aboutsummaryrefslogtreecommitdiff
path: root/lib/Transforms
AgeCommit message (Collapse)Author
2009-10-20Fix PR4313: IPSCCP was not setting the lattice value for the invoke instructionTorok Edwin
when the invoke had multiple return values: it set the lattice value only on the extractvalue. This caused the invoke's lattice value to remain the default (undefined), and later propagated to extractvalue's operand, which incorrectly introduces undefined behavior. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@84637 91177308-0d34-0410-b5e6-96231b3b80d8
2009-10-19Refactor lookup_or_add to contain _MUCH_ less duplicated code. Add support forOwen Anderson
numbering first class aggregate instructions while we're at it. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@84547 91177308-0d34-0410-b5e6-96231b3b80d8
2009-10-19Malloc calls are marked NoAlias, so the code below the isMalloc() check ↵Victor Hernandez
makes it redundant. Removing the isMalloc() check. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@84541 91177308-0d34-0410-b5e6-96231b3b80d8
2009-10-19Simplify some code.Owen Anderson
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@84533 91177308-0d34-0410-b5e6-96231b3b80d8
2009-10-19Fix SplitBlockPredecessors' LoopInfo updating code to handle the caseDan Gohman
where a loop's header is being split and it has predecessors which are not contained by the most-nested loop which contains the loop. This fixes PR5235. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@84505 91177308-0d34-0410-b5e6-96231b3b80d8
2009-10-19Change instnamer to name arguments "arg" instead of "tmp" for clarity, andDan Gohman
to name basic blocks "bb" instead of "BB", for consistency. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@84502 91177308-0d34-0410-b5e6-96231b3b80d8
2009-10-18remove the IndMemRemPass, which only made sense for when malloc/free were ↵Chris Lattner
intrinsic instructions. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@84404 91177308-0d34-0410-b5e6-96231b3b80d8
2009-10-17Use raw_ostream::write_escaped instead of EscapeString.Daniel Dunbar
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@84356 91177308-0d34-0410-b5e6-96231b3b80d8
2009-10-17Simplify some code (first hunk) and fix PR5208 (second hunk) byChris Lattner
updating the callgraph when introducing a call. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@84310 91177308-0d34-0410-b5e6-96231b3b80d8
2009-10-17Remove MallocInst from LLVM Instructions.Victor Hernandez
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@84299 91177308-0d34-0410-b5e6-96231b3b80d8
2009-10-17Autoupgrade malloc insts to malloc calls.Victor Hernandez
Update testcases that rely on malloc insts being present. Also prematurely remove MallocInst handling from IndMemRemoval and RaiseAllocations to help pass tests in this incremental step. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@84292 91177308-0d34-0410-b5e6-96231b3b80d8
2009-10-16HeapAllocSRoA also needs to check if malloc array size can be computed.Victor Hernandez
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@84288 91177308-0d34-0410-b5e6-96231b3b80d8
2009-10-16Move zext and sext casts fed by loads into the same block as theDan Gohman
load, to help SelectionDAG fold them into the loads, unless conditions are unfavorable. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@84271 91177308-0d34-0410-b5e6-96231b3b80d8
2009-10-16Strip trailing white space.Duncan Sands
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@84256 91177308-0d34-0410-b5e6-96231b3b80d8
2009-10-15Fix bug where array malloc with unexpected computation of the size argument ↵Victor Hernandez
resulted in MallocHelper identifying the malloc as a non-array malloc. This broke GlobalOpt's optimization of stores of mallocs to global variables. The fix is to classify malloc's into 3 categories: 1. non-array mallocs 2. array mallocs whose array size can be determined 3. mallocs that cannot be determined to be of type 1 or 2 and cannot be optimized getMallocArraySize() returns NULL for category 3, and all users of this function must avoid their malloc optimization if this function returns NULL. Eventually, currently unexpected codegen for computing the malloc's size argument will be supported in isArrayMalloc() and getMallocArraySize(), extending malloc optimizations to those examples. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@84199 91177308-0d34-0410-b5e6-96231b3b80d8
