2010-05-18 |
Jakob Stoklund... | Teach the machine code verifier to use getSubRegisterRe... |
commit | commitdiff | tree | snapshot |
2010-05-18 |
Daniel Dunbar | MC/Mach-O: Implement support for setting indirect symbo... |
commit | commitdiff | tree | snapshot |
2010-05-18 |
Daniel Dunbar | MC/Mach-O: Remove some FIXMEs. |
commit | commitdiff | tree | snapshot |
2010-05-18 |
Daniel Dunbar | MC/Mach-O: Fail faster/harder when we see .file, which... |
commit | commitdiff | tree | snapshot |
2010-05-18 |
Daniel Dunbar | MC/X86: Implement custom lowering to make sure we match... |
commit | commitdiff | tree | snapshot |
2010-05-18 |
Daniel Dunbar | llc (et al): Add support for --show-encoding and -... |
commit | commitdiff | tree | snapshot |
2010-05-18 |
Dan Gohman | Usage of O_NONBLOCK in bjam is now confirmed as a bug... |
commit | commitdiff | tree | snapshot |
2010-05-18 |
Benjamin Kramer | Simplify MCContext::(Next|Get)Instance |
commit | commitdiff | tree | snapshot |
2010-05-18 |
Sean Hunt | Replace FIRST_* and LAST_* macros with a generic STMT_R... |
commit | commitdiff | tree | snapshot |
2010-05-18 |
Evan Cheng | FIX PR7158. SimplifyVBinOp was asserting when it fails... |
commit | commitdiff | tree | snapshot |
2010-05-17 |
Jakob Stoklund... | ARMBaseRegisterInfo::estimateRSStackSizeLimit() could... |
commit | commitdiff | tree | snapshot |
2010-05-17 |
Evan Cheng | Fix PR7175. Insert copies of a REG_SEQUENCE source... |
commit | commitdiff | tree | snapshot |
2010-05-17 |
Bill Wendling | - Set the "HasCalls" flag after instruction selection... |
commit | commitdiff | tree | snapshot |
2010-05-17 |
Kevin Enderby | Added support in MC for Directional Local Labels. |
commit | commitdiff | tree | snapshot |
2010-05-17 |
Eric Christopher | More data/parsing support for tls directives. Add... |
commit | commitdiff | tree | snapshot |
2010-05-17 |
Evan Cheng | Fix PR7156. If the sources of a REG_SEQUENCE are all... |
commit | commitdiff | tree | snapshot |
2010-05-17 |
Evan Cheng | vmov of immediates are trivially re-materializable. |
commit | commitdiff | tree | snapshot |
2010-05-17 |
Daniel Dunbar | MC/Mach-O/x86: Optimal nop sequences should only be... |
commit | commitdiff | tree | snapshot |
2010-05-17 |
Daniel Dunbar | MC: Add dyn_cast support to MCSection. |
commit | commitdiff | tree | snapshot |
2010-05-17 |
Daniel Dunbar | MC/Mach-O: Reverse order of SymbolData scanning when... |
commit | commitdiff | tree | snapshot |
2010-05-17 |
Jakob Stoklund... | Pull the UsedInInstr.test() calls into calcSpillCost... |
commit | commitdiff | tree | snapshot |
2010-05-17 |
Eric Christopher | Add some section and constant support for darwin TLS. |
commit | commitdiff | tree | snapshot |
2010-05-17 |
Evan Cheng | Careful with reg_sequence coalescing to not to overwrit... |
commit | commitdiff | tree | snapshot |
2010-05-17 |
Bob Wilson | Fix a regression in 464.h264 for thumb1 and thumb2... |
commit | commitdiff | tree | snapshot |
2010-05-17 |
Daniel Dunbar | MC/Mach-O: Fix some differences in symbol flag handling. |
commit | commitdiff | tree | snapshot |
2010-05-17 |
Jakob Stoklund... | Optimize empty DenseMap iteration. |
commit | commitdiff | tree | snapshot |
2010-05-17 |
Jakob Stoklund... | Remove debug option. Add comment on spill order determi... |
commit | commitdiff | tree | snapshot |
2010-05-17 |
Evan Cheng | Turn on -neon-reg-sequence by default. |
commit | commitdiff | tree | snapshot |
2010-05-17 |
Daniel Dunbar | llvm-mc: Support reassignment of variables in one speci... |
commit | commitdiff | tree | snapshot |
2010-05-17 |
Jakob Stoklund... | Avoid allocating the same physreg to multiple virtregs... |
commit | commitdiff | tree | snapshot |
2010-05-17 |
Jakob Stoklund... | Minor optimizations. DenseMap::begin() is surprisingly... |
commit | commitdiff | tree | snapshot |
2010-05-17 |
Jakob Stoklund... | Extract spill cost calculation to a new method, and... |
commit | commitdiff | tree | snapshot |
2010-05-17 |
Zhongxing Xu | Remove unused member variable. |
