| Commit message (Collapse) | Author | Age | Files | Lines | |
|---|---|---|---|---|---|
| * | Fix PR1749 and InstCombine/2007-10-28-EmptyField.ll by handling | Chris Lattner | 2007-10-29 | 1 | -1/+7 |
| | | | | | | | zero-length fields better. llvm-svn: 43427 | ||||
| * | New entry. | Evan Cheng | 2007-10-28 | 1 | -0/+62 |
| | | | | | llvm-svn: 43420 | ||||
| * | Fix off-by-one stack offset computations (dwarf information) for callee-saved | Anton Korobeynikov | 2007-10-26 | 1 | -1/+2 |
| | | | | | | | | | registers in case, when FP pointer was eliminated. This should fixes misc. random EH-related crahses, when stuff is compiled with -fomit-frame-pointer. Thanks Duncan for nailing this bug! llvm-svn: 43381 | ||||
| * | clo/clz aren't supported on mips I. Keep them around for when we'll | Eric Christopher | 2007-10-26 | 1 | -105/+107 |
| | | | | | | | want them later (mips32/64). llvm-svn: 43380 | ||||
| * | Loosen up iv reuse to allow reuse of the same stride but a larger type when ↵ | Evan Cheng | 2007-10-26 | 3 | -19/+25 |
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | truncating from the larger type to smaller type is free. e.g. Turns this loop: LBB1_1: # entry.bb_crit_edge xorl %ecx, %ecx xorw %dx, %dx movw %dx, %si LBB1_2: # bb movl L_X$non_lazy_ptr, %edi movw %si, (%edi) movl L_Y$non_lazy_ptr, %edi movw %dx, (%edi) addw $4, %dx incw %si incl %ecx cmpl %eax, %ecx jne LBB1_2 # bb into LBB1_1: # entry.bb_crit_edge xorl %ecx, %ecx xorw %dx, %dx LBB1_2: # bb movl L_X$non_lazy_ptr, %esi movw %cx, (%esi) movl L_Y$non_lazy_ptr, %esi movw %dx, (%esi) addw $4, %dx incl %ecx cmpl %eax, %ecx jne LBB1_2 # bb llvm-svn: 43375 | ||||
| * | Support non-POSIX hosts by removing use of strncasecmp. | Dale Johannesen | 2007-10-25 | 1 | -44/+53 |
| | | | | | llvm-svn: 43364 | ||||
| * | Disable a couple more things for ppcf128. | Dale Johannesen | 2007-10-23 | 1 | -2/+4 |
| | | | | | llvm-svn: 43267 | ||||
| * | Temporary solution: added a different set of BCTRL_Macho / BCTRL_ELF with ↵ | Evan Cheng | 2007-10-23 | 3 | -8/+18 |
| | | | | | | | right callee-saved defs set for ppc64. llvm-svn: 43248 | ||||
| * | Fix memcpy lowering when addresses are 4-byte aligned but size is not ↵ | Evan Cheng | 2007-10-22 | 3 | -40/+86 |
| | | | | | | | multiple of 4. llvm-svn: 43234 | ||||
| * | Fix the folding of multiplication into addresses on x86, which was broken | Dan Gohman | 2007-10-22 | 1 | -0/+5 |
| | | | | | | | by the recent {U,S}MUL_LOHI changes. llvm-svn: 43230 | ||||
| * | Use ptr type in the immediate field of a BxA instruction so we don't end up ↵ | Evan Cheng | 2007-10-22 | 1 | -1/+2 |
| | | | | | | | selecting 32-bit call instruction for ppc64. llvm-svn: 43228 | ||||
| * | Fix an unfolding bug. | Evan Cheng | 2007-10-22 | 1 | -1/+2 |
| | | | | | llvm-svn: 43212 | ||||
| * | Allow for copysign having f80 second argument. | Dale Johannesen | 2007-10-21 | 1 | -0/+9 |
| | | | | | | | Fixes 5550319. llvm-svn: 43205 | ||||
| * | Resolve unfold tables ambiguity. | Evan Cheng | 2007-10-19 | 1 | -6/+7 |
| | | | | | llvm-svn: 43194 | ||||
| * | Local spiller optimization: | Evan Cheng | 2007-10-19 | 4 | -39/+114 |
