summaryrefslogtreecommitdiffstats
Commit message (Collapse)AuthorAgeFilesLines
* Only clear the thumb bit from function addresses.Rafael Espindola2014-04-033-1/+10
| | | | llvm-svn: 205500
* Simplify ELFObjectFile<ELFT>::getSymbolAddress.Rafael Espindola2014-04-031-30/+11
| | | | | | | | | In particular, we only need to fetch the section if this is a relocatable object. No functionality change. llvm-svn: 205499
* Revert "[Constant Hoisting] Lazily compute the idom and cache the result."Juergen Ributzka2014-04-031-43/+4
| | | | | | | This code is no longer usefull, because we only compute and use the IDom once. There is no benefit in caching it anymore. llvm-svn: 205498
* Make the fail messagesJim Ingham2014-04-037-9/+28
| | | | llvm-svn: 205497
* Print the new eStopReasonExec in stop_reason_to_str.Jim Ingham2014-04-031-0/+2
| | | | llvm-svn: 205496
* Account for scalarization costs in BasicTTI::getMemoryOpCost for extending ↵Hal Finkel2014-04-032-2/+29
| | | | | | | | | | | | | | | | | | | | vector loads When a vector type legalizes to a larger vector type, and the target does not support the associated extending load (or truncating store), then legalization will scalarize the load (or store) resulting in an associated scalarization cost. BasicTTI::getMemoryOpCost needs to account for this. Between this, and r205487, PowerPC on the P7 with VSX enabled shows: MultiSource/Benchmarks/PAQ8p/paq8p: 43% speedup SingleSource/Benchmarks/BenchmarkGame/puzzle: 51% speedup SingleSource/UnitTests/Vectorizer/gcc-loops 28% speedup (some of these are new; some of these, such as PAQ8p, just reverse regressions that VSX support would trigger) llvm-svn: 205495
* I removed SetCallback from the Python API's since it wasn't actually useful, ↵Jim Ingham2014-04-031-1/+2
| | | | | | | | and added SetScriptCallbackFunction, and SetScriptCallbackBody. So add these to the default constructor test case. llvm-svn: 205494
* Revert "Fix a nomenclature error in llvm-nm."Rafael Espindola2014-04-036-47/+41
| | | | | | | | | | | | | | | | This reverts commit r205479. It turns out that nm does use addresses, it is just that every reasonable relocatable ELF object has sections with address 0. I have no idea if those exist in reality, but it at least it shows that llvm-nm should use the name address. The added test was includes an unusual .o file with non 0 section addresses. I created it by hacking ELFObjectWriter.cpp. Really sorry for the churn. llvm-svn: 205493
* Fix test case.Richard Trieu2014-04-031-1/+1
| | | | llvm-svn: 205492
* Allow ASTFile to be reset to nullBen Langmuir2014-04-031-1/+2
| | | | | | Which we do in removeModules(). llvm-svn: 205491
* Fix style.Rui Ueyama2014-04-031-5/+5
| | | | llvm-svn: 205490
* [X86] As per suggestion from Craig Topper and Hal Finkel, overrideLang Hames2014-04-022-40/+39
| | | | | | | | | TargetInstrInfo::findCommutedOpIndices to enable VFMA*231 commutation, rather than abusing commuteInstruction. Thanks very much for the suggestion guys! llvm-svn: 205489
* Add a missing arm64 idef.Jason Molenda2014-04-021-1/+1
| | | | llvm-svn: 205488
* Fix multi-register costs in BasicTTI::getCastInstrCostHal Finkel2014-04-022-1/+23
| | | | | | | | | | | | | For an cast (extension, etc.), the currently logic predicts a low cost if the associated operation (keyed on the destination type) is legal (or promoted). This is not true when the number of values required to legalize the type is changing. For example, <8 x i16> being sign extended by <8 x i32> is not generically cheap on PPC with VSX, even though sign extension to v4i32 is legal, because two output v4i32 values are required compared to the single v8i16 input value, and without custom logic in the target, this conversion will scalarize. llvm-svn: 205487
* Minor style fix.Rui Ueyama2014-04-021-2/+1
| | | | llvm-svn: 205486
* CodeGen: Emit some functions as weak_odr under -fms-compatibilityDavid Majnemer2014-04-028-12/+112
