oota-llvm.git
9 years agoFix unused variable warning
Arnold Schwaighofer [Tue, 9 Jun 2015 00:17:40 +0000 (00:17 +0000)]
Fix unused variable warning

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@239369 91177308-0d34-0410-b5e6-96231b3b80d8

9 years ago[NVPTX] run SROA after NVPTXFavorNonGenericAddrSpaces
Jingyue Wu [Tue, 9 Jun 2015 00:05:56 +0000 (00:05 +0000)]
[NVPTX] run SROA after NVPTXFavorNonGenericAddrSpaces

Summary:
This cleans up most allocas NVPTXLowerKernelArgs emits for byval
parameters.

Test Plan: makes bug21465.ll more stronger to verify no redundant local load/store.

Reviewers: eliben, jholewinski

Reviewed By: eliben, jholewinski

Subscribers: jholewinski, llvm-commits

Differential Revision: http://reviews.llvm.org/D10322

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@239368 91177308-0d34-0410-b5e6-96231b3b80d8

9 years agoMergeFunctions: Impose a total order on the replacement of functions
Arnold Schwaighofer [Tue, 9 Jun 2015 00:03:29 +0000 (00:03 +0000)]
MergeFunctions: Impose a total order on the replacement of functions

We don't want to replace function A by Function B in one module and Function B
by Function A in another module.

If these functions are marked with linkonce_odr we would end up with a function
stub calling B in one module and a function stub calling A in another module. If
the linker decides to pick these two we will have two stubs calling each other.

rdar://21265586

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@239367 91177308-0d34-0410-b5e6-96231b3b80d8

9 years ago[WinEH] Cache declarations of frame intrinsics
Reid Kleckner [Mon, 8 Jun 2015 22:43:32 +0000 (22:43 +0000)]
[WinEH] Cache declarations of frame intrinsics

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@239361 91177308-0d34-0410-b5e6-96231b3b80d8

9 years ago[MC] Use unsigned for the Kind bitfield in MCSymbol
Reid Kleckner [Mon, 8 Jun 2015 22:12:44 +0000 (22:12 +0000)]
[MC] Use unsigned for the Kind bitfield in MCSymbol

Fixes most of the test suite on Windows with clang-cl.

I'm not sure why the test suite was passing with MSVC 2013. Maybe they
changed their behavior and we are emulating their old sign extension
behavior. I think this deserves more investigation, but I want to green
the bot first.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@239357 91177308-0d34-0410-b5e6-96231b3b80d8

9 years agoFix clang-cl self-host -Wc++11-narrowing bug
Reid Kleckner [Mon, 8 Jun 2015 21:57:57 +0000 (21:57 +0000)]
Fix clang-cl self-host -Wc++11-narrowing bug

Use unsigned as the underlying storage type of the AMDGPU address space
enum.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@239355 91177308-0d34-0410-b5e6-96231b3b80d8

9 years ago[AArch64] AsmParser should be case insensitive about accepting vector register names.
Ranjeet Singh [Mon, 8 Jun 2015 21:32:16 +0000 (21:32 +0000)]
[AArch64] AsmParser should be case insensitive about accepting vector register names.

Differential Revision: http://reviews.llvm.org/D10320

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@239353 91177308-0d34-0410-b5e6-96231b3b80d8

9 years ago[InstrInfo] Refactor foldOperandImpl to thread through InsertPt. NFC
Keno Fischer [Mon, 8 Jun 2015 20:09:58 +0000 (20:09 +0000)]
[InstrInfo] Refactor foldOperandImpl to thread through InsertPt. NFC

Summary:
This was a longstanding FIXME and is a necessary precursor to cases
where foldOperandImpl may have to create more than one instruction
(e.g. to constrain a register class). This is the split out NFC changes from
D6262.

Reviewers: pete, ributzka, uweigand, mcrosier

Reviewed By: mcrosier

Subscribers: mcrosier, ted, llvm-commits

Differential Revision: http://reviews.llvm.org/D10174

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@239336 91177308-0d34-0410-b5e6-96231b3b80d8

9 years agoFix a regression in .pop_section.
Rafael Espindola [Mon, 8 Jun 2015 20:08:55 +0000 (20:08 +0000)]
Fix a regression in .pop_section.

It was calling ChangeSection with the wrong current section, eventually leading
to a crash.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@239335 91177308-0d34-0410-b5e6-96231b3b80d8

9 years ago[X86][SSE] Added lzcnt vector tests.
Simon Pilgrim [Mon, 8 Jun 2015 19:58:43 +0000 (19:58 +0000)]
[X86][SSE] Added lzcnt vector tests.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@239333 91177308-0d34-0410-b5e6-96231b3b80d8

9 years agoInclude header file <functional>.
Akira Hatanaka [Mon, 8 Jun 2015 19:03:32 +0000 (19:03 +0000)]
Include header file <functional>.

This is a follow-up to r239325.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@239329 91177308-0d34-0410-b5e6-96231b3b80d8

9 years agoPrefer copy init over direct init. NFC.
Benjamin Kramer [Mon, 8 Jun 2015 18:58:57 +0000 (18:58 +0000)]
Prefer copy init over direct init. NFC.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@239327 91177308-0d34-0410-b5e6-96231b3b80d8

9 years ago[ARM] Pass a callback to FunctionPass constructors to enable skipping execution
Akira Hatanaka [Mon, 8 Jun 2015 18:50:43 +0000 (18:50 +0000)]
[ARM] Pass a callback to FunctionPass constructors to enable skipping execution
on a per-function basis.

