summaryrefslogtreecommitdiffstats
path: root/llvm/lib/Transforms
Commit message (Collapse)AuthorAgeFilesLines
* [objc-arc] Track if we encountered an additive overflow while computing ↵Michael Gottesman2013-08-091-8/+53
| | | | | | | | | | | {TopDown,BottomUp}PathCounts and do nothing if it occurred. I fixed the aforementioned problems that came up on some of the linux boxes. Major thanks to Nick Lewycky for his help debugging! rdar://14590914 llvm-svn: 188122
* Kill some duplicated code for removing unreachable BBs.Peter Collingbourne2013-08-092-172/+140
| | | | | | | | | | | This moves removeUnreachableBlocksFromFn from SimplifyCFGPass.cpp to Utils/Local.cpp and uses it to replace the implementation of llvm::removeUnreachableBlocks, which appears to do a strict subset of what removeUnreachableBlocksFromFn does. Differential Revision: http://llvm-reviews.chandlerc.com/D1334 llvm-svn: 188119
* DataFlowSanitizer: Remove unreachable BBs so IR continues to verifyPeter Collingbourne2013-08-091-0/+3
| | | | | | | | under the args ABI. Differential Revision: http://llvm-reviews.chandlerc.com/D1316 llvm-svn: 188113
* Mark obviously const methods. Also use reference for parameters when possible.Jakub Staszak2013-08-091-8/+8
| | | | llvm-svn: 188103
* Revert "[objc-arc] Track if we encountered an additive overflow while ↵Michael Gottesman2013-08-081-32/+3
| | | | | | | | | | | | computing {TopDown,BottomUp}PathCounts and do nothing if it occured." This reverts commit r187941. The commit was passing on my os x box, but it is failing on some non-osx platforms. I do not have time to look into it now, so I am reverting and will recommit after I figure this out. llvm-svn: 187946
* Fix ARM build.Peter Collingbourne2013-08-081-1/+1
| | | | llvm-svn: 187944
* [objc-arc] Track if we encountered an additive overflow while computing ↵Michael Gottesman2013-08-071-3/+32
| | | | | | | | {TopDown,BottomUp}PathCounts and do nothing if it occured. rdar://14590914 llvm-svn: 187941
* [objc-arc] Change 4 iterator methods which return const_iterators to be ↵Michael Gottesman2013-08-071-4/+4
| | | | | | const methods. llvm-svn: 187940
* Add ISD::FROUND for libm round()Hal Finkel2013-08-071-0/+5
| | | | | | | | | | | | | | | All libm floating-point rounding functions, except for round(), had their own ISD nodes. Recent PowerPC cores have an instruction for round(), and so here I'm adding ISD::FROUND so that round() can be custom lowered as well. For the most part, this is straightforward. I've added an intrinsic and a matching ISD node just like those for nearbyint() and friends. The SelectionDAG pattern I've named frnd (because ISD::FP_ROUND has already claimed fround). This will be used by the PowerPC backend in a follow-up commit. llvm-svn: 187926
* DataFlowSanitizer; LLVM changes.Peter Collingbourne2013-08-073-0/+1008
| | | | | | | | | | | | | DataFlowSanitizer is a generalised dynamic data flow analysis. Unlike other Sanitizer tools, this tool is not designed to detect a specific class of bugs on its own. Instead, it provides a generic dynamic data flow analysis framework to be used by clients to help detect application-specific issues within their own code. Differential Revision: http://llvm-reviews.chandlerc.com/D965 llvm-svn: 187923
* JumpThreading: Turn a select instruction into branching if it allows to ↵Benjamin Kramer2013-08-071-0/+83
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | thread one half of the select. This is a common pattern coming out of simplifycfg generating gross code. a: ; preds = %entry %sel = select i1 %cmp1, double %add, double 0.000000e+00 br label %b b: %cond5 = phi double [ %sel, %a ], [ %sub, %entry ] %cmp6 = fcmp oeq double %cond5, 0.000000e+00 br i1 %cmp6, label %if.then, label %if.end becomes a: br i1 %cmp1, label %b, label %if.then b: %cond5 = phi double [ %sub, %entry ], [ %add, %a ] %cmp6 = fcmp oeq double %cond5, 0.000000e+00 br i1 %cmp6, label %if.then, label %if.end Skipping block b completely if possible. llvm-svn: 187880
* Change the linkage of these global values to 'internal'.Bill Wendling2013-08-061-4/+7
| | | | | | | | | | | | | | | | | | The globals being generated here were given the 'private' linkage type. However, this caused them to end up in different sections with the wrong prefix. E.g., they would be in the __TEXT,__const section with an 'L' prefix instead of an 'l' (lowercase ell) prefix. The problem is that the linker will eat a literal label with 'L'. If a weak symbol is then placed into the __TEXT,__const section near that literal, then it cannot distinguish between the literal and the weak symbol. Part of the problems here was introduced because the address sanitizer converted some C strings into constant initializers with trailing nuls. (Thus putting them in the __const section with the wrong prefix.) The others were variables that the address sanitizer created but simply had the wrong linkage type. llvm-svn: 187827