commit | commitdiff | tree | snapshot |
2010-05-17 |
Jakob Stoklund... | Only use clairvoyance when defining a register, and... |
commit | commitdiff | tree | snapshot |
2010-05-17 |
Jakob Stoklund... | Eliminate a hash table probe when killing virtual regis... |
commit | commitdiff | tree | snapshot |
2010-05-17 |
Jakob Stoklund... | Execute virtreg kills immediately instead of after... |
commit | commitdiff | tree | snapshot |
2010-05-17 |
Jakob Stoklund... | Sprinkle superregister <imp-def> and <imp-kill> operand... |
commit | commitdiff | tree | snapshot |
2010-05-17 |
Jakob Stoklund... | Now that we don't keep live registers across calls... |
commit | commitdiff | tree | snapshot |
2010-05-17 |
Jakob Stoklund... | Boldly attempt consistent capitalization. Functional... |
commit | commitdiff | tree | snapshot |
2010-05-17 |
Eric Christopher | Assume that we'll handle mangling the symbols earlier... |
commit | commitdiff | tree | snapshot |
2010-05-17 |
Jakob Stoklund... | Spill and kill all virtual registers across a call. |
commit | commitdiff | tree | snapshot |
2010-05-17 |
Jakob Stoklund... | Reduce hashtable probes by using DenseMap::insert(... |
commit | commitdiff | tree | snapshot |
2010-05-17 |
Jakob Stoklund... | Make MBB a class member instead of passing it around... |
commit | commitdiff | tree | snapshot |
2010-05-17 |
Evan Cheng | Yes, if the redef is a copy, update the old val# with... |
commit | commitdiff | tree | snapshot |
2010-05-17 |
Evan Cheng | No reason not to run the NEON domain croassing fix... |
commit | commitdiff | tree | snapshot |
2010-05-16 |
Dale Johannesen | Removing as part of previous reversion. |
commit | commitdiff | tree | snapshot |
2010-05-16 |
Dale Johannesen | Revert 103911; it broke a test that expects bitconvert |
commit | commitdiff | tree | snapshot |
2010-05-16 |
Dale Johannesen | Make x86-64 64-bit bitconvert work when SSE is not... |
commit | commitdiff | tree | snapshot |
2010-05-16 |
Anton Korobeynikov | Chris said that the comment char should be escaped... |
commit | commitdiff | tree | snapshot |
2010-05-16 |
Anton Korobeynikov | Add support for thiscall calling convention. |
commit | commitdiff | tree | snapshot |
2010-05-16 |
Anton Korobeynikov | Generalize the ARM DAG combiner of mul with constants... |
commit | commitdiff | tree | snapshot |
2010-05-16 |
Evan Cheng | Model vst lane instructions with REG_SEQUENCE. |
commit | commitdiff | tree | snapshot |
2010-05-16 |
Rafael Espindola | Avoid renaming loadable modules at install time. Now... |
commit | commitdiff | tree | snapshot |
2010-05-16 |
Rafael Espindola | Use $ORIGIN in the rpath of libraries. |
commit | commitdiff | tree | snapshot |
2010-05-15 |
Dale Johannesen | Fix uint64->{float, double} conversion to do rounding... |
commit | commitdiff | tree | snapshot |
2010-05-15 |
Dale Johannesen | Improve assertion messages. |
commit | commitdiff | tree | snapshot |
2010-05-15 |
Anton Korobeynikov | Some cheap DAG combine goodness for multiplication... |
commit | commitdiff | tree | snapshot |
2010-05-15 |
Anton Korobeynikov | "trap" pseudo-op turned out to be apple-local. |
commit | commitdiff | tree | snapshot |
2010-05-15 |
Chris Lattner | improve portability to systems that don't have round... |
commit | commitdiff | tree | snapshot |
2010-05-15 |
Chris Lattner | improve portability to systems that don't have powf... |
commit | commitdiff | tree | snapshot |
2010-05-15 |
Chandler Carruth | Fix an GCC warning that seems to have actually caught... |
commit | commitdiff | tree | snapshot |
2010-05-15 |
Evan Cheng | Model 128-bit vld lane with REG_SEQUENCE. |
commit | commitdiff | tree | snapshot |
2010-05-15 |
Jakob Stoklund... | Calculate liveness on the fly for local registers. |
commit | commitdiff | tree | snapshot |
2010-05-15 |
Nick Lewycky | Teach the always inliner to release its inline cost... |
commit | commitdiff | tree | snapshot |
2010-05-15 |
Nick Lewycky | Clean up, no functional change. |
commit | commitdiff | tree | snapshot |
2010-05-15 |
Jakob Stoklund... | Add reg_nodbg_iterator |