Previously some of the passes were conditionally added to ARM's pass pipeline
based on the target machine's subtarget. This patch makes changes to add those
passes unconditionally and execute them conditonally based on the predicate
functor passed to the pass constructors. This enables running different sets of
passes for different functions in the module.

rdar://problem/20542263

Differential Revision: http://reviews.llvm.org/D8717

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@239325 91177308-0d34-0410-b5e6-96231b3b80d8

9 years agoUse a PointerUnion in MCSymbol for Section and Fragment. NFC.
Pete Cooper [Mon, 8 Jun 2015 18:41:57 +0000 (18:41 +0000)]
Use a PointerUnion in MCSymbol for Section and Fragment.  NFC.

The Fragment and Section, and a bool for HasFragment were all used to create
a PointerUnion.  Just use a pointer union instead.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@239324 91177308-0d34-0410-b5e6-96231b3b80d8

9 years agoRemove includes of MCMachOSymbolFlags.h after it was deleted
Pete Cooper [Mon, 8 Jun 2015 17:25:57 +0000 (17:25 +0000)]
Remove includes of MCMachOSymbolFlags.h after it was deleted

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@239318 91177308-0d34-0410-b5e6-96231b3b80d8

9 years agoMake flags get/set method protected so that all flags have to be managed by subclasses.
Pete Cooper [Mon, 8 Jun 2015 17:17:30 +0000 (17:17 +0000)]
Make flags get/set method protected so that all flags have to be managed by subclasses.

All of ELF, COFF and MachO now manipulate the flags in helpers so we don't need
anyone to read the flags directly, but instead via those helpers.

Reviewed by Rafael Espíndola.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@239317 91177308-0d34-0410-b5e6-96231b3b80d8

9 years agoMove all flags logic to MCSymbolMachO.
Pete Cooper [Mon, 8 Jun 2015 17:17:28 +0000 (17:17 +0000)]
Move all flags logic to MCSymbolMachO.

Also delete the now unused MCMachOSymbolFlags.h header as the only enum in there was moved to MCSymbolMachO.

Similarly to ELF and COFF, manipulating the flags is now done via helpers instead of spread
throughout the codebase.

Reviewed by Rafael Espíndola.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@239316 91177308-0d34-0410-b5e6-96231b3b80d8

9 years agoAdd MCSymbolMachO which will be used to hide the MCSymbolMachO flags.
Pete Cooper [Mon, 8 Jun 2015 17:17:23 +0000 (17:17 +0000)]
Add MCSymbolMachO which will be used to hide the MCSymbolMachO flags.

Reviewed by Rafael Espíndola.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@239315 91177308-0d34-0410-b5e6-96231b3b80d8

9 years agoMove all of the MCSymbol COFF flags logic in to MCSymbolCOFF.
Pete Cooper [Mon, 8 Jun 2015 17:17:19 +0000 (17:17 +0000)]
Move all of the MCSymbol COFF flags logic in to MCSymbolCOFF.

All flags setting/getting is now done in the class with helper methods instead
of users having to get the bits in the correct order.

Reviewed by Rafael Espíndola.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@239314 91177308-0d34-0410-b5e6-96231b3b80d8

9 years agoMove COFF Type in to the MCSymbolCOFF class.
Pete Cooper [Mon, 8 Jun 2015 17:17:16 +0000 (17:17 +0000)]
Move COFF Type in to the MCSymbolCOFF class.

The flags field in MCSymbol only needs to be 16-bits on ELF and MachO.
This moves the 16-bit Type out of there so that it can be reduced in size in a future commit.

Reviewed by Rafael Espíndola.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@239313 91177308-0d34-0410-b5e6-96231b3b80d8

9 years agoAdd MCSymbolCOFF class and use it to get and set the COFF type field.
Pete Cooper [Mon, 8 Jun 2015 17:17:12 +0000 (17:17 +0000)]
Add MCSymbolCOFF class and use it to get and set the COFF type field.

Reviewed by Rafael Espíndola.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@239312 91177308-0d34-0410-b5e6-96231b3b80d8

9 years agoChange MCSymbol IsELF to an enum to support future MCSymbolCOFF and MCSymbolMachO.
Pete Cooper [Mon, 8 Jun 2015 17:17:09 +0000 (17:17 +0000)]
Change MCSymbol IsELF to an enum to support future MCSymbolCOFF and MCSymbolMachO.

Reviewed by Rafael Espíndola.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@239311 91177308-0d34-0410-b5e6-96231b3b80d8

9 years agoX86: Reject register operands with obvious type mismatches.
Matthias Braun [Mon, 8 Jun 2015 16:56:23 +0000 (16:56 +0000)]
X86: Reject register operands with obvious type mismatches.

While we have some code to transform specification like {ax} into
{eax}/{rax} if the operand type isn't 16bit, we should reject cases
where there is no sane way to do this, like the i128 type in the
example.

Related to rdar://21042280

Differential Revision: http://reviews.llvm.org/D10260

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@239309 91177308-0d34-0410-b5e6-96231b3b80d8

9 years agoFix assertion failure in global-merge with unused ConstantExpr
Oliver Stannard [Mon, 8 Jun 2015 16:55:31 +0000 (16:55 +0000)]
Fix assertion failure in global-merge with unused ConstantExpr

The global-merge pass was crashing because it assumes that all ConstantExprs
(reached via the global variables that they use) have at least one user.

I haven't worked out a way to test this, as an unused ConstantExpr cannot be
represented by serialised IR, and global-merge can only be run in llc, which
does not run any passes which can make a ConstantExpr dead.

