summaryrefslogtreecommitdiffstats
Commit message (Collapse)AuthorAgeFilesLines
...
* [AlignmentFromAssumptions] Don't crash just because the target is 32-bitHal Finkel2014-09-112-0/+219
| | | | | | | | | We used to crash processing any relevant @llvm.assume on a 32-bit target (because we'd ask SE to subtract expressions of differing types). I've copied our 'simple.ll' test, but with the data layout from arm-linux-gnueabihf to get some meaningful test coverage here. llvm-svn: 217574
* NULL->nullptrAlexander Musman2014-09-111-1/+1
| | | | llvm-svn: 217573
* [mach-o]: Support deprecated -seg1addr alias for -image_baseTim Northover2014-09-111-0/+2
| | | | | | | Because NO LINKER MAY CHANGE. EVER. Even if it's a complete rewrite from scratch. llvm-svn: 217572
* Build correct vector filled with undef nodesDavid Xu2014-09-112-4/+62
| | | | llvm-svn: 217570
* Fix a couple of -Wsign-compare warnings introduced in r217556Justin Bogner2014-09-111-3/+2
| | | | llvm-svn: 217569
* Make getFlavor function.Rui Ueyama2014-09-111-24/+28
| | | | | | | The dangling "else" at the end of #if looked a bit error-prone. Make it a separate function. No functionality change. llvm-svn: 217568
* utils: Teach lldbDataFormatters how to format ArrayRefsJustin Bogner2014-09-111-0/+33
| | | | llvm-svn: 217567
* If lld is renamed (or symlinked) to "ld" automatically pick the right flavor.Nick Kledzik2014-09-114-6/+74
| | | | | | | | | | | | | The existing system linkers on Darwin and Linux are called "ld". We'd like to eventually drop in lld as "ld" and have it just work. But lld is a universal linker that requires the first option to be -flavor to know which command line mode to emulate (gnu or darwin). This change tests if argv[0] is "ld" and if so, if the tool was built on MacOSX then assume the darwin flavor otherwise the gnu flavor. There are two test cases which copy lld to "ld" and then run it. One for darwin and one for linux. llvm-svn: 217566
* [x86] FileCheck-ize this test.Chandler Carruth2014-09-111-5/+24
| | | | llvm-svn: 217565
* Use CostPool::PoolRef typedef some moreDavid Blaikie2014-09-111-5/+3
| | | | | | Cleanup to 217563 suggested by Lang Hames in post-commit review. llvm-svn: 217564
* shared_ptrify ownershp of PoolEntries in PBQP's CostPoolDavid Blaikie2014-09-101-46/+10
| | | | | | | | | | | | | | Leveraging both intrusive shared_ptr-ing (std::enable_shared_from_this) and shared_ptr<T>-owning-U (to allow external users to hold std::shared_ptr<CostT> while keeping the underlying PoolEntry alive). The intrusiveness could be removed if we had a weak_set that implicitly removed items from the set when their underlying data went away. This /might/ fix an existing memory leak reported by LeakSanitizer in r217504. llvm-svn: 217563
* R600/SI: Fix losing chain when fixing reg class of loads.Matt Arsenault2014-09-102-6/+40
| | | | | | | The lost chain resulting in earlier side effecting nodes being deleted. llvm-svn: 217561
* R600/SI: Report offset in correct units for st64 DS instructionsMatt Arsenault2014-09-101-0/+15
| | | | | | | | | | | Need to convert the 64 element offset into bytes, not just the element size like the normal case instructions. Noticed by inspection. This can't be hit now because st64 instructions aren't emitted during instruction selection, and the post-RA scheduler isn't enabled. llvm-svn: 217560
* [TSan] Use common flags in the same way as all the other sanitizersAlexey Samsonov2014-09-108-116/+37
| | | | llvm-svn: 217559
* [Sanitizer] Get rid of Symbolizer::Get() and Symbolizer::GetOrNull().Alexey Samsonov2014-09-1012-52/+27
| | | | | | | | | We may as well just use Symbolizer::GetOrInit() in all the cases. Don't call Symbolizer::Get() early in tools initialization: these days it doesn't do any important setup work, and we may as well create the symbolizer the first time it's actually needed. llvm-svn: 217558
* Add LLVMgold target to test dependencies.Peter Collingbourne2014-09-101-0/+4
| | | | llvm-svn: 217557
* Thread Safety Analysis: major update to thread safety TIL.DeLesley Hutchins2014-09-108-765/+914
| | | | | | | | | | | | Numerous changes, including: * Changed the way variables and instructions are handled in basic blocks to be more efficient. * Eliminated SExprRef. * Simplified futures. * Fixed documentation. * Compute dominator and post dominator trees. llvm-svn: 217556
