opencl/llvm.git
7 years agoFileCheck: Print a nice error message for missing closing ']' in regex vars.
Adrian Prantl [Fri, 3 Jan 2014 21:49:09 +0000 (21:49 +0000)]
FileCheck: Print a nice error message for missing closing ']' in regex vars.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@198449 91177308-0d34-0410-b5e6-96231b3b80d8

7 years agoRevert "For disassembly when adding a symbolic operand that is a C++ symbol name...
Reid Kleckner [Fri, 3 Jan 2014 19:56:20 +0000 (19:56 +0000)]
Revert "For disassembly when adding a symbolic operand that is a C++ symbol name, also put the human readable name in a comment."

This reverts commit r198441.

This change doesn't build on Windows, and doesn't do the right thing on
Linux and other platforms that don't use a _Z prefix instead of __Z for
C++ names.

It also had no tests, so it wasn't clear how to fix it forward.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@198445 91177308-0d34-0410-b5e6-96231b3b80d8

7 years agoFix typo.
Rafael Espindola [Fri, 3 Jan 2014 19:42:04 +0000 (19:42 +0000)]
Fix typo.

Thanks to Owen Anderson for noticing it.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@198443 91177308-0d34-0410-b5e6-96231b3b80d8

7 years agoFor disassembly when adding a symbolic operand that is a C++
Kevin Enderby [Fri, 3 Jan 2014 19:33:09 +0000 (19:33 +0000)]
For disassembly when adding a symbolic operand that is a C++
symbol name, also put the human readable name in a comment.

Also fix a bug in LLVMDisasmInstruction() that was not flushing
the raw_svector_ostream for the disassembled instruction string
before copying it to the output buffer that was causing truncation
of the output.

rdar://10173828

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@198441 91177308-0d34-0410-b5e6-96231b3b80d8

7 years agoMake the llvm mangler depend only on DataLayout.
Rafael Espindola [Fri, 3 Jan 2014 19:21:54 +0000 (19:21 +0000)]
Make the llvm mangler depend only on DataLayout.

Before this patch any program that wanted to know the final symbol name of a
GlobalValue had to link with Target.

This patch implements a compromise solution where the mangler uses DataLayout.
This way, any tool that already links with Target (llc, clang) gets the exact
behavior as before and new IR files can be mangled without linking with Target.

With this patch the mangler is constructed with just a DataLayout and DataLayout
is extended to include the information the Mangler needs.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@198438 91177308-0d34-0410-b5e6-96231b3b80d8

7 years ago[AArch64][NEON] Added SXTL and SXTL2 instruction aliases
Ana Pazos [Fri, 3 Jan 2014 19:20:31 +0000 (19:20 +0000)]
[AArch64][NEON] Added SXTL and SXTL2 instruction aliases

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@198437 91177308-0d34-0410-b5e6-96231b3b80d8

7 years agoRevert "Revert "Debug Info: Type Units: Simplify type hashing using IR-provided uniqu...
David Blaikie [Fri, 3 Jan 2014 18:59:42 +0000 (18:59 +0000)]
Revert "Revert "Debug Info: Type Units: Simplify type hashing using IR-provided unique names.""

This reverts commit r198398, thus reapplying r198397.

I had accidentally introduced an endianness issue when applying the hash
to the type unit. Using support::ulittle64_t in the reinterpret_cast in
addDwarfTypeUnitType fixes this issue.

Original commit message:

Debug Info: Type Units: Simplify type hashing using IR-provided unique
names.

What's good for LTO metadata size problems ought to be good for non-LTO
debug info size too, so let's rely on the same uniqueness in both cases.
If it's insufficient for non-LTO for whatever reason (since we now won't
be uniquing CU-local types or any C types - but these are likely to not
be the most significant contributors to type bloat) we should consider a
frontend solution that'll help both LTO and non-LTO alike, rather than
using DWARF-level DIE-hashing that only helps non-LTO debug info size.

It's also much simpler this way and benefits C++ even more since we can
deduplicate lexically separate definitions of the same C++ type since
they have the same mangled name.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@198436 91177308-0d34-0410-b5e6-96231b3b80d8

7 years agoFix loop rerolling pass failure with non-consant loop lower bound
David Peixotto [Fri, 3 Jan 2014 17:20:01 +0000 (17:20 +0000)]
Fix loop rerolling pass failure with non-consant loop lower bound

The loop rerolling pass was failing with an assertion failure from a
failed cast on loops like this:

  void foo(int *A, int *B, int m, int n) {
    for (int i = m; i < n; i+=4) {
      A[i+0] = B[i+0] * 4;
      A[i+1] = B[i+1] * 4;
      A[i+2] = B[i+2] * 4;
      A[i+3] = B[i+3] * 4;
    }
  }

The code was casting the SCEV-expanded code for the new
induction variable to a phi-node. When the loop had a non-constant
lower bound, the SCEV expander would end the code expansion with an
add insted of a phi node and the cast would fail.

It looks like the cast to a phi node was only needed to get the
induction variable value coming from the backedge to compute the end
of loop condition. This patch changes the loop reroller to compare
the induction variable to the number of times the backedge is taken
instead of the iteration count of the loop. In other words, we stop
the loop when the current value of the induction variable ==
IterationCount-1. Previously, the comparison was comparing the
induction variable value from the next iteration == IterationCount.

This problem only seems to occur on 32-bit targets. For some reason,
the loop is not rerolled on 64-bit targets.

PR18290

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@198425 91177308-0d34-0410-b5e6-96231b3b80d8

7 years agoMSVC 2010 build fix
Alp Toker [Fri, 3 Jan 2014 07:58:20 +0000 (07:58 +0000)]
MSVC 2010 build fix

Back out the part of r198399 that enabled LLVM_FINAL/LLVM_OVERRIDE on VS 2010.