2009-10-15only try to fold constantexpr operands when the worklist is first populated, Chris Lattner
don't bother every time going around the main worklist. This speeds up a release-asserts opt -std-compile-opts on 403.gcc by about 4% (1.5s). It seems to speed up the most expensive instances of instcombine by ~10%. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@84171 91177308-0d34-0410-b5e6-96231b3b80d8
2009-10-15don't bother calling ConstantFoldInstruction unless there is a use of theChris Lattner
instruction (which disqualifies stores, unreachable, etc) and at least the first operand is a constant. This filters out a lot of obvious cases that can't be folded. Also, switch the IRBuilder to a TargetFolder, which tries harder. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@84170 91177308-0d34-0410-b5e6-96231b3b80d8
2009-10-14Use isVoidTy()Devang Patel
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@84118 91177308-0d34-0410-b5e6-96231b3b80d8
2009-10-14make instcombine's instruction sinking more aggressive in theChris Lattner
presence of PHI nodes. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@84103 91177308-0d34-0410-b5e6-96231b3b80d8
2009-10-13Check void type before using RAUWd.Devang Patel
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@84049 91177308-0d34-0410-b5e6-96231b3b80d8
2009-10-13Do not check use_empty() before replaceAllUsesWith(). This gives ↵Devang Patel
ValueHandles a chance to get properly updated. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@84033 91177308-0d34-0410-b5e6-96231b3b80d8
2009-10-13Use the new CodeMetrics class to compute code size instead ofDan Gohman
manually counting instructions. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@84016 91177308-0d34-0410-b5e6-96231b3b80d8
2009-10-13Update CMake file.Ted Kremenek
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@84001 91177308-0d34-0410-b5e6-96231b3b80d8
2009-10-13Commit the removal of this file, which is now moved to lib/Analysis.Dan Gohman
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@83999 91177308-0d34-0410-b5e6-96231b3b80d8
2009-10-13Move the InlineCost code from Transforms/Utils to Analysis.Dan Gohman
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@83998 91177308-0d34-0410-b5e6-96231b3b80d8
2009-10-13Start refactoring the inline cost estimation code so that it can be usedDan Gohman
for purposes other than inlining. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@83997 91177308-0d34-0410-b5e6-96231b3b80d8
2009-10-13change simplifycfg to not duplicate 'unwind' instructions. HopefullyChris Lattner
this will increase the likelihood of common code getting sunk towards the unwind. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@83996 91177308-0d34-0410-b5e6-96231b3b80d8
2009-10-13Make LoopUnswitch's cost estimation count Instructions, rather thanDan Gohman
BasicBlocks, so that it doesn't blindly procede in the presence of large individual BasicBlocks. This addresses a class of code-size expansion problems. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@83992 91177308-0d34-0410-b5e6-96231b3b80d8
2009-10-12Make licm debug message readable.Evan Cheng
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@83908 91177308-0d34-0410-b5e6-96231b3b80d8
2009-10-12Fix warning.Dale Johannesen
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@83870 91177308-0d34-0410-b5e6-96231b3b80d8
2009-10-12populate instcombine's initial worklist more carefully, causingChris Lattner
it to visit instructions from the start of the function to the end of the function in the first path. This greatly speeds up some pathological cases (e.g. PR5150). Try #3, this time with some unneeded debug info stuff removed which was causing dead pointers to be added to the worklist. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@83818 91177308-0d34-0410-b5e6-96231b3b80d8
2009-10-11revert r83814 for now, it is making the llvm-gcc bootstrap unhappy.Chris Lattner
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@83817 91177308-0d34-0410-b5e6-96231b3b80d8
2009-10-11populate instcombine's initial worklist more carefully, causingChris Lattner
it to visit instructions from the start of the function to the end of the function in the first path. This greatly speeds up some pathological cases (e.g. PR5150). git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@83814 91177308-0d34-0410-b5e6-96231b3b80d8
