| Commit message (Collapse) | Author | Age | Files | Lines | |
|---|---|---|---|---|---|
| * | Enable exception handling int JIT | Nicolas Geoffray | 2008-02-13 | 2 | -5/+23 |
| | | | | | llvm-svn: 47079 | ||||
| * | Fix the PPC JIT regressions by encoding zeroreg as 0 for BLR. | Chris Lattner | 2008-02-13 | 1 | -0/+1 |
| | | | | | llvm-svn: 47067 | ||||
| * | don't try to avoid inserting loads when lowering FORMAL_ARGUMENTS. | Chris Lattner | 2008-02-13 | 3 | -39/+17 |
| | | | | | | | DAGCombine is now quite good at zapifying them. llvm-svn: 47053 | ||||
| * | readme updates | Nate Begeman | 2008-02-13 | 2 | -0/+20 |
| | | | | | llvm-svn: 47051 | ||||
| * | Make register scavenging happy by not using a reg (CR0) that isn't defined | Nate Begeman | 2008-02-13 | 1 | -1/+1 |
| | | | | | llvm-svn: 47045 | ||||
| * | commuteInstr() can now commute non-ssa machine instrs. | Evan Cheng | 2008-02-13 | 2 | -0/+19 |
| | | | | | llvm-svn: 47043 | ||||
| * | Convert SelectionDAG::ComputeMaskedBits to use APInt instead of uint64_t. | Dan Gohman | 2008-02-13 | 9 | -40/+38 |
| | | | | | | | | Add an overload that supports the uint64_t interface for use by clients that haven't been updated yet. llvm-svn: 47039 | ||||
| * | __DATA not __DATA__ is the right segment name on darwin. | Dale Johannesen | 2008-02-12 | 2 | -2/+2 |
| | | | | | | | Spotted by Nick Kledzik. llvm-svn: 47037 | ||||
| * | Remove some dead code | Nate Begeman | 2008-02-12 | 1 | -7/+3 |
| | | | | | llvm-svn: 47036 | ||||
| * | SSE4.1 64b integer insert/extract pattern support | Nate Begeman | 2008-02-12 | 4 | -41/+73 |
| | | | | | | | Move formats into the formats file llvm-svn: 47035 | ||||
| * | Revert r46916 PPCTargetAsmInfo.cpp. | Evan Cheng | 2008-02-12 | 1 | -2/+2 |
| | | | | | llvm-svn: 47020 | ||||
| * | Only using x86-64 rip relative addressing in non-staic mode? | Evan Cheng | 2008-02-12 | 1 | -4/+8 |
| | | | | | llvm-svn: 47019 | ||||
| * | Update comment. | Evan Cheng | 2008-02-12 | 1 | -2/+2 |
| | | | | | llvm-svn: 47002 | ||||
| * | Unbreak various insert_vector_elt and extract_vector_elt tests in presence ↵ | Evan Cheng | 2008-02-12 | 1 | -4/+4 |
| | | | | | | | of SSE4. llvm-svn: 47001 | ||||
| * | Stuff noticed while grepping code | Nate Begeman | 2008-02-11 | 1 | -65/+7 |
| | | | | | llvm-svn: 46979 | ||||
| * | Enable SSE4 codegen and pattern matching. | Nate Begeman | 2008-02-11 | 5 | -25/+242 |
| | | | | | | | Add some notes to the README. llvm-svn: 46949 | ||||
| * | additional missing feature | Nate Begeman | 2008-02-11 | 1 | -0/+1 |
| | | | | | llvm-svn: 46948 | ||||
| * | xmm0 variable blends | Nate Begeman | 2008-02-10 | 1 | -2/+27 |
| | | | | | llvm-svn: 46931 | ||||
| * | Rename MRegisterInfo to TargetRegisterInfo. | Dan Gohman | 2008-02-10 | 54 | -104/+114 |
| | | | | | llvm-svn: 46930 | ||||
| * | Match GCC's behaviour for these sections. | Nick Lewycky | 2008-02-10 | 1 | -2/+2 |
| | | | | | llvm-svn: 46916 | ||||
| * | memopv16i8 had wrong alignment requirement, would have broken pabsb | Nate Begeman | 2008-02-09 | 1 | -129/+193 |
