aboutsummaryrefslogtreecommitdiff
path: root/lib/Transforms
AgeCommit message (Collapse)Author
2011-09-06Add -verify-indvars for imperfect SCEV trip count verification after indvars.Andrew Trick
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@139169 91177308-0d34-0410-b5e6-96231b3b80d8
2011-09-06Use IRBuilder.Devang Patel
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@139156 91177308-0d34-0410-b5e6-96231b3b80d8
2011-09-06Try again at r138809 (make DSE more aggressive in removing dead stores at ↵Owen Anderson
the end of a function), now with less deleting stores before memcpy's. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@139150 91177308-0d34-0410-b5e6-96231b3b80d8
2011-09-06Split the init.trampoline intrinsic, which currently combines GCC'sDuncan Sands
init.trampoline and adjust.trampoline intrinsics, into two intrinsics like in GCC. While having one combined intrinsic is tempting, it is not natural because typically the trampoline initialization needs to be done in one function, and the result of adjust trampoline is needed in a different (nested) function. To get around this llvm-gcc hacks the nested function lowering code to insert an additional parent variable holding the adjust.trampoline result that can be accessed from the child function. Dragonegg doesn't have the luxury of tweaking GCC code, so it stored the result of adjust.trampoline in the memory GCC set aside for the trampoline itself (this is always available in the child function), and set up some new memory (using an alloca) to hold the trampoline. Unfortunately this breaks Go which allocates trampoline memory on the heap and wants to use it even after the parent has exited (!). Rather than doing even more hacks to get Go working, it seemed best to just use two intrinsics like in GCC. Patch mostly by Sanjoy Das. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@139140 91177308-0d34-0410-b5e6-96231b3b80d8
2011-09-05Delete trivial landing pads that just continue unwinding the caughtDuncan Sands
exception. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@139117 91177308-0d34-0410-b5e6-96231b3b80d8
2011-09-04Use Duncan's patch to delete the instructions in reverse order (minus the ↵Bill Wendling
landingpad and terminator). git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@139090 91177308-0d34-0410-b5e6-96231b3b80d8
2011-09-02Update comments to reflect reality.Bill Wendling
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@139023 91177308-0d34-0410-b5e6-96231b3b80d8
2011-09-02Enable SCEV-based unrolling by default.Andrew Trick
This changes loop unrolling to use the same mechanism for trip count computation as indvars. This is a stronger check that tends to unroll more loops. A very common side-effect is that many single iteration loops will be removed sooner. The real goal was simply to remove dependence on canonical IVs. x86 is break even. ARM performance changes to expect (+ is good): External/SPEC/CFP2000/183.equake/183.equake +13% SingleSource/Benchmarks/Dhrystone/fldry +21% MultiSource/Applications/spiff/spiff +3% SingleSource/Benchmarks/Stanford/Puzzle -14% The Puzzle regression is actually an improvement in loop optimization that defeats GVN: rdar://problem/10065079. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@139009 91177308-0d34-0410-b5e6-96231b3b80d8
2011-09-02Compare type size instead of type _store_ size to make sure that BitCastInstJakub Staszak
will be valid. This fixes PR10820. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@139005 91177308-0d34-0410-b5e6-96231b3b80d8
2011-09-01Reduce indentation. No functionality change.Bill Wendling
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@138968 91177308-0d34-0410-b5e6-96231b3b80d8
2011-09-01Change worklist driven deletion to be an iterative process.Bill Wendling
Duncan noticed this! git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@138967 91177308-0d34-0410-b5e6-96231b3b80d8
2011-09-01Fix an issue with the IR sink pass found by inspection. (I'm not sure ↵Eli Friedman
anyone is actually using this, but might as well fix it since I found the issue.) git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@138965 91177308-0d34-0410-b5e6-96231b3b80d8
2011-09-01Resubmit with fix. Properly remove the instructions except for landingpad, ↵Bill Wendling
which should be removed only when its invokes are. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@138932 91177308-0d34-0410-b5e6-96231b3b80d8
2011-09-01Submitted this too early.Bill Wendling
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@138931 91177308-0d34-0410-b5e6-96231b3b80d8
2011-09-01Don't DCE the landingpad instruction.Bill Wendling
The landingpad instruction can be removed only when its invokes are removed. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@138930 91177308-0d34-0410-b5e6-96231b3b80d8
