- Mar 31, 2009
-
-
Bill Wendling authored
llvm-svn: 68099
-
Owen Anderson authored
Remove the "fast" cases for spill and restore point determination, as these were subtlely wrong in obscure cases. Patch the testcase to account for this change. llvm-svn: 68093
-
Bill Wendling authored
llvm-svn: 68092
-
Chris Lattner authored
llvm-svn: 68087
-
Dan Gohman authored
instructions. This fixes lua. llvm-svn: 68083
-
Evan Cheng authored
X86 address mode isel tweak. If the base of the address is also used by a CopyToReg (i.e. it's likely live-out), do not fold the sub-expressions into the addressing mode to avoid computing the address twice. The CopyToReg use will be isel'ed to a LEA, re-use it for address instead. This is not yet enabled. llvm-svn: 68082
-
Douglas Gregor authored
llvm-svn: 68079
-
Dan Gohman authored
only reachable via fall-through edges. This dramatically reduces the number of labels printed, and thus also the number of labels the assembler must parse and remember. llvm-svn: 68073
-
Devang Patel authored
Loop Index Split can eliminate a loop if it can determin if loop body is executed only once. There was a bug in determining IV based value of the iteration for which the loop body is executed. Fix it. llvm-svn: 68071
-
- Mar 30, 2009
-
-
Douglas Gregor authored
llvm-svn: 68067
-
Evan Cheng authored
When optimzing a mul by immediate into two, the resulting mul's should get a x86 specific node to avoid dag combiner from hacking on them further. llvm-svn: 68066
-
Evan Cheng authored
Turn a 2-address instruction into a 3-address one when it's profitable even if the two-address operand is killed. e.g. %reg1024<def> = MOV r1 %reg1025<def> = ADD %reg1024, %reg1026 r0 = MOV %reg1025 If it's not possible / profitable to commute ADD, then turning ADD into a LEA saves a copy. llvm-svn: 68065
-
Chris Lattner authored
llvm-svn: 68060
-
Bill Wendling authored
llvm-svn: 68059
-
Bill Wendling authored
llvm-svn: 68057
-
Chris Lattner authored
anything pointer-like, which may or may not actually be a pointer. llvm-svn: 68056
-
Chris Lattner authored
allow the traits to be specified as well. llvm-svn: 68055
-
Dan Gohman authored
llvm-svn: 68054
-
Dan Gohman authored
valid argument attributes (zeroext and signext are). llvm-svn: 68053
-
John Mosby authored
docs/TestingGuide.html: correction to prev. text (objdir!=srcdir required for running test-suite), removed refs to llvm-test llvm-svn: 68051
-
Bob Wilson authored
llvm-svn: 68050
-
Devang Patel authored
Walking an invalidated iterator is not a good idea. llvm-svn: 68047
-
Mike Stump authored
llvm-svn: 68038
-
Anton Korobeynikov authored
llvm-svn: 68037
-
Anton Korobeynikov authored
llvm-svn: 68036
-
Anton Korobeynikov authored
llvm-svn: 68035
-
Anton Korobeynikov authored
llvm-svn: 68034
-
Anton Korobeynikov authored
llvm-svn: 68033
-
Anton Korobeynikov authored
Do not propagate ELF-specific stuff (data.rel) into other targets. This simplifies code and also ensures correctness. llvm-svn: 68032
-
Anton Korobeynikov authored
llvm-svn: 68031
-
Chris Lattner authored
llvm-svn: 68026
-
Evan Cheng authored
llvm-svn: 68025
-
John Mosby authored
llvm-svn: 68023
-
- Mar 29, 2009
-
-
Misha Brukman authored
llvm-svn: 68016
-
Bill Wendling authored
llvm-svn: 68013
-
Anton Korobeynikov authored
llvm-svn: 68008
-
Anton Korobeynikov authored
llvm-svn: 68007
-
Anton Korobeynikov authored
llvm-svn: 68006
-
Anton Korobeynikov authored
llvm-svn: 68005
-
Anton Korobeynikov authored
Extend the relocation tracker handler, so we can filter on different 'kinds' of relocations required. llvm-svn: 68004
-