumi [Tue, 4 Dec 2012 00:08:14 +0000 (00:08 +0000)]
RuntimeDyld: Fix up r169178. MSVC doesn't like "or".

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@169183 91177308-0d34-0410-b5e6-96231b3b80d8

11 years agordar://12329730 (2nd part)
Shuxin Yang [Tue, 4 Dec 2012 00:04:54 +0000 (00:04 +0000)]
rdar://12329730 (2nd part)

 This change tries to simmplify E1 = " X >> C1 << C2" into :
  - E2 = "X << (C2 - C1)" if C2 > C1, or
  - E2 = "X >> (C1 - C2)" if C1 > C2, or
  - E2 = X if C1 == C2.

 Reviewed by Nadav. Thanks!

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@169182 91177308-0d34-0410-b5e6-96231b3b80d8

11 years agoAdd VirtRegMap::hasKnownPreference().
Jakob Stoklund Olesen [Mon, 3 Dec 2012 23:23:50 +0000 (23:23 +0000)]
Add VirtRegMap::hasKnownPreference().

Virtual registers with a known preferred register are prioritized by
RAGreedy. This function makes the condition explicit without depending
on getRegAllocPref().

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@169179 91177308-0d34-0410-b5e6-96231b3b80d8

11 years agoRuntime dynamic linker for MCJIT should support MIPS BigEndian architecture.
Akira Hatanaka [Mon, 3 Dec 2012 23:12:19 +0000 (23:12 +0000)]
Runtime dynamic linker for MCJIT should support MIPS BigEndian architecture.
This small change adds support for that. It will make all MCJIT tests pass
in make-check on BigEndian platforms.

Patch by Petar Jovanovic.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@169178 91177308-0d34-0410-b5e6-96231b3b80d8

11 years agoClassic JIT is still being supported by MIPS, along with MCJIT.
Akira Hatanaka [Mon, 3 Dec 2012 23:11:12 +0000 (23:11 +0000)]
Classic JIT is still being supported by MIPS, along with MCJIT.
This change adds endian-awareness to MipsJITInfo and emitWordLE in
MipsCodeEmitter has become emitWord now to support both endianness.

Patch by Petar Jovanovic.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@169177 91177308-0d34-0410-b5e6-96231b3b80d8

11 years agoMinor tweaking to SmallVector static size.
Michael Ilseman [Mon, 3 Dec 2012 22:57:47 +0000 (22:57 +0000)]
Minor tweaking to SmallVector static size.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@169176 91177308-0d34-0410-b5e6-96231b3b80d8

11 years agominor renaming, documentation and cleanups.
Nadav Rotem [Mon, 3 Dec 2012 22:57:09 +0000 (22:57 +0000)]
minor renaming, documentation and cleanups.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@169175 91177308-0d34-0410-b5e6-96231b3b80d8

11 years agoFunctions in MipsCodeEmitter.cpp that expand unaligned loads/stores are dead
Akira Hatanaka [Mon, 3 Dec 2012 22:51:22 +0000 (22:51 +0000)]
Functions in MipsCodeEmitter.cpp that expand unaligned loads/stores are dead
code. Removing it.

Patch by Petar Jovanovic.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@169174 91177308-0d34-0410-b5e6-96231b3b80d8

11 years agoUse the new getRegAllocationHints() hook from AllocationOrder.
Jakob Stoklund Olesen [Mon, 3 Dec 2012 22:51:04 +0000 (22:51 +0000)]
Use the new getRegAllocationHints() hook from AllocationOrder.

This simplifies the hinting code quite a bit while making the targets
easier to write at the same time.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@169173 91177308-0d34-0410-b5e6-96231b3b80d8

11 years agoconstify the cost API
Nadav Rotem [Mon, 3 Dec 2012 22:47:12 +0000 (22:47 +0000)]
constify the cost API

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@169172 91177308-0d34-0410-b5e6-96231b3b80d8

11 years agoIF-conversion: teach the cost-model how to grade if-converted loops.
Nadav Rotem [Mon, 3 Dec 2012 22:46:31 +0000 (22:46 +0000)]
IF-conversion: teach the cost-model how to grade if-converted loops.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@169171 91177308-0d34-0410-b5e6-96231b3b80d8

11 years agoImplement ARMBaseRegisterInfo::getRegAllocationHints().
Jakob Stoklund Olesen [Mon, 3 Dec 2012 22:35:35 +0000 (22:35 +0000)]
Implement ARMBaseRegisterInfo::getRegAllocationHints().

