aboutsummaryrefslogtreecommitdiffstats
path: root/lib/Transforms
Commit message (Collapse)AuthorAgeFilesLines
* Fix Android.mk files for plugins on darwin.Stephen Hines2015-06-081-0/+4
| | | | | | | | Darwin uses a few different flags for this than Linux, and we missed this when we updated the test builds recently. Change-Id: I8a0e8bc18ef698a0ed933f96b2bce0653f4307e5 (cherry picked from commit 651de8d5ac0b28b10ad0259efc75cf6152dc13dc)
* Update aosp/master LLVM for rebase to r235153Pirama Arumuga Nainar2015-05-1857-1262/+2492
| | | | | Change-Id: I9bf53792f9fc30570e81a8d80d296c681d005ea7 (cherry picked from commit 0c7f116bb6950ef819323d855415b2f2b0aad987)
* Export all dynamic symbols for opt and bugpointPirama Arumuga Nainar2015-04-301-5/+0
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | bug 15433215 Command line options from an LLVM module are registered and stored in a static global object in libLLVMSupport. Linking the Hello module with its own copy of libLLVMSupport meant this module's option was not registered correctly. This patch updates Hello's build rules to not link with libLLVMSupport and libLLVMCore. Given the above change, the Hello module still needs to find LLVM symbols that it used to pull directly. The second half of this patch adds linker flag --export-dynamic to tools/opt and tools/bugpoint so loaded modules can find LLVM symbols. Other notes: - The linker flag is not added to the device build of opt. No one should be running opt on the device, AFAIK. - Other tools that load modules, such as llc, lli and even clang, would need this linker flag if they were to load modules correctly. Now that we know how to fix the incorrect module loads, they can be fixed if the need arises. - Remove a duplicate from opt's llvm_opt_STATIC_LIBRARIES. All supported tests pass! Cherry pick of https://android-review.googlesource.com/#/c/148017/ from AOSP Change-Id: Ib76507ccfeae37a368718c5a1fb69da337f4f964
* Update aosp/master llvm for rebase to r233350Pirama Arumuga Nainar2015-04-09105-4297/+6352
| | | | Change-Id: I07d935f8793ee8ec6b7da003f6483046594bca49
* Update aosp/master LLVM for rebase to r230699.Stephen Hines2015-03-23140-8113/+17081
| | | | Change-Id: I2b5be30509658cb8266be782de0ab24f9099f9b9
* Update LLVM lit test configuration to be able to work for 3.6.Stephen Hines2014-12-191-0/+24
| | | | | | | | | | | Add support for building more host verification tools: + LLVMHello.so (loadable plugin pass) + lli-child-target + llvm-symbolizer (missing some Android.mk magic) + llvm-vtabledump + verify-uselistorder Change-Id: Id7a9397bf458e97fc330ea0301ebb001750aac7c
* Update aosp/master LLVM for rebase to r222494.Stephen Hines2014-12-02119-5741/+11294
| | | | Change-Id: Ic787f5e0124df789bd26f3f24680f45e678eef2d
* Update LLVM for rebase to r212749.Stephen Hines2014-07-2156-2016/+3592
| | | | | | | Includes a cherry-pick of: r212948 - fixes a small issue with atomic calls Change-Id: Ib97bd980b59f18142a69506400911a6009d9df18
* Update LLVM for 3.5 rebase (r209712).Stephen Hines2014-05-29128-3110/+5955
| | | | Change-Id: I149556c940fb7dc92d075273c87ff584f400941f
* Update to LLVM 3.5a.Stephen Hines2014-04-24139-7674/+9469
| | | | Change-Id: Ifadecab779f128e62e430c2b4f6ddd84953ed617
* Remove the -O1 WAR for SROA.cpp compilation.Stephen Hines2014-04-171-5/+0
| | | | | Bug: 8047767 Change-Id: I036af8d0578f723d3f5cb07f1d26b3c6895d8690
* am b7485134: am 449fc261: Merge "llvm: convert makefiles to support multilib ↵Colin Cross2014-02-181-3/+1
|\ | | | | | | | | | | | | build" * commit 'b7485134a2cbecc47904988b4cfde24019ac4fa1': llvm: convert makefiles to support multilib build
| * llvm: convert makefiles to support multilib buildColin Cross2014-02-061-3/+1
| | | | | | | | | | | | | | | | Convert makefiles to allow for building two architectures at the same time. This will also cause make checkbuild to build the target libraries for all supported architectures. Change-Id: Ia5e6fe5b1186a67753faafd3532ed4cb280a8b10
* | Update LLVM for merge to 3.4.Stephen Hines2014-02-133-5/+6
