summaryrefslogtreecommitdiffstats
path: root/llvm/lib/Transforms
Commit message (Collapse)AuthorAgeFilesLines
* Directly return promoted direct call instead of rely on stripPointerCast.Dehao Chen2017-10-062-11/+15
| | | | | | | | | | | | | | Summary: stripPointerCast is not reliably returning the value that's being type-casted. Instead it may look further at function attributes to further propagate the value. Instead of relying on stripPOintercast, the more reliable solution is to directly use the pointer to the promoted direct call. Reviewers: tejohnson, davidxl Reviewed By: tejohnson Subscribers: llvm-commits, sanjoy Differential Revision: https://reviews.llvm.org/D38603 llvm-svn: 315077
* Revert "[MergeICmps] Disable mergeicmps if the target does not want to ↵Clement Courbet2017-10-061-36/+12
| | | | | | | | | | handle memcmp expansion." Still a few stability issues on windows. This reverts commit 67e3db9bc121ba244e20337aabc7cf341a62b545. llvm-svn: 315058
* Re-land "[MergeICmps] Disable mergeicmps if the target does not want to ↵Clement Courbet2017-10-061-12/+36
| | | | | | | | | | handle memcmp expansion." (fixed unit tests by making comparisons stable) This reverts commit 1b2d359ce256fd6737da4e93833346a0bd6d7583. llvm-svn: 315056
* Roll forward r314928Xinliang David Li2017-10-062-0/+238
| | | | | | | Fixed ThinLTO bootstrap failure : track new bitcast per incomingVal. Added new tests. llvm-svn: 315044
* [PM] Retire disable unit-at-a-time switch.Davide Italiano2017-10-061-33/+24
| | | | | | | | | | | | This is a vestige from the GCC-3 days, which disables IPO passes when set. I don't think anybody actually uses it as there are several IPO passes which still run with this flag set and nobody complained/noticed. This reduces the delta between current and new pass manager and allows us to easily review the difference when we decide to flip the switch (or audit which passes should run, FWIW). llvm-svn: 315043
* [CodeExtractor] Fix multiple bugs under certain shape of extracted regionJakub Kuderski2017-10-061-77/+31
| | | | | | | | | | | | | | | | Summary: If the extracted region has multiple exported data flows toward the same BB which is not included in the region, correct resotre instructions and PHI nodes won't be generated inside the exitStub. The solution is simply put the restore instructions right after the definition of output values instead of putting in exitStub. Unittest for this bug is included. Author: myhsu Reviewers: chandlerc, davide, lattner, silvas, davidxl, wmi, kuhar Subscribers: dberlin, kuhar, mgorny, llvm-commits Differential Revision: https://reviews.llvm.org/D37902 llvm-svn: 315041
* NewGVN: Factor out duplicate parts of OpIsSafeForPHIOfOpsDaniel Berlin2017-10-061-45/+31
| | | | llvm-svn: 315040
* ModuleUtils: Stop using comdat members to generate unique module ids.Peter Collingbourne2017-10-051-1/+1
| | | | | | | | | | | It is possible for two modules to define the same set of external symbols without causing a duplicate symbol error at link time, as long as each of the symbols is a comdat member. So we cannot use them as part of a unique id for the module. Differential Revision: https://reviews.llvm.org/D38602 llvm-svn: 315026
* [InstCombine] improve folds for icmp gt/lt (shr X, C1), C2Sanjay Patel2017-10-051-37/+40