This (reduced to the point of silliness) C code triggers this bug when compiled
for arm-none-eabi at -O1:

  static a = 7;
  static volatile b[10] = {&a};

  c;
  main() {
    c = 0;
    for (; c < 10;)
      printf(b[c]);
  }

Differential Revision: http://reviews.llvm.org/D10314

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@239308 91177308-0d34-0410-b5e6-96231b3b80d8

9 years ago[Hexagon] Adding functionality for searching for compound instruction pairs. Compoun...
Colin LeMahieu [Mon, 8 Jun 2015 16:34:47 +0000 (16:34 +0000)]
[Hexagon] Adding functionality for searching for compound instruction pairs.  Compound instructions reduce slot resource requirements freeing those packet slots up for more instructions.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@239307 91177308-0d34-0410-b5e6-96231b3b80d8

9 years ago[DAGCombiner] Added CTLZ vector constant folding support.
Simon Pilgrim [Mon, 8 Jun 2015 16:19:00 +0000 (16:19 +0000)]
[DAGCombiner] Added CTLZ vector constant folding support.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@239305 91177308-0d34-0410-b5e6-96231b3b80d8

9 years agofix typos; NFC
Sanjay Patel [Mon, 8 Jun 2015 15:21:38 +0000 (15:21 +0000)]
fix typos; NFC

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@239303 91177308-0d34-0410-b5e6-96231b3b80d8

9 years agoARM]: Add support for MMFR4_EL1 in assembler
Javed Absar [Mon, 8 Jun 2015 15:01:11 +0000 (15:01 +0000)]
ARM]: Add support for MMFR4_EL1 in assembler

This patch adds support for system register MMFR4_EL1 (memory model feature register) in the assembler.
This register provides information about the implemented memory model and memory management support.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@239302 91177308-0d34-0410-b5e6-96231b3b80d8

9 years ago[Mips64][mcjit] Add R_MIPS_PC32 relocation
Petar Jovanovic [Mon, 8 Jun 2015 14:10:23 +0000 (14:10 +0000)]
[Mips64][mcjit] Add R_MIPS_PC32 relocation

This patch adds R_MIPS_PC32 relocation for Mips64.

Patch by Vladimir Radosavljevic.

Differential Revision: http://reviews.llvm.org/D10235

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@239301 91177308-0d34-0410-b5e6-96231b3b80d8

9 years agoAVX-512: Implemented 256/128bit VALIGND/Q instructions for SKX and KNL
Igor Breger [Mon, 8 Jun 2015 14:03:17 +0000 (14:03 +0000)]
AVX-512: Implemented 256/128bit VALIGND/Q instructions for SKX and KNL
Implemented DAG lowering for all these forms.
Added tests for DAG lowering and encoding.

Differential Revision: http://reviews.llvm.org/D10310

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@239300 91177308-0d34-0410-b5e6-96231b3b80d8

9 years agoMinor refactoring of GEP handling in isDereferenceablePointer
Artur Pilipenko [Mon, 8 Jun 2015 11:58:13 +0000 (11:58 +0000)]
Minor refactoring of GEP handling in isDereferenceablePointer

For GEP instructions isDereferenceablePointer checks that all indices are constant and within bounds. Replace this index calculation logic to a call to accumulateConstantOffset. Separated from the http://reviews.llvm.org/D9791

Reviewed By: sanjoy

Differential Revision: http://reviews.llvm.org/D9874

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@239299 91177308-0d34-0410-b5e6-96231b3b80d8

9 years ago[LAA] Fix estimation of number of memchecks
Silviu Baranga [Mon, 8 Jun 2015 10:27:06 +0000 (10:27 +0000)]
[LAA] Fix estimation of number of memchecks

Summary:
We need to add a runtime memcheck for pair of accesses (x,y) where at least one of x and y
are writes.

Assuming we have w writes and r reads, currently this number is  estimated as being
w* (w+r-1). This estimation will count (write,write) pairs twice and will overestimate
the number of checks required.

This change adds a getNumberOfChecks method to RuntimePointerCheck, which
will count the number of runtime checks needed (similar in implementation to
needsAnyChecking) and uses it to produce the correct number of runtime checks.

Test Plan:
llvm test suite
spec2k
spec2k6

Performance results: no changes observed (not surprising since the formula for 1 writer is basically the same, which would covers most cases - at least with the current check limit).

Reviewers: anemet

Reviewed By: anemet

Subscribers: mzolotukhin, llvm-commits

Differential Revision: http://reviews.llvm.org/D10217

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@239295 91177308-0d34-0410-b5e6-96231b3b80d8

9 years ago[DAGCombiner] Added CTTZ vector constant folding support.
Simon Pilgrim [Mon, 8 Jun 2015 09:57:09 +0000 (09:57 +0000)]
[DAGCombiner] Added CTTZ vector constant folding support.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@239293 91177308-0d34-0410-b5e6-96231b3b80d8

9 years ago[LoopVectorize] Teach Loop Vectorizor about interleaved memory accesses.
Hao Liu [Mon, 8 Jun 2015 06:39:56 +0000 (06:39 +0000)]
[LoopVectorize] Teach Loop Vectorizor about interleaved memory accesses.
Interleaved memory accesses are grouped and vectorized into vector load/store and shufflevector.
E.g. for (i = 0; i < N; i+=2) {
       a = A[i];         // load of even element
       b = A[i+1];       // load of odd element
       ...               // operations on a, b, c, d
       A[i] = c;         // store of even element
       A[i+1] = d;       // store of odd element
     }

  The loads of even and odd elements are identified as an interleave load group, which will be transfered into vectorized IRs like:
     %wide.vec = load <8 x i32>, <8 x i32>* %ptr
     %vec.even = shufflevector <8 x i32> %wide.vec, <8 x i32> undef, <4 x i32> <i32 0, i32 2, i32 4, i32 6>
     %vec.odd = shufflevector <8 x i32> %wide.vec, <8 x i32> undef, <4 x i32> <i32 1, i32 3, i32 5, i32 7>