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Update config.h files. Add RS SubtargetFeature for +long64 on ARM devices. Adjust Android.mk for added/removed files: + Delinearization.cpp - PathNumbering.cpp - PathProfileInfo.cpp - PathProfileVerifier.cpp - ProfileDataLoader.cpp - ProfileDataLoaderPass.cpp - ProfileEstimatorPass.cpp - ProfileInfo.cpp - ProfileInfoLoader.cpp - ProfileInfoLoaderPass.cpp - ProfileVerifierPass.cpp + LiveRegUnits.cpp - ShrinkWrapping.cpp + StackMaps.cpp - StrongPHIElimination.cpp + DIEHash.cpp + LegacyPassManager.cpp + ELF.cpp + Unicode.cpp - MipsOptimizeMathLibCalls.cpp - MipsELFStreamer.cpp + MipsTargetStreamer.cpp - EdgeProfiling.cpp + DataFlowSanitizer.cpp + DebugIR.cpp - OptimalEdgeProfiling.cpp - PathProfiling.cpp - ProfilingUtils.cpp - BasicBlockPlacement.cpp + LoopRerollPass.cpp + PartiallyInlineLibCalls.cpp + SampleProfile.cpp + GlobalStatus.cpp Change-Id: I17dcf0bf53a1720acd8226ae3e30d84993562a91
* | Merge remote-tracking branch 'upstream/release_34' into merge-20140211Stephen Hines2014-02-1185-4577/+7933
|\ \ | |/ |/| | | | | | | | | | | Conflicts: lib/Linker/LinkModules.cpp lib/Support/Unix/Signals.inc Change-Id: Ia54f291fa5dc828052d2412736e8495c1282aa64
| * Merging r197449:Bill Wendling2013-12-171-1/+31
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | ------------------------------------------------------------------------ r197449 | arnolds | 2013-12-16 17:11:01 -0800 (Mon, 16 Dec 2013) | 7 lines LoopVectorizer: Don't if-convert constant expressions that can trap A phi node operand or an instruction operand could be a constant expression that can trap (division). Check that we don't vectorize such cases. PR16729 radar://15653590 ------------------------------------------------------------------------ git-svn-id: https://llvm.org/svn/llvm-project/llvm/branches/release_34@197453 91177308-0d34-0410-b5e6-96231b3b80d8
| * Merging r197178:Bill Wendling2013-12-141-2/+11
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | ------------------------------------------------------------------------ r197178 | hfinkel | 2013-12-12 12:45:24 -0800 (Thu, 12 Dec 2013) | 9 lines Fix a use-after-free error in GlobalOpt CleanupConstantGlobalUsers GlobalOpt's CleanupConstantGlobalUsers function uses a worklist array to manage constant users to be visited. The pointers in this array need to be weak handles because when we delete a constant array, we may also be holding a pointer to one of its elements (or an element of one of its elements if we're dealing with an array of arrays) in the worklist. Fixes PR17347. ------------------------------------------------------------------------ git-svn-id: https://llvm.org/svn/llvm-project/llvm/branches/release_34@197322 91177308-0d34-0410-b5e6-96231b3b80d8
| * Merging r195505:Manman Ren2013-12-091-51/+0
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | ------------------------------------------------------------------------ r195505 | mren | 2013-11-22 14:06:31 -0800 (Fri, 22 Nov 2013) | 8 lines Debug Info: move StripDebugInfo from StripSymbols.cpp to DebugInfo.cpp. We can share the implementation between StripSymbols and dropping debug info for metadata versions that do not match. Also update the comments to match the implementation. A follow-on patch will drop the "Debug Info Version" module flag in StripDebugInfo. ------------------------------------------------------------------------ git-svn-id: https://llvm.org/svn/llvm-project/llvm/branches/release_34@196816 91177308-0d34-0410-b5e6-96231b3b80d8
| * Merging r196638:Bill Wendling2013-12-081-3/+3
| | | | | | | | | | | | | | | | | | | | | | ------------------------------------------------------------------------ r196638 | arsenm | 2013-12-06 18:58:45 -0800 (Fri, 06 Dec 2013) | 1 line Fix assert with copy from global through addrspacecast ------------------------------------------------------------------------ git-svn-id: https://llvm.org/svn/llvm-project/llvm/branches/release_34@196709 91177308-0d34-0410-b5e6-96231b3b80d8
| * --- Reverse-merging r196668 into '.':Bill Wendling2013-12-081-3/+3