This provides the same functionality as getRawAllocationOrder() for the
even/odd hints, but without the many constant register arrays.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@169169 91177308-0d34-0410-b5e6-96231b3b80d8

11 years agoDefine store instructions with base+immediate offset addressing mode
Jyotsna Verma [Mon, 3 Dec 2012 22:26:28 +0000 (22:26 +0000)]
Define store instructions with base+immediate offset addressing mode
using multiclass.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@169168 91177308-0d34-0410-b5e6-96231b3b80d8

11 years ago[Support] Make FileOutputBuffer work on Windows.
Michael J. Spencer [Mon, 3 Dec 2012 22:09:52 +0000 (22:09 +0000)]
[Support] Make FileOutputBuffer work on Windows.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@169167 91177308-0d34-0410-b5e6-96231b3b80d8

11 years ago[Support][FileSystem] Fix open mode in resize_file on Windows.
Michael J. Spencer [Mon, 3 Dec 2012 22:09:31 +0000 (22:09 +0000)]
[Support][FileSystem] Fix open mode in resize_file on Windows.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@169166 91177308-0d34-0410-b5e6-96231b3b80d8

11 years agoRevert the header sort on this file.
Michael J. Spencer [Mon, 3 Dec 2012 22:07:00 +0000 (22:07 +0000)]
Revert the header sort on this file.

"Windows.h" includes <Windows.h> which defines a bunch of stuff it shouldn't
(even with all the restriction macros). We have no control over this file, so
make it's scope as small as possible.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@169165 91177308-0d34-0410-b5e6-96231b3b80d8

11 years agomoves doInitialization and doFinalization to the Pass class and removes some unreacha...
Pedro Artigas [Mon, 3 Dec 2012 21:56:57 +0000 (21:56 +0000)]
moves doInitialization and doFinalization to the Pass class and removes some unreachable code in MachineModuleInfo

reviewed by Evan Cheng <evan.cheng@apple.com>

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@169164 91177308-0d34-0410-b5e6-96231b3b80d8

11 years agoAdd a getMemorySize() function for DenseSet.
Argyrios Kyrtzidis [Mon, 3 Dec 2012 21:46:21 +0000 (21:46 +0000)]
Add a getMemorySize() function for DenseSet.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@169163 91177308-0d34-0410-b5e6-96231b3b80d8

11 years agoNow that we have a basic if-conversion infrastructure we can rename the
Nadav Rotem [Mon, 3 Dec 2012 21:33:08 +0000 (21:33 +0000)]
Now that we have a basic if-conversion infrastructure we can rename the
"single basic block loop vectorizer" to "innermost loop vectorizer".

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@169158 91177308-0d34-0410-b5e6-96231b3b80d8

11 years agoSince this SmallVector immediately grows on the next line, don't waste stack space...
Michael Ilseman [Mon, 3 Dec 2012 21:29:36 +0000 (21:29 +0000)]
Since this SmallVector immediately grows on the next line, don't waste stack space. SmallVector is still needed due to existing APIs growing their arguments

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@169157 91177308-0d34-0410-b5e6-96231b3b80d8

11 years agoAdd a new hook for providing register allocator hints more flexibly.
Jakob Stoklund Olesen [Mon, 3 Dec 2012 21:17:00 +0000 (21:17 +0000)]
Add a new hook for providing register allocator hints more flexibly.

The TargetRegisterInfo::getRegAllocationHints() function is going to
replace the existing mechanisms for providing target-dependent hints to
the register allocator: ResolveRegAllocHint() and
getRawAllocationOrder().

The new hook is more flexible because it allows the target to provide
multiple preferred candidate registers for each virtual register, and it
is easier to use because targets are not required to return a reference
to a constant array like getRawAllocationOrder().

An optional VirtRegMap argument can be used to provide target-dependent
hints that depend on the provisional assignments of other virtual
registers.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@169154 91177308-0d34-0410-b5e6-96231b3b80d8

11 years agoDefine load instructions with base+immediate offset addressing mode
Jyotsna Verma [Mon, 3 Dec 2012 21:13:13 +0000 (21:13 +0000)]
Define load instructions with base+immediate offset addressing mode
using multiclass.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@169153 91177308-0d34-0410-b5e6-96231b3b80d8

11 years agoAdd initial support for IF-conversion. This patch implements the first 1/3,
Nadav Rotem [Mon, 3 Dec 2012 21:06:35 +0000 (21:06 +0000)]
Add initial support for IF-conversion. This patch implements the first 1/3,
which is the legality of the if-conversion transformation. The next step is to
implement the cost-model for the if-converted code as well as the
vectorization itself.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@169152 91177308-0d34-0410-b5e6-96231b3b80d8

11 years agoDefine unsigned const-ext predicates.
Jyotsna Verma [Mon, 3 Dec 2012 20:39:45 +0000 (20:39 +0000)]
Define unsigned const-ext predicates.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@169149 91177308-0d34-0410-b5e6-96231b3b80d8

11 years agoRemoving unnecessary 'else' statement from the predicates defined in HexagonOperards.td.
Jyotsna Verma [Mon, 3 Dec 2012 20:14:38 +0000 (20:14 +0000)]
Removing unnecessary 'else' statement from the predicates defined in HexagonOperards.td.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@169148 91177308-0d34-0410-b5e6-96231b3b80d8

11 years agoEliminate redundant bitwise operations when using a llvm/ADT/PointerUnion.
Argyrios Kyrtzidis [Mon, 3 Dec 2012 19:59:23 +0000 (19:59 +0000)]
Eliminate redundant bitwise operations when using a llvm/ADT/PointerUnion.

