summaryrefslogtreecommitdiffstats
path: root/llvm
Commit message (Collapse)AuthorAgeFilesLines
* Typos. NFC.Chad Rosier2015-09-171-5/+5
| | | | llvm-svn: 247884
* [mips][microMIPS] Implement TEQ, TGE, TGEU, TLT, TLTU and TNE instructionsZoran Jovanovic2015-09-1711-33/+172
| | | | | | Differential Revision: http://reviews.llvm.org/D9658 llvm-svn: 247880
* AVX-512: shufflevector for i1 vectors <2 x i1> .. <64 x i1>Elena Demikhovsky2015-09-174-40/+481
| | | | | | | | | | AVX-512 does not provide an instruction that shuffles mask register. So I do the following way: mask-2-simd , shuffle simd , simd-2-mask Differential Revision: http://reviews.llvm.org/D12727 llvm-svn: 247876
* GCC AutoFDO profile reader - Initial support.Diego Novillo2015-09-177-13/+580
| | | | | | | | | | | | | | | | | | This adds enough machinery to support reading simple GCC AutoFDO profiles. It now supports reading flat profiles (no function calls). Subsequent patches will add support for: - Inlined calls (in particular, the inline call stack is not traversed to accumulate samples). - Working sets and modules. These are used mostly for GCC's LIPO optimizations, so they're not needed in LLVM atm. I'm not sure that we will ever need them. For now, I've if0'd around the calls. The patch also adds support in GCOV.h for gcov version V704 (generated by GCC's profile conversion tool). llvm-svn: 247874
* Check for errors after reading a sample profile in llvm-profdata.Diego Novillo2015-09-171-1/+3
| | | | | | | After reading the profile, check if the reader returned any errors before showing the profile. llvm-svn: 247873
* Try to fix WebAssembly build after r247864Hans Wennborg2015-09-162-2/+2
| | | | llvm-svn: 247870
* ScalarEvolution: added tmp to avoid use-after-dtor in for loop.Naomi Musgrave2015-09-161-2/+5
| | | | | | | | | | | | | | | | | Summary: For loop destroyed current instance before invoking next. Temporary variable added to prevent use-after-dtor when invoke destructor on current instance. Reviewers: eugenis Subscribers: llvm-commits, sanjoy Differential Revision: http://reviews.llvm.org/D12912 Rename temp var. llvm-svn: 247867
* Make sure we're negating the assembler predicate - no testcaseEric Christopher2015-09-161-1/+1
| | | | | | | because it isn't being used on anything via the assembler right now. llvm-svn: 247866
* Use the cached TargetInstrInfo instead of looking it up again.Eric Christopher2015-09-161-2/+1
| | | | llvm-svn: 247865
* constify the Function parameter to the TTI creation callback andEric Christopher2015-09-1623-33/+36
| | | | | | propagate to all callers/users/etc. llvm-svn: 247864
* [WinEH] Rip out the landingpad-based C++ EH state numbering codeReid Kleckner2015-09-1615-1572/+73
| | | | | | It never really worked, and the new code is working better every day. llvm-svn: 247860
* [CMake] More refactoring of symlink creation.Chris Bieneman2015-09-161-11/+17
| | | | | | This refactoring is to enable clang to re-use this code. llvm-svn: 247850
* [WinEHPrepare] Turn terminatepad into a cleanuppad + call + cleanupretDavid Majnemer2015-09-162-0/+59
| | | | | | | | The MSVC doesn't really support exception specifications so let's just turn these into cleanuppads. Later, we might use terminatepad to more efficiently encode the "noexcept"-ness of a function body. llvm-svn: 247848
* Allow BackpatchWord to be called for non-32-bit aligned words, andTeresa Johnson2015-09-161-8/+19
| | | | | | | | | from outside the BitstreamWriter. Split out of patch D12536 (Function bitcode index in Value Symbol Table and lazy reading support), which will use it to patch in the VST offset. llvm-svn: 247847
* [InstCombine] Optimize icmp slt signum(x), 1 --> icmp slt x, 1Sanjoy Das2015-09-163-0/+97
| | | | | | | | | | | | | | | | | | | | | | Summary: `signum(x)` is sometimes implemented as `(x >> 63) | (-x >>> 63)` (for an `i64` `x`). This change adds a matcher for that pattern, and an instcombine rule to optimize `signum(x) s< 1`. Later, we can also consider optimizing: icmp slt signum(x), 0 --> icmp slt x, 0 icmp sle signum(x), 1 --> true etc. Reviewers: majnemer Subscribers: llvm-commits Differential Revision: http://reviews.llvm.org/D12703 llvm-svn: 247846