DwarfUnit.h legitimately uses them on destructors which unfortunately triggers
Compiler Error C3665 (override specifier not allowed on a destructor/finalizer)
prior to MSVC 2012:

  virtual ~DwarfCompileUnit() LLVM_OVERRIDE;

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@198401 91177308-0d34-0410-b5e6-96231b3b80d8

7 years agoBasicAA: Use reachabilty instead of dominance for checking value equality in phi
Arnold Schwaighofer [Fri, 3 Jan 2014 05:47:03 +0000 (05:47 +0000)]
BasicAA: Use reachabilty instead of dominance for checking value equality in phi
cycles

This allows the value equality check to work even if we don't have a dominator
tree. Also add some more comments.

I was worried about compile time impacts and did not implement reachability but
used the dominance check in the initial patch. The trade-off was that the
dominator tree was required.
The llvm utility function isPotentiallyReachable cuts off the recursive search
after 32 visits. Testing did not show any compile time regressions showing my
worries unjustfied.

No compile time or performance regressions at O3 -flto -mavx on test-suite +
externals.

Addresses review comments from r198290.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@198400 91177308-0d34-0410-b5e6-96231b3b80d8

7 years agoEnable LLVM_FINAL, LLVM_OVERRIDE and LLVM_HAS_VARIADIC_TEMPLATES with more gcc and...
Alp Toker [Fri, 3 Jan 2014 05:00:46 +0000 (05:00 +0000)]
Enable LLVM_FINAL, LLVM_OVERRIDE and LLVM_HAS_VARIADIC_TEMPLATES with more gcc and MSVC versions

The 'sealed' definition of LLVM_FINAL can be dropped once VS 2010 is
decommissioned.

Some of this is speculative so will keep an eye on the waterfall -- ping me if
you see failures.

Incremental work towards C++11 migration.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@198399 91177308-0d34-0410-b5e6-96231b3b80d8

7 years agoRevert "Debug Info: Type Units: Simplify type hashing using IR-provided unique names."
David Blaikie [Fri, 3 Jan 2014 04:49:04 +0000 (04:49 +0000)]
Revert "Debug Info: Type Units: Simplify type hashing using IR-provided unique names."

Reverting due to bot failure I won't have time to investigate until
tomorrow.

This reverts commit r198397.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@198398 91177308-0d34-0410-b5e6-96231b3b80d8

7 years agoDebug Info: Type Units: Simplify type hashing using IR-provided unique names.
David Blaikie [Fri, 3 Jan 2014 04:20:26 +0000 (04:20 +0000)]
Debug Info: Type Units: Simplify type hashing using IR-provided unique names.

What's good for LTO metadata size problems ought to be good for non-LTO
debug info size too, so let's rely on the same uniqueness in both cases.
If it's insufficient for non-LTO for whatever reason (since we now won't
be uniquing CU-local types or any C types - but these are likely to not
be the most significant contributors to type bloat) we should consider a
frontend solution that'll help both LTO and non-LTO alike, rather than
using DWARF-level DIE-hashing that only helps non-LTO debug info size.

It's also much simpler this way and benefits C++ even more since we can
deduplicate lexically separate definitions of the same C++ type since
they have the same mangled name.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@198397 91177308-0d34-0410-b5e6-96231b3b80d8

7 years ago80-column.
Eric Christopher [Fri, 3 Jan 2014 02:17:35 +0000 (02:17 +0000)]
80-column.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@198394 91177308-0d34-0410-b5e6-96231b3b80d8

7 years agoRemove TextSectionSym as it is unused.
Eric Christopher [Fri, 3 Jan 2014 02:16:44 +0000 (02:16 +0000)]
Remove TextSectionSym as it is unused.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@198393 91177308-0d34-0410-b5e6-96231b3b80d8

7 years agoRevert "Reverting r193835 due to weirdness with Go..."
David Blaikie [Fri, 3 Jan 2014 01:30:05 +0000 (01:30 +0000)]
Revert "Reverting r193835 due to weirdness with Go..."

The cgo problem was that it wants dwarf2 which doesn't support direct
constant encoding of the location. So let's add support for dwarf2
encoding (using a location expression) of data member locations.

This reverts commit r198385.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@198389 91177308-0d34-0410-b5e6-96231b3b80d8

7 years agoReverting r193835 due to weirdness with Go...
David Blaikie [Fri, 3 Jan 2014 00:48:38 +0000 (00:48 +0000)]
Reverting r193835 due to weirdness with Go...

Apologies for the noise - we're seeing some Go failures with cgo
interacting with Clang's debug info due to this change.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@198385 91177308-0d34-0410-b5e6-96231b3b80d8

7 years agoFix break introduced in r198377 due to using a local type as a template parameter.
David Blaikie [Fri, 3 Jan 2014 00:00:41 +0000 (00:00 +0000)]
Fix break introduced in r198377 due to using a local type as a template parameter.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@198379 91177308-0d34-0410-b5e6-96231b3b80d8

7 years agoTest coverage for non-default-constructible elements in a StringMap
David Blaikie [Thu, 2 Jan 2014 23:57:28 +0000 (23:57 +0000)]
Test coverage for non-default-constructible elements in a StringMap

This functionality was enabled by r198374. Here's a test to ensure it
works and we don't regress it.

Based on a patch by Maciej Piechotka.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@198377 91177308-0d34-0410-b5e6-96231b3b80d8

7 years agoRemove StringMapEntryInitializer support.
David Blaikie [Thu, 2 Jan 2014 23:28:39 +0000 (23:28 +0000)]
Remove StringMapEntryInitializer support.

It was never specialized so let's just remove that unused
configurability and always do the default.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@198374 91177308-0d34-0410-b5e6-96231b3b80d8

7 years ago[RegAlloc] Make tryInstructionSplit less aggressive.
Quentin Colombet [Thu, 2 Jan 2014 22:47:22 +0000 (22:47 +0000)]
[RegAlloc] Make tryInstructionSplit less aggressive.

The greedy register allocator tries to split a live-range around each
instruction where it is used or defined to relax the constraints on the entire
live-range (this is a last chance split before falling back to spill).
The goal is to have a big live-range that is unconstrained (i.e., that can use
the largest legal register class) and several small local live-range that carry
the constraints implied by each instruction.
E.g.,
Let csti be the constraints on operation i.