For comparison, with this code sample:

PointerUnion<int *, char *> Data;
PointerUnion<int *, char *> foo1() {
Data = new int;
return new int;
}
PointerUnion<int *, char *> foo2() {
Data = new char;
return new char;
}

Before this patch we would get:

define i64 @_Z4foo1v() uwtable ssp {
  %1 = tail call noalias i8* @_Znwm(i64 4)
  %2 = ptrtoint i8* %1 to i64
  %3 = load i64* getelementptr inbounds (%"class.llvm::PointerUnion"* @Data, i64 0, i32 0, i32 0), align 8
  %4 = and i64 %3, 1
  %.masked.i = and i64 %2, -3
  %5 = or i64 %4, %.masked.i
  store i64 %5, i64* getelementptr inbounds (%"class.llvm::PointerUnion"* @Data, i64 0, i32 0, i32 0), align 8
  %6 = tail call noalias i8* @_Znwm(i64 4)
  %7 = ptrtoint i8* %6 to i64
  %8 = and i64 %7, -3
  ret i64 %8
}

define i64 @_Z4foo2v() uwtable ssp {
  %1 = tail call noalias i8* @_Znwm(i64 1)
  %2 = ptrtoint i8* %1 to i64
  %3 = load i64* getelementptr inbounds (%"class.llvm::PointerUnion"* @Data, i64 0, i32 0, i32 0), align 8
  %4 = and i64 %3, 1
  %5 = or i64 %2, %4
  %6 = or i64 %5, 2
  store i64 %6, i64* getelementptr inbounds (%"class.llvm::PointerUnion"* @Data, i64 0, i32 0, i32 0), align 8
  %7 = tail call noalias i8* @_Znwm(i64 1)
  %8 = ptrtoint i8* %7 to i64
  %9 = or i64 %8, 2
  ret i64 %9
}

After the patch:

define i64 @_Z4foo1v() uwtable ssp {
  %1 = tail call noalias i8* @_Znwm(i64 4)
  %2 = ptrtoint i8* %1 to i64
  store i64 %2, i64* getelementptr inbounds (%"class.llvm::PointerUnion"* @Data, i64 0, i32 0, i32 0), align 8
  %3 = tail call noalias i8* @_Znwm(i64 4)
  %4 = ptrtoint i8* %3 to i64
  ret i64 %4
}

declare noalias i8* @_Znwm(i64)

define i64 @_Z4foo2v() uwtable ssp {
  %1 = tail call noalias i8* @_Znwm(i64 1)
  %2 = ptrtoint i8* %1 to i64
  %3 = or i64 %2, 2
  store i64 %3, i64* getelementptr inbounds (%"class.llvm::PointerUnion"* @Data, i64 0, i32 0, i32 0), align 8
  %4 = tail call noalias i8* @_Znwm(i64 1)
  %5 = ptrtoint i8* %4 to i64
  %6 = or i64 %5, 2
  ret i64 %6
}

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@169147 91177308-0d34-0410-b5e6-96231b3b80d8

11 years agoSimplify this test a bit because DWARF emission/dumping on some platforms
Eli Bendersky [Mon, 3 Dec 2012 19:58:12 +0000 (19:58 +0000)]
Simplify this test a bit because DWARF emission/dumping on some platforms
is not yet good enough for more sophistication. The important goal of this
test is to make sure llc doesn't crash on this IR like it used to.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@169146 91177308-0d34-0410-b5e6-96231b3b80d8

11 years agoAdd 'getInt64Field()' method to get the signed integer instead of unsigned.
Bill Wendling [Mon, 3 Dec 2012 19:44:25 +0000 (19:44 +0000)]
Add 'getInt64Field()' method to get the signed integer instead of unsigned.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@169145 91177308-0d34-0410-b5e6-96231b3b80d8

11 years agoASan: add blacklist file to ASan pass options. Clang patch for this will follow.
Alexey Samsonov [Mon, 3 Dec 2012 19:09:26 +0000 (19:09 +0000)]
ASan: add blacklist file to ASan pass options. Clang patch for this will follow.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@169143 91177308-0d34-0410-b5e6-96231b3b80d8

11 years agoFix PR12942: Allow two CUs to be generated from the same source file.
Eli Bendersky [Mon, 3 Dec 2012 18:45:45 +0000 (18:45 +0000)]
Fix PR12942: Allow two CUs to be generated from the same source file.

Thanks Eric for the review.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@169142 91177308-0d34-0410-b5e6-96231b3b80d8

11 years agoTeach the jump threading optimization to stop scanning the basic block when calculati...
Nadav Rotem [Mon, 3 Dec 2012 17:34:44 +0000 (17:34 +0000)]
Teach the jump threading optimization to stop scanning the basic block when calculating the cost after passing the threshold.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@169135 91177308-0d34-0410-b5e6-96231b3b80d8

11 years agoRevert r169039, "Aggregate pass execution time report by pass ID instead of pass...
Jakob Stoklund Olesen [Mon, 3 Dec 2012 17:31:11 +0000 (17:31 +0000)]
Revert r169039, "Aggregate pass execution time report by pass ID instead of pass instance."