  The stores of even and odd elements are identified as an interleave store group, which will be transfered into vectorized IRs like:
     %interleaved.vec = shufflevector <4 x i32> %vec.even, %vec.odd, <8 x i32> <i32 0, i32 4, i32 1, i32 5, i32 2, i32 6, i32 3, i32 7>
     store <8 x i32> %interleaved.vec, <8 x i32>* %ptr

This optimization is currently disabled by defaut. To try it by adding '-enable-interleaved-mem-accesses=true'.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@239291 91177308-0d34-0410-b5e6-96231b3b80d8

9 years ago[llvm-readobj] Use the new symbol API.
Davide Italiano [Mon, 8 Jun 2015 04:55:29 +0000 (04:55 +0000)]
[llvm-readobj] Use the new symbol API.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@239286 91177308-0d34-0410-b5e6-96231b3b80d8

9 years ago[LoopAccessAnalysis] Teach LAA to check the memory dependence between strided accesses.
Hao Liu [Mon, 8 Jun 2015 04:48:37 +0000 (04:48 +0000)]
[LoopAccessAnalysis] Teach LAA to check the memory dependence between strided accesses.

Differential Revision: http://reviews.llvm.org/D9368

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@239285 91177308-0d34-0410-b5e6-96231b3b80d8

9 years agoDo not build LLD from tools/Makefile, as LLD no longer has a Makefile build.
Peter Collingbourne [Mon, 8 Jun 2015 04:04:13 +0000 (04:04 +0000)]
Do not build LLD from tools/Makefile, as LLD no longer has a Makefile build.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@239283 91177308-0d34-0410-b5e6-96231b3b80d8

9 years agoRemove SCEVCache and FindConstantPointers from complete loop unrolling heuristic.
Michael Zolotukhin [Mon, 8 Jun 2015 03:28:06 +0000 (03:28 +0000)]
Remove SCEVCache and FindConstantPointers from complete loop unrolling heuristic.

Summary:
Using some SCEV functionality helped to entirely remove SCEVCache class and FindConstantPointers SCEV visitor.
Also, this makes the code more universal - I'll take advandate of it in next patches where I start handling additional types of instructions.

Test Plan: Tests would be submitted in subsequent patches.

Reviewers: atrick, chandlerc

Reviewed By: atrick, chandlerc

Subscribers: atrick, llvm-commits

Differential Revision: http://reviews.llvm.org/D10205

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@239282 91177308-0d34-0410-b5e6-96231b3b80d8

9 years agoFix Windows build.
Peter Collingbourne [Mon, 8 Jun 2015 02:43:32 +0000 (02:43 +0000)]
Fix Windows build.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@239279 91177308-0d34-0410-b5e6-96231b3b80d8

9 years agollvm-ar: Move archive writer to Object.
Peter Collingbourne [Mon, 8 Jun 2015 02:32:01 +0000 (02:32 +0000)]
llvm-ar: Move archive writer to Object.

No functional change intended, other than some minor changes to certain
diagnostics.

Differential Revision: http://reviews.llvm.org/D10296

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@239278 91177308-0d34-0410-b5e6-96231b3b80d8

9 years ago[TableGen] Remove unnecessary default constructed SMLoc that just existed to return...
Craig Topper [Mon, 8 Jun 2015 01:35:45 +0000 (01:35 +0000)]
[TableGen] Remove unnecessary default constructed SMLoc that just existed to return. Instead just call the default constructor in the return.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@239277 91177308-0d34-0410-b5e6-96231b3b80d8

9 years ago[TableGen] Use range-based for loop. NFC
Craig Topper [Mon, 8 Jun 2015 01:35:43 +0000 (01:35 +0000)]
[TableGen] Use range-based for loop. NFC

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@239276 91177308-0d34-0410-b5e6-96231b3b80d8

9 years ago[TableGen] Remove unnecessary extern declaration that's already in an included header...
Craig Topper [Mon, 8 Jun 2015 01:35:42 +0000 (01:35 +0000)]
[TableGen] Remove unnecessary extern declaration that's already in an included header file.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@239275 91177308-0d34-0410-b5e6-96231b3b80d8

9 years ago[TableGen] Use the SMLoc header file instead of SourceMgr header file in a couple...
Craig Topper [Mon, 8 Jun 2015 01:35:40 +0000 (01:35 +0000)]
[TableGen] Use the SMLoc header file instead of SourceMgr header file in a couple places. NFC

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@239274 91177308-0d34-0410-b5e6-96231b3b80d8

9 years ago[objdump] Moving PrintImmHex out of MachODump and in to llvm-objdump and setting...
Colin LeMahieu [Sun, 7 Jun 2015 21:07:17 +0000 (21:07 +0000)]
[objdump] Moving PrintImmHex out of MachODump and in to llvm-objdump and setting instprinter appropriately.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@239265 91177308-0d34-0410-b5e6-96231b3b80d8

9 years ago[X86] Added tzcnt vector tests.
Simon Pilgrim [Sun, 7 Jun 2015 21:01:34 +0000 (21:01 +0000)]
[X86] Added tzcnt vector tests.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@239264 91177308-0d34-0410-b5e6-96231b3b80d8

9 years ago[MC] Function naming NFC.
Colin LeMahieu [Sun, 7 Jun 2015 20:29:37 +0000 (20:29 +0000)]
[MC] Function naming NFC.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@239263 91177308-0d34-0410-b5e6-96231b3b80d8

9 years agoSeparateConstOffsetFromGEP: Pass address space to isLegalAddressingMode
Matt Arsenault [Sun, 7 Jun 2015 20:17:44 +0000 (20:17 +0000)]
SeparateConstOffsetFromGEP: Pass address space to isLegalAddressingMode

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@239262 91177308-0d34-0410-b5e6-96231b3b80d8