V1=
op1 V1(cst1)
op2 V1(cst2)

V1 live-range is constrained on the intersection of cst1 and cst2.

tryInstructionSplit relaxes those constraints by aggressively splitting each
def/use point:
V1=
V2 = V1
V3 = V2
op1 V3(cst1)
V4 = V2
op2 V4(cst2)

Because of how the coalescer infrastructure works, each new variable (V3, V4)
that is alive at the same time as V1 (or its copy, here V2) interfere with V1.
Thus, we end up with an uncoalescable copy for each split point.

To make tryInstructionSplit less aggressive, we check if the split point
actually relaxes the constraints on the whole live-range. If it does not, we do
not insert it.
Indeed, it will not help the global allocation problem:
- V1 will have the same constraints.
- V1 will have the same interference + possibly the newly added split variable
  VS.
- VS will produce an uncoalesceable copy if alive at the same time as V1.

<rdar://problem/15570057>

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@198369 91177308-0d34-0410-b5e6-96231b3b80d8

7 years ago[PPC] Fix comment to match function name
Hal Finkel [Thu, 2 Jan 2014 22:09:39 +0000 (22:09 +0000)]
[PPC] Fix comment to match function name

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@198362 91177308-0d34-0410-b5e6-96231b3b80d8

7 years agoRemove comments on CU skeleton construction, they're probably
Eric Christopher [Thu, 2 Jan 2014 22:04:47 +0000 (22:04 +0000)]
Remove comments on CU skeleton construction, they're probably
obvious.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@198361 91177308-0d34-0410-b5e6-96231b3b80d8

7 years ago[PPC] Fix the scheduling of CR logicals on the P7
Hal Finkel [Thu, 2 Jan 2014 21:38:26 +0000 (21:38 +0000)]
[PPC] Fix the scheduling of CR logicals on the P7

CR logicals (crand, crxor, etc.) on the P7 need to be in the first slot of each
dispatch group. The old itinerary entry was just wrong (but has not mattered
because we don't generate these instructions).

This will matter when, in an upcoming commit, we start generating these
instructions.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@198359 91177308-0d34-0410-b5e6-96231b3b80d8

7 years agoElaborate on comment for skeleton CU construction.
Eric Christopher [Thu, 2 Jan 2014 21:38:18 +0000 (21:38 +0000)]
Elaborate on comment for skeleton CU construction.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@198358 91177308-0d34-0410-b5e6-96231b3b80d8

7 years agoRevert seemingly unnecessary section sym for the data section.
Eric Christopher [Thu, 2 Jan 2014 21:38:13 +0000 (21:38 +0000)]
Revert seemingly unnecessary section sym for the data section.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@198357 91177308-0d34-0410-b5e6-96231b3b80d8

7 years ago[PPC] Use the correct immediate operands on 64-bit instructions
Hal Finkel [Thu, 2 Jan 2014 21:26:59 +0000 (21:26 +0000)]
[PPC] Use the correct immediate operands on 64-bit instructions

Several of the 64-bit fixed-point instructions with immediate operands were
using the 32-bit (i32) operand nodes instead of the corresponding 64-bit (i64)
operand definitions (u16imm instead of u16imm64, for example).

This error has had no effect so far, but would have caused type-checking
violations with an upcoming change.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@198356 91177308-0d34-0410-b5e6-96231b3b80d8

7 years agoDisable compare sinking in CodeGenPrepare when multiple condition registers are available
Hal Finkel [Thu, 2 Jan 2014 21:13:43 +0000 (21:13 +0000)]
Disable compare sinking in CodeGenPrepare when multiple condition registers are available

As noted in the comment above CodeGenPrepare::OptimizeInst, which aggressively
sinks compares to reduce pressure on the condition register(s), for targets
such as PowerPC with multiple condition registers, this may not be the right
thing to do. This adds an HasMultipleConditionRegisters boolean to TLI, and
CodeGenPrepare::OptimizeInst is skipped when HasMultipleConditionRegisters is
true.

This functionality will be used by the PowerPC backend in an upcoming commit.
Especially when the PowerPC backend starts tracking individual condition
register bits as separate allocatable entities (which will happen in this
upcoming commit), this sinking from CodeGenPrepare::OptimizeInst is
significantly suboptimial.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@198354 91177308-0d34-0410-b5e6-96231b3b80d8

7 years agoindvars: cleanup the IV visitor. It does more than gather sext/zext info.
Andrew Trick [Thu, 2 Jan 2014 21:12:11 +0000 (21:12 +0000)]
indvars: cleanup the IV visitor. It does more than gather sext/zext info.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@198353 91177308-0d34-0410-b5e6-96231b3b80d8

7 years agoFix all the verifier tests I added for address spaces.
Matt Arsenault [Thu, 2 Jan 2014 21:09:05 +0000 (21:09 +0000)]
Fix all the verifier tests I added for address spaces.

I originally had these using opt -verify, and I never removed the
-verify when converting them to use llvm-as instead, so these were
failing because of using the -verify argument which llvm-as doesn't have
instead of what it's actually supposed to be testing.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@198352 91177308-0d34-0410-b5e6-96231b3b80d8

7 years agoFix up a couple of review comments:
Eric Christopher [Thu, 2 Jan 2014 21:03:28 +0000 (21:03 +0000)]
Fix up a couple of review comments:

Use an if statement instead of a pair of ternary operators checking
the same condition.
Use a cheap method call rather than returning the local symbol.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@198351 91177308-0d34-0410-b5e6-96231b3b80d8

7 years agoSimplify conditional.
Eric Christopher [Thu, 2 Jan 2014 21:03:22 +0000 (21:03 +0000)]
Simplify conditional.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@198350 91177308-0d34-0410-b5e6-96231b3b80d8

7 years agoAllow addrspacecast in global aliases
Matt Arsenault [Thu, 2 Jan 2014 20:55:01 +0000 (20:55 +0000)]
Allow addrspacecast in global aliases

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@198349 91177308-0d34-0410-b5e6-96231b3b80d8

7 years ago[TableGen] Correctly generate implicit anonymous prototype defs in multiclasses
Hal Finkel [Thu, 2 Jan 2014 20:47:09 +0000 (20:47 +0000)]
[TableGen] Correctly generate implicit anonymous prototype defs in multiclasses

Even within a multiclass, we had been generating concrete implicit anonymous
defs when parsing values (generally in value lists). This behavior was
incorrect, and led to errors when multiclass parameters were used in the
parameter list of the implicit anonymous def.