* More test for "void *" argument as index of a dictionary literal.Fariborz Jahanian2014-09-101-0/+5
| | | | llvm-svn: 217555
* R600: Custom lower fremMatt Arsenault2014-09-103-0/+123
| | | | llvm-svn: 217553
* Remove a couple of fixed paths that snuck into my test from 217550Ben Langmuir2014-09-101-2/+2
| | | | | | I forgot to fix these again the second time I copy-and-pasted. llvm-svn: 217552
* Rework how resetting breakpoints in changed modules works. Try to match up oldJim Ingham2014-09-109-50/+396
| | | | | | | | locations with new ones if possible. Next up some test cases... llvm-svn: 217551
* Avoid a couple of assertions when preprocessing with modulesBen Langmuir2014-09-108-22/+61
| | | | | | | | | | | | | | | | 1. We were hitting the NextIsPrevious assertion because we were trying to merge decl chains that were independent of each other because we had no Sema object to allow them to find existing decls. This is fixed by delaying loading the "preloaded" decls until Sema is available. 2. We were trying to get identifier info from an annotation token, which asserts. The fix is to special-case the module annotations in the preprocessed output printer. Fixed in a single commit because when you hit 1 you almost invariably hit 2 as well. llvm-svn: 217550
* llgs: implement qThreadStopInfo.Todd Fiala2014-09-103-2/+114
| | | | | | | | | | | This change implements this ticket: http://llvm.org/bugs/show_bug.cgi?id=20899 Adds the qThreadStopInfo RSP command for llgs and includes a test that verifies both debugserver and llgs respond with something reasonable on a multithreaded app. llvm-svn: 217549
* Add doInitialization/doFinalization to DataLayoutPass.Rafael Espindola2014-09-1017-34/+36
| | | | | | | | | | | | | With this a DataLayoutPass can be reused for multiple modules. Once we have doInitialization/doFinalization, it doesn't seem necessary to pass a Module to the constructor. Overall this change seems in line with the idea of making DataLayout a required part of Module. With it the only way of having a DataLayout used is to add it to the Module. llvm-svn: 217548
* Update for llvm api change.Rafael Espindola2014-09-101-3/+3
| | | | llvm-svn: 217547
* Only override the target architecture on -m32 and friends if it isJoerg Sonnenberger2014-09-102-1/+42
| | | | | | | | actually different. Fixes a surprising link error with nodejs on rpi, where armv6-netbsd-eabihf turned into armv5e-netbsd-eabihf, which doesn't lacks the necessary VFP support. llvm-svn: 217546
* Enable use of __builtin_assume_aligned when self-hostingHal Finkel2014-09-101-1/+1
| | | | | | | Clang/LLVM trunk now have support for __builtin_assume_aligned, turn this && into an || so we can use it ourselves. llvm-svn: 217545
* [AlignmentFromAssumptions] Don't divide by zero for unknown starting alignmentHal Finkel2014-09-102-2/+158
| | | | | | | | | | The routine that determines an alignment given some SCEV returns zero if the answer is unknown. In a case where we could determine the increment of an AddRec but not the starting alignment, we would compute the integer modulus by zero (which is illegal and traps). Prevent this by returning early if either the start or increment alignment is unknown (zero). llvm-svn: 217544
* Objective-C. Accept 'nil' as indexing argument toFariborz Jahanian2014-09-102-1/+12
| | | | | | | dictionary literals since the API which implements them accepts it too. // rdar://18254621 llvm-svn: 217543
* [UBSan] Add noinline attribute to handlers that should never return.Alexey Samsonov2014-09-102-15/+48
| | | | | | | | | | | | | FIx a problem reported by Jakub Jelinek: don't do early-exit from fatal UBSan handlers: even if source location is disabled (i.e. acquired by some other thread), we should continue the execution to make sure that: a) some thread will print the error report before calling Die(). b) handler marked as noreturn will indeed not return. Explicitly add "Die()" calls at the end of all fatal handlers to be sure UBSan handlers don't introduce UB themselves. llvm-svn: 217542
* [sphinx cleanup]Dan Liew2014-09-101-1/+1