9 years agoMake NaryReassociate pass the address space to isLegalAddressingMode
Matt Arsenault [Sun, 7 Jun 2015 20:17:42 +0000 (20:17 +0000)]
Make NaryReassociate pass the address space to isLegalAddressingMode

No test since the kinds of transforms this prevents seem to not really
be relevant for SI's different addressing modes.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@239261 91177308-0d34-0410-b5e6-96231b3b80d8

9 years agoAdd isLegalAddressingMode address space argument to TTI
Matt Arsenault [Sun, 7 Jun 2015 20:12:03 +0000 (20:12 +0000)]
Add isLegalAddressingMode address space argument to TTI

Update to match the TLI version, and remove the TLI version's
default argument.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@239260 91177308-0d34-0410-b5e6-96231b3b80d8

9 years ago[cmake] [OCaml] Make ocamldoc targets depend on output files
Peter Zotov [Sun, 7 Jun 2015 19:22:22 +0000 (19:22 +0000)]
[cmake] [OCaml] Make ocamldoc targets depend on output files

OCaml doc builds fail without .cmi files, and .cmi files are collected
in ocaml_outputs. Therefore, make doc targets depend on ocaml_outputs as
well.

Fixes: https://llvm.org/bugs/show_bug.cgi?id=23777
Patch by Michał Górny <mgorny@gentoo.org>

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@239259 91177308-0d34-0410-b5e6-96231b3b80d8

9 years ago[X86] Added BitScanForward/BitScanReverse memory folding + tests
Simon Pilgrim [Sun, 7 Jun 2015 18:34:25 +0000 (18:34 +0000)]
[X86] Added BitScanForward/BitScanReverse memory folding + tests

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@239257 91177308-0d34-0410-b5e6-96231b3b80d8

9 years agoRemove global std::string. NFC.
Benjamin Kramer [Sun, 7 Jun 2015 16:36:28 +0000 (16:36 +0000)]
Remove global std::string. NFC.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@239254 91177308-0d34-0410-b5e6-96231b3b80d8

9 years agoFixed line endings
Simon Pilgrim [Sun, 7 Jun 2015 16:09:48 +0000 (16:09 +0000)]
Fixed line endings

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@239253 91177308-0d34-0410-b5e6-96231b3b80d8

9 years ago[DAGCombiner] Added CTPOP vector constant folding support.
Simon Pilgrim [Sun, 7 Jun 2015 15:37:14 +0000 (15:37 +0000)]
[DAGCombiner] Added CTPOP vector constant folding support.

Added tests to the existing SSE/AVX test files.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@239252 91177308-0d34-0410-b5e6-96231b3b80d8

9 years ago[AsmWriter] Rewrite module asm printing using StringRef::split.
Benjamin Kramer [Sun, 7 Jun 2015 13:59:33 +0000 (13:59 +0000)]
[AsmWriter] Rewrite module asm printing using StringRef::split.

No change in output intended.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@239251 91177308-0d34-0410-b5e6-96231b3b80d8

9 years agoFix doxygen comments. NFC
Filipe Cabecinhas [Sun, 7 Jun 2015 06:40:24 +0000 (06:40 +0000)]
Fix doxygen comments. NFC

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@239250 91177308-0d34-0410-b5e6-96231b3b80d8

9 years ago[TableGen] Remove unnecessary include. NFC.
Craig Topper [Sun, 7 Jun 2015 06:01:22 +0000 (06:01 +0000)]
[TableGen] Remove unnecessary include. NFC.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@239249 91177308-0d34-0410-b5e6-96231b3b80d8

9 years ago[TableGen] Remove unnecessary forward declaration. NFC
Craig Topper [Sun, 7 Jun 2015 06:01:19 +0000 (06:01 +0000)]
[TableGen] Remove unnecessary forward declaration. NFC

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@239248 91177308-0d34-0410-b5e6-96231b3b80d8

9 years ago[TableGen] Write one of the Record constructors in terms of the other constructor...
Craig Topper [Sun, 7 Jun 2015 06:01:17 +0000 (06:01 +0000)]
[TableGen] Write one of the Record constructors in terms of the other constructor to reduce code duplication. NFC

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@239247 91177308-0d34-0410-b5e6-96231b3b80d8

9 years ago[TableGen] Remove unused function. NFC
Craig Topper [Sun, 7 Jun 2015 06:01:15 +0000 (06:01 +0000)]
[TableGen] Remove unused function. NFC

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@239246 91177308-0d34-0410-b5e6-96231b3b80d8

9 years ago[TableGen] Use 'isa' to check if something is an UnsetInit rather than getting the...
Craig Topper [Sun, 7 Jun 2015 06:01:13 +0000 (06:01 +0000)]
[TableGen] Use 'isa' to check if something is an UnsetInit rather than getting the UnsetInit singleton and comparing pointers. NFC.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@239245 91177308-0d34-0410-b5e6-96231b3b80d8

9 years ago[Object/ELF] Don't confuse isDefined() and isCommon.
Davide Italiano [Sun, 7 Jun 2015 02:45:19 +0000 (02:45 +0000)]
[Object/ELF] Don't confuse isDefined() and isCommon.

Reported by Rafael Espindola. Pointy-hat to me.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@239241 91177308-0d34-0410-b5e6-96231b3b80d8

9 years agoTeaching llvm-mc how to understand the defsym command line option. This allows integ...
Colin LeMahieu [Sun, 7 Jun 2015 01:46:24 +0000 (01:46 +0000)]
Teaching llvm-mc how to understand the defsym command line option.  This allows integer-constant symbols to be defined on the command line and used during assembly.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@239240 91177308-0d34-0410-b5e6-96231b3b80d8

9 years ago[Object/ELF] Provide helpers for symbol types.
Davide Italiano [Sat, 6 Jun 2015 22:54:09 +0000 (22:54 +0000)]
[Object/ELF] Provide helpers for symbol types.