If we had some multiclass:

multiclass mc<string n> {

 ... : SomeClass<SomeOtherClass<n> >

The capture of the multiclass parameter 'n' would not work correctly, and
depending on how the implicit SomeOtherClass was used, either TableGen would
ignore something it shouldn't, or would crash.

To fix this problem, when inside a multiclass, we generate prototype anonymous
defs for implicit anonymous defs (just as we do for explicit anonymous defs).
Within the multiclass, the current record prototype is populated with a node
that is essentially: !cast<SomeOtherClass>(!strconcat(NAME, anon_value_name)).
This is then resolved to the correct concrete anonymous def, in the usual way,
when NAME is resolved during multiclass instantiation.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@198348 91177308-0d34-0410-b5e6-96231b3b80d8

7 years ago[TableGen] Handle ValueType in CodeGenDAGPatterns GetNumNodeResults
Hal Finkel [Thu, 2 Jan 2014 20:47:05 +0000 (20:47 +0000)]
[TableGen] Handle ValueType in CodeGenDAGPatterns GetNumNodeResults

A ValueType in a pattern dag is a type cast, and GetNumNodeResults should
handle it (the type cast has only one result).

This comes up, for example, during the type checking of pattern fragments, for
example, AArch64's Neon_combine_2d fragment is:
  dag Operands = (ops node:$Rm, node:$Rn);
  dag Fragment = (v2f64 (concat_vectors (v1f64 node:$Rm), (v1f64 node:$Rn)));

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@198347 91177308-0d34-0410-b5e6-96231b3b80d8

7 years agoDelete unread globals through addrspacecast
Matt Arsenault [Thu, 2 Jan 2014 20:01:43 +0000 (20:01 +0000)]
Delete unread globals through addrspacecast

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@198346 91177308-0d34-0410-b5e6-96231b3b80d8

7 years agoFix addrspacecast with metadata globals
Matt Arsenault [Thu, 2 Jan 2014 19:53:49 +0000 (19:53 +0000)]
Fix addrspacecast with metadata globals

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@198345 91177308-0d34-0410-b5e6-96231b3b80d8

7 years ago[CMake] Add missing set_output_directory after Takumi's change in r198205.
Jordan Rose [Thu, 2 Jan 2014 19:47:45 +0000 (19:47 +0000)]
[CMake] Add missing set_output_directory after Takumi's change in r198205.

Plugins need to go in build/Debug/lib as well (rather than build/lib/Debug).

Also, fix the SHLIBDIR path for Xcode, which by default includes Xcode build
settings rather than a simple %(build_mode)s parameter.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@198344 91177308-0d34-0410-b5e6-96231b3b80d8

7 years agoRemove redundant fold call introduced in r195944. Thanks very much to Juergen
Lang Hames [Thu, 2 Jan 2014 19:38:41 +0000 (19:38 +0000)]
Remove redundant fold call introduced in r195944. Thanks very much to Juergen
for pointing this out.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@198341 91177308-0d34-0410-b5e6-96231b3b80d8

7 years ago[TableGen] Use the same anonymous name as the prefix on all multiclass defs
Hal Finkel [Thu, 2 Jan 2014 19:35:33 +0000 (19:35 +0000)]
[TableGen] Use the same anonymous name as the prefix on all multiclass defs

TableGen had been generating a different name for an anonymous multiclass's
NAME for every def in the multiclass. This had an unfortunate side effect: it
was impossible to reference one def within the multiclass from another (in the
parameter list, for example). By making sure we only generate an anonymous name
once per multiclass (which, as it turns out, requires only changing the name
parameter to reference type), we can now concatenate NAME within the multiclass
with a def name in order to generate a reference to that def.

This does not matter so much, in and of itself, but is necessary for a
follow-up commit that will fix variable capturing in implicit anonymous
multiclass defs (and that is important).

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@198340 91177308-0d34-0410-b5e6-96231b3b80d8

7 years agoindvars: insert truncate at loop boundary to avoid redundant IVs.
Andrew Trick [Thu, 2 Jan 2014 19:29:38 +0000 (19:29 +0000)]
indvars: insert truncate at loop boundary to avoid redundant IVs.

When widening an IV to remove s/zext, we generally try to eliminate
the original narrow IV. However, LCSSA phi nodes outside the loop were
still using the original IV. Clean this up more aggressively to avoid
redundancy in generated code.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@198338 91177308-0d34-0410-b5e6-96231b3b80d8

7 years agoMark REX64_PREFIX as In64BitMode, remove hack from X86RecognizableInstr.
Craig Topper [Thu, 2 Jan 2014 19:12:10 +0000 (19:12 +0000)]
Mark REX64_PREFIX as In64BitMode, remove hack from X86RecognizableInstr.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@198336 91177308-0d34-0410-b5e6-96231b3b80d8

7 years agoCMake separate projects: use correct name for LIBRARY_OUTPUT_DIRECTORY_${suffix}.
Douglas Gregor [Thu, 2 Jan 2014 19:07:19 +0000 (19:07 +0000)]
CMake separate projects: use correct name for LIBRARY_OUTPUT_DIRECTORY_${suffix}.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@198335 91177308-0d34-0410-b5e6-96231b3b80d8

7 years agoMake llvm::Regex non-copyable but movable.
David Blaikie [Thu, 2 Jan 2014 19:04:59 +0000 (19:04 +0000)]
Make llvm::Regex non-copyable but movable.

