wangtao
/
llvm-project
forked from OSchip/llvm-project
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity
llvm-project/llvm/lib/Transforms/Scalar
History
Michael Gottesman b40db26eae Teach TailRecursionElimination to handle certain cases of nocapture escaping allocas. 
Without the changes introduced into this patch, if TRE saw any allocas at all,
TRE would not perform TRE *or* mark callsites with the tail marker.

Because TRE runs after mem2reg, this inadequacy is not a death sentence. But
given a callsite A without escaping alloca argument, A may not be able to have
the tail marker placed on it due to a separate callsite B having a write-back
parameter passed in via an argument with the nocapture attribute.

Assume that B is the only other callsite besides A and B only has nocapture
escaping alloca arguments (*NOTE* B may have other arguments that are not passed
allocas). In this case not marking A with the tail marker is unnecessarily
conservative since:

  1. By assumption A has no escaping alloca arguments itself so it can not
     access the caller's stack via its arguments.

  2. Since all of B's escaping alloca arguments are passed as parameters with
     the nocapture attribute, we know that B does not stash said escaping
     allocas in a manner that outlives B itself and thus could be accessed
     indirectly by A.

With the changes introduced by this patch:

  1. If we see any escaping allocas passed as a capturing argument, we do
     nothing and bail early.

  2. If we do not see any escaping allocas passed as captured arguments but we
     do see escaping allocas passed as nocapture arguments:

       i. We do not perform TRE to avoid PR962 since the code generator produces
          significantly worse code for the dynamic allocas that would be created
          by the TRE algorithm.

       ii. If we do not return twice, mark call sites without escaping allocas
           with the tail marker. *NOTE* This excludes functions with escaping
           nocapture allocas.

  3. If we do not see any escaping allocas at all (whether captured or not):

       i. If we do not have usage of setjmp, mark all callsites with the tail
          marker.

       ii. If there are no dynamic/variable sized allocas in the function,
           attempt to perform TRE on all callsites in the function.

Based off of a patch by Nick Lewycky.

rdar://14324281.

llvm-svn: 186057
 		2013-07-11 04:40:01 +00:00
..
ADCE.cpp Use SmallVectorImpl::iterator/const_iterator instead of SmallVector to avoid specifying the vector size. 2013-07-04 01:31:24 +00:00
BasicBlockPlacement.cpp Move all of the header files which are involved in modelling the LLVM IR 2013-01-02 11:36:10 +00:00
CMakeLists.txt Remove the simplify-libcalls pass (finally) 2013-06-20 19:48:07 +00:00
CodeGenPrepare.cpp Remove a useless declarations (found by scan-build) 2013-07-05 15:58:12 +00:00
ConstantProp.cpp Move all of the header files which are involved in modelling the LLVM IR 2013-01-02 11:36:10 +00:00
CorrelatedValuePropagation.cpp CVP: If we have a PHI with an incoming select, try to skip the select. 2013-02-24 15:34:43 +00:00
DCE.cpp Move all of the header files which are involved in modelling the LLVM IR 2013-01-02 11:36:10 +00:00
DeadStoreElimination.cpp Change GetPointerBaseWithConstantOffset's DataLayout argument from a 2013-01-31 02:00:45 +00:00
EarlyCSE.cpp Move all of the header files which are involved in modelling the LLVM IR 2013-01-02 11:36:10 +00:00
GVN.cpp Use SmallVectorImpl::iterator/const_iterator instead of SmallVector to avoid specifying the vector size. 2013-07-04 01:31:24 +00:00
GlobalMerge.cpp Access the TargetLoweringInfo from the TargetMachine object instead of caching it. The TLI may change between functions. No functionality change. 2013-06-19 21:07:11 +00:00
IndVarSimplify.cpp IndVarSimplify: check if loop invariant expansion can trap 2013-06-04 17:51:58 +00:00
JumpThreading.cpp Move all of the header files which are involved in modelling the LLVM IR 2013-01-02 11:36:10 +00:00
LICM.cpp LICM: Hoist insertvalue/extractvalue out of loops. 2013-01-09 18:12:03 +00:00
LLVMBuild.txt LLVMBuild: Remove trailing newline, which irked me. 2011-12-12 19:48:00 +00:00
LoopDeletion.cpp Use SmallVectorImpl::iterator/const_iterator instead of SmallVector to avoid specifying the vector size. 2013-07-04 01:31:24 +00:00
LoopIdiomRecognize.cpp PR14904: Segmentation fault running pass 'Recognize loop idioms' 2013-01-10 23:32:01 +00:00
LoopInstSimplify.cpp Remove #includes from the commonly used LoopInfo.h. 2013-02-09 01:04:28 +00:00
LoopRotation.cpp Rotate multi-exit loops even if the latch was simplified. 2013-05-06 17:58:18 +00:00
LoopStrengthReduce.cpp LSR: Fix the parameters used to compute the scaling factor cost. 2013-06-19 19:59:41 +00:00
LoopUnrollPass.cpp Switch CodeMetrics itself over to use TTI to determine if an instruction 2013-01-21 13:04:33 +00:00
LoopUnswitch.cpp Switch CodeMetrics itself over to use TTI to determine if an instruction 2013-01-21 13:04:33 +00:00
LowerAtomic.cpp Move all of the header files which are involved in modelling the LLVM IR 2013-01-02 11:36:10 +00:00
Makefile
…
MemCpyOptimizer.cpp Use SmallVectorImpl::iterator/const_iterator instead of SmallVector to avoid specifying the vector size. 2013-07-03 15:07:05 +00:00
Reassociate.cpp Reassociate: Remove unnecessary default operator=. 2013-07-06 15:10:13 +00:00
Reg2Mem.cpp Move all of the header files which are involved in modelling the LLVM IR 2013-01-02 11:36:10 +00:00
SCCP.cpp dbgs() << Instruction doesn't print a newline on the end any more. Update these 2013-06-26 00:30:18 +00:00
SROA.cpp Use SmallVectorImpl::iterator/const_iterator instead of SmallVector to avoid specifying the vector size. 2013-07-03 15:07:05 +00:00
Scalar.cpp Remove the simplify-libcalls pass (finally) 2013-06-20 19:48:07 +00:00
ScalarReplAggregates.cpp Use SmallVectorImpl::iterator/const_iterator instead of SmallVector to avoid specifying the vector size. 2013-07-03 15:07:05 +00:00
SimplifyCFGPass.cpp Make SimplifyCFG simply depend upon TargetTransformInfo and pass it 2013-01-07 03:53:25 +00:00
Sink.cpp Move all of the header files which are involved in modelling the LLVM IR 2013-01-02 11:36:10 +00:00
StructurizeCFG.cpp Move StructurizeCFG out of R600 to generic Transforms. 2013-06-19 20:18:24 +00:00
TailRecursionElimination.cpp Teach TailRecursionElimination to handle certain cases of nocapture escaping allocas. 2013-07-11 04:40:01 +00:00