summaryrefslogtreecommitdiffstats
path: root/llvm/lib
Commit message (Collapse)AuthorAgeFilesLines
...
* [Hexagon] Add a bitcast to required type in LowerHvxMulKrzysztof Parzyszek2018-01-051-1/+2
| | | | llvm-svn: 321893
* [Hexagon] Add pattern for vsplat to v8i8Krzysztof Parzyszek2018-01-051-0/+5
| | | | llvm-svn: 321892
* [Hexagon] Set boolean contents in HexagonISelLoweringKrzysztof Parzyszek2018-01-051-0/+2
| | | | llvm-svn: 321891
* Fix -Wsign-compare warnings on WindowsReid Kleckner2018-01-052-5/+6
| | | | | | These arise because enums are 'int' by default. llvm-svn: 321887
* Limit size of non-GlobalValue nameSerge Guelton2018-01-051-0/+9
| | | | | | | | | | | | | Otherwise, in some extreme test case, very long names are created and the compiler consumes large amount of memory. Size limit is set to a relatively high value not to disturb debugging. Compiler flag -non-global-value-max-name-size=<value> can be used to customize the size. Differential Revision: https://reviews.llvm.org/D41296 llvm-svn: 321886
* [InstCombine] add folds for min(~a, b) --> ~max(a, b)Sanjay Patel2018-01-051-22/+12
| | | | | | | | | | | | | | | | | Besides the bug of omitting the inverse transform of max(~a, ~b) --> ~min(a, b), the use checking and operand creation were off. We were potentially creating repeated identical instructions of existing values. This led to infinite looping after I added the extra folds. By using the simpler m_Not matcher and not creating new 'not' ops for a and b, we avoid that problem. It's possible that not using IsFreeToInvert() here is more limiting than the simpler matcher, but there are no tests for anything more exotic. It's also possible that we should relax the use checking further to handle a case like PR35834: https://bugs.llvm.org/show_bug.cgi?id=35834 ...but we can make that a follow-up if it is needed. llvm-svn: 321882
* [MSF] Fix FPM interval calcluationZachary Turner2018-01-051-8/+6
| | | | | | | | | | | | | We have some code to try to determine how many pieces an MSF Free Page Map is split into, and this code had an off by one error which would cause the calculation to be incorrect when there were exactly 4096*k + 1 blocks in an MSF file. Original investigation and patch outline by Colden Cullen. Differential Revision: https://reviews.llvm.org/D41742 llvm-svn: 321880
* [Option] Add 'findNearest' method to catch typosBrian Gesiak2018-01-051-0/+63
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Summary: Add a method `OptTable::findNearest`, which allows users of OptTable to check user input for misspelled options. In addition, have llvm-mt check for misspelled options. For example, if a user invokes `llvm-mt /oyt:foo`, the error message will indicate that while an option named `/oyt:` does not exist, `/out:` does. The method ports the functionality of the `LookupNearestOption` method from LLVM CommandLine to libLLVMOption. This allows tools like Clang and Swift, which do not use CommandLine, to use this functionality to suggest similarly spelled options. As room for future improvement, the new method as-is cannot yet properly suggest nearby "joined" options -- that is, for an option string "-FozBar", where "-Foo" is the correct option name and "Bar" is the value being passed along with the misspelled option, this method will calculate an edit distance of 4, by deleting "Bar" and changing "z" to "o". It should instead calculate an edit distance of just 1, by changing "z" to "o" and recognizing "Bar" as a value. This commit includes a disabled test that expresses this limitation. Test Plan: `check-llvm` Reviewers: yamaguchi, v.g.vassilev, teemperor, ruiu, jroelofs Reviewed By: jroelofs Subscribers: jroelofs, llvm-commits Differential Revision: https://reviews.llvm.org/D41732 llvm-svn: 321877
* [BasicAA] Fix linearization of shifts beyond the bitwidth.Davide Italiano2018-01-051-0/+13