| | | | | | | | | | | | | | | | | | Turn a store folding instruction into a load folding instruction. e.g. xorl %edi, %eax movl %eax, -32(%ebp) movl -36(%ebp), %eax orl %eax, -32(%ebp) => xorl %edi, %eax orl -36(%ebp), %eax mov %eax, -32(%ebp) This enables the unfolding optimization for a subsequent instruction which will also eliminate the newly introduced store instruction. llvm-svn: 43192 | ||||
| * | split LowerMEMCPY into LowerMEMCPYCall and LowerMEMCPYInline in the ARM backend. | Rafael Espindola | 2007-10-19 | 2 | -29/+70 |
| | | | | | llvm-svn: 43176 | ||||
| * | Add support for byval function whose argument is not 32 bit aligned. | Rafael Espindola | 2007-10-19 | 1 | -8/+17 |
| | | | | | | | | | | | To do this it is necessary to add a "always inline" argument to the memcpy node. For completeness I have also added this node to memmove and memset. I have also added getMem* functions, because the extra argument makes it cumbersome to use getNode and because I get confused by it :-) llvm-svn: 43172 | ||||
| * | comment fixes | Chris Lattner | 2007-10-19 | 1 | -3/+3 |
| | | | | | llvm-svn: 43168 | ||||
| * | Add an easy microoptimization I noticed. | Chris Lattner | 2007-10-19 | 1 | -0/+19 |
| | | | | | llvm-svn: 43164 | ||||
| * | More ppcf128 issues (maybe the last)? | Dale Johannesen | 2007-10-19 | 1 | -0/+7 |
| | | | | | llvm-svn: 43160 | ||||
| * | - Added getOpcodeAfterMemoryUnfold(). It doesn't unfold an instruction, but ↵ | Evan Cheng | 2007-10-18 | 14 | -28/+48 |
| | | | | | | | | | only returns the opcode of the instruction post unfolding. - Fix some copy+paste bugs. llvm-svn: 43153 | ||||
| * | Use SmallVectorImpl instead of SmallVector with hardcoded size in MRegister ↵ | Evan Cheng | 2007-10-18 | 14 | -62/+62 |
| | | | | | | | public interface. llvm-svn: 43150 | ||||
| * | Fix a misnamed parameter. | Christopher Lamb | 2007-10-18 | 1 | -1/+1 |
| | | | | | llvm-svn: 43145 | ||||
| * | Fix a typo | Christopher Lamb | 2007-10-18 | 1 | -1/+1 |
| | | | | | llvm-svn: 43144 | ||||
| * | Work around downrev gccs which do not inherit visibility of the | Gordon Henriksen | 2007-10-18 | 1 | -2/+2 |
| | | | | | | | Registry<>::iterator member class. llvm-svn: 43122 | ||||
| * | legalizing the ret operation on f64 shouldn't introduce a new | Chris Lattner | 2007-10-18 | 1 | -4/+6 |
| | | | | | | | i64 bit convert needlessly. llvm-svn: 43116 | ||||
| * | Switching TargetMachineRegistry to use the new generic Registry. | Gordon Henriksen | 2007-10-17 | 2 | -39/+18 |
| | | | | | llvm-svn: 43094 | ||||
| * | Change fp to sint legalization on x86-32 to do 2 x i32 | Chris Lattner | 2007-10-17 | 1 | -2/+12 |
| | | | | | | | loads instead of 1 x i64 loads. This doesn't change any functionality yet. llvm-svn: 43068 | ||||
| * | fix some funny indentation, add comments. | Chris Lattner | 2007-10-17 | 1 | -60/+61 |
| | | | | | llvm-svn: 43066 | ||||
| * | Check for invalid cc's in f80 select. | Dale Johannesen | 2007-10-16 | 1 | -0/+2 |
| | | | | | llvm-svn: 43033 | ||||
| * | Fix a bug handling frame references in ppc inline asm when the frame offset | Chris Lattner | 2007-10-16 | 1 | -24/+33 |
| | | | | | | | doesn't fit into 16 bits. llvm-svn: 43032 | ||||
| * | Correction to tail call optimization code. The new return address | Arnold Schwaighofer | 2007-10-16 | 1 | -11/+19 |
