| Commit message (Collapse) | Author | Age | Files | Lines | ||
|---|---|---|---|---|---|---|
| ... | ||||||
| * | - Add AVX form of all SSE2 logical instructions | Bruno Cardoso Lopes | 2010-06-30 | 2 | -22/+99 | |
| | | | | | | | - Add VEX encoding bits to x86 MRM0r-MRM7r llvm-svn: 107238 | |||||
| * | Do not construct DIE for already processed MDNode. | Devang Patel | 2010-06-30 | 1 | -1/+2 | |
| | | | | | llvm-svn: 107237 | |||||
| * | Use skipInstruction() as a simpler way of iterating over instructions using ↵ | Jakob Stoklund Olesen | 2010-06-30 | 1 | -10/+2 | |
| | | | | | | | SrcReg llvm-svn: 107234 | |||||
| * | Use clEnumValN macro to work around keyword clash | Jakob Stoklund Olesen | 2010-06-30 | 1 | -1/+1 | |
| | | | | | llvm-svn: 107233 | |||||
| * | Add variables into a scope before constructing scope DIE otherwise variables ↵ | Devang Patel | 2010-06-30 | 1 | -2/+2 | |
| | | | | | | | won't be included DIE tree. llvm-svn: 107228 | |||||
| * | Begin implementation of an inline spiller. | Jakob Stoklund Olesen | 2010-06-29 | 3 | -1/+147 | |
| | | | | | | | | | | | | InlineSpiller inserts loads and spills immediately instead of deferring to VirtRegMap. This is possible now because SlotIndexes allows instructions to be inserted and renumbered. This is work in progress, and is mostly a copy of TrivialSpiller so far. It works very well for functions that don't require spilling. llvm-svn: 107227 | |||||
| * | Add *several* AVX integer packed binop instructions | Bruno Cardoso Lopes | 2010-06-29 | 1 | -35/+109 | |
| | | | | | llvm-svn: 107225 | |||||
| * | Fix ScalarEvolution's tripcount computation for chains of loops | Dan Gohman | 2010-06-29 | 1 | -41/+61 | |
| | | | | | | | | where each loop's induction variable's start value is the exit value of a preceding loop. llvm-svn: 107224 | |||||
| * | Revert r107205 and r107207. | Bill Wendling | 2010-06-29 | 17 | -46/+14 | |
| | | | | | llvm-svn: 107215 | |||||
| * | Print InlinedAt location. | Devang Patel | 2010-06-29 | 1 | -7/+21 | |
| | | | | | llvm-svn: 107214 | |||||
| * | Add another bswap idiom that isn't matched. | Eric Christopher | 2010-06-29 | 1 | -0/+8 | |
| | | | | | llvm-svn: 107213 | |||||
| * | Move SSE2 Packed Integer instructions around, and create specific sections ↵ | Bruno Cardoso Lopes | 2010-06-29 | 1 | -83/+113 | |
| | | | | | | | for each of them llvm-svn: 107211 | |||||
| * | Print InlinedAt location. | Devang Patel | 2010-06-29 | 1 | -12/+23 | |
| | | | | | llvm-svn: 107208 | |||||
| * | Add AVX Move Aligned/Unaligned packed integers | Bruno Cardoso Lopes | 2010-06-29 | 1 | -12/+53 | |
| | | | | | llvm-svn: 107206 | |||||
| * | Introducing the "linker_weak" linkage type. This will be used for Objective-C | Bill Wendling | 2010-06-29 | 17 | -14/+46 | |
| | | | | | | | | | | | | | | | | | | | | metadata types which should be marked as "weak", but which the linker will remove upon final linkage. For example, the "objc_msgSend_fixup_alloc" symbol is defined like this: .globl l_objc_msgSend_fixup_alloc .weak_definition l_objc_msgSend_fixup_alloc .section __DATA, __objc_msgrefs, coalesced .align 3 l_objc_msgSend_fixup_alloc: .quad _objc_msgSend_fixup .quad L_OBJC_METH_VAR_NAME_1 This is different from the "linker_private" linkage type, because it can't have the metadata defined with ".weak_definition". llvm-svn: 107205 | |||||