commit | commitdiff | tree | snapshot |
2010-05-15 |
Evan Cheng | v4i64 and v8i64 are only synthesizable when NEON is... |
commit | commitdiff | tree | snapshot |
2010-05-15 |
Evan Cheng | Allow TargetLowering::getRegClassFor() to be called... |
commit | commitdiff | tree | snapshot |
2010-05-15 |
Evan Cheng | Model 64-bit lane vld with REG_SEQUENCE. |
commit | commitdiff | tree | snapshot |
2010-05-15 |
Evan Cheng | A partial re-def instruction may be a copy. |
commit | commitdiff | tree | snapshot |
2010-05-14 |
Evan Cheng | Teach two-address pass to do some coalescing while... |
commit | commitdiff | tree | snapshot |
2010-05-14 |
Dan Gohman | Add some comments about undef. |
commit | commitdiff | tree | snapshot |
2010-05-14 |
Evan Cheng | Model VST*_UPD and VST*oddUPD pair with REG_SEQUENCE. |
commit | commitdiff | tree | snapshot |
2010-05-14 |
Dan Gohman | Fast ISel trivially coalesces away no-op casts, so... |
commit | commitdiff | tree | snapshot |
2010-05-14 |
Jakob Stoklund... | Don't bother spilling before a return |
commit | commitdiff | tree | snapshot |
2010-05-14 |
Jakob Stoklund... | RegAllocLocal can count copies too |
commit | commitdiff | tree | snapshot |
2010-05-14 |
Bill Wendling | SystemZ really does mean "has calls" and not just ... |
commit | commitdiff | tree | snapshot |
2010-05-14 |
Jakob Stoklund... | Track allocatable instead of reserved regs, and never... |
commit | commitdiff | tree | snapshot |
2010-05-14 |
Dan Gohman | Don't set kill flags for instructions which the schedul... |
commit | commitdiff | tree | snapshot |
2010-05-14 |
Dan Gohman | BR is a barrier. |
commit | commitdiff | tree | snapshot |
2010-05-14 |
Bill Wendling | Several tail call tests apparently rely upon this being... |
commit | commitdiff | tree | snapshot |
2010-05-14 |
Jakob Stoklund... | Avoid scanning the long tail of physreg operands on... |
commit | commitdiff | tree | snapshot |
2010-05-14 |
Devang Patel | Do not forget to mark prcessed arguments. |
commit | commitdiff | tree | snapshot |
2010-05-14 |
Jakob Stoklund... | Count coalesced copies |
commit | commitdiff | tree | snapshot |
2010-05-14 |
Jakob Stoklund... | Allow virtreg redefines when verifying for RegAllocFast |
commit | commitdiff | tree | snapshot |
2010-05-14 |
Bill Wendling | This should happen if there are no calls, not if it... |
commit | commitdiff | tree | snapshot |
2010-05-14 |
Bill Wendling | Revert r103804. The comment is correct. |
commit | commitdiff | tree | snapshot |
2010-05-14 |
Jim Grosbach | Remove trailing whitespace |
commit | commitdiff | tree | snapshot |
2010-05-14 |
Jim Grosbach | 80 column and trailing whitespace cleanup |
commit | commitdiff | tree | snapshot |
2010-05-14 |
Jim Grosbach | add cmd line option to leave dbgvalues in during post... |
commit | commitdiff | tree | snapshot |
2010-05-14 |
Bill Wendling | Fix comment. |
commit | commitdiff | tree | snapshot |
2010-05-14 |
Bill Wendling | Rename "HasCalls" in MachineFrameInfo to "AdjustsStack... |
commit | commitdiff | tree | snapshot |
2010-05-14 |
Devang Patel | Test case for r103800. |
commit | commitdiff | tree | snapshot |
2010-05-14 |
Dan Gohman | Lowering of atomic instructions can result in operands... |
commit | commitdiff | tree | snapshot |
2010-05-14 |
Devang Patel | Add support to preserve type info for the variables... |
commit | commitdiff | tree | snapshot |
2010-05-14 |
Jakob Stoklund... | When verifying two-address instructions, check the... |
commit | commitdiff | tree | snapshot |
2010-05-14 |
Kevin Enderby | Fix so "int3" is correctly accepted, added "into" and... |
commit | commitdiff | tree | snapshot |
2010-05-14 |
Evan Cheng | Model VLD*_UPD and VLD*odd_UPD pair with REG_SEQUENCE. |
commit | commitdiff | tree | snapshot |
2010-05-14 |
Daniel Dunbar | MC/Mach-O/x86_64: Darwin's special "signed_N" relocatio... |
commit | commitdiff | tree | snapshot |
2010-05-14 |
Jakob Stoklund... | Simplify the handling of physreg defs and uses in RegAl... |
commit | commitdiff | tree | snapshot |
next |