2011-08-31Make sure we aren't deleting the landingpad instruction.Bill Wendling
The landingpad instruction is required in the landing pad block. Because we're not deleting terminating instructions, the invoke may still jump to here (see Transforms/SCCP/2004-11-16-DeadInvoke.ll). Remove all uses of the landingpad instruction, but keep it around until code-gen can remove the basic block. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@138890 91177308-0d34-0410-b5e6-96231b3b80d8
2011-08-30Remove the old tail duplication pass. It is not used and is unable to updateRafael Espindola
ssa, so it has to be run really early in the pipeline. Any replacement should probably use the SSAUpdater. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@138841 91177308-0d34-0410-b5e6-96231b3b80d8
2011-08-30Speculatively revert r138809 in an attempt to fix DragonEgg.Owen Anderson
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@138829 91177308-0d34-0410-b5e6-96231b3b80d8
2011-08-30When walking backwards to eliminate final stores to allocas at the end of a ↵Owen Anderson
function, encountering an unrelated store should not cause us to give up like encountering a load does. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@138809 91177308-0d34-0410-b5e6-96231b3b80d8
2011-08-29Fixes following the CR by Chris and Duncan:Nadav Rotem
Optimize chained bitcasts of the form A->B->A. Undo r138722 and change isEliminableCastPair to allow this case. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@138756 91177308-0d34-0410-b5e6-96231b3b80d8
2011-08-28Bitcasts are transitive. Bitcast-Bitcast-X becomes Bitcast-X.Nadav Rotem
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@138722 91177308-0d34-0410-b5e6-96231b3b80d8
2011-08-26Don't sink landingpad instructions during ind-var simplification.Bill Wendling
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@138651 91177308-0d34-0410-b5e6-96231b3b80d8
2011-08-26Address review comments.Benjamin Kramer
- Reword comments. - Allow undefined behavior interfering with undefined behavior. - Add address space checks. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@138619 91177308-0d34-0410-b5e6-96231b3b80d8
2011-08-26SimplifyCFG: If we have a PHI node that can evaluate to NULL and do a load ↵Benjamin Kramer
or store to the address returned by the PHI node then we can consider this incoming value as dead and remove the edge pointing there, unless there are instructions that can affect control flow executed in between. In theory this could be extended to other instructions, eg. division by zero, but it's likely that it will "miscompile" some code because people depend on div by zero not trapping. NULL pointer dereference usually leads to a crash so we should be on the safe side. This shrinks the size of a Release clang by 16k on x86_64. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@138618 91177308-0d34-0410-b5e6-96231b3b80d8
2011-08-25LSR wants to split the landing pad's critical edge. Let it do it, but use theBill Wendling
proper function to do it. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@138550 91177308-0d34-0410-b5e6-96231b3b80d8
2011-08-25When inserting new instructions, use getFirstInsertionPt instead ofBill Wendling
getFirstNonPHI so that it will skip over the landingpad instructions as well. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@138537 91177308-0d34-0410-b5e6-96231b3b80d8
2011-08-24Skip the landingpad instruction when determining the insertion point.Bill Wendling
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@138481 91177308-0d34-0410-b5e6-96231b3b80d8
2011-08-24Use getFirstInsertionPt instead of getFirstNonPHI so that it skips to the properBill Wendling
insertion place. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@138473 91177308-0d34-0410-b5e6-96231b3b80d8
2011-08-24Fix a crashing bug in SplitBlock when it is called on a block with noRafael Espindola
dominator information even though dominators were previously computed. Patch by Nick Sumner. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@138449 91177308-0d34-0410-b5e6-96231b3b80d8
2011-08-22Add a comment.Dan Gohman
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@138243 91177308-0d34-0410-b5e6-96231b3b80d8
2011-08-22Constant pointers to objects don't need reference counting.Dan Gohman
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@138242 91177308-0d34-0410-b5e6-96231b3b80d8
2011-08-19If we're splitting the landing pad block and assigning it only one predecessor,Bill Wendling
then don't split it a second time, since that block will be dead. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@138153 91177308-0d34-0410-b5e6-96231b3b80d8
2011-08-19The landingpad instruction isn't dead simply because it's value isn't used.Bill Wendling
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@138102 91177308-0d34-0410-b5e6-96231b3b80d8