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | We can always eliminate the shift in: icmp gt/lt (shr X, C1), C2 --> icmp gt/lt X, C' This patch was supposed to just be an efficiency improvement because we were doing this 3-step process to fold: IC: Visiting: %c = icmp ugt i4 %s, 1 IC: ADD: %s = lshr i4 %x, 1 IC: ADD: %1 = udiv i4 %x, 2 IC: Old = %c = icmp ugt i4 %1, 1 New = <badref> = icmp uge i4 %x, 4 IC: ADD: %c = icmp uge i4 %x, 4 IC: ERASE %2 = icmp ugt i4 %1, 1 IC: Visiting: %c = icmp uge i4 %x, 4 IC: Old = %c = icmp uge i4 %x, 4 New = <badref> = icmp ugt i4 %x, 3 IC: ADD: %c = icmp ugt i4 %x, 3 IC: ERASE %2 = icmp uge i4 %x, 4 IC: Visiting: %c = icmp ugt i4 %x, 3 IC: DCE: %1 = udiv i4 %x, 2 IC: ERASE %1 = udiv i4 %x, 2 IC: DCE: %s = lshr i4 %x, 1 IC: ERASE %s = lshr i4 %x, 1 IC: Visiting: ret i1 %c When we could go directly to canonical icmp form: IC: Visiting: %c = icmp ugt i4 %s, 1 IC: Old = %c = icmp ugt i4 %s, 1 New = <badref> = icmp ugt i4 %x, 3 IC: ADD: %c = icmp ugt i4 %x, 3 IC: ERASE %1 = icmp ugt i4 %s, 1 IC: ADD: %s = lshr i4 %x, 1 IC: DCE: %s = lshr i4 %x, 1 IC: ERASE %s = lshr i4 %x, 1 IC: Visiting: %c = icmp ugt i4 %x, 3 ...but then I noticed that the folds were incomplete too: https://godbolt.org/g/aB2hLE Here are attempts to prove the logic with Alive: https://rise4fun.com/Alive/92o Name: lshr_ult Pre: ((C2 << C1) u>> C1) == C2 %sh = lshr i8 %x, C1 %r = icmp ult i8 %sh, C2 => %r = icmp ult i8 %x, (C2 << C1) Name: ashr_slt Pre: ((C2 << C1) >> C1) == C2 %sh = ashr i8 %x, C1 %r = icmp slt i8 %sh, C2 => %r = icmp slt i8 %x, (C2 << C1) Name: lshr_ugt Pre: (((C2+1) << C1) u>> C1) == (C2+1) %sh = lshr i8 %x, C1 %r = icmp ugt i8 %sh, C2 => %r = icmp ugt i8 %x, ((C2+1) << C1) - 1 Name: ashr_sgt Pre: (C2 != 127) && ((C2+1) << C1 != -128) && (((C2+1) << C1) >> C1) == (C2+1) %sh = ashr i8 %x, C1 %r = icmp sgt i8 %sh, C2 => %r = icmp sgt i8 %x, ((C2+1) << C1) - 1 Name: ashr_exact_sgt Pre: ((C2 << C1) >> C1) == C2 %sh = ashr exact i8 %x, C1 %r = icmp sgt i8 %sh, C2 => %r = icmp sgt i8 %x, (C2 << C1) Name: ashr_exact_slt Pre: ((C2 << C1) >> C1) == C2 %sh = ashr exact i8 %x, C1 %r = icmp slt i8 %sh, C2 => %r = icmp slt i8 %x, (C2 << C1) Name: lshr_exact_ugt Pre: ((C2 << C1) u>> C1) == C2 %sh = lshr exact i8 %x, C1 %r = icmp ugt i8 %sh, C2 => %r = icmp ugt i8 %x, (C2 << C1) Name: lshr_exact_ult Pre: ((C2 << C1) u>> C1) == C2 %sh = lshr exact i8 %x, C1 %r = icmp ult i8 %sh, C2 => %r = icmp ult i8 %x, (C2 << C1) We did something similar for 'shl' in D28406. Differential Revision: https://reviews.llvm.org/D38514 llvm-svn: 315021
* Annotate VP prof on indirect call if it is ICPed in the profiled binary.Dehao Chen2017-10-051-1/+3
| | | | | | | | | | | | | | Summary: In SamplePGO, when an indirect call is promoted in the profiled binary, before profile annotation, it will be promoted and inlined. For the original indirect call, the current implementation will not mark VP profile on it. This is an issue when profile becomes stale. This patch annotates VP prof on indirect calls during annotation. Reviewers: tejohnson Reviewed By: tejohnson Subscribers: sanjoy, llvm-commits Differential Revision: https://reviews.llvm.org/D38477 llvm-svn: 315016
* [PassManager] Improve the interaction between -O2 and ThinLTO.Davide Italiano2017-10-051-13/+11
| | | | | | | Run GDCE slightly later so that we don't have to repeat it twice when preparing for Thin. Thanks to Mehdi for the suggestion. llvm-svn: 314999
* [PassManager] Run global optimizations after the inliner.Davide Italiano2017-10-051-0/+14
| | | | | | | | | | | | | | | The inliner performs some kind of dead code elimination as it goes, but there are cases that are not really caught by it. We might at some point consider teaching the inliner about them, but it is OK for now to run GlobalOpt + GlobalDCE in tandem as their benefits generally outweight the cost, making the whole pipeline faster. This fixes PR34652. Differential Revision: https://reviews.llvm.org/D38154 llvm-svn: 314997
* [LV] Fix PR34743 - handle casts that sink after interleaved loadsAyal Zaks2017-10-051-1/+4
| | | | | | | | | | | When ignoring a load that participates in an interleaved group, make sure to move a cast that needs to sink after it. Testcase derived from reproducer of PR34743. Differential Revision: https://reviews.llvm.org/D38338 llvm-svn: 314986
* Revert "Re-land "[MergeICmps] Disable mergeicmps if the target does not want ↵Clement Courbet2017-10-051-25/+10