| | | | | | | | | | pabs{b,w,d} are not two address fix extract-to-mem sse4 ops add sse4 vector sign extend nodes llvm-svn: 46915 | ||||
| * | Skeleton of insert and extract matching, more to come | Nate Begeman | 2008-02-09 | 1 | -2/+62 |
| | | | | | llvm-svn: 46902 | ||||
| * | Tablegen support for insert & extract element matching | Nate Begeman | 2008-02-09 | 1 | -0/+16 |
| | | | | | llvm-svn: 46901 | ||||
| * | It's not always safe to fold movsd into xorpd, etc. Check the alignment of ↵ | Evan Cheng | 2008-02-08 | 14 | -33/+101 |
| | | | | | | | the load address first to make sure it's 16 byte aligned. llvm-svn: 46893 | ||||
| * | 64-bit (MMX) vectors do not need restrictive alignment. | Dale Johannesen | 2008-02-08 | 1 | -6/+4 |
| | | | | | | | 128-bit vectors need it only when SSE is on. llvm-svn: 46890 | ||||
| * | Avoid needlessly casting away const qualifiers. | Dan Gohman | 2008-02-08 | 4 | -8/+8 |
| | | | | | llvm-svn: 46877 | ||||
| * | Added missing entries in X86 load / store folding tables. | Evan Cheng | 2008-02-08 | 1 | -0/+8 |
| | | | | | llvm-svn: 46866 | ||||
| * | Follow Chris' suggestion; change the PseudoSourceValue accessors | Dan Gohman | 2008-02-07 | 2 | -15/+15 |
| | | | | | | | | to return pointers instead of references, since this is always what is needed. llvm-svn: 46857 | ||||
| * | Add SourceValue information for outgoing argument stores on x86. | Dan Gohman | 2008-02-07 | 1 | -2/+4 |
| | | | | | llvm-svn: 46854 | ||||
| * | Fix a x86-64 codegen deficiency. Allow gv + offset when using rip addressing ↵ | Evan Cheng | 2008-02-07 | 5 | -49/+30 |
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | mode. Before: _main: subq $8, %rsp leaq _X(%rip), %rax movsd 8(%rax), %xmm1 movss _X(%rip), %xmm0 call _t xorl %ecx, %ecx movl %ecx, %eax addq $8, %rsp ret Now: _main: subq $8, %rsp movsd _X+8(%rip), %xmm1 movss _X(%rip), %xmm0 call _t xorl %ecx, %ecx movl %ecx, %eax addq $8, %rsp ret Notice there is another idiotic codegen issue that needs to be fixed asap: xorl %ecx, %ecx movl %ecx, %eax llvm-svn: 46850 | ||||
| * | In some cases, e.g. ADD32ri, no transformation is made. Guide against it. | Evan Cheng | 2008-02-07 | 1 | -0/+2 |
| | | | | | llvm-svn: 46849 | ||||
| * | Re-apply the memory operand changes, with a fix for the static | Dan Gohman | 2008-02-06 | 6 | -94/+102 |
| | | | | | | | | | initializer problem, a minor tweak to the way the DAGISelEmitter finds load/store nodes, and a renaming of the new PseudoSourceValue objects. llvm-svn: 46827 | ||||
| * | Move to getCALLSEQ_END to ensure CALLSEQ_END node produces a flag. This is ↵ | Evan Cheng | 2008-02-05 | 1 | -2/+7 |
| | | | | | | | consistent with the definition in td file. llvm-svn: 46775 | ||||
| * | Implement sseregparm. | Dale Johannesen | 2008-02-05 | 2 | -4/+18 |
| | | | | | llvm-svn: 46764 | ||||
| * | Ident mnemonics appropriately | Nate Begeman | 2008-02-05 | 1 | -4/+4 |
| | | | | | llvm-svn: 46746 | ||||