| | | | | | | | | | | | | | | | | | | Summary: MSVC always emits inline functions marked with the extern storage class specifier. The result is something similar to the opposite of __attribute__((gnu_inline)). This extension is also available in C. This fixes PR19264. Reviewers: rnk, rsmith CC: cfe-commits Differential Revision: http://llvm-reviews.chandlerc.com/D3207 llvm-svn: 205485
* Add test case for [Constant Hoisting] Erase dead cast instructions (r204538).Juergen Ributzka2014-04-021-0/+16
| | | | llvm-svn: 205484
* ARM Linux supportRenato Golin2014-04-021-0/+4
| | | | llvm-svn: 205483
* Add decorator for pr19311 on FreeBSDEd Maste2014-04-021-0/+1
| | | | | | | A breakpoint at the resume address for "continue" is skipped. This is presumably the root cause for the Linux failure as well. llvm-svn: 205482
* [CodeGen] Teach the peephole optimizer to remember (and exploit) all foldingLang Hames2014-04-022-35/+73
| | | | | | | | opportunities in the current basic block, rather than just the last one seen. <rdar://problem/16478629> llvm-svn: 205481
* Workaround for collision between enum members in LLVM's MachO.h and system ↵Jim Ingham2014-04-0221-28/+143
| | | | | | | | | | headers on Mac OS X (in particular mach/machine.h). <rdar://problem/16494607> llvm-svn: 205480
* Fix a nomenclature error in llvm-nm.Rafael Espindola2014-04-024-31/+47
| | | | | | | | | | | What llvm-nm prints depends on the file format. On ELF for example, if the file is relocatable, it prints offsets. If it is not, it prints addresses. Since it doesn't really need to care what it is that it is printing, use the generic term value. Fix or implement getSymbolValue to keep llvm-nm working. llvm-svn: 205479
* Add ability to disable building LLVM utilsPete Cooper2014-04-021-6/+15
| | | | | | Patch by Chris Bieneman llvm-svn: 205478
* Objective-C [IRGen]. Add encoding for 'weak' attribute of aFariborz Jahanian2014-04-022-0/+11
| | | | | | 'readonly' property. // rdar://16136439 llvm-svn: 205477
* [PowerPC] Make PPCTTI::getMemoryOpCost call BasicTTI::getMemoryOpCostHal Finkel2014-04-021-3/+3
| | | | | | | | | | PPCTTI::getMemoryOpCost will now make use of BasicTTI::getMemoryOpCost to calculate the base cost of the memory access, and then adjust on top of that. There is no functionality change from this modification, but it will become important so that PPCTTI can take advantage of scalarization information for which BasicTTI::getMemoryOpCost will account in the near future. llvm-svn: 205476
* Concatenate strings that are unnecessarily separated.Rui Ueyama2014-04-021-5/+2
| | | | llvm-svn: 205475
* Add comments and test case for [DAG] Keep the opaque constant flag when ↵Juergen Ributzka2014-04-022-1/+29
| | | | | | performing unary constant folding operations (r204737). llvm-svn: 205474
* typoAdrian Prantl2014-04-021-1/+1
| | | | llvm-svn: 205473
* [X86] Make the VFMA*231 variants commutable and relax the alignment restrictionsLang Hames2014-04-022-106/+147
| | | | | | | | | | | on FMA3 memory operands. FMA3 instructions are VEX encoded, so they can load from unaligned memory. Testcase to follow, along with related patch. <rdar://problem/16478629> llvm-svn: 205472
* Revert "Reapply "LTO: add API to set strategy for -internalize""Duncan P. N. Exon Smith2014-04-026-86/+17
| | | | | | | | | | | This reverts commit r199244. Conflicts: include/llvm-c/lto.h include/llvm/LTO/LTOCodeGenerator.h lib/LTO/LTOCodeGenerator.cpp llvm-svn: 205471
* Minor cleanup.Rui Ueyama2014-04-021-1/+1
| | | | llvm-svn: 205470
* Remove unused fields/member functions from unit test.Rui Ueyama2014-04-021-13/+4
| | | | llvm-svn: 205469
* Add comments and test case for [X86TTI] Make constant base pointers for ↵Juergen Ributzka2014-04-022-0/+27
| | | | | | GetElementPtr opaque (r204739). llvm-svn: 205468
* If a using-declaration names a class member, but appears outside a class, tryRichard Smith2014-04-025-3/+150