| | | | | | | Thanks to Simon Pilgrim for the reduced testcase. Fixes PR35821. llvm-svn: 321873
* [ARM] Issue an erorr when non-general-purpose registers are used in address ↵Momchil Velikov2018-01-051-8/+31
| | | | | | | | | | | | operands Currently the assembler would accept, e.g. `ldr r0, [s0, #12]` and similar. This patch add checks that only general-purpose registers are used in address operands, shifted registers, and shift amounts. Differential revision: https://reviews.llvm.org/D39910 llvm-svn: 321866
* [DebugInfo] Don't crash when given invalid DWARFv5 line table prologue.Jonas Devlieghere2018-01-052-7/+10
| | | | | | | | | | | | | This patch replaces an assertion with an explicit check for the validity of the FORM parameters. The assertion was triggered when the DWARFv5 line table contained a zero address size. This fixes OSS-Fuzz Issue 4644 https://bugs.chromium.org/p/oss-fuzz/issues/detail?id=4644 Differential revision: https://reviews.llvm.org/D41615 llvm-svn: 321863
* [DAGCombine] Fix for PR37563Sam Parker2018-01-051-2/+12
| | | | | | | | | | | | While searching for loads to be narrowed, equal sized loads were not added to the list, resulting in anyext loads not being converted to zext loads. https://bugs.llvm.org/show_bug.cgi?id=35763 Differential Revision: https://reviews.llvm.org/D41628 llvm-svn: 321862
* [ORC] Re-revert r321838: Tests are still failing.Lang Hames2018-01-053-345/+2
| | | | llvm-svn: 321858
* [GISel]: Don't create G_MUL with 1 during translation of GEPAditya Nandakumar2018-01-051-7/+12
| | | | | | | When element size is 1, it's just wasteful to create MUL with 1. https://reviews.llvm.org/D41738 llvm-svn: 321857
* [ORC] Re-apply r321838 - Addition of new ORC core APIs.Lang Hames2018-01-053-2/+345
| | | | | | | | The original commit broke the builders due to a think-o in an assertion: AsynchronousSymbolQuery's constructor needs to check the callback member variables, not the constructor arguments. llvm-svn: 321853
* Debug Info: Support DW_AT_calling_convention on composite types.Adrian Prantl2018-01-052-2/+14
| | | | | | | | | | | | | | | | | | | | | | | | | | | | This implements the DWARF 5 feature described at http://www.dwarfstd.org/ShowIssue.php?issue=141215.1 This allows a consumer to understand whether a composite data type is trivially copyable and thus should be passed by value instead of by reference. The canonical example is being able to distinguish the following two types: // S is not trivially copyable because of the explicit destructor. struct S { ~S() {} }; // T is a POD type. struct T { ~T() = default; }; This patch adds two new (DI)flags to LLVM metadata: TypePassByValue and TypePassByReference. <rdar://problem/36034922> Differential Revision: https://reviews.llvm.org/D41743 llvm-svn: 321844
* Revert r321838 -- It broke some of the builders.Lang Hames2018-01-053-344/+2
| | | | llvm-svn: 321842
* WholeProgramDevirt: Simplify ORE getter mechanism for old PM. NFCI.Peter Collingbourne2018-01-051-34/+17
| | | | llvm-svn: 321841
* [ORC] Add new core ORC APIs (Core.h/Core.cpp): VSO, AsynchronousSymbolQuery andLang Hames2018-01-053-2/+344
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | SymbolSource. These new APIs are a first stab at tackling some current shortcomings of ORC, especially in performance and threading support. VSO (Virtual Shared Object) is a symbol table representing the symbol definitions of a set of modules that behave as if they had been statically linked together into a shared object or dylib. Symbol definitions, either pre-defined addresses or lazy definitions, can be added and queries for symbol addresses made. The table applies the same linkage strength rules that static linkers do when constructing a dylib or shared object: duplicate definitions result in errors, strong definitions override weak or common ones. This class should improve