| | | | | | | | | | | | | | | | | | U lib/Transforms/InstCombine/InstCombineLoadStoreAlloca.cpp U test/Transforms/InstCombine/addrspacecast.ll git-svn-id: https://llvm.org/svn/llvm-project/llvm/branches/release_34@196705 91177308-0d34-0410-b5e6-96231b3b80d8
| * Merging r196638:Bill Wendling2013-12-071-3/+3
| | | | | | | | | | | | | | | | | | | | | | ------------------------------------------------------------------------ r196638 | arsenm | 2013-12-06 18:58:45 -0800 (Fri, 06 Dec 2013) | 1 line Fix assert with copy from global through addrspacecast ------------------------------------------------------------------------ git-svn-id: https://llvm.org/svn/llvm-project/llvm/branches/release_34@196668 91177308-0d34-0410-b5e6-96231b3b80d8
| * Merging r196611:Bill Wendling2013-12-061-1/+4
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | ------------------------------------------------------------------------ r196611 | dexonsmith | 2013-12-06 13:48:36 -0800 (Fri, 06 Dec 2013) | 5 lines Don't use isNullValue to evaluate ConstantExpr ConstantExpr can evaluate to false even when isNullValue gives false. Fixes PR18143. ------------------------------------------------------------------------ git-svn-id: https://llvm.org/svn/llvm-project/llvm/branches/release_34@196614 91177308-0d34-0410-b5e6-96231b3b80d8
| * Merging r196508:Bill Wendling2013-12-061-5/+1
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | ------------------------------------------------------------------------ r196508 | arnolds | 2013-12-05 07:14:40 -0800 (Thu, 05 Dec 2013) | 12 lines SLPVectorizer: An in-tree vectorized entry cannot also be a scalar external use We were creating external uses for scalar values in MustGather entries that also had a ScalarToTreeEntry (they also are present in a vectorized tuple). This meant we would keep a value 'alive' as a scalar and vectorized causing havoc. This is not necessary because when we create a MustGather vector we explicitly create external uses entries for the insertelement instructions of the MustGather vector elements. Fixes PR18129. radar://15582184 ------------------------------------------------------------------------ git-svn-id: https://llvm.org/svn/llvm-project/llvm/branches/release_34@196571 91177308-0d34-0410-b5e6-96231b3b80d8
| * Merging r196129:Bill Wendling2013-12-021-4/+12
| | | | | | | | | | | | | | | | | | | | | | ------------------------------------------------------------------------ r196129 | kkhoo | 2013-12-02 10:43:59 -0800 (Mon, 02 Dec 2013) | 1 line Conservative fix for PR17827 - don't optimize a shift + and + compare sequence where the shift is logical unless the comparison is unsigned ------------------------------------------------------------------------ git-svn-id: https://llvm.org/svn/llvm-project/llvm/branches/release_34@196132 91177308-0d34-0410-b5e6-96231b3b80d8
| * Merging r196004:Bill Wendling2013-12-011-3/+4
| | | | | | | | | | | | | | | | | | | | | | | | | | ------------------------------------------------------------------------ r196004 | void | 2013-11-30 19:36:07 -0800 (Sat, 30 Nov 2013) | 3 lines Use 'unsigned char' to get this past gcc error message: error: invalid conversion from 'unsigned char' to '{anonymous}::Sequence' ------------------------------------------------------------------------ git-svn-id: https://llvm.org/svn/llvm-project/llvm/branches/release_34@196005 91177308-0d34-0410-b5e6-96231b3b80d8
| * Merging r195787:Bill Wendling2013-12-011-0/+9
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | ------------------------------------------------------------------------ r195787 | arnolds | 2013-11-26 14:11:23 -0800 (Tue, 26 Nov 2013) | 8 lines LoopVectorizer: Truncate i64 trip counts of i32 phis if necessary In signed arithmetic we could end up with an i64 trip count for an i32 phi. Because it is signed arithmetic we know that this is only defined if the i32 does not wrap. It is therefore safe to truncate the i64 trip count to a i32 value. Fixes PR18049. ------------------------------------------------------------------------ git-svn-id: https://llvm.org/svn/llvm-project/llvm/branches/release_34@195991 91177308-0d34-0410-b5e6-96231b3b80d8