* [WinEH] Pull Adjectives and CatchObj out of the catchpad arg listReid Kleckner2015-09-1610-53/+82
| | | | | | | | | | Clang now passes the adjectives as an argument to catchpad. Getting the CatchObj working is simply a matter of threading another static alloca through codegen, first as an alloca, then as a frame index, and finally as a frame offset. llvm-svn: 247844
* [WinEHPrepare] Provide a cloning mode which doesn't demoteDavid Majnemer2015-09-162-9/+190
| | | | | | | | | | | We are experimenting with a new approach to saving and restoring SSA values used across funclets: let the register allocator do the dirty work for us. However, this means that we need to be able to clone commoned blocks without relying on demotion. llvm-svn: 247835
* [WinEHPrepare] Refactor explicit EH preparationDavid Majnemer2015-09-161-10/+48
| | | | | | | | Split the preparation machinery into several functions, we will want to selectively enable/disable different parts of it for an alternative mechanism for dealing with cross-funclet uses. llvm-svn: 247834
* Disable the second verification run when performing LTO throughTeresa Johnson2015-09-162-1/+35
| | | | | | | | gold in NDEBUG mode. Follow on patch for r247729 - LTO: Disable extra verify runs in release builds. llvm-svn: 247824
* [WinEH] Skip state numbering when no EH pads are presentReid Kleckner2015-09-163-6/+25
| | | | | | | | | Otherwise we'd try to emit the thunk that passes the LSDA to __CxxFrameHandler3. We don't emit the LSDA if there were no landingpads, so we'd end up with an assembler error when trying to write the COFF object. llvm-svn: 247820
* Improve "default_triple" specification: make it at the directory level for ↵Mehdi Amini2015-09-165-5/+2
| | | | | | | test/tools/llvm-mc From: Mehdi Amini <mehdi.amini@apple.com> llvm-svn: 247819
* [WebAssembly] Check in an initial CFG Stackifier passDan Gohman2015-09-1615-6/+929
| | | | | | | | | | | | This pass implements a simple algorithm for conversion from CFG to wasm's structured control flow. It doesn't yet handle multiple-entry loops; that will be added in a future patch. It also adds initial support for switch statements. Differential Revision: http://reviews.llvm.org/D12735 llvm-svn: 247818
* propagate fast-math-flags on DAG nodesSanjay Patel2015-09-1617-205/+344
| | | | | | | | | | | | | | | | | | | After D10403, we had FMF in the DAG but disabled by default. Nick reported no crashing errors after some stress testing, so I enabled them at r243687. However, Escha soon notified us of a bug not covered by any in-tree regression tests: if we don't propagate the flags, we may fail to CSE DAG nodes because differing FMF causes them to not match. There is one test case in this patch to prove that point. This patch hopes to fix or leave a 'TODO' for all of the in-tree places where we create nodes that are FMF-capable. I did this by putting an assert in SelectionDAG.getNode() to find any FMF-capable node that was being created without FMF ( D11807 ). I then ran all regression tests and test-suite and confirmed that everything passes. This patch exposes remaining work to get DAG FMF to be fully functional: (1) add the flags to non-binary nodes such as FCMP, FMA and FNEG; (2) add the flags to intrinsics; (3) use the flags as conditions for transforms rather than the current global settings. Differential Revision: http://reviews.llvm.org/D12095 llvm-svn: 247815
* Add assembler fatal error for undefined assembler labels in COFF writerReid Kleckner2015-09-162-0/+9
| | | | llvm-svn: 247814
* don't repeat function names in comments; NFCSanjay Patel2015-09-161-29/+24
| | | | llvm-svn: 247813
* [sanitizer] Add MSan support for AArch64Adhemerval Zanella2015-09-161-0/+34
