- Apr 20, 2010
-
-
Jakob Stoklund Olesen authored
register is not killed in the loop. This fixes 188.ammp on ARM where the post-ra scheduler would grab a register that looked available but wasn't. A testcase would be huge and fragile, sorry. llvm-svn: 101930
-
Evan Cheng authored
llvm-svn: 101914
-
Dan Gohman authored
into SelectionDAGBuilder. This avoids a separate pass over the instructions, and has the side effect of providing debug location information to the copy. llvm-svn: 101906
-
Dan Gohman authored
they end up doing nothing. llvm-svn: 101904
-
Dan Gohman authored
llvm-svn: 101902
-
Dan Gohman authored
llvm-svn: 101901
-
Dan Gohman authored
llvm-svn: 101900
-
Gabor Greif authored
llvm-svn: 101899
-
Chris Lattner authored
in other types. fix this by only bumping zero-byte globals up to a single byte if the *entire global* is zero size, fixing PR6340. This also fixes empty arrays etc to be handled correctly, and only does this on subsection-via-symbols targets (aka darwin) which is the only place where this matters. llvm-svn: 101879
-
Dan Gohman authored
llvm-svn: 101860
-
Bill Wendling authored
it as it's not dead. llvm-svn: 101855
-
Dan Gohman authored
in the case where a basic block is split. llvm-svn: 101850
-
Dan Gohman authored
SelectionDAGBuilder, where it doesn't have to be as complicated. llvm-svn: 101848
-
Dan Gohman authored
responsible for figuring out what that's supposed to be on its own. llvm-svn: 101844
-
Dan Gohman authored
where multiple blocks are emitted; functions which do this need to return the new BB so that their callers can stay current. llvm-svn: 101843
-
Dan Gohman authored
llvm-svn: 101832
-
Dan Gohman authored
and End arguments by-value rather than by-reference. llvm-svn: 101830
-
Dan Gohman authored
an argument to things that need it. llvm-svn: 101825
-
Dan Gohman authored
need it, just pass around the parent block of the current instruction explicitly. llvm-svn: 101822
-
- Apr 19, 2010
-
-
Evan Cheng authored
llvm-svn: 101808
-
Dan Gohman authored
fetch one from the MachineFunction. llvm-svn: 101807
-
Evan Cheng authored
llvm-svn: 101806
-
Devang Patel authored
llvm-svn: 101805
-
Dan Gohman authored
than just getting one through a TargetLowering. llvm-svn: 101802
-
Dan Gohman authored
SelectionDAG-specific. llvm-svn: 101801
-
- Apr 17, 2010
-
-
Dan Gohman authored
llvm-svn: 101655
-
Dan Gohman authored
llvm-svn: 101640
-
Dan Gohman authored
llvm-svn: 101639
-
Dan Gohman authored
llvm-svn: 101638
-
Dan Gohman authored
llvm-svn: 101637
-
Dan Gohman authored
const_casts, and it reinforces the design of the Target classes being immutable. SelectionDAGISel::IsLegalToFold is now a static member function, because PIC16 uses it in an unconventional way. There is more room for API cleanup here. And PIC16's AsmPrinter no longer uses TargetLowering. llvm-svn: 101635
-
Evan Cheng authored
the live-in sets of BBs in the loop. Otherwise later pass may end up using the registers and override the invariant. rdar://7852937 No reasonablly sized test case possible. llvm-svn: 101626
-
Evan Cheng authored
llvm-svn: 101622
-
Evan Cheng authored
llvm-svn: 101621
-
Evan Cheng authored
llvm-svn: 101620
-
Bob Wilson authored
just remove them all. Radar 7873207 (working around the root problem of Radar 7759363). llvm-svn: 101604
-
Jakob Stoklund Olesen authored
Revert "Use a simpler data structure to calculate the least recently used register in RegAllocLocal." This reverts commit 101392. It broke a buildbot. llvm-svn: 101595
-
Eric Christopher authored
Probably the best way to know that all getOperand() calls have been handled is to replace that API instead of updating. llvm-svn: 101579
-
Devang Patel authored
llvm-svn: 101575
-
Jakob Stoklund Olesen authored
This makes the local register allocator about 20% faster. llvm-svn: 101574
-