Based on a patch by Maciej Piechotka.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@198334 91177308-0d34-0410-b5e6-96231b3b80d8

7 years agoRevert "Debug info: Add enumerators to the __apple_names accelerator table."
Adrian Prantl [Thu, 2 Jan 2014 18:48:24 +0000 (18:48 +0000)]
Revert "Debug info: Add enumerators to the __apple_names accelerator table."

This reverts r197927 until the discussion on llvm-commits comes to a
conclusion.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@198333 91177308-0d34-0410-b5e6-96231b3b80d8

7 years agoRemove unused HasFROperands field from disassembler.
Craig Topper [Thu, 2 Jan 2014 18:44:21 +0000 (18:44 +0000)]
Remove unused HasFROperands field from disassembler.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@198332 91177308-0d34-0410-b5e6-96231b3b80d8

7 years agoUse LLVM_STATIC_ASSERT rather than a hand-rolled implementation.
David Blaikie [Thu, 2 Jan 2014 18:29:40 +0000 (18:29 +0000)]
Use LLVM_STATIC_ASSERT rather than a hand-rolled implementation.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@198330 91177308-0d34-0410-b5e6-96231b3b80d8

7 years agoMark PUSHFS64/PUSHGS64/POPFS64/POPGS64 as In64BitMode and remove the hack from the...
Craig Topper [Thu, 2 Jan 2014 18:20:48 +0000 (18:20 +0000)]
Mark PUSHFS64/PUSHGS64/POPFS64/POPGS64 as In64BitMode and remove the hack from the disassembler table builder.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@198327 91177308-0d34-0410-b5e6-96231b3b80d8

7 years agoRemove unnecessary stirng comparison from disassembler.
Craig Topper [Thu, 2 Jan 2014 17:41:40 +0000 (17:41 +0000)]
Remove unnecessary stirng comparison from disassembler.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@198325 91177308-0d34-0410-b5e6-96231b3b80d8

7 years agoMark all x86 Int_ and _Int patterns as isCodeGenOnly so the disassembler table builde...
Craig Topper [Thu, 2 Jan 2014 17:28:14 +0000 (17:28 +0000)]
Mark all x86 Int_ and _Int patterns as isCodeGenOnly so the disassembler table builder doesn't need to string match them to exclude them.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@198323 91177308-0d34-0410-b5e6-96231b3b80d8

7 years agoRemove CMake-Xcode hack that symlinked llvm-config into a common place.
Douglas Gregor [Thu, 2 Jan 2014 16:30:55 +0000 (16:30 +0000)]
Remove CMake-Xcode hack that symlinked llvm-config into a common place.

When building Clang separately from LLVM with CMake, one should set
the path of llvm-config via the cache variable LLVM_CONFIG.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@198316 91177308-0d34-0410-b5e6-96231b3b80d8

7 years ago[arm] Add softvfp to supported FPU names.
Logan Chien [Thu, 2 Jan 2014 15:50:02 +0000 (15:50 +0000)]
[arm] Add softvfp to supported FPU names.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@198313 91177308-0d34-0410-b5e6-96231b3b80d8

7 years agoMake the ARM ABI selectable via SubtargetFeature.
Rafael Espindola [Thu, 2 Jan 2014 13:40:08 +0000 (13:40 +0000)]
Make the ARM ABI selectable via SubtargetFeature.

This patch makes it possible to select the ABI with -mattr. It will be used to
forward clang's -target-abi option to llvm's CodeGen.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@198304 91177308-0d34-0410-b5e6-96231b3b80d8

7 years agoSupport LLVM_STATIC_ASSERT() in clang pre-C++11 mode
Alp Toker [Thu, 2 Jan 2014 04:07:41 +0000 (04:07 +0000)]
Support LLVM_STATIC_ASSERT() in clang pre-C++11 mode

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@198292 91177308-0d34-0410-b5e6-96231b3b80d8

7 years agoRemove unused function argument.
Craig Topper [Thu, 2 Jan 2014 03:58:45 +0000 (03:58 +0000)]
Remove unused function argument.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@198291 91177308-0d34-0410-b5e6-96231b3b80d8

7 years agoBasicAA: Fix value equality and phi cycles
Arnold Schwaighofer [Thu, 2 Jan 2014 03:31:36 +0000 (03:31 +0000)]
BasicAA: Fix value equality and phi cycles

When there are cycles in the value graph we have to be careful interpreting
"Value*" identity as "value" equivalence. We interpret the value of a phi node
as the value of its operands.
When we check for value equivalence now we make sure that the "Value*" dominates
all cycles (phis).

%0 = phi [%noaliasval, %addr2]
%l = load %ptr
%addr1 = gep @a, 0, %l
%addr2 = gep @a, 0, (%l + 1)
store %ptr ...

Before this patch we would return NoAlias for (%0, %addr1) which is wrong
because the value of the load is from different iterations of the loop.

Tested on x86_64 -mavx at O3 and O3 -flto with no performance or compile time
regressions.

PR18068
radar://15653794

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@198290 91177308-0d34-0410-b5e6-96231b3b80d8

7 years agoRename 'assert' to something less loaded in CompileAssertHasType
Alp Toker [Wed, 1 Jan 2014 23:34:16 +0000 (23:34 +0000)]
Rename 'assert' to something less loaded in CompileAssertHasType

Suggested by Aaron Ballman.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@198288 91177308-0d34-0410-b5e6-96231b3b80d8

7 years agoRemove the 's' DataLayout specification
Rafael Espindola [Wed, 1 Jan 2014 22:29:43 +0000 (22:29 +0000)]
Remove the 's' DataLayout specification

During the years there have been some attempts at figuring out how to
align byval arguments. A look at the commit log suggests that they
were

* Use the ABI alignment.
* When that was not sufficient for x86-64, I added the 's' specification to
  DataLayout.
* When that was not sufficient Evan added the virtual getByValTypeAlignment.
* When even that was not sufficient, we just got the FE to add the alignment
  to the byval.

This patch is just a simple cleanup that removes my first attempt at fixing the
problem. I also added an AArch64 implementation of getByValTypeAlignment to
make sure this patch is a nop. I also left the 's' parsing for backward
compatibility.

