Commits

Show all
Author Commit Message Labels Comments Date
Dan Gohman
Add more information to the EFLAGS note.
Tags
2 tags
stuart
Check for warnings about inappropriate weak_imports.
Dan Gohman
Add a note about EFLAGS optimization.
Devang Patel
Ignore debug info while evaluating function.
Dan Gohman
Don't record the increment instruction; just recompute it from the Phi
Tags
checker-0.170
Chris Lattner
Fix PR3724 by searching for the largest free block when
Devang Patel
Remove llvm.dbg.global_variables also.
Dan Gohman
Fix a few more places where induction variable types were used
Dan Gohman
Use ReplacedTy instead of recomputing the same value.
Dan Gohman
Use LoopInfo's getLoopLatch() instead of doing what it does manualy.
Dan Gohman
Don't use an induction variable type as a memory access type.
Dan Gohman
Factor out the code that determines the memory access type
Chris Lattner
Clarify wording, patch by Stefanus Du Toit!
Devang Patel
Add helper pass to remove llvm.dbg.declare intrinsics.
Dan Gohman
Move the sorting of the StrideOrder array earlier so that it doesn't
Dan Gohman
Delete the isOnlyStride argument, which is unused.
Dan Gohman
Tidy some LSR debug output: announce the loop it's about to process
Evan Cheng
ARM target now also recognize triplets like thumbv6-apple-darwin and set thumb mode and arch subversion. Eventually thumb triplets will go way and replaced with function notes.
Chris Lattner
Fix PR3763 by using proper APInt methods instead of uint64_t's.
Gabor Greif
make the Op<N> methods protected, how it was intended to be
Evan Cheng
ARM isLegalAddressImmediate should check if type is a simple type now that optimizer can create values of funky scalar types.
Evan Cheng
Yet another case where the spiller marked two uses of the same register on the same instruction as kill. This fixes PR3706.
Evan Cheng
Re-apply 66315, but restrict it to Darwin only.
Bill Wendling
Revert r66415. It's causing failures during bootstrap builds:
Bob Wilson
Add a constructor for BuildVectorSDNode and revert the temporary workaround
Duncan Sands
This debug info special case should no longer
Chris Lattner
just remove the use_empty() check entirely, the only reason it
Gabor Greif
in builds without asserts we do not need to allocate the Next pointer in "ghostly" sentinels
Nick Lewycky
Regenerate.
Nick Lewycky
Remove configurey-fu to autodetect hash_map and hash_set now that they are
  1. Prev
  2. Next