2010-07-01 |
Dan Gohman | Enable on-demand fast-isel. |
tree | commitdiff |
2010-07-01 |
Dan Gohman | Reapply r106422, splitting the code for materializing... |
tree | commitdiff |
2010-07-01 |
Dan Gohman | Don't use operator[] here, because it's not desirable... |
tree | commitdiff |
2010-07-01 |
Mikhail Glushenkov | Trailing whitespace. |
tree | commitdiff |
2010-07-01 |
Jakob Stoklund Olesen | Add memory operand folding support to InlineSpiller. |
tree | commitdiff |
2010-06-30 |
Jakob Stoklund Olesen | Add support for rematerialization to InlineSpiller. |
tree | commitdiff |
2010-06-30 |
Bill Wendling | Use the catch-all selectors we already found when conve... |
tree | commitdiff |
2010-06-30 |
Jim Grosbach | Handle array and vector typed parameters in sjljehprepa... |
tree | commitdiff |
2010-06-30 |
Jim Grosbach | grammar tweak in comment. |
tree | commitdiff |
2010-06-30 |
Jakob Stoklund Olesen | Some fool committed without testing (or even building... |
tree | commitdiff |
2010-06-30 |
Jakob Stoklund Olesen | Remember to track spill slot uses in VirtRegMap when... |
tree | commitdiff |
2010-06-30 |
Duncan Sands | Remove an unused variable. The call to getRoot has... |
tree | commitdiff |
2010-06-30 |
Gabor Greif | use ArgOperand API |
tree | commitdiff |
2010-06-30 |
Gabor Greif | use ArgOperand API |
tree | commitdiff |
2010-06-30 |
Gabor Greif | use CallSite::arg_end instead of CallInst::op_end |
tree | commitdiff |
2010-06-30 |
John Mosby | Remove trailing whitespace, no functionality changes. |
tree | commitdiff |
2010-06-30 |
Devang Patel | Do not construct DIE for already processed MDNode. |
tree | commitdiff |
2010-06-30 |
Jakob Stoklund Olesen | Use skipInstruction() as a simpler way of iterating... |
tree | commitdiff |
2010-06-30 |
Jakob Stoklund Olesen | Use clEnumValN macro to work around keyword clash |
tree | commitdiff |
2010-06-30 |
Devang Patel | Add variables into a scope before constructing scope... |
tree | commitdiff |
2010-06-29 |
Jakob Stoklund Olesen | Begin implementation of an inline spiller. |
tree | commitdiff |
2010-06-29 |
Bill Wendling | Revert r107205 and r107207. |
tree | commitdiff |
2010-06-29 |
Devang Patel | Print InlinedAt location. |
tree | commitdiff |
2010-06-29 |
Devang Patel | Print InlinedAt location. |
tree | commitdiff |
2010-06-29 |
Bill Wendling | Introducing the "linker_weak" linkage type. This will... |
tree | commitdiff |
2010-06-29 |
Devang Patel | Do not hardcode DW_AT_stmt_list value. |
tree | commitdiff |
2010-06-29 |
Jakob Stoklund Olesen | Fix the handling of partial redefines in the fast regis... |
tree | commitdiff |
2010-06-29 |
Bob Wilson | Fix a register scavenger crash when dealing with undefi... |
tree | commitdiff |
2010-06-29 |
Duncan Sands | It seems clear that this should return Changed. |
tree | commitdiff |
2010-06-29 |
Rafael Espindola | Add a VT argument to getMinimalPhysRegClass and replace... |
tree | commitdiff |
2010-06-29 |
Duncan Sands | getMachineBasicBlockAddress returns a uintptr_t - don... |
tree | commitdiff |
2010-06-29 |
Gabor Greif | use ArgOperand APIs |
tree | commitdiff |
2010-06-29 |
Duncan Sands | Remove initialized but otherwise unused variables. |
tree | commitdiff |
2010-06-29 |
Jim Grosbach | When processing loops for scheduling latencies (used... |
tree | commitdiff |
2010-06-29 |