| * Merging r195791:Bill Wendling2013-11-271-16/+11
| | | | | | | | | | | | | | | | | | | | | | | | | | | | ------------------------------------------------------------------------ r195791 | nadav | 2013-11-26 14:24:25 -0800 (Tue, 26 Nov 2013) | 4 lines PR1860 - We can't save a list of ExtractElement instructions to CSE because some of these instructions may be removed and optimized in future iterations. Instead we save a list of basic blocks that we need to CSE. ------------------------------------------------------------------------ git-svn-id: https://llvm.org/svn/llvm-project/llvm/branches/release_34@195818 91177308-0d34-0410-b5e6-96231b3b80d8
| * Merging r195773:Bill Wendling2013-11-271-0/+8
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | ------------------------------------------------------------------------ r195773 | nadav | 2013-11-26 09:29:19 -0800 (Tue, 26 Nov 2013) | 6 lines PR18060 - When we RAUW values with ExtractElement instructions in some cases we generate PHI nodes with multiple entries from the same basic block but with different values. Enabling CSE on ExtractElement instructions make sure that all of the RAUWed instructions are the same. ------------------------------------------------------------------------ git-svn-id: https://llvm.org/svn/llvm-project/llvm/branches/release_34@195817 91177308-0d34-0410-b5e6-96231b3b80d8
| * Merging r195769:Bill Wendling2013-11-271-11/+11
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | ------------------------------------------------------------------------ r195769 | dyatkovskiy | 2013-11-26 08:11:03 -0800 (Tue, 26 Nov 2013) | 27 lines PR17925 bugfix. Short description. This issue is about case of treating pointers as integers. We treat pointers as different if they references different address space. At the same time, we treat pointers equal to integers (with machine address width). It was a point of false-positive. Consider next case on 32bit machine: void foo0(i32 addrespace(1)* %p) void foo1(i32 addrespace(2)* %p) void foo2(i32 %p) foo0 != foo1, while foo1 == foo2 and foo0 == foo2. As you can see it breaks transitivity. That means that result depends on order of how functions are presented in module. Next order causes merging of foo0 and foo1: foo2, foo0, foo1 First foo0 will be merged with foo2, foo0 will be erased. Second foo1 will be merged with foo2. Depending on order, things could be merged we don't expect to. The fix: Forbid to treat any pointer as integer, except for those, who belong to address space 0. ------------------------------------------------------------------------ git-svn-id: https://llvm.org/svn/llvm-project/llvm/branches/release_34@195810 91177308-0d34-0410-b5e6-96231b3b80d8
| * Merging r195528:Bill Wendling2013-11-251-2/+37
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | ------------------------------------------------------------------------ r195528 | chandlerc | 2013-11-22 16:48:34 -0800 (Fri, 22 Nov 2013) | 7 lines Migrate metadata information from scalar to vector instructions during SLP vectorization. Based on the code in BBVectorizer. Fixes PR17741. Patch by Raul Silvera, reviewed by Hal and Nadav. Reformatted by my driving of clang-format. =] ------------------------------------------------------------------------ git-svn-id: https://llvm.org/svn/llvm-project/llvm/branches/release_34@195608 91177308-0d34-0410-b5e6-96231b3b80d8
| * Merging r195493:Bill Wendling2013-11-251-0/+14
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | ------------------------------------------------------------------------ r195493 | arsenm | 2013-11-22 11:24:39 -0800 (Fri, 22 Nov 2013) | 6 lines StructurizeCFG: Fix verification failure with some loops. If the beginning of the loop was also the entry block of the function, branches were inserted to the entry block which isn't allowed. If this occurs, create a new dummy function entry block that branches to the start of the loop. ------------------------------------------------------------------------ git-svn-id: https://llvm.org/svn/llvm-project/llvm/branches/release_34@195606 91177308-0d34-0410-b5e6-96231b3b80d8
