| Commit message (Collapse) | Author | Age | Files | Lines | ||
|---|---|---|---|---|---|---|
| ... | ||||||
| * | long double 9 of N. This finishes up the X86-32 bits | Dale Johannesen | 2007-08-09 | 1 | -1/+13 | |
| | | | | | | | | | (constants are still not handled). Adds ConvertActions to control fp-to-fp conversions (these are currently defaulted for all other targets, so no changes there). llvm-svn: 40958 | |||||
| * | Fix arguments for some Altivec instructions. From SWB. | Dale Johannesen | 2007-08-09 | 1 | -9/+15 | |
| | | | | | llvm-svn: 40957 | |||||
| * | Fix spelling of mtvscr and mfvscr. | Dale Johannesen | 2007-08-07 | 1 | -2/+2 | |
| | | | | | llvm-svn: 40908 | |||||
| * | Long double patch 8 of N: make it partially work in | Dale Johannesen | 2007-08-07 | 3 | -100/+100 | |
| | | | | | | | | | | | | SSE mode (all but conversions <-> other FP types, I think): >>Do not mark all-80-bit operations as "Requires[FPStack]" (which really means "not SSE"). >>Refactor load-and-extend to facilitate this. >>Update comments. >>Handle long double in SSE when computing FP_REG_KILL. llvm-svn: 40906 | |||||
| * | Initial JIT support for ARM by Raul Fernandes Herbster. | Evan Cheng | 2007-08-07 | 9 | -275/+889 | |
| | | | | | llvm-svn: 40887 | |||||
| * | Long double patch 7 of N, unless I lost count:). | Dale Johannesen | 2007-08-07 | 3 | -1/+37 | |
| | | | | | | | | | Last x87 bits for full functionality (not thoroughly tested, and long doubles do not work in SSE modes at all - use -mcpu=i486 for now) llvm-svn: 40886 | |||||
| * | Replace 4-line function with 10-line version per review comment. | Dale Johannesen | 2007-08-06 | 1 | -4/+10 | |
| | | | | | llvm-svn: 40881 | |||||
| * | Move lengthy conditional down 1 level per review comment. | Dale Johannesen | 2007-08-06 | 2 | -5/+6 | |
| | | | | | llvm-svn: 40878 | |||||
| * | Get X86 long double calling convention to work | Dale Johannesen | 2007-08-06 | 4 | -3/+34 | |
| | | | | | | | | (on Darwin, anyway). Fix some table omissions for LD arithmetic. llvm-svn: 40877 | |||||
| * | Make 80-bit store maintain simulated FP stack correctly. | Dale Johannesen | 2007-08-06 | 1 | -1/+2 | |
| | | | | | llvm-svn: 40868 | |||||
| * | Fix minor doxygen nits. | Reid Spencer | 2007-08-05 | 1 | -3/+3 | |
| | | | | | llvm-svn: 40854 | |||||
| * | Long double patch 4 of N: initial x87 implementation. | Dale Johannesen | 2007-08-05 | 8 | -6/+162 | |
| | | | | | | | Lots of problems yet but some simple things work. llvm-svn: 40847 | |||||
| * | This is the patch to provide clean intrinsic function overloading support in ↵ | Chandler Carruth | 2007-08-04 | 1 | -2/+2 | |
| | | | | | | | | | LLVM. It cleans up the intrinsic definitions and generally smooths the process for more complicated intrinsic writing. It will be used by the upcoming atomic intrinsics as well as vector and float intrinsics in the future. This also changes the syntax for llvm.bswap, llvm.part.set, llvm.part.select, and llvm.ct* intrinsics. They are automatically upgraded by both the LLVM ASM reader and the bitcode reader. The test cases have been updated, with special tests added to ensure the automatic upgrading is supported. llvm-svn: 40807 | |||||
| * | Make x86 long double alignment 32 for everything but | Dale Johannesen | 2007-08-03 | 1 | -1/+3 | |