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@169134 91177308-0d34-0410-b5e6-96231b3b80d8

11 years agoSort the #include lines for the include/... tree with the script.
Chandler Carruth [Mon, 3 Dec 2012 17:02:12 +0000 (17:02 +0000)]
Sort the #include lines for the include/... tree with the script.

AKA: Recompile *ALL* the source code!

This one went much better. No manual edits here. I spot-checked for
silliness and grep-checked for really broken edits and everything seemed
good. It all still compiles. Yell if you see something that looks goofy.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@169133 91177308-0d34-0410-b5e6-96231b3b80d8

11 years agoTeach the include sorter to not choke on include guards. ;]
Chandler Carruth [Mon, 3 Dec 2012 17:01:46 +0000 (17:01 +0000)]
Teach the include sorter to not choke on include guards. ;]

Kind of important when prepping the include/... tree version of the sort
changes.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@169132 91177308-0d34-0410-b5e6-96231b3b80d8

11 years agoUse the new script to sort the includes of every file under lib.
Chandler Carruth [Mon, 3 Dec 2012 16:50:05 +0000 (16:50 +0000)]
Use the new script to sort the includes of every file under lib.

Sooooo many of these had incorrect or strange main module includes.
I have manually inspected all of these, and fixed the main module
include to be the nearest plausible thing I could find. If you own or
care about any of these source files, I encourage you to take some time
and check that these edits were sensible. I can't have broken anything
(I strictly added headers, and reordered them, never removed), but they
may not be the headers you'd really like to identify as containing the
API being implemented.

Many forward declarations and missing includes were added to a header
files to allow them to parse cleanly when included first. The main
module rule does in fact have its merits. =]

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@169131 91177308-0d34-0410-b5e6-96231b3b80d8

11 years agoclarify that this isn't lld.
Chris Lattner [Mon, 3 Dec 2012 16:08:32 +0000 (16:08 +0000)]
clarify that this isn't lld.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@169130 91177308-0d34-0410-b5e6-96231b3b80d8

11 years agoAdd a completely hack-ish tool to sort includes according to the coding
Chandler Carruth [Mon, 3 Dec 2012 14:23:44 +0000 (14:23 +0000)]
Add a completely hack-ish tool to sort includes according to the coding
standards.

I am a terrible Python programmer. Patches more the welcome. Please tell
me how this should look if it should look differently. It's just a tiny
little script so it didn't make sense to go through pre-commit review,
especially as someone who actually knows python may want to just rip it
apart and do it The Right Way.

I will be preparing a commit shortly that uses this script to
canonicalize *all* of the #include lines in LLVM. Really, all of them.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@169125 91177308-0d34-0410-b5e6-96231b3b80d8

11 years agoRemove some buggy and apparantly unnecessary code from SROA.
Chandler Carruth [Mon, 3 Dec 2012 10:59:55 +0000 (10:59 +0000)]
Remove some buggy and apparantly unnecessary code from SROA.

The partitioning logic attempted to handle uses of an alloca with an
offset starting before the alloca so long as the use had some overlap
with the alloca itself. However, there was a bug where we tested
'(uint64_t)Offset >= AllocSize' without first checking whether 'Offset'
was positive. As a consequence, essentially every negative offset (that
is, starting *before* the alloca does) would be thrown out, even if it
was overlapping. The subsequent code to throw out negative offsets which
were actually non-overlapping was essentially dead. The code to *handle*
overlapping negative offsets was actually dead!

I've just removed all of this, and taught SROA to discard any uses which
start prior to the alloca from the beginning. It has the lovely property
of simplifying the code. =] All the tests still pass, and in fact no new
tests are needed as this is already covered by our testsuite. Fixing the
code so that negative offsets work the way the comments indicate they
were supposed to work causes regressions. That's how I found this.

Anyways, this is all progress in the correct direction -- tightening up
SROA to be maximally aggressive. Some day, I really hope to turn
out-of-bounds accesses to an alloca into 'unreachable'.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@169120 91177308-0d34-0410-b5e6-96231b3b80d8

11 years agofix stats for added checks
Nuno Lopes [Mon, 3 Dec 2012 10:15:03 +0000 (10:15 +0000)]
fix stats for added checks

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@169119 91177308-0d34-0410-b5e6-96231b3b80d8

11 years agoDefine signed const-ext predicates.
Jyotsna Verma [Mon, 3 Dec 2012 06:54:50 +0000 (06:54 +0000)]
Define signed const-ext predicates.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@169117 91177308-0d34-0410-b5e6-96231b3b80d8

11 years agoRemove 'deplibs' keyword, since it's no longer used.
Bill Wendling [Mon, 3 Dec 2012 06:34:06 +0000 (06:34 +0000)]
Remove 'deplibs' keyword, since it's no longer used.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@169116 91177308-0d34-0410-b5e6-96231b3b80d8

11 years agoClean up, bring up-to-date and apply consistent formatting.
Eli Bendersky [Mon, 3 Dec 2012 04:10:58 +0000 (04:10 +0000)]
Clean up, bring up-to-date and apply consistent formatting.