Bob Wilson | Reapply my if-conversion cleanup from svn r106939 with... |
tree | commitdiff |
2010-06-28 |
Bob Wilson | Unlike other targets, ARM now uses BUILD_VECTORs post... |
tree | commitdiff |
2010-06-28 |
Devang Patel | Use DW_FORM_addr for DW_AT_entry_pc. |
tree | commitdiff |
2010-06-28 |
Dale Johannesen | In asm's, output operands with matching input constraints |
tree | commitdiff |
2010-06-28 |
Devang Patel | Include inlined function in list of processed subprograms. |
tree | commitdiff |
2010-06-28 |
Jim Grosbach | new, no longer brain-dead, r106907 |
tree | commitdiff |
2010-06-28 |
Jakob Stoklund Olesen | After physreg coalescing, physical registers might... |
tree | commitdiff |
2010-06-28 |
Jakob Stoklund Olesen | Add more special treatment for inline asm in RegAllocFast. |
tree | commitdiff |
2010-06-28 |
Devang Patel | Preserve deleted function's local variables' debug... |
tree | commitdiff |
2010-06-28 |
Gabor Greif | simplify: we have solid argument iterator range |
tree | commitdiff |
2010-06-28 |
Daniel Dunbar | Revert r106907, "make sure to handle dbg_value instruct... |
tree | commitdiff |
2010-06-28 |
Devang Patel | Remove dead code. |
tree | commitdiff |
2010-06-26 |
Rafael Espindola | When splitting a VAARG, remember its alignment. |
tree | commitdiff |
2010-06-26 |
Bob Wilson | Revert my if-conversion cleanup since it caused a bunch... |
tree | commitdiff |
2010-06-26 |
Benjamin Kramer | VNInfos don't need to be destructed anymore. |
tree | commitdiff |
2010-06-26 |
Bob Wilson | Clean up some problems with extra CFG edges being intro... |
tree | commitdiff |
2010-06-25 |
Jim Grosbach | make sure to handle dbg_value instructions in the middl... |
tree | commitdiff |
2010-06-25 |
Jakob Stoklund Olesen | Don't track kills in VNInfo. Use interval ends instead. |
tree | commitdiff |
2010-06-25 |
Evan Cheng | Change if-conversion block size limit checks to add... |
tree | commitdiff |
2010-06-25 |
Devang Patel | Collect debug info for optimized variables of inlined... |
tree | commitdiff |
2010-06-25 |
Jim Grosbach | 80 column and typo fix |
tree | commitdiff |
2010-06-25 |
Dale Johannesen | The hasMemory argument is irrelevant to how the argument |
tree | commitdiff |
2010-06-25 |
Bill Wendling | - Reapply r106066 now that the bzip2 build regression... |
tree | commitdiff |
2010-06-25 |
Bill Wendling | We should remove the live range from the destination... |
tree | commitdiff |
2010-06-25 |
Dale Johannesen | Cosmetic. |
tree | commitdiff |
2010-06-25 |
Duncan Sands | Remove variables which are assigned to but for which... |
tree | commitdiff |
2010-06-25 |
Gabor Greif | use ArgOperand accessors |
tree | commitdiff |
2010-06-25 |
Gabor Greif | use ArgOperand API (the simple part) |
tree | commitdiff |
2010-06-25 |
Gabor Greif | use ArgOperand API |
tree | commitdiff |
2010-06-25 |
Gabor Greif | use ArgOperand API |
tree | commitdiff |
2010-06-25 |
Gabor Greif | use ArgOperand API and CallSite to access arguments... |
tree | commitdiff |
2010-06-25 |
Gabor Greif | use ArgOperand API and CallSite to access arguments... |
tree | commitdiff |
2010-06-25 |
Gabor Greif | use ArgOperand API |
tree | commitdiff |
2010-06-25 |
Gabor Greif | prune an include |
tree | commitdiff |
2010-06-25 |
Dale Johannesen | Fix a case where an earlyclobber operand of an asm |
tree | commitdiff |