symbol lookup speed by providing centralized symbol tables (as compared to the findSymbol implementation in the in-tree ORC layers, which maintain one symbol table per object file / module added). AsynchronousSymbolQuery is a query for the addresses of a set of symbols. Query results are returned via a callback once they become available. Querying for a set of symbols, rather than one symbol at a time (as the current lookup scheme does) the JIT has the opportunity to make better use of available resources (e.g. by spawning multiple jobs to materialize the requested symbols if possible). Returning results via a callback makes queries asynchronous, so queries from multiple threads of JIT'd code can proceed simultaneously. SymbolSource represents a source of symbol definitions. It is used when adding lazy symbol definitions to a VSO. Symbol definitions can be materialized when needed or discarded if a stronger definition is found. Materializing on demand via SymbolSources should (eventually) allow us to remove the lazy materializers from JITSymbol, which will in turn allow the removal of many current error checks and reduce the number of RPC round-trips involved in materializing remote symbols. Adding a discard function allows sources to discard symbol definitions (or mark them as available_externally), reducing the amount of redundant code generated by the JIT for ODR symbols. llvm-svn: 321838
* Revert "[JumpThreading] Preservation of DT and LVI across the pass"Reid Kleckner2018-01-045-498/+89
| | | | | | | This reverts r321825, it causes crashes in Chromium. Reproducer forthcoming. llvm-svn: 321832
* [JumpThreading] Preservation of DT and LVI across the passBrian M. Rzycki2018-01-045-89/+498
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Summary: See D37528 for a previous (non-deferred) version of this patch and its description. Preserves dominance in a deferred manner using a new class DeferredDominance. This reduces the performance impact of updating the DominatorTree at every edge insertion and deletion. A user may call DDT->flush() within JumpThreading for an up-to-date DT. This patch currently has one flush() at the end of runImpl() to ensure DT is preserved across the pass. LVI is also preserved to help subsequent passes such as CorrelatedValuePropagation. LVI is simpler to maintain and is done immediately (not deferred). The code to perfom the preversation was minimally altered and was simply marked as preserved for the PassManager to be informed. This extends the analysis available to JumpThreading for future enhancements. One example is loop boundary threading. Reviewers: dberlin, kuhar, sebpop Reviewed By: kuhar, sebpop Subscribers: hiraditya, llvm-commits Differential Revision: https://reviews.llvm.org/D40146 llvm-svn: 321825
* [AArch64] Improve code generation of vector buildEvandro Menezes2018-01-041-3/+22
| | | | | | | | | Instead of using, for example, `dup v0.4s, wzr`, which transfers between register files, use the more efficient `movi v0.4s, #0` instead. Differential revision: https://reviews.llvm.org/D41515 llvm-svn: 321824
* [X86] Correct the execution domain for AVX1 VBROADCASTF128 to be FP instead ↵Craig Topper2018-01-041-1/+2
| | | | | | of integer. llvm-svn: 321821
* [DAGCombine] Ensure SDNode use iterator is incremented properly.Amara Emerson2018-01-041-2/+2
| | | | | | Fixes an ASAN bug found by oss-fuzz. llvm-svn: 321813
* Teach InlineCost about address spacesBjorn Pettersson2018-01-041-8/+15