I will send a short email to llvmdev about the change for anyone maintaining
an out of tree target.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@198287 91177308-0d34-0410-b5e6-96231b3b80d8

7 years ago[Sparc] Handle atomic loads/stores in sparc backend.
Venkatraman Govindaraju [Wed, 1 Jan 2014 22:11:54 +0000 (22:11 +0000)]
[Sparc] Handle atomic loads/stores in sparc backend.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@198286 91177308-0d34-0410-b5e6-96231b3b80d8

7 years agoRemove modifierType/Base from X86 disassembler tables as they are no longer used...
Craig Topper [Wed, 1 Jan 2014 21:52:57 +0000 (21:52 +0000)]
Remove modifierType/Base from X86 disassembler tables as they are no longer used. Removes ~11.5K from static tables.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@198284 91177308-0d34-0410-b5e6-96231b3b80d8

7 years ago[SparcV9]: Custom lower UMULO/SMULO so that the arguments are send to __multi3()...
Venkatraman Govindaraju [Wed, 1 Jan 2014 20:22:45 +0000 (20:22 +0000)]
[SparcV9]: Custom lower UMULO/SMULO so that the arguments are send to __multi3() in correct order.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@198281 91177308-0d34-0410-b5e6-96231b3b80d8

7 years ago[SparcV9]: Use SRL instead of SLL to clear top 32-bits in ctpop:i32. SLL does not...
Venkatraman Govindaraju [Wed, 1 Jan 2014 19:00:10 +0000 (19:00 +0000)]
[SparcV9]: Use SRL instead of SLL to clear top 32-bits in ctpop:i32. SLL does not clear top 32 bit, only SRL does.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@198280 91177308-0d34-0410-b5e6-96231b3b80d8

7 years agoX86Disassembler.cpp: Prune stray @return on translateFPRegister(). [-Wdocumentation]
NAKAMURA Takumi [Wed, 1 Jan 2014 16:19:26 +0000 (16:19 +0000)]
X86Disassembler.cpp: Prune stray @return on translateFPRegister(). [-Wdocumentation]

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@198279 91177308-0d34-0410-b5e6-96231b3b80d8

7 years agoRemove need for MODIFIER_OPCODE in the disassembler tables. AddRegFrms are really...
Craig Topper [Wed, 1 Jan 2014 15:29:32 +0000 (15:29 +0000)]
Remove need for MODIFIER_OPCODE in the disassembler tables. AddRegFrms are really more like OrRegFrm so we don't need a difference since we can just mask bits.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@198278 91177308-0d34-0410-b5e6-96231b3b80d8

7 years agoAVX-512: Added intrinsics for vcvt, vcvtt, vrndscale, vcmp
Elena Demikhovsky [Wed, 1 Jan 2014 15:12:34 +0000 (15:12 +0000)]
AVX-512: Added intrinsics for vcvt, vcvtt, vrndscale, vcmp
Printing rounding control.
Enncoding for EVEX_RC (rounding control).

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@198277 91177308-0d34-0410-b5e6-96231b3b80d8

7 years agoSecond attempt at Removing special form of AddRegFrm used by FP instructions. These...
Craig Topper [Wed, 1 Jan 2014 14:22:37 +0000 (14:22 +0000)]
Second attempt at Removing special form of AddRegFrm used by FP instructions. These instructions can be handled by MRMXr instead.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@198276 91177308-0d34-0410-b5e6-96231b3b80d8

7 years agoUpdate the copyright credits -- Happy new year 2014!
NAKAMURA Takumi [Wed, 1 Jan 2014 08:27:31 +0000 (08:27 +0000)]
Update the copyright credits -- Happy new year 2014!

FIXME: Dragonegg may be updated at non-trivial changes.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@198274 91177308-0d34-0410-b5e6-96231b3b80d8

7 years agoSilence g++ 4.9 build issue in unit tests
Alp Toker [Wed, 1 Jan 2014 06:57:01 +0000 (06:57 +0000)]
Silence g++ 4.9 build issue in unit tests

Stopgap measure until we can just use static_assert().

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@198273 91177308-0d34-0410-b5e6-96231b3b80d8

7 years agoAdd two fp test cases I missed in my previous commit.
Craig Topper [Tue, 31 Dec 2013 23:15:19 +0000 (23:15 +0000)]
Add two fp test cases I missed in my previous commit.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@198269 91177308-0d34-0410-b5e6-96231b3b80d8

7 years agoAdd more X86 FP stack disassembler test cases.
Craig Topper [Tue, 31 Dec 2013 22:51:53 +0000 (22:51 +0000)]
Add more X86 FP stack disassembler test cases.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@198268 91177308-0d34-0410-b5e6-96231b3b80d8

7 years agoFold vector selects with undef elements in the condition. Fixes PR18319.
Nick Lewycky [Tue, 31 Dec 2013 19:30:47 +0000 (19:30 +0000)]
Fold vector selects with undef elements in the condition. Fixes PR18319.
Patch by Ilia Filippov!

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@198267 91177308-0d34-0410-b5e6-96231b3b80d8

7 years agoRevert r198238 and add FP disassembler tests. It didn't work and I didn't realized...
Craig Topper [Tue, 31 Dec 2013 17:21:44 +0000 (17:21 +0000)]
Revert r198238 and add FP disassembler tests. It didn't work and I didn't realized we had no FP disassembler test cases.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@198265 91177308-0d34-0410-b5e6-96231b3b80d8

7 years agoRemove old comment referring to an argument that no longer exists.
Craig Topper [Tue, 31 Dec 2013 15:29:14 +0000 (15:29 +0000)]
Remove old comment referring to an argument that no longer exists.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@198263 91177308-0d34-0410-b5e6-96231b3b80d8

7 years agoFix misaligned indentation in "if" block in MipsMCCodeEmitter.cpp
Mark Seaborn [Tue, 31 Dec 2013 13:05:15 +0000 (13:05 +0000)]
Fix misaligned indentation in "if" block in MipsMCCodeEmitter.cpp

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@198262 91177308-0d34-0410-b5e6-96231b3b80d8