2010-06-24 |
Jakob Stoklund Olesen | Make sure all eliminated kills are removed from VNInfo... |
tree | commitdiff |
2010-06-24 |
Dan Gohman | Add some comments. |
tree | commitdiff |
2010-06-24 |
Dan Gohman | Teach EmitLiveInCopies to omit copies for unused virtua... |
tree | commitdiff |
2010-06-24 |
Bill Wendling | It's possible that a flag is added to the SDNode that... |
tree | commitdiff |
2010-06-24 |
Devang Patel | DBG_VALUE machine instruction pointing to undefined... |
tree | commitdiff |
2010-06-24 |
Jakob Stoklund Olesen | Don't return a std::vector in the Spiller interface... |
tree | commitdiff |
2010-06-24 |
Jakob Stoklund Olesen | Remove the now unused LiveIntervals::getVNInfoSourceReg(). |
tree | commitdiff |
2010-06-24 |
Jakob Stoklund Olesen | Teach AdjustCopiesBackFrom to also use CoalescerPair... |
tree | commitdiff |
2010-06-24 |
Jakob Stoklund Olesen | Remove the -fast-spill option. |
tree | commitdiff |
2010-06-24 |
Bill Wendling | Loosen up the requirements in the Horrible Hack(tm... |
tree | commitdiff |
2010-06-24 |
Jakob Stoklund Olesen | Replace a big gob of old coalescer logic with the new... |
tree | commitdiff |
2010-06-24 |
Jakob Stoklund Olesen | Print the LSBs of a SlotIndex symbolically using letter... |
tree | commitdiff |
2010-06-24 |
Dan Gohman | Simplify this code; switch lowering shouldn't produce... |
tree | commitdiff |
2010-06-24 |
Jakob Stoklund Olesen | Be more strict about subreg-to-subreg copies in Coalesc... |
tree | commitdiff |
2010-06-24 |
Jakob Stoklund Olesen | Verify that VNI kills are pointing to existing instruct... |
tree | commitdiff |
2010-06-24 |
Dan Gohman | Eliminate the other half of the BRCOND optimization... |
tree | commitdiff |
2010-06-24 |
Dan Gohman | Eliminate the first have of the optimization which... |
tree | commitdiff |
2010-06-24 |
Dan Gohman | Reapply r106634, now that the bug it exposed is fixed. |
tree | commitdiff |
2010-06-24 |
Dan Gohman | Optimize the "bit test" code path for switch lowering... |
tree | commitdiff |
2010-06-24 |
Jakob Stoklund Olesen | Revert "Replace a big gob of old coalescer logic with... |
tree | commitdiff |
2010-06-24 |
Jakob Stoklund Olesen | Replace a big gob of old coalescer logic with the new... |
tree | commitdiff |
2010-06-23 |
Bill Wendling | MorphNodeTo doesn't preserve the memory operands. Becau... |
tree | commitdiff |
2010-06-23 |
Daniel Dunbar | Revert r106263, "Fold the ShrinkDemandedOps pass into... |
tree | commitdiff |
2010-06-23 |
Jim Grosbach | Some targets don't require the fencing MEMBARRIER instr... |
tree | commitdiff |
2010-06-23 |
Jakob Stoklund Olesen | Add a few VNInfo data structure checks. |
tree | commitdiff |
2010-06-23 |
Daniel Dunbar | Revert r106066, "Create a more targeted fix for not... |
tree | commitdiff |
2010-06-22 |
Jakob Stoklund Olesen | Also convert SUBREG_TO_REG to a KILL when relevant... |
tree | commitdiff |
2010-06-22 |
Dan Gohman | Move PHIElimination's SplitCriticalEdge for MachineBasi... |
tree | commitdiff |
2010-06-22 |
Jakob Stoklund Olesen | Remove the SimpleJoin optimization from SimpleRegisterC... |
tree | commitdiff |
2010-06-22 |
Dan Gohman | Use pre-increment instead of post-increment when the... |
tree | commitdiff |
2010-06-22 |
Dan Gohman | When unfolding a load, avoid assuming which instruction... |
tree | commitdiff |
next |