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Summary: I basically copied this patch from here: https://reviews.llvm.org/D1251 But I skipped some of the refactoring to make the patch more clean. The new outer3/inner3 test case in ptr-diff.ll triggers the following assert without this patch: lib/IR/Constants.cpp:1834: static llvm::Constant *llvm::ConstantExpr::getCompare(unsigned short, llvm::Constant *, llvm::Constant *, bool): Assertion `C1->getType() == C2->getType() && "Op types should be identical!"' failed. The other new test cases makes sure that there is code coverage for all modifications in InlineCost.cpp (getting different values due to not fetching sizes for address space zero). I only guarantee code coverage for those tests. The tests are not written in a way that they would break if not having the corrections in InlineCost.cpp. I found it quite hard to fine tune the tests into getting different results based on the pointer sizes (except for the test case where we hit an assert if not teaching InlineCost about address spaces). Reviewers: chandlerc, arsenm, haicheng Reviewed By: arsenm Subscribers: wdng, eraman, llvm-commits, haicheng Differential Revision: https://reviews.llvm.org/D40455 llvm-svn: 321809
* Add assertion on DT availability during LI update in UpdateAnalysisInformationAnna Thomas2018-01-041-0/+1
| | | | | | | | | | | | | This came up during discussions in llvm-commits for rL321653: Check for unreachable preds before updating LI in UpdateAnalysisInformation The assert provides hints to passes to require both DT and LI if we plan on updating LI through this function. Tests run: make check llvm-svn: 321805
* [InstCombine] safely create a constant of the right type (PR35794)Sanjay Patel2018-01-041-4/+4
| | | | llvm-svn: 321801
* [ARM] Fix endianness of Thumb .inst.w directiveOliver Stannard2018-01-041-2/+4
| | | | | | | | | Wide Thumb2 instructions should be emitted into the object file as pairs of 16-bit words of the appropriate endianness, not one 32-bit word. Differential revision: https://reviews.llvm.org/D41185 llvm-svn: 321799
* [Hexagon] Replace INSERTRP/EXTRACTRP with INSERT/EXTRACT in HexagonISDKrzysztof Parzyszek2018-01-043-41/+25
| | | | llvm-svn: 321798
* [ARM GlobalISel] Select G_PHIDiana Picus2018-01-041-16/+33
| | | | | | | | Select G_PHI to PHI and manually constrain the result register. This is very similar to how COPY is handled, so extract and reuse some of that code. llvm-svn: 321797
* [ARM GlobalISel] Legalize scalar G_PHIDiana Picus2018-01-041-0/+7
| | | | | | | Mark G_PHI as Legal for s32 and p0, and also for s64 if we have hard float. Widen any smaller types. llvm-svn: 321795
* [ARM GlobalISel] Fix selection of pointer constantsDiana Picus2018-01-041-8/+26
| | | | | | | | | | | | | | | | | | | | We used to handle G_CONSTANT with pointer type by forcing the type of the result register to s32 and then letting TableGen handle it. Unfortunately, setting the type only works for generic virtual registers, that haven't yet been constrained to a register class (e.g. those used only by a COPY later on). If the result register has already been constrained as a use of a previously selected instruction, then setting the type will assert. It would be nice to be able to teach TableGen to select pointer constants the same as integer constants, but since it's such an edge case (at the moment the only pointer constant that we're generally interested in is 0, and that is mostly used for comparisons and selects, which are also not supported by TableGen) it's probably not worth the effort right now. Instead, handle pointer constants with some trivial handwritten code. llvm-svn: 321793
* [GVNHoist] Fix: PR35222 gvn-hoist incorrectly erases load in case of a loopAditya Kumar2018-01-041-1/+1
| | | | | | | | | | | Reviewers: dberlin sebpop eli.friedman Differential Revision: https://reviews.llvm.org/D41453 llvm-svn: 321789
* Changes in the branch relaxation algorithm.Elena Demikhovsky2018-01-041-57/+107
| | | | | | | | | The existing version worked incorrectly when inversion of a branch condintion is impossible. Changed the "fixupConditionalBranch()" function - a new BB (a trampoline) is created to keep the original branch condition. Differential Revision: https://reviews.llvm.org/D41634 llvm-svn: 321785