| | | | | | | | Darwin (which makes size within a struct==96) llvm-svn: 40796 | |||||
| * | long double patch 2 of N. Handle it in TargetData. | Dale Johannesen | 2007-08-03 | 6 | -7/+19 | |
| | | | | | | | | (I've tried to get the info right for all targets, but I'm not expert on all of them - check yours.) llvm-svn: 40792 | |||||
| * | add an observation | Chris Lattner | 2007-08-03 | 1 | -0/+27 | |
| | | | | | llvm-svn: 40772 | |||||
| * | More explicit keywords. | Dan Gohman | 2007-08-02 | 6 | -6/+6 | |
| | | | | | llvm-svn: 40757 | |||||
| * | Fix the alignment requirements of several unpck and shuf instructions. | Dan Gohman | 2007-08-02 | 2 | -13/+21 | |
| | | | | | | | | | Generalize isPSHUFDMask and add a unary SHUFPD pattern so that SHUFPD's memory operand alignment can be tested as well, with a fix to avoid breaking MMX's use of isPSHUFDMask. llvm-svn: 40756 | |||||
| * | Fix pastos in vector arithmetic intrinsics. | Dan Gohman | 2007-08-02 | 1 | -4/+4 | |
| | | | | | llvm-svn: 40754 | |||||
| * | Mark the SSE and MMX load instructions that | Dan Gohman | 2007-08-02 | 2 | -0/+6 | |
| | | | | | | | | | | | X86InstrInfo::isReallyTriviallyReMaterializable knows how to handle with the isReMaterializable flag so that it is given a chance to handle them. Without hoisting constant-pool loads from loops this isn't very visible, though it does keep CodeGen/X86/constant-pool-remat-0.ll from making a copy of the constant pool on the stack. llvm-svn: 40736 | |||||
| * | Switch some multiplication instructions over to the new scheme for testing. | Evan Cheng | 2007-08-02 | 2 | -12/+53 | |
| | | | | | llvm-svn: 40723 | |||||
| * | Can't handle offset and scale if rip-relative addressing is to be used. | Evan Cheng | 2007-08-01 | 1 | -6/+10 | |
| | | | | | llvm-svn: 40703 | |||||
| * | Mac OS X X86-64 low 4G address not available. | Evan Cheng | 2007-08-01 | 1 | -4/+4 | |
| | | | | | llvm-svn: 40702 | |||||
| * | Mac OS X X86-64 low 4G address not available. | Evan Cheng | 2007-08-01 | 4 | -1/+21 | |
| | | | | | llvm-svn: 40701 | |||||
| * | Some out operands were incorrectly specified as input operands. | Evan Cheng | 2007-08-01 | 2 | -17/+17 | |
| | | | | | llvm-svn: 40697 | |||||
| * | Missing Requires. | Evan Cheng | 2007-08-01 | 1 | -2/+2 | |
| | | | | | llvm-svn: 40691 | |||||
| * | Be more precise. | Evan Cheng | 2007-08-01 | 1 | -1/+1 | |
| | | | | | llvm-svn: 40689 | |||||
| * | Change a .size directive to use a tab instead of a space, for consistency. | Dan Gohman | 2007-08-01 | 1 | -1/+1 | |
| | | | | | llvm-svn: 40672 | |||||
| * | Indexed loads each has 2 outputs. | Evan Cheng | 2007-08-01 | 1 | -10/+10 | |
| | | | | | llvm-svn: 40658 | |||||
| * | Change the x86 assembly output to use tab characters to separate the | Dan Gohman | 2007-07-31 | 5 | -1117/+1117 | |
| | | | | | | | | | | mnemonics from their operands instead of single spaces. This makes the assembly output a little more consistent with various other compilers (f.e. GCC), and slightly easier to read. Also, update the regression tests accordingly. llvm-svn: 40648 | |||||
| * | Redo and generalize previously removed opt for pinsrw: (vextract (v4i32 bc ↵ | Evan Cheng | 2007-07-31 | 6 | -235/+248 | |