2009-10-11remove some harmful code that would turn an insertelement on an undefChris Lattner
into a shuffle even if it was used by another insertelement. If the visitation order of instcombine was wrong, this would turn a chain of insertelements into a chain of shufflevectors, which was quite painful. Since CollectShuffleElements handles these cases, the code can just be nuked. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@83810 91177308-0d34-0410-b5e6-96231b3b80d8
2009-10-11teach instcombine to simplify xor's harder, catching theChris Lattner
new testcase. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@83799 91177308-0d34-0410-b5e6-96231b3b80d8
2009-10-11cleanupsChris Lattner
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@83797 91177308-0d34-0410-b5e6-96231b3b80d8
2009-10-11cleanup, no functionality change.Chris Lattner
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@83795 91177308-0d34-0410-b5e6-96231b3b80d8
2009-10-11generalize a transformation even more: we don't care whether theChris Lattner
input the the mul is a zext from bool, just that it is all zeros other than the low bit. This fixes some phase ordering issues that would cause us to miss some xforms in mul.ll when the worklist is visited differently. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@83794 91177308-0d34-0410-b5e6-96231b3b80d8
2009-10-11simplify a transformation by making it more general.Chris Lattner
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@83792 91177308-0d34-0410-b5e6-96231b3b80d8
2009-10-11temporarily revert previous patchChris Lattner
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@83791 91177308-0d34-0410-b5e6-96231b3b80d8
2009-10-11populate instcombine's initial worklist more carefully, causingChris Lattner
it to visit instructions from the start of the function to the end of the function in the first path. This greatly speeds up some pathological cases (e.g. PR5150). git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@83790 91177308-0d34-0410-b5e6-96231b3b80d8
2009-10-11Remove CleanupDbgInfo, instcombine does this and its not worth duplicating itTorok Edwin
here. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@83789 91177308-0d34-0410-b5e6-96231b3b80d8
2009-10-11LICM shouldn't sink/delete debug information. Fix this and add a testcase.Torok Edwin
For now the metadata of sinked/hoisted instructions is still wrong, but that'll be fixed when instructions will have debug metadata directly attached. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@83786 91177308-0d34-0410-b5e6-96231b3b80d8
2009-10-11when folding duplicate conditions, delete the Chris Lattner
now-probably-dead instruction tree feeding it. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@83778 91177308-0d34-0410-b5e6-96231b3b80d8
2009-10-11implement rdar://7293527, a trivial instcombine that llvm-gccChris Lattner
gets but clang doesn't, because it is implemented in GCC's fold routine. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@83761 91177308-0d34-0410-b5e6-96231b3b80d8
2009-10-11implement a transformation in jump threading that is currentlyChris Lattner
done by condprop, but do it in a much more general form. The basic idea is that we can do a limited form of tail duplication in the case when we have a branch on a phi. Moving the branch up in to the predecessor block makes instruction selection much easier and encourages chained jump threadings. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@83759 91177308-0d34-0410-b5e6-96231b3b80d8
2009-10-11restructure some code, no functionality change.Chris Lattner
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@83756 91177308-0d34-0410-b5e6-96231b3b80d8
2009-10-11factor some code better and move a function, no functionality change.Chris Lattner
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@83755 91177308-0d34-0410-b5e6-96231b3b80d8
2009-10-11make jump threading on a phi with undef inputs happen.Chris Lattner
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@83754 91177308-0d34-0410-b5e6-96231b3b80d8
2009-10-11rewrite LCSSA to use SSAUpdate, to only return true if it modifiesChris Lattner
the IR, and to implement the FIXME'd optimization. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@83748 91177308-0d34-0410-b5e6-96231b3b80d8
generated by cgit v1.2.3-18-g5258 (git 2.32.0) at 2025-07-19 01:43:22 +0000