* LoopVectorize: Allow vectorization of loops with lifetime markersArnold Schwaighofer2013-08-061-8/+18
| | | | | | Patch by Marc Jessome! llvm-svn: 187825
* Adjust file to the coding standard.Jakub Staszak2013-08-061-53/+49
| | | | llvm-svn: 187808
* Unbreak Debug build on WindowsSerge Pavlov2013-08-061-0/+1
| | | | llvm-svn: 187786
* Factor FlattenCFG out from SimplifyCFGTom Stellard2013-08-069-571/+689
| | | | | | Patch by: Mei Ye llvm-svn: 187764
* Fix missing -*- C++ -*-sMatt Arsenault2013-08-061-1/+1
| | | | llvm-svn: 187758
* Introduce an optimisation for special case lists with large numbers of ↵Peter Collingbourne2013-08-051-13/+38
| | | | | | | | | | | | | | | | | | | | | | | | | | literal entries. Our internal regex implementation does not cope with large numbers of anchors very efficiently. Given a ~3600-entry special case list, regex compilation can take on the order of seconds. This patch solves the problem for the special case of patterns matching literal global names (i.e. patterns with no regex metacharacters). Rather than forming regexes from literal global name patterns, add them to a StringSet which is checked before matching against the regex. This reduces regex compilation time by an order of roughly thousands when reading the aforementioned special case list, according to a completely unscientific study. No test cases. I figure that any new tests for this code should check that regex metacharacters are properly recognised. However, I could not find any documentation which documents the fact that the syntax of global names in special case lists is based on regexes. The extent to which regex syntax is supported in special case lists should probably be decided on/documented before writing tests. Differential Revision: http://llvm-reviews.chandlerc.com/D1150 llvm-svn: 187732
* 80-colsAlexey Samsonov2013-08-051-1/+2
| | | | llvm-svn: 187725
* SLPVectorizer: Fix PR16777. PHInodes may use multiple extracted values that ↵Nadav Rotem2013-08-021-10/+13
| | | | | | | | come from different blocks. Thanks Alexey Samsonov. llvm-svn: 187663
* Fix dereferencing end iterator in SimplifyCFG. Patch by Ye Mei.Alexey Samsonov2013-08-021-5/+8
| | | | llvm-svn: 187646
* Teach getOrEnforceKnownAlignment about address spacesMatt Arsenault2013-08-011-4/+5
| | | | llvm-svn: 187629
* Move the optlevel check to the frontend.Nadav Rotem2013-08-011-1/+1
| | | | llvm-svn: 187628
* Only enable SLP-vectorization on O3 builds.Nadav Rotem2013-08-011-1/+1
| | | | llvm-svn: 187595
* 80-colNadav Rotem2013-07-311-9/+14
| | | | llvm-svn: 187535
* Preserve fast-math flags when folding (fsub x, (fneg y)) to (fadd x, y).Owen Anderson2013-07-301-4/+11
| | | | llvm-svn: 187462
* Change behavior of calling bitcasted alias functions.Matt Arsenault2013-07-301-9/+9
| | | | | | | | It will now only convert the arguments / return value and call the underlying function if the types are able to be bitcasted. This avoids using fp<->int conversions that would occur before. llvm-svn: 187444
* SLPVectorier: update the debug location for the new instructions.Nadav Rotem2013-07-291-2/+22
| | | | llvm-svn: 187363
* Teach the AllocaPromoter which is wrapped around the SSAUpdaterChandler Carruth2013-07-291-15/+51
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | infrastructure to do promotion without a domtree the same smarts about looking through GEPs, bitcasts, etc., that I just taught mem2reg about. This way, if SROA chooses to promote an alloca which still has some noisy instructions this code can cope with them. I've not used as principled of an approach here for two reasons: 1) This code doesn't really need it as we were already set up to zip through the instructions used by the alloca. 