| | | | | | to suggest a different syntax to get the same effect. llvm-svn: 205467
* Remove unused include.Rui Ueyama2014-04-021-2/+0
| | | | llvm-svn: 205466
* ARM: fixup tests to specify the target more explicitlySaleem Abdulrasool2014-04-0220-21/+35
| | | | | | | | | | | | | This changes the tests that were targeting ARM EABI to explicitly specify the environment rather than relying on the default. This breaks with the new Windows on ARM support when running the tests on Windows where the default environment is no longer EABI. Take the opportunity to avoid a pointless redirect (helps when trying to debug with providing a command line invocation which can be copy and pasted) and removing a few greps in favour of FileCheck. llvm-svn: 205465
* Add test case for [Stackmaps][X86TTI] Fix think-o in getIntImmCost ↵Juergen Ributzka2014-04-021-0/+17
| | | | | | calculation (r204738). llvm-svn: 205464
* Simplify communication between Resolver and Input Graph.Rui Ueyama2014-04-026-89/+45
| | | | | | | | | | | | | | | | | | | | | | | | Resolver is sending too much information to Input Graph than Input Graph actually needs. In order to collect the detailed information, which wouldn't be consumed by anyone, we have a good amount of code in Resolver, Input Graph and Input Elements. This patch is to simplify it. No functionality change. Specifically, this patch replaces ResolverState enum with a boolean value. The enum defines many bits to notify the progress about linking to Input Graph using bit masks, however, what Input Graph actually does is to compare a given value with 0. The details of the bit mask is simply being ignored, so the efforts to collect such data is wasted. This patch also changes the name of the notification interface from setResolverState to notifyProgress, to make it sounds more like message passing style. It's not a setter but something to notify of an update, so the new name should be more appropriate than before. Differential Revision: http://llvm-reviews.chandlerc.com/D3267 llvm-svn: 205463
* Don’t #include "llvm/Support/MachO.h” as it isn’t needed here.Greg Clayton2014-04-021-2/+0
| | | | llvm-svn: 205462
* Don’t include “llvm/Support/MachO.h” as it isn’t needed here.Greg Clayton2014-04-021-2/+0
| | | | llvm-svn: 205461
* Don’t #include “llvm/Support/MachO.h” in a header file if we can avoid ↵Greg Clayton2014-04-022-2/+2
| | | | | | it. llvm-svn: 205460
* ARM: update subtarget information for Windows on ARMSaleem Abdulrasool2014-04-0213-16/+157
| | | | | | | Update the subtarget information for Windows on ARM. This enables using the MC layer to target Windows on ARM. llvm-svn: 205459
* Make a few more range-based loops use explicit types.Jim Grosbach2014-04-022-2/+2
| | | | | | No functional change. llvm-svn: 205458
* Use explit types in foreach iteratorsTobias Grosser2014-04-021-18/+17
| | | | llvm-svn: 205457
* Add back an assert that was lost in the ELFObjectFile.h split.Rafael Espindola2014-04-021-0/+5
| | | | llvm-svn: 205456
* TargetLibraryInfo: Disable memcpy and memset on R600Tom Stellard2014-04-022-1/+64
| | | | | | There are no implementations of these for R600. llvm-svn: 205455
* Remove dead code.Rui Ueyama2014-04-022-6/+0
| | | | llvm-svn: 205454
* Simplify resolveFrameIndex() signature.Jim Grosbach2014-04-0210-26/+17
| | | | | | | | Just pass a MachineInstr reference rather than an MBB iterator. Creating a MachineInstr& is the first thing every implementation did anyway. llvm-svn: 205453
* ARM: cortex-m0 doesn't support unaligned memory access.Jim Grosbach2014-04-022-1/+19
| | | | | | | | | | | | Unlike other v6+ processors, cortex-m0 never supports unaligned accesses. From the v6m ARM ARM: "A3.2 Alignment support: ARMv6-M always generates a fault when an unaligned access occurs." rdar://16491560 llvm-svn: 205452
* Make some range based loop types more explicit.Jim Grosbach2014-04-022-6/+6
| | | | | | No functional change, but more readable code. llvm-svn: 205451
OpenPOWER on IntegriCloud