* support phi ranges for machine-level IRBob Wilson2018-01-041-16/+10
| | | | | | | | | | | | Add iterator ranges for machine instruction phis, similar to the IR-level phi ranges added in r303964. I updated a few places to use this. Besides general code simplification, this change will allow removing a non-upstream change from Swift's copy of LLVM (in a better way than my previous attempt in http://reviews.llvm.org/D19080). https://reviews.llvm.org/D41672 llvm-svn: 321783
* Do not look up symbol names when n_strx == 0Michael Trent2018-01-031-0/+4
| | | | | | | | | | | | | | | | | | | | | | Summary: Historical tools for working with mach-o binaries verify the nlist field n_strx has a non-zero value before using that value to retrieve symbol names. Under some cirumstances, llvm-nm will attempt to display the symbol name at position 0, even though symbol names at that position are not well defined. This change addresses this problem by returning an empty string when n_strx is zero. rdar://problem/35750548 Reviewers: enderby, davide Reviewed By: enderby, davide Subscribers: davide, llvm-commits, JDevlieghere Differential Revision: https://reviews.llvm.org/D41657 llvm-svn: 321773
* [DAGCombine] Handle out of range EXTRACT_VECTOR_ELT indices Simon Pilgrim2018-01-032-1/+5
| | | | | | | | Handle this in DAGCombiner::visitEXTRACT_VECTOR_ELT the same as we already do in SelectionDAG::getNode and use APInt instead of getZExtValue. This should also fix oss-fuzz #4910 llvm-svn: 321767
* [ExpandMemcmp] rename variables and add hook to override pref for number of ↵Sanjay Patel2018-01-031-14/+17
| | | | | | | | | | | loads per block; NFC The preference only applies to 'memcmp() == 0' expansion, so try to make that clearer. x86 will likely benefit by increasing the default value from '1' to '2' as seen in PR33325: https://bugs.llvm.org/show_bug.cgi?id=33325 ...so that is the planned follow-up to this clean-up step. llvm-svn: 321756
* [X86] Remove 'else' after 'return' I forgot to cleanup before committing D41691.Craig Topper2018-01-031-4/+7
| | | | llvm-svn: 321755
* AMDGPU: Remove dead fileMatt Arsenault2018-01-031-12/+0
| | | | llvm-svn: 321752
* StructurizeCFG: Fix broken backedge detectionMatt Arsenault2018-01-031-82/+28
| | | | | | | | | | | | | | | | | | | | | | | | | | | | The work order was changed in r228186 from SCC order to RPO with an arbitrary sorting function. The sorting function attempted to move inner loop nodes earlier. This was was apparently relying on an assumption that every block in a given loop / the same loop depth would be seen before visiting another loop. In the broken testcase, a block outside of the loop was encountered before moving onto another block in the same loop. The testcase would then structurize such that one blocks unconditional successor could never be reached. Revert to plain RPO for the analysis phase. This fixes detecting edges as backedges that aren't really. The processing phase does use another visited set, and I'm unclear on whether the order there is as important. An arbitrary order doesn't work, and triggers some infinite loops. The reversed RPO list seems to work and is closer to the order that was used before, minus the arbitary custom sorting. A few of the changed tests now produce smaller code, and a few are slightly worse looking. llvm-svn: 321751
* [InstCombine] Check for out of range shift values using APInt before calling ↵Simon Pilgrim2018-01-031-4/+4
| | | | | | | | getZExtValue Reduced from oss-fuzz #4871 test case llvm-svn: 321748
* [X86] Remove useless custom inserter for 64-bit TAILJMP and TCRETURN opcodesCraig Topper2018-01-032-12/+1