2) I view the code here as more of a hack, and hopefully a temporary one. The SSAUpdater path in SROA is a real sore point for me. It doesn't make a lot of architectural sense for many reasons: - We're likely to end up needing the domtree anyways in a subsequent pass, so why not compute it earlier and use it. - In the future we'll likely end up needing the domtree for parts of the inliner itself. - If we need to we could teach the inliner to preserve the domtree. Part of the re-work of the pass manager will allow this to be very powerful even in large SCCs with many functions. - Ultimately, computing a domtree has gotten significantly faster since the original SSAUpdater-using code went into ScalarRepl. We no longer use domfrontiers, and much of domtree is lazily done based on queries rather than eagerly. - At this point keeping the SSAUpdater-based promotion saves a total of 0.7% on a build of the 'opt' tool for me. That's not a lot of performance given the complexity! So I'm leaving this a bit ugly in the hope that eventually we just remove all of this nonsense. I can't even readily test this because this code isn't reachable except through SROA. When I re-instate the patch that fast-tracks allocas already suitable for promotion, I'll add a testcase there that failed before this change. Before that, SROA will fix any test case I give it. llvm-svn: 187347
* Don't vectorize when the attribute NoImplicitFloat is used.Nadav Rotem2013-07-291-0/+5
| | | | llvm-svn: 187340
* Fix -Wdocumentation warnings.Rafael Espindola2013-07-281-4/+4
| | | | llvm-svn: 187336
* Update comments for SSAUpdater to use the modern doxygen commentChandler Carruth2013-07-281-41/+3
| | | | | | | | | | | | standards for LLVM. Remove duplicated comments on the interface from the implementation file (implementation comments are left there of course). Also clean up, re-word, and fix a few typos and errors in the commenst spotted along the way. This is in preparation for changes to these files and to keep the uninteresting tidying in a separate commit. llvm-svn: 187335
* Temporarily revert r187323 until I update SSAUpdater to match mem2reg.Chandler Carruth2013-07-281-81/+12
| | | | | | I forgot that we had two totally independent things here. :: sigh :: llvm-svn: 187327
* Now that mem2reg understands how to cope with a slightly wider set ofChandler Carruth2013-07-281-12/+81
| | | | | | | | | | | | uses of an alloca, we can pre-compute promotability while analyzing an alloca for splitting in SROA. That lets us short-circuit the common case of a bunch of trivially promotable allocas. This cuts 20% to 30% off the run time of SROA for typical frontend-generated IR sequneces I'm seeing. It gets the new SROA to within 20% of ScalarRepl for such code. My current benchmark for these numbers is PR15412, but it fits the general pattern of IR emitted by Clang so it should be widely applicable. llvm-svn: 187323
* Thread DataLayout through the callers and into mem2reg. This will beChandler Carruth2013-07-284-12/+20
| | | | | | | useful in a subsequent patch, but causes an unfortunate amount of noise, so I pulled it out into a separate patch. llvm-svn: 187322
* Update the commentNadav Rotem2013-07-271-0/+1
| | | | llvm-svn: 187316
* Don't use all the #ifdefs to hide the stats counters and instead rely onChandler Carruth2013-07-271-18/+0
| | | | | | | | | their being optimized out in debug mode. Realistically, this just isn't going to be the slow part anyways. This also fixes unused variable warnings that are breaking LLD build bots. =/ I didn't see these at first, and kept losing track of the fact that they were broken. llvm-svn: 187297
* Merge the removal of dead instructions and lifetime markers with theChandler Carruth2013-07-271-41/+32
| | | | | | | | | | | analysis of the alloca. We don't need to visit all the users twice for this. We build up a kill list during the analysis and then just process it afterward. This recovers the tiny bit of performance lost by moving to the visitor based analysis system as it removes one entire use-list walk from mem2reg. In some cases, this is now faster than mem2reg was previously. llvm-svn: 187296
* Reimplement isPotentiallyReachable to make nocapture deduction much stronger.Nick Lewycky2013-07-275-97/+5
| | | | | | | | | | Adds unit tests for it too. Split BasicBlockUtils into an analysis-half and a transforms-half, and put the analysis bits into a new Analysis/CFG.{h,cpp}. Promote isPotentiallyReachable into llvm::isPotentiallyReachable and move it into Analysis/CFG. llvm-svn: 187283
* SimplifyCFG: Use parallel-and and parallel-or mode to consolidate branch ↵Tom Stellard2013-07-274-36/+495
| | | | | | | | | | | | | | conditions Merge consecutive if-regions if they contain identical statements. Both transformations reduce number of branches. The transformation is guarded by a target-hook, and is currently enabled only for +R600, but the correctness has been tested on X86 target using a variety of CPU benchmarks. Patch by: Mei Ye llvm-svn: 187278
* SLP Vectorier: Don't vectorize really short chains because they are already ↵Nadav Rotem2013-07-261-2/+6
| | | | | | handled by the SelectionDAG store-vectorizer, which does a better job in deciding when to vectorize. llvm-svn: 187267
