- Feb 07, 2009
-
-
Mon P Wang authored
changes the address space of the pointer. llvm-svn: 64035
-
Dale Johannesen authored
No functional change. llvm-svn: 64026
-
Dan Gohman authored
ScheduleDAG's TLI member to use const. llvm-svn: 64018
-
Nick Lewycky authored
two AddString methods to use it. llvm-svn: 64005
-
Mike Stump authored
llvm-svn: 64003
-
Dale Johannesen authored
Adjust derived classes to pass UnknownLoc where a DebugLoc does not make sense. Pick one of DebugLoc and non-DebugLoc variants to survive for all such classes. llvm-svn: 64000
-
Evan Cheng authored
llvm-svn: 63999
-
Dale Johannesen authored
llvm-svn: 63995
-
Evan Cheng authored
Don't sink the instruction if TargetRegisterInfo::isSafeToMoveRegClassDefs doesn't think it's safe. This works around PR1911. llvm-svn: 63994
-
Dale Johannesen authored
llvm-svn: 63993
-
Dale Johannesen authored
Many targets build placeholder nodes for special operands, e.g. GlobalBaseReg on X86 and PPC for the PIC base. There's no sensible way to associate debug info with these. I've left them built with getNode calls with explicit DebugLoc::getUnknownLoc operands. I'm not too happy about this but don't see a good improvement; I considered adding a getPseudoOperand or something, but it seems to me that'll just make it harder to read. llvm-svn: 63992
-
Dan Gohman authored
llvm-svn: 63989
-
Dan Gohman authored
llvm-svn: 63988
-
Dale Johannesen authored
getCALLSEQ_{END,START} to permit passing no DebugLoc there. UNDEF doesn't logically have DebugLoc; add getUNDEF to encapsulate this. llvm-svn: 63978
-
- Feb 06, 2009
-
-
Dale Johannesen authored
llvm-svn: 63971
-
Dale Johannesen authored
llvm-svn: 63969
-
Bill Wendling authored
A simple test program shows that debugging works. :-) llvm-svn: 63968
-
Bill Wendling authored
llvm-svn: 63967
-
Dan Gohman authored
that already hold the lock can call an entry point that doesn't re-acquire the lock. llvm-svn: 63965
-
Dan Gohman authored
addresses to symbols. llvm-svn: 63962
-
Dale Johannesen authored
llvm-svn: 63951
-
Dan Gohman authored
SelectionDAGISel::CreateScheduler, and make it just create the scheduler. Leave running the scheduler to the higher-level code. This makes the higher-level code a little more explicit and easier to follow, and will help enable some future refactoring. llvm-svn: 63944
-
Sanjiv Gupta authored
llvm-svn: 63943
-
Dan Gohman authored
llvm-svn: 63941
-
Evan Cheng authored
llvm-svn: 63938
-
Dan Gohman authored
that previously included this header should include SchedulerRegistry.h instead. llvm-svn: 63937
-
Evan Cheng authored
Add TargetInstrInfo::isSafeToMoveRegisterClassDefs. It returns true if it's safe to move an instruction which defines a value in the register class. Replace pre-splitting specific IgnoreRegisterClassBarriers with this new hook. llvm-svn: 63936
-
Dan Gohman authored
that used this header to select a scheduling policy should use SchedulerRegistry.h instead (llvm-gcc and clang were updated a while ago). llvm-svn: 63934
-
Devang Patel authored
llvm-svn: 63923
-
Chris Lattner authored
llvm-svn: 63916
-
Dale Johannesen authored
llvm-svn: 63909
-
Dale Johannesen authored
its corresponding getTargetNode. Lots of caller changes. llvm-svn: 63904
-
Owen Anderson authored
llvm-svn: 63902
-
Owen Anderson authored
Refactor my previous change to maintain the distinction between AliasAnalysis and BasicAliasAnalysis. This involves some wider changes because it folds away some never-used methods. llvm-svn: 63900
-
Devang Patel authored
Ignore dbg intrinsics while propagating conditional expression info. Take 2. llvm-svn: 63898
-
- Feb 05, 2009
-
-
Evan Cheng authored
isAsCheapAsMove instructions can have register src operands. Check if they are really re-materializable. This fixes sse.expandfft and sse.stepfft. llvm-svn: 63890
-
Dale Johannesen authored
llvm-svn: 63889
-
Devang Patel authored
Revert rev. 63876. It is causing llvm-gcc bootstrap failure. llvm-svn: 63888
-
Dale Johannesen authored
llvm-svn: 63886
-
Devang Patel authored
llvm-svn: 63880
-