| | | | | | | | | | | | was stored to the acutal stack slot before the parameters were lowered to their stack slot. This could cause arguments to be overwritten by the return address if the called function had less parameters than the caller function. The update should remove the last failing test case of llc-beta: SPASS. llvm-svn: 43027 | ||||
| * | Change LowerFP_TO_SINT to create the specific code it needs instead of | Chris Lattner | 2007-10-15 | 1 | -4/+11 |
| | | | | | | | | | | | unconditionally creating an i64 bitcast. With the future legalizer design, operation legalization can't introduce new nodes with illegal types. This fixes the rest of olden on ppc32. llvm-svn: 43005 | ||||
| * | LowerFP_TO_SINT must not create a stack object if it's not needed. | Evan Cheng | 2007-10-15 | 1 | -6/+6 |
| | | | | | llvm-svn: 43004 | ||||
| * | Handle PPC long double in CBackend. | Dale Johannesen | 2007-10-15 | 1 | -0/+9 |
| | | | | | llvm-svn: 42972 | ||||
| * | Unbreak x86-64. | Evan Cheng | 2007-10-14 | 1 | -4/+3 |
| | | | | | llvm-svn: 42962 | ||||
| * | Revert 42908 for now. | Evan Cheng | 2007-10-14 | 8 | -51/+50 |
| | | | | | llvm-svn: 42960 | ||||
| * | Fix type mismatch error in PPC Altivec (only causes | Dale Johannesen | 2007-10-14 | 1 | -2/+2 |
| | | | | | | | a problem when asserts are on). From vecLib. llvm-svn: 42959 | ||||
| * | Clarify that fastcc has a problem with nested function | Duncan Sands | 2007-10-13 | 1 | -1/+1 |
| | | | | | | | trampolines, rather than with nested functions themselves. llvm-svn: 42955 | ||||
| * | Change unfoldMemoryOperand(). User is now responsible for passing in the | Evan Cheng | 2007-10-13 | 2 | -12/+19 |
| | | | | | | | | register used by the unfolded instructions. User can also specify whether to unfold the load, the store, or both. llvm-svn: 42946 | ||||
| * | Correcting the corrections. Bad bad baaad emacs! | Arnold Schwaighofer | 2007-10-12 | 1 | -1/+1 |
| | | | | | llvm-svn: 42935 | ||||
| * | Corrected many typing errors. And removed 'nest' parameter handling | Arnold Schwaighofer | 2007-10-12 | 9 | -37/+38 |
| | | | | | | | | for fastcc from X86CallingConv.td. This means that nested functions are not supported for calling convention 'fastcc'. llvm-svn: 42934 | ||||
| * | Due to the new tail call optimization, trampolines can no | Duncan Sands | 2007-10-12 | 1 | -1/+0 |
| | | | | | | | longer be created for fastcc functions. llvm-svn: 42925 | ||||
| * | Update. | Evan Cheng | 2007-10-12 | 1 | -18/+2 |
| | | | | | llvm-svn: 42922 | ||||
| * | Change the names used for internal labels to use the current | Dan Gohman | 2007-10-12 | 8 | -50/+51 |
| | | | | | | | | | | function symbol name instead of a codegen-assigned function number. Thanks Evan! :-) llvm-svn: 42908 | ||||
| * | Mark vector ctpop, cttz, and ctlz as Expand on x86. | Dan Gohman | 2007-10-12 | 1 | -0/+3 |
| | | | | | llvm-svn: 42905 | ||||
| * | Mark vector pow, ctpop, cttz, and ctlz as Expand on PowerPC. | Dan Gohman | 2007-10-12 | 1 | -1/+4 |
| | | | | | llvm-svn: 42904 | ||||
| * | Fold load / store into MOV32to32_ and MOV16to16_. | Evan Cheng | 2007-10-12 | 1 | -0/+4 |
| | | | | | llvm-svn: 42895 | ||||
| * | Flag MOV32to32_ with EXTRACT_SUBREG. They should not be scheduled apart. | Evan Cheng | 2007-10-12 | 1 | -7/+7 |
| | | | | | llvm-svn: 42894 | ||||
| * | Set ISD::FPOW to Expand. | Dan Gohman | 2007-10-11 | 6 | -4/+21 |
| | | | | | llvm-svn: 42881 | ||||