This document is a long-time pet peeve :-) More fixes to come.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@169115 91177308-0d34-0410-b5e6-96231b3b80d8

11 years agoAllow merging multiple store sequences on the same chain.
Nadav Rotem [Sun, 2 Dec 2012 17:14:09 +0000 (17:14 +0000)]
Allow merging multiple store sequences on the same chain.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@169111 91177308-0d34-0410-b5e6-96231b3b80d8

11 years agoBring vim keyword lists up to date.
Benjamin Kramer [Sun, 2 Dec 2012 16:40:38 +0000 (16:40 +0000)]
Bring vim keyword lists up to date.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@169110 91177308-0d34-0410-b5e6-96231b3b80d8

11 years agoFix a bug in FileCheck that wouldn't let define variables as follows:
Eli Bendersky [Sun, 2 Dec 2012 16:02:41 +0000 (16:02 +0000)]
Fix a bug in FileCheck that wouldn't let define variables as follows:

; CHECK: [[VAR:[a-z]]]

The problem was that to find the end of the regex var definition, it was
simplistically looking for the next ]] and finding the incorrect one. A
better approach is to count nesting of brackets (taking escaping into
account). This way the brackets that are part of the regex can be discovered
and skipped properly, and the ]] ending is detected in the right place.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@169109 91177308-0d34-0410-b5e6-96231b3b80d8

11 years agoFix an invalid regex in the test
Eli Bendersky [Sun, 2 Dec 2012 15:46:02 +0000 (15:46 +0000)]
Fix an invalid regex in the test

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@169108 91177308-0d34-0410-b5e6-96231b3b80d8

11 years agoSimplify the coding standards for #include ordering. The ordering is now
Chandler Carruth [Sun, 2 Dec 2012 11:53:27 +0000 (11:53 +0000)]
Simplify the coding standards for #include ordering. The ordering is now
trivially achievable with an editor. I'll likely check in a silly python
script to help with this too.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@169107 91177308-0d34-0410-b5e6-96231b3b80d8

11 years agoUnbreak Sphinx build
Justin Holewinski [Sun, 2 Dec 2012 02:47:48 +0000 (02:47 +0000)]
Unbreak Sphinx build

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@169106 91177308-0d34-0410-b5e6-96231b3b80d8

11 years agoAnother fix attempt to Makefile.sphinx - copy the PNGs from tutorial/ as well,
Eli Bendersky [Sat, 1 Dec 2012 22:21:04 +0000 (22:21 +0000)]
Another fix attempt to Makefile.sphinx - copy the PNGs from tutorial/ as well,
and recursive copying is not required for the tutorial/ directory.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@169105 91177308-0d34-0410-b5e6-96231b3b80d8

11 years agoAttempt to fix Makefile.sphinx to not generate errors while trying to copy
Eli Bendersky [Sat, 1 Dec 2012 22:11:59 +0000 (22:11 +0000)]
Attempt to fix Makefile.sphinx to not generate errors while trying to copy
files from tutorial/.svn

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@169104 91177308-0d34-0410-b5e6-96231b3b80d8

11 years agoUpdate FileCheck's documentation to mention recently added feature of
Eli Bendersky [Sat, 1 Dec 2012 22:03:57 +0000 (22:03 +0000)]
Update FileCheck's documentation to mention recently added feature of
matching a variable defined on the same line.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@169103 91177308-0d34-0410-b5e6-96231b3b80d8

11 years ago[python] Add markup option to disassembler
Gregory Szorc [Sat, 1 Dec 2012 21:57:30 +0000 (21:57 +0000)]
[python] Add markup option to disassembler

Patch contributed by Wladimir J. van der Laan <laanwj@gmail.com>

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@169102 91177308-0d34-0410-b5e6-96231b3b80d8

11 years agoSupport referencing variables defined on the same line.
Eli Bendersky [Sat, 1 Dec 2012 21:54:48 +0000 (21:54 +0000)]
Support referencing variables defined on the same line.