| * Merging r195492:Bill Wendling2013-11-251-11/+22
| | | | | | | | | | | | | | | | | | | | | | ------------------------------------------------------------------------ r195492 | arsenm | 2013-11-22 11:24:37 -0800 (Fri, 22 Nov 2013) | 1 line StructurizeCFG: Fix inverting a branch on an argument ------------------------------------------------------------------------ git-svn-id: https://llvm.org/svn/llvm-project/llvm/branches/release_34@195605 91177308-0d34-0410-b5e6-96231b3b80d8
| * Merging r195477:Bill Wendling2013-11-251-17/+37
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | ------------------------------------------------------------------------ r195477 | rafael | 2013-11-22 09:58:12 -0800 (Fri, 22 Nov 2013) | 13 lines Add a fixed version of r195470 back. The fix is simply to use CurI instead of I when handling aliases to avoid accessing a invalid iterator. original message: Convert linkonce* to weak* instead of strong. Also refactor the logic into a helper function. This is an important improve on mingw where the linker complains about mixed weak and strong symbols. Converting to weak ensures that the symbol is not dropped, but keeps in a comdat, making the linker happy. ------------------------------------------------------------------------ git-svn-id: https://llvm.org/svn/llvm-project/llvm/branches/release_34@195603 91177308-0d34-0410-b5e6-96231b3b80d8
| * Merging r195162:Bill Wendling2013-11-201-0/+30
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | ------------------------------------------------------------------------ r195162 | arnolds | 2013-11-19 14:20:20 -0800 (Tue, 19 Nov 2013) | 12 lines SLPVectorizer: Fix stale for Value pointer array We are slicing an array of Value pointers and process those slices in a loop. The problem is that we might invalidate a later slice by vectorizing a former slice. Use a WeakVH to track the pointer. If the pointer is deleted or RAUW'ed we can tell. The test case will only fail when running with libgmalloc. radar://15498655 ------------------------------------------------------------------------ git-svn-id: https://llvm.org/svn/llvm-project/llvm/branches/release_34@195222 91177308-0d34-0410-b5e6-96231b3b80d8
| * Merging r195161:Bill Wendling2013-11-201-11/+12
| | | | | | | | | | | | | | | | | | | | | | ------------------------------------------------------------------------ r195161 | arnolds | 2013-11-19 14:20:18 -0800 (Tue, 19 Nov 2013) | 1 line SLPVectorizer: Fix whitespace errors ------------------------------------------------------------------------ git-svn-id: https://llvm.org/svn/llvm-project/llvm/branches/release_34@195221 91177308-0d34-0410-b5e6-96231b3b80d8
| * Merging r195118:Bill Wendling2013-11-201-10/+19
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | ------------------------------------------------------------------------ r195118 | chandlerc | 2013-11-19 01:03:18 -0800 (Tue, 19 Nov 2013) | 22 lines Fix an issue where SROA computed different results based on the relative order of slices of the alloca which have exactly the same size and other properties. This was found by a perniciously unstable sort implementation used to flush out buggy uses of the algorithm. The fundamental idea is that findCommonType should return the best common type it can find across all of the slices in the range. There were two bugs here previously: 1) We would accept an integer type smaller than a byte-width multiple, and if there were different bit-width integer types, we would accept the first one. This caused an actual failure in the testcase updated here when the sort order changed. 2) If we found a bad combination of types or a non-load, non-store use before an integer typed load or store we would bail, but if we found the integere typed load or store, we would use it. The correct behavior is to always use an integer typed operation which covers the partition if one exists. While a clever debugging sort algorithm found problem #1 in our existing test cases, I have no useful test case ideas for #2. I spotted in by inspection when looking at this code. ------------------------------------------------------------------------ git-svn-id: https://llvm.org/svn/llvm-project/llvm/branches/release_34@195217 91177308-0d34-0410-b5e6-96231b3b80d8