| * | Add AVX ld/st XCSR register. | Bruno Cardoso Lopes | 2010-06-29 | 2 | -15/+26 | |
| | | | | | | | Add VEX encoding bits for MRMXm x86 form llvm-svn: 107204 | |||||
| * | Do not hardcode DW_AT_stmt_list value. | Devang Patel | 2010-06-29 | 2 | -7/+17 | |
| | | | | | | | Inspired by Artur Pietrek. llvm-svn: 107202 | |||||
| * | Add support for encoding VDUP (ARM core register) instructions. | Bob Wilson | 2010-06-29 | 1 | -0/+17 | |
| | | | | | llvm-svn: 107201 | |||||
| * | Fix the handling of partial redefines in the fast register allocator. | Jakob Stoklund Olesen | 2010-06-29 | 1 | -17/+39 | |
| | | | | | | | | | | | | A partial redefine needs to be treated like a tied operand, and the register must be reloaded while processing use operands. This fixes a bug where partially redefined registers were processed as normal defs with a reload added. The reload could clobber another use operand if it was a kill that allowed register reuse. llvm-svn: 107193 | |||||
| * | Fix a register scavenger crash when dealing with undefined subregs. | Bob Wilson | 2010-06-29 | 1 | -0/+18 | |
| | | | | | | | | The LowerSubregs pass needs to preserve implicit def operands attached to EXTRACT_SUBREG instructions when it replaces those instructions with copies. llvm-svn: 107189 | |||||
| * | Add AVX non-temporal stores | Bruno Cardoso Lopes | 2010-06-29 | 1 | -10/+52 | |
| | | | | | llvm-svn: 107178 | |||||
| * | Fix whitespace style. | Dan Gohman | 2010-06-29 | 1 | -2/+2 | |
| | | | | | llvm-svn: 107175 | |||||
| * | Move non-temporal movs to their own section | Bruno Cardoso Lopes | 2010-06-29 | 1 | -34/+38 | |
| | | | | | llvm-svn: 107168 | |||||
| * | Add support for encoding NEON VMOV (from core register to scalar) instructions. | Bob Wilson | 2010-06-29 | 1 | -6/+19 | |
| | | | | | | | | The encoding is the same as VMOV (from scalar to core register) except that the operands are in different places. llvm-svn: 107167 | |||||
| * | Add sqrt, rsqrt and rcp AVX instructions | Bruno Cardoso Lopes | 2010-06-29 | 1 | -1/+67 | |
| | | | | | llvm-svn: 107166 | |||||
| * | skip dbg_value instructions | Jim Grosbach | 2010-06-29 | 1 | -0/+2 | |
| | | | | | llvm-svn: 107154 | |||||
| * | The t2MOVi16 and t2MOVTi16 instructions do not set CPSR. Trying to add | Bob Wilson | 2010-06-29 | 1 | -2/+2 | |
| | | | | | | | | a CPSR operand to them causes an assertion failure, so apparently these instructions haven't been getting a lot of use. llvm-svn: 107147 | |||||
| * | use ArgOperand API | Gabor Greif | 2010-06-29 | 1 | -6/+6 | |
| | | | | | llvm-svn: 107145 | |||||
| * | Return Changed. This required setting Changed if dbg metadata | Duncan Sands | 2010-06-29 | 1 | -5/+7 | |
| | | | | | | | | | is stripped off. Currently set unconditionally, since the API does not provide a way of working out if anything was actually stripped off. llvm-svn: 107142 | |||||
| * | It seems clear that this should return Changed. | Duncan Sands | 2010-06-29 | 1 | -1/+1 | |
| | | | | | llvm-svn: 107141 | |||||
| * | Add a VT argument to getMinimalPhysRegClass and replace the copy related uses | Rafael Espindola | 2010-06-29 | 6 | -9/+15 | |
| | | | | | | | | | | of getPhysicalRegisterRegClass with it. If we want to make a copy (or estimate its cost), it is better to use the smallest class as more efficient operations might be possible. llvm-svn: 107140 | |||||