See http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20121126/157198.html
and related discussions.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@169101 91177308-0d34-0410-b5e6-96231b3b80d8

11 years agoRemove bugzilla link.
James Molloy [Sat, 1 Dec 2012 14:44:23 +0000 (14:44 +0000)]
Remove bugzilla link.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@169091 91177308-0d34-0410-b5e6-96231b3b80d8

11 years agoDocumentation: convert WritingAnLLVMBackend.html to reST
Dmitri Gribenko [Sat, 1 Dec 2012 12:13:48 +0000 (12:13 +0000)]
Documentation: convert WritingAnLLVMBackend.html to reST

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@169087 91177308-0d34-0410-b5e6-96231b3b80d8

11 years agoAdd .arcconfig to the repository. Useful if someone wants to use phabricator's comman...
Benjamin Kramer [Sat, 1 Dec 2012 12:07:58 +0000 (12:07 +0000)]
Add .arcconfig to the repository. Useful if someone wants to use phabricator's command line tool.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@169085 91177308-0d34-0410-b5e6-96231b3b80d8

11 years agoSROA: Avoid struct and array types early to avoid creating an overly large integer...
Benjamin Kramer [Sat, 1 Dec 2012 11:53:32 +0000 (11:53 +0000)]
SROA: Avoid struct and array types early to avoid creating an overly large integer type.

Fixes PR14465.

Differential Revision: http://llvm-reviews.chandlerc.com/D148

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@169084 91177308-0d34-0410-b5e6-96231b3b80d8

11 years agoRevert previous check in r168581, r169079 as they are still in code review status.
Zhou Sheng [Sat, 1 Dec 2012 10:54:28 +0000 (10:54 +0000)]
Revert previous check in r168581, r169079 as they are still in code review status.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@169083 91177308-0d34-0410-b5e6-96231b3b80d8

11 years agoThe patch is to improve the memory footprint of pass GlobalOpt.
Zhou Sheng [Sat, 1 Dec 2012 04:38:53 +0000 (04:38 +0000)]
The patch is to improve the memory footprint of pass GlobalOpt.
Also check in a case to repeat the issue, on which 'opt -globalopt' consumes 1.6GB memory.
The big memory footprint cause is that current GlobalOpt one by one hoists and stores the leaf element constant into the global array, in each iteration, it recreates the global array initializer constant and leave the old initializer alone. This may result in many obsolete constants left.
For example:  we have global array @rom = global [16 x i32] zeroinitializer
After the first element value is hoisted and installed:   @rom = global [16 x i32] [ 1, 0, 0, ... ]
After the second element value is installed:  @rom = global [16 x 32] [ 1, 2, 0, 0, ... ]        // here the previous initializer is obsolete
...
When the transform is done, we have 15 obsolete initializers left useless.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@169079 91177308-0d34-0410-b5e6-96231b3b80d8

11 years agoVMCore/DebugInfo.cpp: DICompileUnit::getSubprograms(): Check numOperands().
NAKAMURA Takumi [Sat, 1 Dec 2012 02:23:45 +0000 (02:23 +0000)]
VMCore/DebugInfo.cpp: DICompileUnit::getSubprograms(): Check numOperands().

2012-11-30-misched-dbg.ll had crashed. Then (MDNode)N was "!{}".
I am not sure it would be ill-formed or not.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@169074 91177308-0d34-0410-b5e6-96231b3b80d8

11 years agoMC/AsmParser: Avoid unnecessary use of SourceMgr::FindBufferForLoc()
Daniel Dunbar [Sat, 1 Dec 2012 01:38:48 +0000 (01:38 +0000)]
MC/AsmParser: Avoid unnecessary use of SourceMgr::FindBufferForLoc()

 - Each macro instantiation introduces a new buffer, and FindBufferForLoc() is
   linear, so previously macro instantiation could be N^2 for some pathological
   inputs.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@169073 91177308-0d34-0410-b5e6-96231b3b80d8

11 years agomisched: Fix RegisterPressureTracker handling of DebugVals.
Andrew Trick [Sat, 1 Dec 2012 01:22:49 +0000 (01:22 +0000)]
misched: Fix RegisterPressureTracker handling of DebugVals.

Assertion failed: (TopRPTracker.getPos() == RegionBegin && "bad initial Top tracker").
rdar://12790302.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@169072 91177308-0d34-0410-b5e6-96231b3b80d8

11 years agomisched: Fix the DAG builder to handle an undef operand at ExitSU.
Andrew Trick [Sat, 1 Dec 2012 01:22:44 +0000 (01:22 +0000)]
misched: Fix the DAG builder to handle an undef operand at ExitSU.

Assertion failed: (VNI && "No value to read by operand")
rdar://12790267.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@169071 91177308-0d34-0410-b5e6-96231b3b80d8

11 years agomisched: Fix LiveInterval update to better handle DebugVal.
Andrew Trick [Sat, 1 Dec 2012 01:22:41 +0000 (01:22 +0000)]
misched: Fix LiveInterval update to better handle DebugVal.

Assertion failed: (itr != mi2iMap.end() && "Instruction not found in maps.")
rdar://12777252.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@169070 91177308-0d34-0410-b5e6-96231b3b80d8

11 years agomisched: fix RegionBegin when DebugValues get shuffled to the top.
Andrew Trick [Sat, 1 Dec 2012 01:22:38 +0000 (01:22 +0000)]
misched: fix RegionBegin when DebugValues get shuffled to the top.

assert (RemainingInstrs == 0 && "Instruction count mismatch!")

rdar://12776937.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@169069 91177308-0d34-0410-b5e6-96231b3b80d8

11 years agoSimplify REG_SEQUENCE lowering.
Jakob Stoklund Olesen [Sat, 1 Dec 2012 01:06:44 +0000 (01:06 +0000)]
Simplify REG_SEQUENCE lowering.