| | | | | | | | | | to handle memcmp expansion.""" broken test on windows This reverts commit c91479518344fd1fc071c5bd5848f6eb83e53dca. llvm-svn: 314985
* revert r314698 - [InstCombine] remove one-use restriction for icmp (shr ↵Sanjay Patel2017-10-051-6/+6
| | | | | | | | | | | exact X, C1), C2 --> icmp X, (C2<<C1) There is a bot failure that appears to be related to this change: http://lab.llvm.org:8011/builders/clang-cmake-armv7-a15-selfhost-neon/builds/2117 ...so reverting to confirm that and attempting to keep the bot green while investigating. llvm-svn: 314984
* [LV] Fix PR34711 - widen instruction ranges when sinking castsAyal Zaks2017-10-051-23/+22
| | | | | | | | | | | | | | | | Instead of trying to keep LastWidenRecipe updated after creating each recipe, have tryToWiden() retrieve the last recipe of the current VPBasicBlock and check if it's a VPWidenRecipe when attempting to extend its range. This ensures that such extensions, optimized to maintain the original instruction order, do so only when the instructions are to maintain their relative order. The latter does not always hold, e.g., when a cast needs to sink to unravel first order recurrence (r306884). Testcase derived from reproducer of PR34711. Differential Revision: https://reviews.llvm.org/D38339 llvm-svn: 314981
* Re-land "[MergeICmps] Disable mergeicmps if the target does not want to ↵Clement Courbet2017-10-051-10/+25
| | | | | | handle memcmp expansion."" llvm-svn: 314980
* Revert "[MergeICmps] Disable mergeicmps if the target does not want to ↵Clement Courbet2017-10-051-15/+6
| | | | | | | | | | | handle memcmp expansion." Breaks clang-stage1-cmake-RA-incremental/llvm/test/Transforms/MergeICmps/X86/tuple-four-int8.ll This reverts commit 3038c459d67f8898ffa295d54a013b280690abfa. llvm-svn: 314972
* [InstCombine] Fix a vector splat handling bug in transformZExtICmp.Craig Topper2017-10-051-3/+1
| | | | | | | | | | We were using an i1 type and then zero extending to a vector. Instead just create the 0/1 directly as a ConstantInt with the correct type. No need to ask ConstantExpr to zero extend for us. This bug is a bit tricky to hit because it requires us to visit a zext of an icmp that would normally be simplified to true/false, but that icmp hasnt' been visited yet. In the test case this zext and icmp were created by visiting a udiv and due to worklist ordering we got to the zext first. Fixes PR34841. llvm-svn: 314971
* [MergeICmps] Disable mergeicmps if the target does not want to handle memcmp ↵Clement Courbet2017-10-051-6/+15
| | | | | | | | | | | | | | expansion. Summary: This is to avoid e.g. merging two cheap icmps if the target is not going to expand to something nice later. Reviewers: dberlin, spatel Subscribers: davide, nemanjai Differential Revision: https://reviews.llvm.org/D38232 llvm-svn: 314970
* Revert r314928 to investigate thinLTO bootstrap failureXinliang David Li2017-10-052-234/+0
| | | | llvm-svn: 314961
* [InstCombine] Improve support for ashr in foldICmpAndShiftCraig Topper2017-10-041-9/+12
| | | | | | | | We can support ashr similar to lshr, if we know that none of the shifted in bits are used. In that case SimplifyDemandedBits would normally convert it to lshr. But that conversion doesn't happen if the shift has additional users. Differential Revision: https://reviews.llvm.org/D38521 llvm-svn: 314945
* Fix a -Wparentheses warning. NFC.Hans Wennborg2017-10-041-1/+1
| | | | llvm-svn: 314936
* [LoopDeletion] Move deleteDeadLoop to to LoopUtils. NFCMarcello Maggioni2017-10-042-134/+124
| | | | llvm-svn: 314934
* [SimplifyCFG] put the optional assumption cache pointer in the options ↵Sanjay Patel2017-10-042-52/+45
| | | | | | | | | | | | struct; NFCI This is a follow-up to https://reviews.llvm.org/D38138. I fixed the capitalization of some functions because we're changing those lines anyway and that helped verify that we weren't accidentally dropping any options by using default param values. llvm-svn: 314930