These were, originally, in a different form in lld.
They can be reused for other tools, e.g. llvm-readobj.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@239231 91177308-0d34-0410-b5e6-96231b3b80d8

9 years ago[InstCombine, InstSimplify] Move xforms from Combine to Simplify
David Majnemer [Sat, 6 Jun 2015 22:40:21 +0000 (22:40 +0000)]
[InstCombine, InstSimplify] Move xforms from Combine to Simplify

There were several SelectInst combines that always returned an existing
instruction instead of modifying an old one or creating a new one.
These are prime candidates for moving to InstSimplify.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@239229 91177308-0d34-0410-b5e6-96231b3b80d8

9 years agoUse early return idiom. NFC
Filipe Cabecinhas [Sat, 6 Jun 2015 20:44:53 +0000 (20:44 +0000)]
Use early return idiom. NFC

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@239228 91177308-0d34-0410-b5e6-96231b3b80d8

9 years ago[MC] Common symbols weren't being checked for redeclaration which allowed an assembly...
Colin LeMahieu [Sat, 6 Jun 2015 20:12:40 +0000 (20:12 +0000)]
[MC] Common symbols weren't being checked for redeclaration which allowed an assembly file to generate an assertion in setCommon(): !isCommon().  This change allows redeclaration as long as the size and alignment match exactly, otherwise report a fatal error.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@239227 91177308-0d34-0410-b5e6-96231b3b80d8

9 years ago[LoopUnroll] Fix truncation bug in canUnrollCompletely.
Sanjoy Das [Sat, 6 Jun 2015 05:24:10 +0000 (05:24 +0000)]
[LoopUnroll] Fix truncation bug in canUnrollCompletely.

Summary:
canUnrollCompletely takes `unsigned` values for `UnrolledCost` and
`RolledDynamicCost` but is passed in `uint64_t`s that are silently
truncated.  Because of this, when `UnrolledSize` is a large integer
that has a small remainder with UINT32_MAX, LLVM tries to completely
unroll loops with high trip counts.

Reviewers: mzolotukhin, chandlerc

Subscribers: llvm-commits

Differential Revision: http://reviews.llvm.org/D10293

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@239218 91177308-0d34-0410-b5e6-96231b3b80d8

9 years ago[CVP] Don't assume Constants of type i1 can be known to be true or false
David Majnemer [Sat, 6 Jun 2015 04:56:51 +0000 (04:56 +0000)]
[CVP] Don't assume Constants of type i1 can be known to be true or false

CVP wants to analyze the condition operand of a select along an edge.
It succeeds in getting back a Constant but not a ConstantInt.  Instead,
it gets a ConstantExpr.  It then assumes that the Constant must be equal
to false because it isn't equal to true.

Instead, perform an additional comparison.

This fixes PR23752.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@239217 91177308-0d34-0410-b5e6-96231b3b80d8

9 years ago[InstCombine] Don't miscompile select to poison
David Majnemer [Sat, 6 Jun 2015 02:30:43 +0000 (02:30 +0000)]
[InstCombine] Don't miscompile select to poison

If we have (select a, b, c), it is sometimes valid to simplify this to a
single select operand.  However, doing so is only valid if the
computation doesn't inject poison into the computation.

It might be helpful to consider the following example:
  (select (icmp ne %i, INT_MAX), (add nsw %i, 1), INT_MIN)

The select is equivalent to (add %i, 1) but not (add nsw %i, 1).

Self hosting on x86_64 revealed that this occurs very, very rarely so
bailing out is hopefully pretty reasonable.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@239215 91177308-0d34-0410-b5e6-96231b3b80d8

9 years agoHandle 16 bit PC relative relocations.
Rafael Espindola [Sat, 6 Jun 2015 02:29:56 +0000 (02:29 +0000)]
Handle 16 bit PC relative relocations.

Fixes pr23771.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@239214 91177308-0d34-0410-b5e6-96231b3b80d8

9 years agoTargetParser: Fix comments in enum(s) introduced in r239150. [-Wdocumentation]
NAKAMURA Takumi [Sat, 6 Jun 2015 01:41:35 +0000 (01:41 +0000)]
TargetParser: Fix comments in enum(s) introduced in r239150. [-Wdocumentation]

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@239211 91177308-0d34-0410-b5e6-96231b3b80d8

9 years ago[TableGen] Change OpInit::getNumOperands and getOperand to use unsigned integers...
Craig Topper [Sat, 6 Jun 2015 01:34:04 +0000 (01:34 +0000)]
[TableGen] Change OpInit::getNumOperands and getOperand to use unsigned integers. NFC

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@239210 91177308-0d34-0410-b5e6-96231b3b80d8

9 years ago[TableGen] Remove trailing whitespace, add space between 'if' and paren, other format...
Craig Topper [Sat, 6 Jun 2015 01:34:01 +0000 (01:34 +0000)]
[TableGen] Remove trailing whitespace, add space between 'if' and paren, other formatting fixes. NFC

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@239209 91177308-0d34-0410-b5e6-96231b3b80d8

9 years ago[TableGen] Remove unnecessary temporary. NFC
Craig Topper [Sat, 6 Jun 2015 01:34:00 +0000 (01:34 +0000)]
[TableGen] Remove unnecessary temporary. NFC

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@239208 91177308-0d34-0410-b5e6-96231b3b80d8

9 years ago[TableGen] Fold variable declaration/initialization into if condition for a couple...
Craig Topper [Sat, 6 Jun 2015 01:33:58 +0000 (01:33 +0000)]
[TableGen] Fold variable declaration/initialization into if condition for a couple short lived variables. NFC