| * | getMachineBasicBlockAddress returns a uintptr_t - don't truncate | Duncan Sands | 2010-06-29 | 1 | -1/+1 | |
| | | | | | | | | to unsigned only to extend back to a pointer sized value on the next line. llvm-svn: 107139 | |||||
| * | The variable ValueSize is set to 1 on both code paths, and then | Duncan Sands | 2010-06-29 | 1 | -5/+1 | |
| | | | | | | | ignored! Remove it. llvm-svn: 107138 | |||||
| * | The variable "Value" is carefully set to Layout.getSymbolAddress, | Duncan Sands | 2010-06-29 | 1 | -4/+0 | |
| | | | | | | | but then not actually used - maybe a bug? Remove the variable. llvm-svn: 107137 | |||||
| * | Remove unused calls to Lexer.getLoc and the pointless variable HasFillExpr. | Duncan Sands | 2010-06-29 | 1 | -7/+0 | |
| | | | | | llvm-svn: 107136 | |||||
| * | Remove pointless variable LastDef. | Duncan Sands | 2010-06-29 | 1 | -2/+0 | |
| | | | | | llvm-svn: 107135 | |||||
| * | Remove unused variable Loc and pointless variables unified_syntax | Duncan Sands | 2010-06-29 | 1 | -15/+4 | |
| | | | | | | | and thumb_mode. llvm-svn: 107133 | |||||
| * | use ArgOperand APIs | Gabor Greif | 2010-06-29 | 1 | -3/+3 | |
| | | | | | llvm-svn: 107132 | |||||
| * | Remove an unused and a pointless variable. | Duncan Sands | 2010-06-29 | 1 | -3/+0 | |
| | | | | | llvm-svn: 107131 | |||||
| * | Remove pointless and unused variables. | Duncan Sands | 2010-06-29 | 1 | -9/+0 | |
| | | | | | llvm-svn: 107130 | |||||
| * | encode operand initializations (at fixed index) | Gabor Greif | 2010-06-29 | 1 | -12/+8 | |
| | | | | | | | | | | in terms of Op<> and ArgOffset. This works for values of {0, 1} for ArgOffset. Please note that ArgOffset will become 0 soon and will go away eventually. llvm-svn: 107129 | |||||
| * | Remove a pointless variable. | Duncan Sands | 2010-06-29 | 1 | -5/+1 | |
| | | | | | llvm-svn: 107128 | |||||
| * | Remove initialized but otherwise unused variables. | Duncan Sands | 2010-06-29 | 7 | -9/+0 | |
| | | | | | llvm-svn: 107127 | |||||
| * | Use a more obvious way to avoid compiling functions which are only used when ↵ | Benjamin Kramer | 2010-06-29 | 1 | -3/+2 | |
| | | | | | | | XDEBUG is enabled. llvm-svn: 107125 | |||||
| * | Jump through some silly hoops to make GCC accept that a function may not always | Chandler Carruth | 2010-06-29 | 1 | -0/+3 | |
| | | | | | | | be called. llvm-svn: 107124 | |||||
| * | PR7503: uxtb16 is not available for ARMv7-M. Patch by Brian G. Lucas. | Evan Cheng | 2010-06-29 | 1 | -2/+2 | |
| | | | | | llvm-svn: 107122 | |||||
| * | Change if-cvt options to something that actually as useable. | Evan Cheng | 2010-06-29 | 1 | -4/+6 | |
| | | | | | llvm-svn: 107121 | |||||
| * | When processing loops for scheduling latencies (used for live outs on loop | Jim Grosbach | 2010-06-29 | 1 | -1/+4 | |
| | | | | | | | | back-edges), make sure not to include dbg_value instructions in the count. Closing in on the end of rdar://7797940 llvm-svn: 107119 | |||||
| * | Just as its not safe to blindly transfer the nsw bit from an add | Dan Gohman | 2010-06-29 | 1 | -5/+9 | |
| | | | | | | | | | instruction to an add scev, it's not safe to blindly transfer the inbounds flag from a gep instruction to an nsw on the scev for the gep. llvm-svn: 107117 | |||||
| * | Refactoring of arithmetic instruction classes with unary operator | Bruno Cardoso Lopes | 2010-06-29 | 1 | -118/+60 | |
| | | | | | llvm-svn: 107116 | |||||