| * Debug info: Let LowerDbgDeclare perfom the dbg.declare -> dbg.valueAdrian Prantl2013-11-181-1/+5
| | | | | | | | | | | | | | | | | | | | lowering only for load/stores to scalar allocas. The resulting values confuse the backend and don't add anything because we can describe array-allocas with a dbg.declare intrinsic just fine. rdar://problem/15464571 git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@195052 91177308-0d34-0410-b5e6-96231b3b80d8
| * [ASan] Fix PR17867 - make sure ASan doesn't crash if use-after-scope and ↵Alexey Samsonov2013-11-181-6/+6
| | | | | | | | | | | | use-after-return are combined. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@195014 91177308-0d34-0410-b5e6-96231b3b80d8
| * LoopVectorizer: Extend the induction variable to a larger typeArnold Schwaighofer2013-11-181-0/+6
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | In some case the loop exit count computation can overflow. Extend the type to prevent most of those cases. The problem is loops like: int main () { int a = 1; char b = 0; lbl: a &= 4; b--; if (b) goto lbl; return a; } The backedge count is 255. The induction variable type is i8. If we add one to 255 to get the exit count we overflow to zero. To work around this issue we extend the type of the induction variable to i32 in the case of i8 and i16. PR17532 git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@195008 91177308-0d34-0410-b5e6-96231b3b80d8
| * Utils/LoopUnroll.cpp: Tweak (StringRef)OldName to be valid until it is used, ↵NAKAMURA Takumi2013-11-171-1/+3
| | | | | | | | | | | | | | | | since r194601. eraseFromParent() invalidates OldName. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@194970 91177308-0d34-0410-b5e6-96231b3b80d8
| * Add a loop rerolling flag to the PassManagerBuilderHal Finkel2013-11-171-1/+2
| | | | | | | | | | | | | | | | | | This adds a boolean member variable to the PassManagerBuilder to control loop rerolling (just like we have for unrolling and the various vectorization options). This is necessary for control by the frontend. Loop rerolling remains disabled by default at all optimization levels. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@194966 91177308-0d34-0410-b5e6-96231b3b80d8
| * Add the cold attribute to error-reporting call sitesHal Finkel2013-11-171-0/+72
| | | | | | | | | | | | | | | | | | | | | | Generally speaking, control flow paths with error reporting calls are cold. So far, error reporting calls are calls to perror and calls to fprintf, fwrite, etc. with stderr as the stream. This can be extended in the future. The primary motivation is to improve block placement (the cold attribute affects the static branch prediction heuristics). git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@194943 91177308-0d34-0410-b5e6-96231b3b80d8
| * Fix ndebug-build unused variable in loop rerollingHal Finkel2013-11-171-1/+1
| | | | | | | | git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@194941 91177308-0d34-0410-b5e6-96231b3b80d8
| * Add a loop rerolling passHal Finkel2013-11-164-0/+1196
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | This adds a loop rerolling pass: the opposite of (partial) loop unrolling. The transformation aims to take loops like this: for (int i = 0; i < 3200; i += 5) { a[i] += alpha * b[i]; a[i + 1] += alpha * b[i + 1]; a[i + 2] += alpha * b[i + 2]; a[i + 3] += alpha * b[i + 3]; a[i + 4] += alpha * b[i + 4]; } and turn them into this: for (int i = 0; i < 3200; ++i) { a[i] += alpha * b[i]; } and loops like this: for (int i = 0; i < 500; ++i) { x[3*i] = foo(0); x[3*i+1] = foo(0); x[3*i+2] = foo(0); } and turn them into this: for (int i = 0; i < 1500; ++i) { x[i] = foo(0); } There are two motivations for this transformation: 1. Code-size reduction (especially relevant, obviously, when compiling for code size). 