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@239207 91177308-0d34-0410-b5e6-96231b3b80d8

9 years ago[TableGen] Remove unnecessary outer 'if' and merge it's conditions into the inner...
Craig Topper [Sat, 6 Jun 2015 01:33:55 +0000 (01:33 +0000)]
[TableGen] Remove unnecessary outer 'if' and merge it's conditions into the inner 'if's. NFC

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@239206 91177308-0d34-0410-b5e6-96231b3b80d8

9 years ago[TableGen] Fold variable declarations with their assignments. NFC
Craig Topper [Sat, 6 Jun 2015 00:44:45 +0000 (00:44 +0000)]
[TableGen] Fold variable declarations with their assignments. NFC

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@239205 91177308-0d34-0410-b5e6-96231b3b80d8

9 years ago[TableGen] Correct the documentation for 'foreach' in the Language Intro.
Craig Topper [Sat, 6 Jun 2015 00:44:42 +0000 (00:44 +0000)]
[TableGen] Correct the documentation for 'foreach' in the Language Intro.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@239204 91177308-0d34-0410-b5e6-96231b3b80d8

9 years ago[dsymutil] Fix misspelled CHECK line.
Frederic Riss [Fri, 5 Jun 2015 23:46:18 +0000 (23:46 +0000)]
[dsymutil] Fix misspelled CHECK line.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@239200 91177308-0d34-0410-b5e6-96231b3b80d8

9 years ago[dsymutil] Add support for linking the debug_frame section.
Frederic Riss [Fri, 5 Jun 2015 23:06:11 +0000 (23:06 +0000)]
[dsymutil] Add support for linking the debug_frame section.

Linking the debug frame section is actually very easy as we just have to
patch the start address in the FDE header and then copy the rest of the
FDE without even looking at it. The only small complexity comes from the
handling of the CIEs that we should unique across object file. This is
also really easy by using a StringMap keyed on the raw contents of the
CIE.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@239198 91177308-0d34-0410-b5e6-96231b3b80d8

9 years agoMove the code in TargetPassConfig::addPass that inserts machine printer pass to
Akira Hatanaka [Fri, 5 Jun 2015 21:58:14 +0000 (21:58 +0000)]
Move the code in TargetPassConfig::addPass that inserts machine printer pass to
the overloaded version of addPass which takes Pass*.

This change enables inserting the machine printer pass when the overloaded
version of addPass that takes Pass* is called to add a pass, instead of the
one which takes AnalysisID. I need this to prevent make-check tests from
failing when I commit another patch later.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@239192 91177308-0d34-0410-b5e6-96231b3b80d8

9 years ago[dsymutil] Rename a variable to appease some bots.
Frederic Riss [Fri, 5 Jun 2015 21:21:57 +0000 (21:21 +0000)]
[dsymutil] Rename a variable to appease some bots.

Anyway having the type and the name of the member being the same
thing wasn't the wisest of the choices.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@239190 91177308-0d34-0410-b5e6-96231b3b80d8

9 years ago[dsymutil] Have the YAML deserialization rewrite the object address of symbols.
Frederic Riss [Fri, 5 Jun 2015 21:12:07 +0000 (21:12 +0000)]
[dsymutil] Have the YAML deserialization rewrite the object address of symbols.

The main use of the YAML debug map format is for testing inside LLVM. If we have IR
files in the tests used to generate object files, then we obviously don't know the
addresses of the symbols inside the object files beforehand.

This change lets the YAML import lookup the addresses in the object files and rewrite
them. This will allow to have test that really don't need any binary input.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@239189 91177308-0d34-0410-b5e6-96231b3b80d8

9 years ago[dsymutil] Apply clang-format. NFC
Frederic Riss [Fri, 5 Jun 2015 20:27:07 +0000 (20:27 +0000)]
[dsymutil] Apply clang-format. NFC

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@239186 91177308-0d34-0410-b5e6-96231b3b80d8

9 years ago[dsymutil] Out-line the YAML serialization code. NFC
Frederic Riss [Fri, 5 Jun 2015 20:27:04 +0000 (20:27 +0000)]
[dsymutil] Out-line the YAML serialization code. NFC

It will get a bit bigger in an upcoming commit. No need to have all
of that in the header.

Also move parseYAMLDebugMap() to the same place as the serialization
code. This way it will be able to share a private Context object with
it.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@239185 91177308-0d34-0410-b5e6-96231b3b80d8

9 years agoRevert "[InstCombine] Rephrase fix to SimplifyWithOpReplaced"
Renato Golin [Fri, 5 Jun 2015 18:24:12 +0000 (18:24 +0000)]
Revert "[InstCombine] Rephrase fix to SimplifyWithOpReplaced"

This reverts commit r239141. This commit was an attempt to reintroduce
a previous patch that broke many self-hosting bots with clang timeouts,
but it still has slowdown issues, at least  on ARM, increasing the
compilation time (stage 2, clang's) by 5x.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@239175 91177308-0d34-0410-b5e6-96231b3b80d8

9 years agoRefactor padding writing into a helper function.
Rafael Espindola [Fri, 5 Jun 2015 18:21:00 +0000 (18:21 +0000)]
Refactor padding writing into a helper function.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@239174 91177308-0d34-0410-b5e6-96231b3b80d8

9 years ago[InstCombine][NFC] Add a ``break;`` statement.
Sanjoy Das [Fri, 5 Jun 2015 18:04:46 +0000 (18:04 +0000)]
[InstCombine][NFC] Add a ``break;`` statement.

This change is NFC because both the ``break;`` and the fall through end
up returning immediately. However, this helps clarify intent and also
ensures correctness in case more ``case`` blocks are added later.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@239172 91177308-0d34-0410-b5e6-96231b3b80d8

9 years ago[InstCombine] Fix PR23751.
Sanjoy Das [Fri, 5 Jun 2015 18:04:42 +0000 (18:04 +0000)]
[InstCombine] Fix PR23751.