* SLP Vectorizer: Disable the vectorization of non power of two chains, such ↵Nadav Rotem2013-07-261-16/+0
| | | | | | as <3 x float>, because we dont have a good cost model for these types. llvm-svn: 187265
* Fix variable name.Owen Anderson2013-07-261-2/+2
| | | | llvm-svn: 187253
* When InstCombine tries to fold away (fsub x, (fneg y)) into (fadd x, y), it isOwen Anderson2013-07-261-1/+13
| | | | | | | also worthwhile for it to look through FP extensions and truncations, whose application commutes with fneg. llvm-svn: 187249
* Correct case of m_UIToFp to m_UIToFP to match instruction name, add m_SIToFP ↵Stephen Lin2013-07-261-4/+4
| | | | | | for consistency. llvm-svn: 187225
* Re-implement the analysis of uses in mem2reg to be significantly moreChandler Carruth2013-07-261-87/+157
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | robust. It now uses an InstVisitor and worklist to actually walk the uses of the Alloca transitively and detect the pattern which we can directly promote: loads & stores of the whole alloca and instructions we can completely ignore. Also, with this new implementation teach both the predicate for testing whether we can promote and the promotion engine itself to use the same code so we no longer have strange divergence between the two code paths. I've added some silly test cases to demonstrate that we can handle slightly more degenerate code patterns now. See the below for why this is even interesting. Performance impact: roughly 1% regression in the performance of SROA or ScalarRepl on a large C++-ish test case where most of the allocas are basically ready for promotion. The reason is because of silly redundant work that I've left FIXMEs for and which I'll address in the next commit. I wanted to separate this commit as it changes the behavior. Once the redundant work in removing the dead uses of the alloca is fixed, this code appears to be faster than the old version. =] So why is this useful? Because the previous requirement for promotion required a *specific* visit pattern of the uses of the alloca to verify: we *had* to look for no more than 1 intervening use. The end goal is to have SROA automatically detect when an alloca is already promotable and directly hand it to the mem2reg machinery rather than trying to partition and rewrite it. This is a 25% or more performance improvement for SROA, and a significant chunk of the delta between it and ScalarRepl. To get there, we need to make mem2reg actually capable of promoting allocas which *look* promotable to SROA without have SROA do tons of work to massage the code into just the right form. This is actually the tip of the iceberg. There are tremendous potential savings we can realize here by de-duplicating work between mem2reg and SROA. llvm-svn: 187191
* [PowerPC] Support powerpc64le as a syntax-checking target.Bill Schmidt2013-07-261-1/+2
| | | | | | | | | | | | | | | | | | | | | | | This patch provides basic support for powerpc64le as an LLVM target. However, use of this target will not actually generate little-endian code. Instead, use of the target will cause the correct little-endian built-in defines to be generated, so that code that tests for __LITTLE_ENDIAN__, for example, will be correctly parsed for syntax-only testing. Code generation will otherwise be the same as powerpc64 (big-endian), for now. The patch leaves open the possibility of creating a little-endian PowerPC64 back end, but there is no immediate intent to create such a thing. The LLVM portions of this patch simply add ppc64le coverage everywhere that ppc64 coverage currently exists. There is nothing of any import worth testing until such time as little-endian code generation is implemented. In the corresponding Clang patch, there is a new test case variant to ensure that correct built-in defines for little-endian code are generated. llvm-svn: 187179
* Respect llvm.used in Internalize.Rafael Espindola2013-07-253-21/+41
| | | | | | | | | | | | | | | The language reference says that: "If a symbol appears in the @llvm.used list, then the compiler, assembler, and linker are required to treat the symbol as if there is a reference to the symbol that it cannot see" Since even the linker cannot see the reference, we must assume that the reference can be using the symbol table. For example, a user can add __attribute__((used)) to a debug helper function like dump and use it from a debugger. llvm-svn: 187103
* Check that TD isn't NULL before dereferencing it down this path.Nick Lewycky2013-07-251-1/+1
| | | | llvm-svn: 187099
* Make these methods const correct.Rafael Espindola2013-07-251-2/+2
| | | | | | Thanks to Nick Lewycky for noticing it. llvm-svn: 187098
OpenPOWER on IntegriCloud