* Recommit r314561 after fixing msan build failureXinliang David Li2017-10-042-0/+234
| | | | | | | (trial 2) Incoming val defined by terminator instruction which also requires bitcasts can not be handled. llvm-svn: 314928
* Recommit : Use the basic cost if a GEP is not used as addressing modeJun Bum Lim2017-10-042-2/+2
| | | | | | | | | | | | | | Recommitting r314517 with the fix for handling ConstantExpr. Original commit message: Currently, getGEPCost() returns TCC_FREE whenever a GEP is a legal addressing mode in the target. However, since it doesn't check its actual users, it will return FREE even in cases where the GEP cannot be folded away as a part of actual addressing mode. For example, if an user of the GEP is a call instruction taking the GEP as a parameter, then the GEP may not be folded in isel. llvm-svn: 314923
* [NFC] clang-format lib/Transforms/Scalar/MergeICmps.cppClement Courbet2017-10-041-39/+21
| | | | llvm-svn: 314906
* [IRCE] Temporarily disable unsigned latch conditions by defaultMax Kazantsev2017-10-041-0/+21
| | | | | | | | | | | | We have found some corner cases connected to range intersection where IRCE makes a bad thing when the latch condition is unsigned. The fix for that will go as a follow up. This patch temporarily disables IRCE for unsigned latch conditions until the issue is fixed. The unsigned latch conditions were introduced to IRCE by rL310027. Differential Revision: https://reviews.llvm.org/D38529 llvm-svn: 314881
* [InstCombine] Use isSignBitCheck to simplify an if statement. Directly ↵Craig Topper2017-10-031-12/+8
| | | | | | | | create new sign bit compares instead of manipulating the constant. NFCI Since we no longer had the direct constant compares, manipulating the constant seemeded less clear. llvm-svn: 314830
* Revert r314806 "[SLP] Vectorize jumbled memory loads."Hans Wennborg2017-10-031-185/+84
| | | | | | | | | | | | | | | | | | | | | | | | | All the buildbots are red, e.g. http://lab.llvm.org:8011/builders/clang-cmake-aarch64-lld/builds/2436/ > Summary: > This patch tries to vectorize loads of consecutive memory accesses, accessed > in non-consecutive or jumbled way. An earlier attempt was made with patch D26905 > which was reverted back due to some basic issue with representing the 'use mask' of > jumbled accesses. > > This patch fixes the mask representation by recording the 'use mask' in the usertree entry. > > Change-Id: I9fe7f5045f065d84c126fa307ef6ebe0787296df > > Reviewers: mkuper, loladiro, Ayal, zvi, danielcdh > > Reviewed By: Ayal > > Subscribers: hans, mzolotukhin > > Differential Revision: https://reviews.llvm.org/D36130 llvm-svn: 314824
* Revert the change that accidentally went in r314806.Dehao Chen2017-10-031-4/+0
| | | | llvm-svn: 314807
* [SLP] Vectorize jumbled memory loads.Mohammad Shahid2017-10-031-84/+185
| | | | | | | | | | | | | | | | | | | | | | Summary: This patch tries to vectorize loads of consecutive memory accesses, accessed in non-consecutive or jumbled way. An earlier attempt was made with patch D26905 which was reverted back due to some basic issue with representing the 'use mask' of jumbled accesses. This patch fixes the mask representation by recording the 'use mask' in the usertree entry. Change-Id: I9fe7f5045f065d84c126fa307ef6ebe0787296df Reviewers: mkuper, loladiro, Ayal, zvi, danielcdh Reviewed By: Ayal Subscribers: hans, mzolotukhin Differential Revision: https://reviews.llvm.org/D36130 llvm-svn: 314806
* [InstCombine] Change a bunch of methods to take APInts by reference instead ↵Craig Topper2017-10-032-134/+134
| | | | | | | | of pointer. This allows us to remove a bunch of dereferences and only have a few dereferences at the call sites. llvm-svn: 314762
