- Mar 14, 2008
-
-
Evan Cheng authored
Livein copy scheduling fixes: do not coalesce physical register copies, correctly determine the safe location to insert the copies. llvm-svn: 48348
-
Nate Begeman authored
llvm-svn: 48347
-
Dan Gohman authored
llvm-svn: 48346
-
Dan Gohman authored
llvm-svn: 48345
-
- Mar 13, 2008
-
-
Dan Gohman authored
llvm-svn: 48344
-
Owen Anderson authored
pointer bitcast when performing return slot optimization. llvm-svn: 48343
-
Tanya Lattner authored
Patch by Julien Lerouge. Thanks! llvm-svn: 48342
-
Evan Cheng authored
llvm-svn: 48341
-
Devang Patel authored
Thanks Daniel Dunbar! llvm-svn: 48340
-
Ted Kremenek authored
to be run on other declarations of blocks of code (e.g., Objective-C methods.) llvm-svn: 48339
-
Evan Cheng authored
llvm-svn: 48338
-
Evan Cheng authored
llvm-svn: 48337
-
Evan Cheng authored
Don't try to sink 3-address instruction if convertToThreeAddress created more than one instructions. llvm-svn: 48336
-
Evan Cheng authored
llvm-svn: 48335
-
Evan Cheng authored
llvm-svn: 48334
-
Evan Cheng authored
TwoAddressInstructionPass enhancement. After it converts a two address instruction into a 3-address one, sink it past the instruction that kills the read-mod-write register if its definition is used past the kill. This reduces the number of live register by one. llvm-svn: 48333
-
Chris Lattner authored
llvm-svn: 48332
-
Chris Lattner authored
This still isn't right, but is slightly nicer. llvm-svn: 48331
-
Chris Lattner authored
Set the start of DeclStmt range. Right now the end is meaningless though. llvm-svn: 48330
-
Christopher Lamb authored
Get rid of a pseudo instruction and replace it with subreg based operation on real instructions, ridding the asm printers of the hack used to do this previously. In the process, update LowerSubregs to be careful about eliminating copies that have side affects. Note: the coalescer will have to be careful about this too, when it starts coalescing insert_subreg nodes. llvm-svn: 48329
-
Chris Lattner authored
llvm-svn: 48328
-
Chris Lattner authored
llvm-svn: 48327
-
Chris Lattner authored
llvm-svn: 48326
-
Chris Lattner authored
llvm-svn: 48325
-
Anders Carlsson authored
llvm-svn: 48324
-
Chris Lattner authored
llvm-svn: 48323
-
Ted Kremenek authored
we gracefully back out and return NULL for the CFG, allowing clients to skip analyzing functions with these CFGs. We will add support later. Modified base ASTConsumer "CFGVisitor" to detect when a CFG is not constructed and to emit a warning. llvm-svn: 48322
-
Evan Cheng authored
llvm-svn: 48321
-
Nick Lewycky authored
successors. This makes it support nounwind. llvm-svn: 48320
-
Evan Cheng authored
llvm-svn: 48319
-
Dan Gohman authored
llvm-svn: 48318
-
Dan Gohman authored
llvm-svn: 48317
-
Dan Gohman authored
llvm-svn: 48316
-
Dan Gohman authored
Pass* to PMDataManager*. PMDataManager is more specific than Pass, so this more accurately describes the objects that are being stored. This eliminates the need for several dynamic_casts to PMDataManager*. It does introduce one dynamic_cast though, in dumpPasses(). Give this one a comment describing why a dynamic_cast is being used. llvm-svn: 48315
-
Dan Gohman authored
a Pass*. PMDataManager* is what it actually holds, so this makes it clearer. llvm-svn: 48314
-
Dan Gohman authored
llvm-svn: 48313
-
Dan Gohman authored
llvm-svn: 48312
-
Evan Cheng authored
llvm-svn: 48311
-
Dale Johannesen authored
calls here. This was done earlier for params in the varargs part of the params; any float params that survive to here are in the non-varargs part, and must not be promoted. llvm-svn: 48310
-
Steve Naroff authored
- For explicit ivar refers, make sure the cast is propagated to the AST. - Don't free the base (since it is still in use). This fixes the recent regression to test/Rewriter/objc-ivar-receiver-1.m. llvm-svn: 48309
-