2. Providing greater choice to the loop vectorizer (and generic unroller) to choose the unrolling factor (and a better ability to vectorize). The loop vectorizer can take vector lengths and register pressure into account when choosing an unrolling factor, for example, and a pre-unrolled loop limits that choice. This is especially problematic if the manual unrolling was optimized for a machine different from the current target. The current implementation is limited to single basic-block loops only. The rerolling recognition should work regardless of how the loop iterations are intermixed within the loop body (subject to dependency and side-effect constraints), but the significant restriction is that the order of the instructions in each iteration must be identical. This seems sufficient to capture all current use cases. This pass is not currently enabled by default at any optimization level. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@194939 91177308-0d34-0410-b5e6-96231b3b80d8
| * Apply the InstCombine fptrunc sqrt optimization to llvm.sqrtHal Finkel2013-11-161-6/+11
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | InstCombine, in visitFPTrunc, applies the following optimization to sqrt calls: (fptrunc (sqrt (fpext x))) -> (sqrtf x) but does not apply the same optimization to llvm.sqrt. This is a problem because, to enable vectorization, Clang generates llvm.sqrt instead of sqrt in fast-math mode, and because this optimization is being applied to sqrt and not applied to llvm.sqrt, sometimes the fast-math code is slower. This change makes InstCombine apply this optimization to llvm.sqrt as well. This fixes the specific problem in PR17758, although the same underlying issue (optimizations applied to libcalls are not applied to intrinsics) exists for other optimizations in SimplifyLibCalls. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@194935 91177308-0d34-0410-b5e6-96231b3b80d8
| * InstCombine: fold (A >> C) == (B >> C) --> (A^B) < (1 << C) for constant Cs.Benjamin Kramer2013-11-161-0/+18
| | | | | | | | | | | | This is common in bitfield code. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@194925 91177308-0d34-0410-b5e6-96231b3b80d8
| * LoopVectorizer: Use abi alignment for accesses with no alignmentArnold Schwaighofer2013-11-151-0/+4
| | | | | | | | | | | | | | | | | | | | | | When we vectorize a scalar access with no alignment specified, we have to set the target's abi alignment of the scalar access on the vectorized access. Using the same alignment of zero would be wrong because most targets will have a bigger abi alignment for vector types. This probably fixes PR17878. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@194876 91177308-0d34-0410-b5e6-96231b3b80d8
| * ArgumentPromotion: correctly transfer TBAA tags and alignments.Manman Ren2013-11-151-3/+5
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | We used to use std::map<IndicesVector, LoadInst*> for OriginalLoads, and when we try to promote two arguments, they will both write to OriginalLoads causing created loads for the two arguments to have the same original load. And the same tbaa tag and alignment will be put to the created loads for the two arguments. The fix is to use std::map<std::pair<Argument*, IndicesVector>, LoadInst*> for OriginalLoads, so each Argument will write to different parts of the map. PR17906 git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@194846 91177308-0d34-0410-b5e6-96231b3b80d8
| * [asan] use GlobalValue::PrivateLinkage for coverage guard to save quite a ↵Kostya Serebryany2013-11-151-1/+1
| | | | | | | | | | | | bit of code size git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@194800 91177308-0d34-0410-b5e6-96231b3b80d8
| * Reapply "[asan] Poor man's coverage that works with ASan"Bob Wilson2013-11-151-0/+52
| | | | | | | | | | | | | | | | I was able to successfully run a bootstrapped LTO build of clang with r194701, so this change does not seem to be the cause of our failing buildbots. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@194789 91177308-0d34-0410-b5e6-96231b3b80d8
generated by cgit v1.1 at 2024-05-17 11:15:19 +0000