The TwoAddressInstructionPass takes the machine code out of SSA form by
expanding REG_SEQUENCE instructions into copies. It is no longer
necessary to rewrite the registers used by a REG_SEQUENCE instruction
because the new coalescer algorithm can do it now.

REG_SEQUENCE is just converted to a sequence of sub-register copies now.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@169067 91177308-0d34-0410-b5e6-96231b3b80d8

11 years agoUpdate the emacs mode to recognize fadd, fsum, fmul, fdiv, frem, fcmp, icmp
Michael Ilseman [Sat, 1 Dec 2012 00:42:27 +0000 (00:42 +0000)]
Update the emacs mode to recognize fadd, fsum, fmul, fdiv, frem, fcmp, icmp

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@169064 91177308-0d34-0410-b5e6-96231b3b80d8

11 years agoAdd some first skeleton work for the DWARF5 Fission proposal. Emit
Eric Christopher [Fri, 30 Nov 2012 23:59:06 +0000 (23:59 +0000)]
Add some first skeleton work for the DWARF5 Fission proposal. Emit
part of the compile unit CU and start separating out information into
the various sections that will be pulled out later.

WIP.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@169061 91177308-0d34-0410-b5e6-96231b3b80d8

11 years agoConvert COPY instructions into KILLs if they have implicit defs.
Jakob Stoklund Olesen [Fri, 30 Nov 2012 23:53:00 +0000 (23:53 +0000)]
Convert COPY instructions into KILLs if they have implicit defs.

MachineCopyPropagation doesn't understand super-register liveness well
enough to be able to remove implicit defs of super-registers.

This fixes a problem in ARM/2012-01-26-CopyPropKills.ll that is exposed
by an future TwoAddressInstructionPass change. The KILL instructions are
removed before the machine code is emitted.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@169060 91177308-0d34-0410-b5e6-96231b3b80d8

11 years agoAdd support for fission attributes/forms/operations -> string.
Eric Christopher [Fri, 30 Nov 2012 23:20:43 +0000 (23:20 +0000)]
Add support for fission attributes/forms/operations -> string.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@169056 91177308-0d34-0410-b5e6-96231b3b80d8

11 years agoVim mode updated to recognize fast-math flags
Michael Ilseman [Fri, 30 Nov 2012 23:18:21 +0000 (23:18 +0000)]
Vim mode updated to recognize fast-math flags

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@169055 91177308-0d34-0410-b5e6-96231b3b80d8

11 years agoRemoved redundancy in wording
Michael Ilseman [Fri, 30 Nov 2012 23:14:52 +0000 (23:14 +0000)]
Removed redundancy in wording

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@169053 91177308-0d34-0410-b5e6-96231b3b80d8

11 years agoClean up the documentation to have a common description area for fast-math flags.
Michael Ilseman [Fri, 30 Nov 2012 23:12:42 +0000 (23:12 +0000)]
Clean up the documentation to have a common description area for fast-math flags.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@169052 91177308-0d34-0410-b5e6-96231b3b80d8

11 years agoBe more clear on what parts of code I own.
Bill Wendling [Fri, 30 Nov 2012 23:00:25 +0000 (23:00 +0000)]
Be more clear on what parts of code I own.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@169050 91177308-0d34-0410-b5e6-96231b3b80d8

11 years agoreversed the logic of the log2 detection routine to reduce the number of nested ifs
Pedro Artigas [Fri, 30 Nov 2012 22:47:15 +0000 (22:47 +0000)]
reversed the logic of the log2 detection routine to reduce the number of nested ifs

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@169049 91177308-0d34-0410-b5e6-96231b3b80d8

11 years agominor cleanups
Nadav Rotem [Fri, 30 Nov 2012 22:37:11 +0000 (22:37 +0000)]
minor cleanups

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@169048 91177308-0d34-0410-b5e6-96231b3b80d8

11 years agoFix a bug in APFloat.cpp: declare APFloat after fltSemantics it
Alexey Samsonov [Fri, 30 Nov 2012 22:27:54 +0000 (22:27 +0000)]
Fix a bug in APFloat.cpp: declare APFloat after fltSemantics it
uses. APFloat::convert() takes the pointer to the fltSemantics
variable, which is later accessed it in ~APFloat() desctructor.
That is, semantics must still be alive at the moment we delete
APFloat.

Found by experimental AddressSanitizer use-after-scope checker.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@169047 91177308-0d34-0410-b5e6-96231b3b80d8

11 years agoAdd me as LTO code owner.
Bill Wendling [Fri, 30 Nov 2012 22:15:08 +0000 (22:15 +0000)]
Add me as LTO code owner.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@169046 91177308-0d34-0410-b5e6-96231b3b80d8

11 years agoReplace r168930 with a more reasonable patch.
Bill Wendling [Fri, 30 Nov 2012 22:08:55 +0000 (22:08 +0000)]
Replace r168930 with a more reasonable patch.