| | | | | | | | (v4f32 s2v (f32 load ))), 0) -> (i32 load ) llvm-svn: 40628 | |||||
| * | This isn't safe when there are uses of load's chain result. | Evan Cheng | 2007-07-31 | 1 | -11/+4 | |
| | | | | | llvm-svn: 40617 | |||||
| * | Use tabs more consistently in assembler pseudo-ops. | Dan Gohman | 2007-07-30 | 2 | -5/+5 | |
| | | | | | llvm-svn: 40594 | |||||
| * | Fix the comment for getClosestTargetForJIT to reflect the fact that | Dan Gohman | 2007-07-30 | 1 | -4/+3 | |
| | | | | | | | it does not have a Module parameter. llvm-svn: 40590 | |||||
| * | More explicit keywords. | Dan Gohman | 2007-07-30 | 1 | -1/+1 | |
| | | | | | llvm-svn: 40589 | |||||
| * | Vector fneg must be expanded into fsub -0.0, X. | Evan Cheng | 2007-07-30 | 3 | -2/+21 | |
| | | | | | llvm-svn: 40586 | |||||
| * | Change the x86 backend to use extract_subreg for truncation operations. ↵ | Christopher Lamb | 2007-07-29 | 6 | -113/+37 | |
| | | | | | | | Passes DejaGnu, SingleSource and MultiSource. llvm-svn: 40578 | |||||
| * | Add register info needed to use subreg sets on X86. | Christopher Lamb | 2007-07-28 | 1 | -2/+51 | |
| | | | | | llvm-svn: 40572 | |||||
| * | Trampoline codegen support for X86-32. | Duncan Sands | 2007-07-27 | 5 | -12/+130 | |
| | | | | | llvm-svn: 40566 | |||||
| * | Re-apply 40504, but with a fix for the segfault it caused in oggenc: | Dan Gohman | 2007-07-27 | 3 | -75/+42 | |
| | | | | | | | | | | Make the alignedload and alignedstore patterns always require 16-byte alignment. This way when they are used in the "Fs" instructions, in which a vector instruction is used for a scalar purpose, they can still require the full vector alignment. And add a regression test for this. llvm-svn: 40555 | |||||
| * | Support for trampolines, except for X86 codegen which is | Duncan Sands | 2007-07-27 | 8 | -3/+34 | |
| | | | | | | | still under discussion. llvm-svn: 40549 | |||||
| * | Reverting 40504 for now. It's breaking oggenc. | Evan Cheng | 2007-07-27 | 3 | -35/+70 | |
| | | | | | llvm-svn: 40547 | |||||
| * | Make sure epilogue esp adjustment is placed before any terminator and pop ↵ | Evan Cheng | 2007-07-26 | 1 | -2/+3 | |
| | | | | | | | instructions. llvm-svn: 40538 | |||||
| * | Don't pollute the meaning of isUnpredicatedTerminator. | Evan Cheng | 2007-07-26 | 1 | -7/+12 | |
| | | | | | llvm-svn: 40537 | |||||
| * | Minor bug. | Evan Cheng | 2007-07-26 | 1 | -3/+3 | |
| | | | | | llvm-svn: 40535 | |||||
| * | In the .loc directive, print the fields as "debug" fields, so they | Dan Gohman | 2007-07-26 | 1 | -1/+1 | |
| | | | | | | | don't get decorated as if for immediate fields for instructions. llvm-svn: 40529 | |||||
| * | Fix a whitespace difference between CMPSSrr and CMPSDrr. | Dan Gohman | 2007-07-26 | 1 | -2/+1 | |
| | | | | | llvm-svn: 40528 | |||||
| * | Add target independent MachineInstr's to represent subreg insert/extract in ↵ | Christopher Lamb | 2007-07-26 | 1 | -0/+12 | |
| | | | | | | | MBB's. PR1350 llvm-svn: 40518 | |||||
| * | Same goes for constantpool, etc. | Evan Cheng | 2007-07-26 | 1 | -5/+5 | |
| | | | | | llvm-svn: 40517 | |||||
| * | Add selection DAG nodes for subreg insert/extract. PR1350 | Christopher Lamb | 2007-07-26 | 1 | -0/+5 | |
| | | | | | llvm-svn: 40516 | |||||