* [InstCombine] Replace an equality compare of two APInt pointers with a ↵Craig Topper2017-10-031-1/+1
| | | | | | | | compare of the APInts themselves. Apparently this works by virtue of the fact that the pointers are pointers to the APInts stored inside of the ConstantInt objects. But I really don't think we should be relying on that. llvm-svn: 314761
* [PassManager] Retire cl::opt that have been set for a while. NFCI.Davide Italiano2017-10-021-33/+21
| | | | llvm-svn: 314740
* [InstCombine] remove one-use restriction for icmp (shr exact X, C1), C2 --> ↵Sanjay Patel2017-10-021-6/+6
| | | | | | icmp X, (C2<<C1) llvm-svn: 314698
* Update getMergedLocation to check the instruction type and merge properly.Dehao Chen2017-10-025-30/+31
| | | | | | | | | | | | | | Summary: If the merged instruction is call instruction, we need to set the scope to the closes common scope between 2 locations, otherwise it will cause trouble when the call is getting inlined. Reviewers: dblaikie, aprantl Reviewed By: dblaikie, aprantl Subscribers: llvm-commits, sanjoy Differential Revision: https://reviews.llvm.org/D37877 llvm-svn: 314694
* [InstCombine] Use APInt for all the math in foldICmpDivConstantCraig Topper2017-10-011-95/+46
| | | | | | | | | | | | | | Summary: This currently uses ConstantExpr to do its math, but as noted in a TODO it can all be done directly on APInt. Reviewers: spatel, majnemer Reviewed By: majnemer Subscribers: llvm-commits Differential Revision: https://reviews.llvm.org/D38440 llvm-svn: 314640
* Revert r314579: "Recommi r314561 after fixing over-debug assertion".Daniel Jasper2017-10-012-225/+0
| | | | | | | | | | | | And follow-up r314585. Leads to segfaults. I'll forward reproduction instructions to the patch author. Also, for a recommit, still add the original patch description. Otherwise, it becomes really tedious to find out what a patch actually does. The fact that it is a recommit with a fix is somewhat secondary. llvm-svn: 314622
* Separate the logic when handling indirect calls in SamplePGO ThinLTO compile ↵Dehao Chen2017-10-011-12/+26
| | | | | | | | | | | | | | | | phase and other phases. Summary: In SamplePGO ThinLTO compile phase, we will not invoke ICP as it may introduce confusion to the 2nd annotation. This patch extracted that logic and makes it clearer before profile annotation. In the mean time, we need to make function importing process both inlined callsites as well as not promoted indirect callsites. Reviewers: tejohnson Reviewed By: tejohnson Subscribers: sanjoy, mehdi_amini, llvm-commits, inglorion Differential Revision: https://reviews.llvm.org/D38094 llvm-svn: 314619
* NewGVN: Fix PR 34473, by not using ExactlyEqualsExpression for findingDaniel Berlin2017-09-301-6/+6
| | | | | | phi of ops users. llvm-svn: 314612
* NewGVN: Evaluate phi of ops expressions before creating phi nodeDaniel Berlin2017-09-301-48/+72
| | | | llvm-svn: 314611
* NewGVN: Allow dependent PHI of opsDaniel Berlin2017-09-301-57/+100
| | | | llvm-svn: 314610
* NewGVN: Make OpIsSafeForPhiOfOps non-recursiveDaniel Berlin2017-09-301-7/+38
| | | | llvm-svn: 314609
* Refactor the SamplePGO profile annotation logic to extract ↵Dehao Chen2017-09-301-58/+65
| | | | | | inlineCallInstruction. (NFC) llvm-svn: 314601
* Revert r314435: "[JumpThreading] Preserve DT and LVI across the pass"Daniel Jasper2017-09-303-253/+65
| | | | | | | Causes a segfault on a builtbot (and in our internal bootstrapping of Clang). See Eli's response on the commit thread. llvm-svn: 314589
* Fix buildbot failure -- tighten type check for matching phiXinliang David Li2017-09-301-1/+1
| | | | llvm-svn: 314585
* Recommi r314561 after fixing over-debug assertionXinliang David Li2017-09-302-0/+225
| | | | llvm-svn: 314579
* Revert 314561 due to debug build assertion failureXinliang David Li2017-09-292-223/+0
| | | | llvm-svn: 314563
OpenPOWER on IntegriCloud