PR23751 was caused by a missing ``break;`` in r234388.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@239171 91177308-0d34-0410-b5e6-96231b3b80d8

9 years agoRevert r238473, "Thumb2: Modify codegen for memcpy intrinsic to prefer LDM/STM."
Peter Collingbourne [Fri, 5 Jun 2015 18:01:28 +0000 (18:01 +0000)]
Revert r238473, "Thumb2: Modify codegen for memcpy intrinsic to prefer LDM/STM."
as it caused miscompilations and assertion failures (PR23768,
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20150601/280380.html).

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@239169 91177308-0d34-0410-b5e6-96231b3b80d8

9 years agoSave a map lookup. NFC.
Rafael Espindola [Fri, 5 Jun 2015 17:54:25 +0000 (17:54 +0000)]
Save a map lookup. NFC.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@239168 91177308-0d34-0410-b5e6-96231b3b80d8

9 years agoDAGCombiner: don't duplicate (fmul x, c) in visitFNEG if fneg is free
Fiona Glaser [Fri, 5 Jun 2015 17:52:34 +0000 (17:52 +0000)]
DAGCombiner: don't duplicate (fmul x, c) in visitFNEG if fneg is free

For targets with a free fneg, this fold is always a net loss if it
ends up duplicating the multiply, so definitely avoid it.

This might be true for some targets without a free fneg too, but
I'll leave that for future investigation.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@239167 91177308-0d34-0410-b5e6-96231b3b80d8

9 years agoRangify more for loops in LegacyPassManager.cpp.
Yaron Keren [Fri, 5 Jun 2015 17:48:47 +0000 (17:48 +0000)]
Rangify more for loops in LegacyPassManager.cpp.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@239166 91177308-0d34-0410-b5e6-96231b3b80d8

9 years agoRemove stray semi-colon, NFC
Duncan P. N. Exon Smith [Fri, 5 Jun 2015 17:22:00 +0000 (17:22 +0000)]
Remove stray semi-colon, NFC

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@239165 91177308-0d34-0410-b5e6-96231b3b80d8

9 years ago[Unroll] Rework the naming and structure of the new unroll heuristics.
Chandler Carruth [Fri, 5 Jun 2015 17:01:43 +0000 (17:01 +0000)]
[Unroll] Rework the naming and structure of the new unroll heuristics.

The new naming is (to me) much easier to understand. Here is a summary
of the new state of the world:

- '*Threshold' is the threshold for full unrolling. It is measured
  against the estimated unrolled cost as computed by getUserCost in TTI
  (or CodeMetrics, etc). We will exceed this threshold when unrolling
  loops where unrolling exposes a significant degree of simplification
  of the logic within the loop.
- '*PercentDynamicCostSavedThreshold' is the percentage of the loop's
  estimated dynamic execution cost which needs to be saved by unrolling
  to apply a discount to the estimated unrolled cost.
- '*DynamicCostSavingsDiscount' is the discount applied to the estimated
  unrolling cost when the dynamic savings are expected to be high.

When actually analyzing the loop, we now produce both an estimated
unrolled cost, and an estimated rolled cost. The rolled cost is notably
a dynamic estimate based on our analysis of the expected execution of
each iteration.

While we're still working to build up the infrastructure for making
these estimates, to me it is much more clear *how* to make them better
when they have reasonably descriptive names. For example, we may want to
apply estimated (from heuristics or profiles) dynamic execution weights
to the *dynamic* cost estimates. If we start doing that, we would also
need to track the static unrolled cost and the dynamic unrolled cost, as
only the latter could reasonably be weighted by profile information.

This patch is sadly not without functionality change for the new unroll
analysis logic. Buried in the heuristic management were several things
that surprised me. For example, we never subtracted the optimized
instruction count off when comparing against the unroll heursistics!
I don't know if this just got lost somewhere along the way or what, but
with the new accounting of things, this is much easier to keep track of
and we use the post-simplification cost estimate to compare to the
thresholds, and use the dynamic cost reduction ratio to select whether
we can exceed the baseline threshold.

The old values of these flags also don't necessarily make sense. My
impression is that none of these thresholds or discounts have been tuned
yet, and so they're just arbitrary placehold numbers. As such, I've not
bothered to adjust for the fact that this is now a discount and not
a tow-tier threshold model. We need to tune all these values once the
logic is ready to be enabled.

Differential Revision: http://reviews.llvm.org/D9966

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@239164 91177308-0d34-0410-b5e6-96231b3b80d8

9 years ago[dsymutil] Handle the -oso-prepend-path option when the input is a YAML debug map
Frederic Riss [Fri, 5 Jun 2015 16:35:44 +0000 (16:35 +0000)]
[dsymutil] Handle the -oso-prepend-path option when the input is a YAML debug map

All the tests using a YAML debug map will need this.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@239163 91177308-0d34-0410-b5e6-96231b3b80d8

9 years ago[bpf] rename triple names bpf_be -> bpfeb
Alexei Starovoitov [Fri, 5 Jun 2015 16:11:14 +0000 (16:11 +0000)]
[bpf] rename triple names bpf_be -> bpfeb

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@239162 91177308-0d34-0410-b5e6-96231b3b80d8

9 years ago[Hexagon] Reapply r239097 with tests corrected for shuffling and duplexing.
Colin LeMahieu [Fri, 5 Jun 2015 16:00:11 +0000 (16:00 +0000)]
[Hexagon] Reapply r239097 with tests corrected for shuffling and duplexing.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@239161 91177308-0d34-0410-b5e6-96231b3b80d8