2011-08-19Make a bunch of symbols private.Benjamin Kramer
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@138025 91177308-0d34-0410-b5e6-96231b3b80d8
2011-08-19C API functions must be able to see their extern "C" definitions, or it will ↵Benjamin Kramer
be impossible to call them from C. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@138022 91177308-0d34-0410-b5e6-96231b3b80d8
2011-08-19Track a retain+release nesting level independently of theDan Gohman
known-incremented level, because the two concepts can be used to prove the saftey of a retain+release removal in different ways. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@138016 91177308-0d34-0410-b5e6-96231b3b80d8
2011-08-19Intelligently split the landing pad block.Bill Wendling
We have to be careful when splitting the landing pad block, because the landingpad instruction is required to remain as the first non-PHI of an invoke's unwind edge. To retain this, we split the block into two blocks, moving the predecessors within the loop to one block and the remaining predecessors to the other. The landingpad instruction is cloned into the new blocks. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@138015 91177308-0d34-0410-b5e6-96231b3b80d8
2011-08-19Add SplitLandingPadPredecessors().Bill Wendling
SplitLandingPadPredecessors is similar to SplitBlockPredecessors in that it splits the current block and attaches a set of predecessors to the new basic block. However, it differs from SplitBlockPredecessors in that it's specifically designed to handle landing pad blocks. Two new basic blocks are created: one that is has the vector of predecessors as its predecessors and one that has the remaining predecessors as its predecessors. Those two new blocks then receive a cloned copy of the landingpad instruction from the original block. The landingpad instructions are joined in a PHI, etc. Like SplitBlockPredecessors, it updates the LLVM IR, AliasAnalysis, DominatorTree, DominanceFrontier, LoopInfo, and LCCSA analyses. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@138014 91177308-0d34-0410-b5e6-96231b3b80d8
2011-08-18Use 'getFirstInsertionPt' when trying to insert new instructions during LICM.Bill Wendling
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@138008 91177308-0d34-0410-b5e6-96231b3b80d8
2011-08-18Make it clear that this code is iterating in reverse order through the array.Dan Gohman
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@137985 91177308-0d34-0410-b5e6-96231b3b80d8
2011-08-18Revert r137871. The loop simplify pass should require all exits from a loop thatBill Wendling
aren't from an indirect branch need to be dominated by the loop header. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@137981 91177308-0d34-0410-b5e6-96231b3b80d8
2011-08-18Split out the updating of PHI nodes after splitting the BB into a separateBill Wendling
function. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@137979 91177308-0d34-0410-b5e6-96231b3b80d8
2011-08-18Use this fantzy ArrayRef thing to pass in the list of predecessors.Bill Wendling
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@137978 91177308-0d34-0410-b5e6-96231b3b80d8
2011-08-18The edge from DISubprogram to DICompileUnit has been removed in recent versionsNick Lewycky
of debug info. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@137972 91177308-0d34-0410-b5e6-96231b3b80d8
2011-08-18Use static instead of anonymous namespace.Bill Wendling
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@137959 91177308-0d34-0410-b5e6-96231b3b80d8
2011-08-18Split out the analysis updating code into a helper function. No intendedBill Wendling
functionality change. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@137926 91177308-0d34-0410-b5e6-96231b3b80d8
2011-08-18Dramatically speedup codegen prepare by a) avoiding use of dominator tree ↵Devang Patel
and b) doing a separate pass over dbg.value instructions. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@137908 91177308-0d34-0410-b5e6-96231b3b80d8
2011-08-17Do not use DebugInfoFinder. Extract debug info directly from llvm.dbg.cu ↵Devang Patel
named mdnode. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@137890 91177308-0d34-0410-b5e6-96231b3b80d8
2011-08-17Atomic load/store handling for the passes using memdep (GVN, DSE, memcpyopt).Eli Friedman
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@137888 91177308-0d34-0410-b5e6-96231b3b80d8
2011-08-17Disable PRE for landing pads.Bill Wendling
PRE needs the landing pads to have their critical edges split. Doing this for a landing pad is non-trivial. Abandon the attempt to perform PRE when we come across a landing pad. (Reviewed by Owen!) git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@137876 91177308-0d34-0410-b5e6-96231b3b80d8
generated by cgit v1.2.3-18-g5258 (git 2.32.0) at 2025-05-03 00:51:09 +0000