7 years agoAdd missing MRM_XX forms to the old JIT emitter for consistency.
Craig Topper [Tue, 31 Dec 2013 03:26:24 +0000 (03:26 +0000)]
Add missing MRM_XX forms to the old JIT emitter for consistency.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@198258 91177308-0d34-0410-b5e6-96231b3b80d8

7 years agoRemove MRMInitReg form now that it's last use is gone.
Craig Topper [Tue, 31 Dec 2013 03:19:03 +0000 (03:19 +0000)]
Remove MRMInitReg form now that it's last use is gone.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@198257 91177308-0d34-0410-b5e6-96231b3b80d8

7 years agoSilence g++ 4.9 build issue
Alp Toker [Tue, 31 Dec 2013 03:16:55 +0000 (03:16 +0000)]
Silence g++ 4.9 build issue

lib/Support/ThreadLocal.cpp:53:15: error: typedef 'SIZE_TOO_BIG' locally defined but not used [-Werror=unused-local-typedefs]
   typedef int SIZE_TOO_BIG[sizeof(pthread_key_t) <= sizeof(data) ? 1 : -1];

Done the C++11 way, switching on and using LLVM_STATIC_ASSERT() instead of LLVM_ATTRIBUTE_UNUSED.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@198255 91177308-0d34-0410-b5e6-96231b3b80d8

7 years agoHandle MOV32r0 in expandPostRAPseudo instead of MCInst lowering. No functional change...
Craig Topper [Tue, 31 Dec 2013 03:05:38 +0000 (03:05 +0000)]
Handle MOV32r0 in expandPostRAPseudo instead of MCInst lowering. No functional change intended.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@198254 91177308-0d34-0410-b5e6-96231b3b80d8

7 years agoMerge case statements to remove redundant code.
Craig Topper [Mon, 30 Dec 2013 19:47:49 +0000 (19:47 +0000)]
Merge case statements to remove redundant code.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@198241 91177308-0d34-0410-b5e6-96231b3b80d8

7 years agoRemove special form of AddRegFrm used by FP instructions. These instructions can...
Craig Topper [Mon, 30 Dec 2013 19:16:48 +0000 (19:16 +0000)]
Remove special form of AddRegFrm used by FP instructions. These instructions can be handled by MRMXr instead.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@198238 91177308-0d34-0410-b5e6-96231b3b80d8

7 years agoARM IAS: account for predicated pre-UAL mnemonics
Saleem Abdulrasool [Mon, 30 Dec 2013 18:38:01 +0000 (18:38 +0000)]
ARM IAS: account for predicated pre-UAL mnemonics

Checking the trailing letter of the mnemonic is insufficient.  Be more thorough
in the scanning of the instruction to ensure that we correctly work with the
predicated mnemonics.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@198235 91177308-0d34-0410-b5e6-96231b3b80d8

7 years agoRefactor and reduce code duplication for non-split dwarf strings.
Eric Christopher [Mon, 30 Dec 2013 18:32:31 +0000 (18:32 +0000)]
Refactor and reduce code duplication for non-split dwarf strings.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@198233 91177308-0d34-0410-b5e6-96231b3b80d8

7 years agoRemove EscapeFilter. It's funcionality can be covered by correctly using ExtendedFilt...
Craig Topper [Mon, 30 Dec 2013 17:37:10 +0000 (17:37 +0000)]
Remove EscapeFilter. It's funcionality can be covered by correctly using ExtendedFilter and ExactFilter. No functional change.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@198226 91177308-0d34-0410-b5e6-96231b3b80d8

7 years agoRevert r198208 and reapply:
Eric Christopher [Mon, 30 Dec 2013 17:22:27 +0000 (17:22 +0000)]
Revert r198208 and reapply:

      r198196: Use a pointer to keep track of the skeleton unit for each normal unit and construct it up front.
      r198199: Reapply r198196 with a fix to zero initialize the skeleton pointer.
      r198202: Fix aranges and split dwarf by ensuring that the symbol and relocation back to the compile unit from the aranges section is to the skeleton unit and not the one in the dwo.

with a fix to use integer 0 for DW_AT_low_pc since the relocation to the text section symbol was causing issues with COFF. Accordingly remove addLocalLabelAddress and machinery since we're not currently using it.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@198222 91177308-0d34-0410-b5e6-96231b3b80d8

7 years agoSimplify filter accepts function to just return 'condition' instead of branching...
Craig Topper [Mon, 30 Dec 2013 17:22:20 +0000 (17:22 +0000)]
Simplify filter accepts function to just return 'condition' instead of branching to return true/false. No functional change.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@198221 91177308-0d34-0410-b5e6-96231b3b80d8

7 years agoRevert r198199 (and r198202). It broke 3 DebugInfo tests for targeting i686-cygming.
NAKAMURA Takumi [Mon, 30 Dec 2013 09:26:10 +0000 (09:26 +0000)]
Revert r198199 (and r198202). It broke 3 DebugInfo tests for targeting i686-cygming.

  r198196: Use a pointer to keep track of the skeleton unit for each normal unit and construct it up front.
  r198199: Reapply r198196 with a fix to zero initialize the skeleton pointer.
  r198202: Fix aranges and split dwarf by ensuring that the symbol and relocation back to the compile unit from the aranges section is to the skeleton unit and not the one in the dwo.