| | | | | | Fix sphinx warning introduced by r217537 llvm-svn: 217541
* [AArch64] Revert r216141 for cycloneGerolf Hoflehner2014-09-101-1/+1
| | | | | | | | | | The increase of the interleave factor to 4 has side-effects like performance losses eg. due to reminder loops being executed more frequently and may increase code size. It requires more analysis and careful heuristic tuning. Expect double digit gains in small benchmarks like lowercase.c and losses in puzzle.c. llvm-svn: 217540
* [PassManager] Adding Hidden attribute to EnableMLSM optionGerolf Hoflehner2014-09-101-2/+2
| | | | llvm-svn: 217539
* [MergedLoadStoreMotion] Move pass enabling option to PassManagerBuilderGerolf Hoflehner2014-09-102-7/+8
| | | | llvm-svn: 217538
* Fix docs reference to inexistent class.Nico Weber2014-09-101-2/+2
| | | | | | Patch sent via telegraph by TNorthover. Thanks! llvm-svn: 217537
* Remember to eraseFromParent after replaceAllUsesWith.Rafael Espindola2014-09-102-2/+5
| | | | llvm-svn: 217536
* [UBSan] Add halt_on_error runtime flagAlexey Samsonov2014-09-104-10/+13
| | | | llvm-svn: 217535
* MS ABI: Emit dynamic initializers for selectany globals in ↵Hans Wennborg2014-09-102-6/+25
| | | | | | | | | | comdat-associative sections (PR20889) This prevents initializers for comdat-folded globals from running multiple times. Differential Revision: http://reviews.llvm.org/D5281 llvm-svn: 217534
* Cleanup: Use the appropriate API for accessing the DIVariable of aAdrian Prantl2014-09-101-1/+1
| | | | | | DBG_VALUE intrinsic. llvm-svn: 217533
* Remove unused local typedef.Eric Christopher2014-09-101-1/+0
| | | | llvm-svn: 217532
* [AArch64] Temporarily desactivate the PBQP test, while I investigate some ↵Arnaud A. de Grandmaison2014-09-101-14/+0
| | | | | | leaks in the allocator llvm-svn: 217531
* Objective-C arc. Fixes a crash when issuing diagnostic forFariborz Jahanian2014-09-102-1/+25
| | | | | | passing parameter to an audited CF API. rdar://18222007 llvm-svn: 217530
* Make CallingConv::ID an alias of "unsigned".Alexey Samsonov2014-09-103-11/+10
| | | | | | | | | | | | | | | | | | | | Summary: Make CallingConv::ID a plain unsigned instead of enum with a fixed set of valus. LLVM IR allows arbitraty calling conventions (you are free to write cc12345), and loading them as enum is an undefined behavior. This was reported by UBSan. Test Plan: llvm regression test suite Reviewers: nicholas Reviewed By: nicholas Subscribers: llvm-commits Differential Revision: http://reviews.llvm.org/D5248 llvm-svn: 217529
* Rename getMaximumUnrollFactor -> getMaxInterleaveFactor; also rename option ↵Sanjay Patel2014-09-1092-147/+147
| | | | | | | | | | | names controlling this variable. "Unroll" is not the appropriate name for this variable. Clang already uses the term "interleave" in pragmas and metadata for this. Differential Revision: http://reviews.llvm.org/D5066 llvm-svn: 217528
* Removed misleading comment.Gerolf Hoflehner2014-09-101-1/+0
| | | | llvm-svn: 217527
* Added missing blankGerolf Hoflehner2014-09-101-1/+1
| | | | llvm-svn: 217526
* Allow to generate a loop without the GuardBBJohannes Doerfert2014-09-104-43/+91
| | | | | | | | | | | | This allows us to omit the GuardBB in front of created loops if we can show the loop trip count is at least one. It also simplifies the dominance relation inside the new created region. A GuardBB (even with a constant branch condition) might trigger false dominance errors during function verification. Differential Revision: http://reviews.llvm.org/D5297 llvm-svn: 217525
* [sphinx cleanup]Dan Liew2014-09-101-1/+1
| | | | | | Fix sphinx warning introduced by r217501. llvm-svn: 217524
* Make compiler-rt tests work with relocatable SDKs on OS XKuba Brecka2014-09-102-0/+43
| | | | | | Reviewed at http://reviews.llvm.org/D4047 llvm-svn: 217523
* Recognize .lib files as linker input explicitlyEhsan Akhgari2014-09-101-0/+1
| | | | | | | | | | | | | | | | Summary: Currently, this is done implicitly in Driver::BuildInputs by considering any invalid input type as linker input. Test Plan: I don't think this behavior is observable for the reason stated above. Reviewers: hansw Subscribers: cfe-commits Differential Revision: http://reviews.llvm.org/D5294 llvm-svn: 217522
OpenPOWER on IntegriCloud