Commits

Show all
Author Commit Message Labels Comments Date
Evan Cheng
Only try to sink immediate when TLI is not null. It needs to check if immediate would fit in target addressing field.
Tags
2 tags
Nick Lewycky
Regenerate.
Nick Lewycky
Remove tests for flex/bison which are no longer needed by LLVM.
Nick Lewycky
Don't sign extend the char when expanding char -> int during load(bitcast(char[4] to i32*)) evaluation.
Daniel Dunbar
Add makefile debugging target; use "make print-FOO" to print the value/definition/origin of FOO.
Richard Pennington
bug 3610: Floating point vaarg not softened.
Ted Kremenek
Add version of StringsEqualNoCase that takes two null-terminated C-strings and compares up to 'len' characters. I tend to screw up string comparison functions, so anyone who is interested please review this\!
Anton Korobeynikov
Drop bunch of half-working stuff in the ext_weak linkage support. Now we're using one gross, but quite robust hack :) (previous ones did not work, for example, when ext_weak symbol was used deep inside constant expression in the initializer). The proper fix of this problem will require some quite huge asmprinter changes and that's why was postponed. This fixes PR3629 by the way :)
Nick Lewycky
Pass different flags on different platforms when building PIC.
Evan Cheng
Add AddrModeMatcher.cpp
Evan Cheng
If two-address def is dead and the instruction does not define other registers, and it doesn't produce side effects, just delete the instruction.
Tags
checker-0.161
Evan Cheng
Teach LSR sink to sink the immediate portion of the common expression back into uses if they fit in address modes of all the uses.
Bill Wendling
Make sure this doesn't access .end() too.
Gabor Greif
fix and clean up a comment
Bill Wendling
Make sure we don't dereference the .end() of the container.
Chris Lattner
rename a function to indicate that it checks for profitability as well as legality. Make load sinking and gep sinking more careful: we only do it when it won't pessimize loads from the stack. This has the added benefit of not producing code that is unanalyzable to SROA.
Bill Wendling
Propagate more debug loc infos. This also includes some code cleaning.
Bill Wendling
We need to propagate the debug location information even when dealing with the prologue/epilogue.
Misha Brukman
* Fixed spelling * Linters now return their information instead of printing it, to enable easier unittesting * Added support for finding tabs in files, added to C++ linter
Dan Gohman
Fix a bug that David Greene found in the DAGCombiner's logic that checks whether it's safe to transform a store of a bitcast value into a store of the original value.
Misha Brukman
Removed trailing whitespace.
Misha Brukman
Removed trailing whitespace.
Misha Brukman
Removed trailing whitespace.
Misha Brukman
Keep the newline character at the end of the lines whose trailing whitespace we are deleting; otherwise, everything ends up on a single line.
Misha Brukman
Only strip the newline character at the end of the lines that we're considering for length and for trailing whitespace; otherwise, the whitespace themselves will also be removed.
Misha Brukman
Fixed lint errors: * Alphabetized #includes * Removed trailing whitespace * Wrapped or shortened lines over 80 chars
Bill Wendling
We have logic in there to emit a default debugging label at the beginning of a function. Emitting another label after the prologue messes up the debugging. We are doing that because the first DebugLoc object it sees is different from the previous, which was nothing. Check for this situation, and don't emit one if it's the first.
Evan Cheng
Fix strange logic in CollectIVUsers used to determine whether all uses are addresses, part 1. This fixes an obvious logic bug. Previously if the only in-loop use is a PHI, it would return AllUsesAreAddresses as true.
Dan Gohman
Simplify code and reduce indentation. No functionality change.
Dan Gohman
Fix 80-column violations.
  1. Prev
  2. Next