They could be reproducible with explicit target.

  llvm/lib/MC/WinCOFFObjectWriter.cpp:224: bool {anonymous}::COFFSymbol::should_keep() const: Assertion `Section->Number != -1 && "Sections with relocations must be real!"' failed.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@198208 91177308-0d34-0410-b5e6-96231b3b80d8

7 years ago[CMake] Generate ${BUILD_MODE}/llvm-lit for each ${CMAKE_CONFIGURATION_TYPES}.
NAKAMURA Takumi [Mon, 30 Dec 2013 07:02:12 +0000 (07:02 +0000)]
[CMake] Generate ${BUILD_MODE}/llvm-lit for each ${CMAKE_CONFIGURATION_TYPES}.

llvm-lit can be invoked;

  $ Release/bin/llvm-lit

instead of;

  $ bin/llvm-lit --param buid_mode=Release

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@198206 91177308-0d34-0410-b5e6-96231b3b80d8

7 years ago[CMake][VS][XCode] Restruct the output directory layout more comfortable, ${BINARY_DI...
NAKAMURA Takumi [Mon, 30 Dec 2013 06:48:30 +0000 (06:48 +0000)]
[CMake][VS][XCode] Restruct the output directory layout more comfortable, ${BINARY_DIR}/${BUILD_MODE}/(bin|lib)

We have been seeing nasty directory layout with CMake multiconfig, such as,
  bin/Release/clang.exe
  lib/clang/3.x/...
  lib/Release/clang/3.x/.. (duplicated)

Move the layout similar to autoconf's;
  Release/bin/clang.exe
  Release/lib/clang/3.x/...

Checked on Visual Studio 10. Could you guys please confirm my change on XCode(and other multiconfig builders)?

Note: Don't set variables CMAKE_*_OUTPUT_DIRECTORY any more, or a certain builder, for eaxample, msbuild.exe, would be confused.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@198205 91177308-0d34-0410-b5e6-96231b3b80d8

7 years ago Visual C++ does not support -ffunction-sections -fdata-sections.
Yaron Keren [Mon, 30 Dec 2013 05:31:53 +0000 (05:31 +0000)]
 Visual C++ does not support -ffunction-sections -fdata-sections.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@198203 91177308-0d34-0410-b5e6-96231b3b80d8

7 years agoFix aranges and split dwarf by ensuring that the symbol and relocation
Eric Christopher [Mon, 30 Dec 2013 05:25:49 +0000 (05:25 +0000)]
Fix aranges and split dwarf by ensuring that the symbol and relocation
back to the compile unit from the aranges section is to the skeleton
unit and not the one in the dwo.

Do this by adding a method to grab a forwarded on local sym and local
section by querying the skeleton if one exists and using that. Add
a few tests to verify the relocations are back to the correct section.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@198202 91177308-0d34-0410-b5e6-96231b3b80d8

7 years agoKeep comment with 'Subtarget' ivar.
Bill Wendling [Mon, 30 Dec 2013 05:17:29 +0000 (05:17 +0000)]
Keep comment with 'Subtarget' ivar.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@198201 91177308-0d34-0410-b5e6-96231b3b80d8

7 years agoReapply r198196 with a fix to zero initialize the skeleton pointer.
Eric Christopher [Mon, 30 Dec 2013 03:40:32 +0000 (03:40 +0000)]
Reapply r198196 with a fix to zero initialize the skeleton pointer.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@198199 91177308-0d34-0410-b5e6-96231b3b80d8

7 years agoPort r198087 and r198089 (strip dead code by default) from make to cmake.
Nico Weber [Mon, 30 Dec 2013 03:36:05 +0000 (03:36 +0000)]
Port r198087 and r198089 (strip dead code by default) from make to cmake.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@198198 91177308-0d34-0410-b5e6-96231b3b80d8

7 years agoTemporarily revert "Use a pointer to keep track of the skeleton unit for
Eric Christopher [Mon, 30 Dec 2013 03:12:31 +0000 (03:12 +0000)]
Temporarily revert "Use a pointer to keep track of the skeleton unit for
each normal unit" as it seems to be causing problems in the asan tests.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@198197 91177308-0d34-0410-b5e6-96231b3b80d8

7 years agoUse a pointer to keep track of the skeleton unit for each normal unit
Eric Christopher [Mon, 30 Dec 2013 03:02:12 +0000 (03:02 +0000)]
Use a pointer to keep track of the skeleton unit for each normal unit
and construct it up front. Add address ranges at the end and a helper
routine so that we're not needlessly using an indirction in the case
of split dwarf.

Update testcases according to the new ordering of attributes on
the compile unit.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@198196 91177308-0d34-0410-b5e6-96231b3b80d8

7 years agoFor AArch64 Neon, simplify scalar dup by lane0 for fp.
Jiangning Liu [Mon, 30 Dec 2013 02:44:35 +0000 (02:44 +0000)]
For AArch64 Neon, simplify scalar dup by lane0 for fp.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@198194 91177308-0d34-0410-b5e6-96231b3b80d8

7 years ago[AArch64]Add code to spill/fill Q register tuples such as QPair/QTriple/QQuad.
Hao Liu [Mon, 30 Dec 2013 02:38:12 +0000 (02:38 +0000)]
[AArch64]Add code to spill/fill Q register tuples such as QPair/QTriple/QQuad.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@198193 91177308-0d34-0410-b5e6-96231b3b80d8

7 years ago[AArch64]Can't select shift left 0 of type v1i64
Hao Liu [Mon, 30 Dec 2013 02:12:46 +0000 (02:12 +0000)]
[AArch64]Can't select shift left 0 of type v1i64

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@198192 91177308-0d34-0410-b5e6-96231b3b80d8

7 years agoFix a bug in DAGcombiner about zero-extend after setcc.
Kevin Qin [Mon, 30 Dec 2013 02:05:13 +0000 (02:05 +0000)]
Fix a bug in DAGcombiner about zero-extend after setcc.

For AArch64 backend, if DAGCombiner see "sext(setcc)", it will
combine them together to a single setcc with extended value type.
Then if it see "zext(setcc)", it assumes setcc is Vxi1, and try to
create "(and (vsetcc), (1, 1, ...)". While setcc isn't Vxi1,
DAGcombiner will create wrong node and get wrong code emitted.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@198190 91177308-0d34-0410-b5e6-96231b3b80d8

7 years ago[AArch64]Fix the problem that can't select mul of v1i64/v2i64 types.
Hao Liu [Mon, 30 Dec 2013 01:38:41 +0000 (01:38 +0000)]
[AArch64]Fix the problem that can't select mul of v1i64/v2i64 types.
E.g. Can't select such IR:
     %tmp = mul <2 x i64> %a, %b

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@198188 91177308-0d34-0410-b5e6-96231b3b80d8