| | | | | | | | | This patch adds support for msan on aarch64-linux for both 39 and 42-bit VMA. The support is enabled by defining the SANITIZER_AARCH64_VMA compiler flag to either 39 or 42 at build time for both clang/llvm and compiler-rt. The default VMA is 39 bits. llvm-svn: 247807
* [SPARC] Both GNU and Solaris as support eq as condition code for integer ops.Joerg Sonnenberger2015-09-162-0/+5
| | | | llvm-svn: 247804
* [SPARC] Recognize st/stx operations with %fsr argument too.Joerg Sonnenberger2015-09-163-0/+32
| | | | llvm-svn: 247794
* Test commit: Fixed a few typos in the comments.David L Kreitzer2015-09-161-6/+6
| | | | llvm-svn: 247793
* [ARM] Register ARMPreAllocLoadStoreOpt pass with LLVM pass manager.Chad Rosier2015-09-161-2/+14
| | | | llvm-svn: 247791
* llvm-mc-fuzzer: A fuzzing tool for the MC layer.Daniel Sanders2015-09-163-1/+155
| | | | | | | | | | | | | | | Summary: Only the disassembler is supported in this patch but it has already found a few issues in the Mips disassembler (mostly invalid instructions being successfully disassembled). Reviewers: kcc Subscribers: russell.gallop, silvas, kcc, llvm-commits Differential Revision: http://reviews.llvm.org/D12723 llvm-svn: 247786
* [X86] Do not generate 64-bit pops of 32-bit GPRs.Michael Kuperstein2015-09-162-4/+24
| | | | | | | | | | | When trying emit a stack adjustments using pops, frame lowering selects an arbitrary free GPR. It should always select one from an appropriate class... This fixes PR24649. Patch by: amjad.aboud@intel.com Differential Revision: http://reviews.llvm.org/D12609 llvm-svn: 247785
* [X86] Fix emitEpilogue() to make less assumptions about popsMichael Kuperstein2015-09-163-8/+20
| | | | | | | | | | | | | | | This is the mirror image of r242395. When X86FrameLowering::emitEpilogue() looks for where to insert the %esp addition that deallocates stack space used for local allocations, it assumes that any sequence of pop instructions from function exit backwards consists purely of restoring callee-save registers. This may be false, since from some point backward, the pops may be clean-up of stack space allocated for arguments to a call. Patch by: amjad.aboud@intel.com Differential Revision: http://reviews.llvm.org/D12688 llvm-svn: 247784
* [mips][microMIPS] Implement PREFX, LHUE, LBE, LBUE, LHE, LWE, SBE, SHE and ↵Zoran Jovanovic2015-09-1613-4/+141
| | | | | | | | SWE instructions Differential Revision: http://reviews.llvm.org/D9189 llvm-svn: 247780
* Whitespace.NAKAMURA Takumi2015-09-161-2/+1
| | | | llvm-svn: 247779
* Copy back Inputs/gmlt.ll. Also DebugInfo/X86/gmlt.test uses it.NAKAMURA Takumi2015-09-161-0/+153
| | | | llvm-svn: 247777
* Fix test gmlt.test by moving its Inputs where expected.Mehdi Amini2015-09-161-0/+0
| | | | | | | I couldn't see the failure as the test is XFAIL'ed on Darwin. From: Mehdi Amini <mehdi.amini@apple.com> llvm-svn: 247776
* Make the default triple optional by allowing an empty stringMehdi Amini2015-09-16116-25/+57
| | | | | | | | | | | | | | | When building LLVM as a (potentially dynamic) library that can be linked against by multiple compilers, the default triple is not really meaningful. We allow to explicitely set it to an empty string when configuring LLVM. In this case, said "target independent" tests in the test suite that are using the default triple are disabled by matching the newly available feature "default_triple". Reviewers: probinson, echristo Differential Revision: http://reviews.llvm.org/D12660 From: Mehdi Amini <mehdi.amini@apple.com> llvm-svn: 247775
* [NaryReassociate] Improve test CHECKMehdi Amini2015-09-162-3/+6
| | | | | | | | | | | Add `CHECK` directives for the function calls. Differential Revision: http://reviews.llvm.org/D12885 Patch by: Volkan Keles <vkeles@apple.com> From: Mehdi Amini <mehdi.amini@apple.com> llvm-svn: 247774
* Use range-based for loops. NFCCraig Topper2015-09-161-47/+31
| | | | llvm-svn: 247772
* Fix a spelling error in the description of a statistic. NFCCraig Topper2015-09-161-1/+1
| | | | llvm-svn: 247771
* [Unroll] Fix a bug in UnrolledInstAnalyzer::visitLoad.Michael Zolotukhin2015-09-162-1/+30