The original patch removed a bunch of code that the SjLjEHPrepare pass placed
into the entry block if all of the landing pads were removed during the
CodeGenPrepare class. The more natural way of doing things is to run the CGP
*before* we run the SjLjEHPrepare pass.

Make it so!

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@169044 91177308-0d34-0410-b5e6-96231b3b80d8

11 years agoAddresses many style issues with prior checkin (r169025)
Pedro Artigas [Fri, 30 Nov 2012 22:07:05 +0000 (22:07 +0000)]
Addresses many style issues with prior checkin (r169025)

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@169043 91177308-0d34-0410-b5e6-96231b3b80d8

11 years agoAdd a -time-compilations=<N> option to llc.
Jakob Stoklund Olesen [Fri, 30 Nov 2012 21:42:47 +0000 (21:42 +0000)]
Add a -time-compilations=<N> option to llc.

This causes llc to repeat the module compilation N times, making it
possible to get more accurate information from -time-passes when
compiling small modules.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@169040 91177308-0d34-0410-b5e6-96231b3b80d8

11 years agoAggregate pass execution time report by pass ID instead of pass instance.
Jakob Stoklund Olesen [Fri, 30 Nov 2012 21:42:45 +0000 (21:42 +0000)]
Aggregate pass execution time report by pass ID instead of pass instance.

This avoids unidentified duplicates in the pass execution time report
when a pass runs more than once in the pass manager pipeline.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@169039 91177308-0d34-0410-b5e6-96231b3b80d8

11 years agolit: Add a simple test suite for checking test runner parallelism.
Daniel Dunbar [Fri, 30 Nov 2012 21:34:04 +0000 (21:34 +0000)]
lit: Add a simple test suite for checking test runner parallelism.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@169038 91177308-0d34-0410-b5e6-96231b3b80d8

11 years agotest/CodeGen/PowerPC/vec_mul.ll: Add a triple. Thanks, Hal.
Chad Rosier [Fri, 30 Nov 2012 19:15:10 +0000 (19:15 +0000)]
test/CodeGen/PowerPC/vec_mul.ll: Add a triple. Thanks, Hal.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@169026 91177308-0d34-0410-b5e6-96231b3b80d8

11 years agoAdd fast math inst combine X*log2(Y*0.5)-->X*log2(Y)-X
Pedro Artigas [Fri, 30 Nov 2012 19:09:41 +0000 (19:09 +0000)]
Add fast math inst combine X*log2(Y*0.5)-->X*log2(Y)-X

reviewed by Michael Ilseman <milseman@apple.com>

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@169025 91177308-0d34-0410-b5e6-96231b3b80d8

11 years agoCodegen failure for vmull with small vectors
Sebastian Pop [Fri, 30 Nov 2012 19:08:04 +0000 (19:08 +0000)]
Codegen failure for vmull with small vectors

Codegen was failing with an assertion because of unexpected vector
operands when legalizing the selection DAG for a MUL instruction.

The asserting code was legalizing multiplies for vectors of size 128
bits. It uses a custom lowering to try and detect cases where it can
use a VMULL instruction instead of a VMOVL + VMUL.  The code was
looking for input operands to the MUL that had been sign or zero
extended. If it found the extended operands it would drop the
sign/zero extension and use the original vector size as input to a
VMULL instruction.

The code assumed that the original input vector was 64 bits so that
after dropping the extension it would fit directly into a D register
and could be used as an operand of a VMULL instruction. The input
code that trigger the failure used a vector of <4 x i8> that was
sign extended to <4 x i32>. It was not safe to drop the sign
extension in this case because the original vector is only 32 bits
wide. The fix is to insert a sign extension for the vector to reach
the required 64 bit size. In this particular example, the vector would
need to be sign extented to a <4 x i16>.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@169024 91177308-0d34-0410-b5e6-96231b3b80d8

11 years agotest/CodeGen/PowerPC/vec_mul.ll: Fix register operands.
Chad Rosier [Fri, 30 Nov 2012 18:29:01 +0000 (18:29 +0000)]
test/CodeGen/PowerPC/vec_mul.ll: Fix register operands.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@169020 91177308-0d34-0410-b5e6-96231b3b80d8

11 years agoUse multiclass for the load instructions with MEMri operand.
Jyotsna Verma [Fri, 30 Nov 2012 17:31:52 +0000 (17:31 +0000)]
Use multiclass for the load instructions with MEMri operand.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@169018 91177308-0d34-0410-b5e6-96231b3b80d8

11 years agoRemove the use of LPPassManager. We can remove LPM because we dont need to run any...
Nadav Rotem [Fri, 30 Nov 2012 17:27:53 +0000 (17:27 +0000)]
Remove the use of LPPassManager. We can remove LPM because we dont need to run any additional loop passes on the new vector loop.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@169016 91177308-0d34-0410-b5e6-96231b3b80d8