| | | | | | | | | | This custom inserter was added in r124272 at which time it added about bunch of Defs for Win64. In r150708, those defs were removed leaving only the "return BB". So I think this means the custom inserter is a NOP these days. This patch removes the remaining code and stops tagging the instructions for custom insertion Differential Revision: https://reviews.llvm.org/D41671 llvm-svn: 321747
* [X86] Use ANY_EXTEND instead of SIGN_EXTEND in lowerMasksToRegCraig Topper2018-01-031-1/+1
| | | | | | | | Currently we use SIGN_EXTEND in lowerMasksToReg as part of calling convention setup, but we don't require a specific value for the upper bits. This patch changes it to ANY_EXTEND which will be lowered as SIGN_EXTEND if it ends up sticking around. llvm-svn: 321746
* Remove left-over debug printout from r321692Hans Wennborg2018-01-031-1/+0
| | | | | | | Besides the unsightly print-out, it was causing some buildbots to fail, e.g. http://lab.llvm.org:8011/builders/clang-x86-windows-msvc2015/builds/9311 llvm-svn: 321711
* [InstSimplify] Missed optimization in math expression: squashing exp(log), ↵Dmitry Venikov2018-01-031-6/+33
| | | | | | | | | | | | | | | | log(exp) Summary: This patch enables folding following expressions under -ffast-math flag: exp(log(x)) -> x, exp2(log2(x)) -> x, log(exp(x)) -> x, log2(exp2(x)) -> x Reviewers: spatel, hfinkel, davide Reviewed By: spatel, hfinkel, davide Subscribers: scanon, llvm-commits Differential Revision: https://reviews.llvm.org/D41381 llvm-svn: 321710
* [ARM][NFC] Avoid recreating MCSubtargetInfo in ARMAsmBackendAlex Bradbury2018-01-036-36/+31
| | | | | | | | | | | | | After D41349, we can now directly access MCSubtargetInfo from createARM*AsmBackend. This patch makes use of this, avoiding the need to create a fresh MCSubtargetInfo (which was previously always done with a blank CPU and feature string). Given the total size of the change remains pretty tiny and we're removing the old explicit destructor, I changed the STI field to a reference rather than a pointer. Differential Revision: https://reviews.llvm.org/D41693 llvm-svn: 321707
* [AArch64][SVE] Asm: Add restricted register classes for SVE predicate vectors.Sander de Smalen2018-01-033-11/+52
| | | | | | | | | | | | | | | | | Summary: Add a register class for SVE predicate operands that can only be p0-p7 (as opposed to p0-p15) Patch [1/3] in a series to add predicated ADD/SUB instructions for SVE. Reviewers: rengolin, mcrosier, evandro, fhahn, echristo, olista01, SjoerdMeijer, javed.absar Reviewed By: fhahn Subscribers: aemerson, javed.absar, tschuett, kristof.beyls, llvm-commits Differential Revision: https://reviews.llvm.org/D41441 llvm-svn: 321699
* Fix build of WebAssembly and AVR backends after r321692Alex Bradbury2018-01-033-7/+8
| | | | | | | As experimental backends, I didn't have them configured to build in my local build config. llvm-svn: 321696
* Thread MCSubtargetInfo through Target::createMCAsmBackendAlex Bradbury2018-01-0325-72/+89
| | | | | | | | | | | | | | | | | | | | | Currently it's not possible to access MCSubtargetInfo from a TgtMCAsmBackend. D20830 threaded an MCSubtargetInfo reference through MCAsmBackend::relaxInstruction, but this isn't the only function that would benefit from access. This patch removes the Triple and CPUString arguments from createMCAsmBackend and replaces them with MCSubtargetInfo. This patch just changes the interface without making any intentional functional changes. Once in, several cleanups are possible: * Get rid of the awkward MCSubtargetInfo handling in ARMAsmBackend * Support 16-bit instructions when valid in MipsAsmBackend::writeNopData * Get rid of the CPU string parsing in X86AsmBackend and just use a SubtargetFeature for HasNopl * Emit 16-bit nops in RISCVAsmBackend::writeNopData if the compressed instruction set extension is enabled (see D41221) This change initially exposed PR35686, which has since been resolved in r321026. Differential Revision: https://reviews.llvm.org/D41349 llvm-svn: 321692
OpenPOWER on IntegriCloud