| | | | | | | | We only checked that a global is initialized with constants, which is incorrect. We should be checking that GlobalVariable *is* a constant, not just initialized with it. llvm-svn: 247769
* llvm/CodeGen/CommandFlags.h: Prune doubleslash in #include.NAKAMURA Takumi2015-09-161-1/+1
| | | | | | | | | | | While packaging 3.7 for Fedora, the debug info splitting process fell over this, so fix it upstream seems like a good plan. This should be put in the 3.7 branch as well. Noticed by Dave Airlie <airlied@redhat.com> llvm-svn: 247757
* [IndVars] Fix PR24783.Sanjoy Das2015-09-152-1/+32
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | In `IndVarSimplify::ExpandSCEVIfNeeded`, `SCEVExpander::findExistingExpansion` may return an `llvm::Value` that differs in type from the SCEV it was asked to find an expansion for (but computes the same value). In such cases, we fall back on `expandCodeFor`; and rely on LLVM to CSE the two equivalent expressions (different only by a no-op cast) into a single computation. I tried a few other approaches to fixing PR24783, all of which turned out to be more complex than this current version: 1. Move the `ExpandSCEVIfNeeded` logic into `expandCodeFor`. This got problematic because currently we do not pass in the `Loop *` into `expandCodeFor`. Changing the interface to do this is a more invasive change, and really does not make much semantic sense unless the SCEV being passed in is an add recurrence. There is also the problem of `expandCodeFor` being used in places other than `indvars` -- there may be performance / correctness issues elsewhere if `expandCodeFor` is moved from always generating IR from scratch to cache-like model. 2. Have `findExistingExpansion` only return expression with the correct type. This would make `isHighCostExpansionHelper` and thus `isHighCostExpansion` more conservative than necessary. 3. Insert casts on the value returned by `findExistingExpansion` if needed using `InsertNoopCastOfTo`. This is complicated because `InsertNoopCastOfTo` depends on internal state of its `SCEVExpander` (specifically `Builder.GetInserPoint()`), and this may not be set up when `ExpandSCEVIfNeeded` is called. 4. Manually insert casts on the value returned by `findExistingExpansion` if needed using `InsertNoopCastOfTo` via `CastInst::Create`. This is probably workable, but figuring out the location where the cast instruction needs to be inserted has enough edge cases (arguments, constants, invokes, LCSSA must be preserved) makes me feel what I have right now is simplest solution. llvm-svn: 247749
* [IndVars] Rename variable; NFC.Sanjoy Das2015-09-151-2/+2
| | | | llvm-svn: 247748
* [SCEVExpander] Fix comment formatting; NFC.Sanjoy Das2015-09-151-9/+12
| | | | llvm-svn: 247747
* [llvm-cxxdump] Remove duplicate code check.Davide Italiano2015-09-152-6/+3
| | | | | | | We already fail with 'No such file or directory' when we try to open the file -- if that doesn't exist. Also add a test to verify this behavior. llvm-svn: 247744
* Reapply "LTO: Disable extra verify runs in release builds"Duncan P. N. Exon Smith2015-09-155-21/+58
| | | | | | | This reverts commit r247730, effectively reapplying r247729. This time I have an lld commit ready to follow. llvm-svn: 247735
* [ASan] Don't instrument globals in .preinit_array/.init_array/.fini_arrayAlexey Samsonov2015-09-155-25/+49
| | | | | | | | | | | | These sections contain pointers to function that should be invoked during startup/shutdown by __libc_csu_init and __libc_csu_fini. Instrumenting these globals will append redzone to them, which will be filled with zeroes. This will cause null pointer dereference at runtime. Merge ASan regression tests for globals that should be ignored by instrumentation pass. llvm-svn: 247734
* Revert "LTO: Disable extra verify runs in release builds"Duncan P. N. Exon Smith2015-09-155-58/+21
| | | | | | | This temporarily reverts commit r247729, as it caused lld build failures. I'll recommit once I have an lld patch ready-to-go. llvm-svn: 247730
OpenPOWER on IntegriCloud