| * | Dwarf requires variable entries to be in the source order. Right now, since ↵ | Evan Cheng | 2008-02-04 | 8 | -8/+0 |
| | | | | | | | we are recording variable information at isel time this means parameters would appear in the reverse order. The short term fix is to issue recordVariable() at asm printing time instead. llvm-svn: 46724 | ||||
| * | This method should be virtual | Nate Begeman | 2008-02-04 | 1 | -1/+1 |
| | | | | | llvm-svn: 46723 | ||||
| * | Eliminate some redundant code. | Nate Begeman | 2008-02-04 | 1 | -7/+1 |
| | | | | | llvm-svn: 46720 | ||||
| * | The rest of the SSE4.1 intrinsic patterns that are obvious to me. Getting | Nate Begeman | 2008-02-04 | 1 | -8/+45 |
| | | | | | | | Evan's help with the rest. llvm-svn: 46697 | ||||
| * | Some more SSE 4.1 intrinsic patterns. | Nate Begeman | 2008-02-04 | 1 | -31/+94 |
| | | | | | llvm-svn: 46696 | ||||
| * | SSE 4.1 Intrinsics and detection | Nate Begeman | 2008-02-03 | 6 | -1/+119 |
| | | | | | llvm-svn: 46681 | ||||
| * | explicitly include Compiler.h instead of getting it from tblgen in the ↵ | Chris Lattner | 2008-02-03 | 4 | -0/+4 |
| | | | | | | | middle of a class. llvm-svn: 46676 | ||||
| * | don't do ReplaceUses on a result that doesn't exist. | Chris Lattner | 2008-02-03 | 1 | -2/+4 |
| | | | | | llvm-svn: 46673 | ||||
| * | Get rid of the annoying blank lines before labels. | Evan Cheng | 2008-02-02 | 9 | -11/+1 |
| | | | | | llvm-svn: 46667 | ||||
| * | Don't use uninitialized values. Fixes vec_align.ll on X86 Linux. | Nick Lewycky | 2008-02-02 | 1 | -1/+1 |
| | | | | | llvm-svn: 46666 | ||||
| * | Unbreak ppc debug support. | Evan Cheng | 2008-02-02 | 1 | -0/+3 |
| | | | | | llvm-svn: 46665 | ||||
| * | SDIsel processes llvm.dbg.declare by recording the variable debug ↵ | Evan Cheng | 2008-02-02 | 10 | -2/+21 |
| | | | | | | | | | | information descriptor and its corresponding stack frame index in MachineModuleInfo. This only works if the local variable is "homed" in the stack frame. It does not work for byval parameter, etc. Added ISD::DECLARE node type to represent llvm.dbg.declare intrinsic. Now the intrinsic calls are lowered into a SDNode and lives on through out the codegen passes. For now, since all the debugging information recording is done at isel time, when a ISD::DECLARE node is selected, it has the side effect of also recording the variable. This is a short term solution that should be fixed in time. llvm-svn: 46659 | ||||
| * | Frame index can be negative. | Evan Cheng | 2008-02-02 | 3 | -3/+3 |
| | | | | | llvm-svn: 46655 | ||||
| * | CBackend: Implement unaligned load/store. | Lauro Ramos Venancio | 2008-02-01 | 1 | -14/+35 |
| | | | | | llvm-svn: 46646 | ||||
| * | Remove the nasty LABEL hack with a much less evil one. Now ↵ | Evan Cheng | 2008-02-01 | 2 | -20/+0 |
| | | | | | | | llvm.dbg.func.start implies a stoppoint is set. SelectionDAGISel records a new source line but does not create a ISD::LABEL node for this special stoppoint. Asm printer will magically print this label. This ensures nothing is emitted before. llvm-svn: 46635 | ||||

