Commit message (Collapse) | Author | Age | Files | Lines | ||
---|---|---|---|---|---|---|
... | ||||||
* | Handle visibility printing with all generality. Remove bunch of duplicate code. | Anton Korobeynikov | 2008-08-08 | 8 | -48/+36 | |
| | | | | llvm-svn: 54540 | |||||
* | Use chars, where possible | Anton Korobeynikov | 2008-08-08 | 1 | -30/+30 | |
| | | | | llvm-svn: 54539 | |||||
* | Convert PPC/Linux to new section printing stuff | Anton Korobeynikov | 2008-08-08 | 3 | -89/+74 | |
| | | | | llvm-svn: 54538 | |||||
* | Switch PPC/Darwin to new section handling stuff | Anton Korobeynikov | 2008-08-08 | 3 | -153/+107 | |
| | | | | llvm-svn: 54537 | |||||
* | Cleanup | Anton Korobeynikov | 2008-08-08 | 1 | -72/+70 | |
| | | | | llvm-svn: 54536 | |||||
* | Reduce the entries in a phi before testing it for deadness, because removing ↵ | Owen Anderson | 2008-08-08 | 1 | -7/+7 | |
| | | | | | | the entries might make it dead. llvm-svn: 54535 | |||||
* | Undo most of r54519. | Evan Cheng | 2008-08-08 | 6 | -21/+17 | |
| | | | | llvm-svn: 54534 | |||||
* | Don't call getAnalysisUsage unless -debug-pass is enabled. This speeds | Chris Lattner | 2008-08-08 | 3 | -31/+40 | |
| | | | | | | up the passmgr by avoiding useless work. llvm-svn: 54528 | |||||
* | Add new parameter Fast to createJIT to enable the fast codegen path. | Evan Cheng | 2008-08-08 | 7 | -17/+24 | |
| | | | | llvm-svn: 54523 | |||||
* | Add skeleton of simple basic block instruction selector. | Evan Cheng | 2008-08-08 | 3 | -3/+81 | |
| | | | | llvm-svn: 54522 | |||||
* | It's not legal to output a GV in a coalesced section if it's used in an ARM ↵ | Evan Cheng | 2008-08-08 | 6 | -24/+41 | |
| | | | | | | PIC relative constantpool. llvm-svn: 54519 | |||||
* | Fix indentation. | Evan Cheng | 2008-08-08 | 1 | -8/+8 | |
| | | | | llvm-svn: 54518 | |||||
* | Don't crash printing the asm for a ConstantExpr PtrToInt just because the int | Nick Lewycky | 2008-08-08 | 1 | -4/+5 | |
| | | | | | | | | is narrower than the pointer. This testcase emits: .byte (((17) - 16) & 255) llvm-svn: 54517 | |||||
* | Support added for ctlz intrinsic, test case added. | Bruno Cardoso Lopes | 2008-08-08 | 3 | -18/+19 | |
| | | | | llvm-svn: 54516 | |||||
* | Speed up the passmgr by avoiding heap thrashing on vectors. | Chris Lattner | 2008-08-08 | 2 | -14/+15 | |
| | | | | llvm-svn: 54515 | |||||
* | Match raw "psp" triple target, as done by the homebrew toolchain. | Bruno Cardoso Lopes | 2008-08-08 | 1 | -1/+2 | |
| | | | | llvm-svn: 54514 | |||||
* | Added Mips support for DYNAMIC_STACKALLOC | Bruno Cardoso Lopes | 2008-08-07 | 3 | -23/+52 | |
| | | | | | | | | | Fixed bug in adjustMipsStackFrame, which was breaking while trying to access a dead stack object index. Also added one more alignment before fixing the callee saved registers stack offset adjustment. llvm-svn: 54485 | |||||
* | Add the remaining fp_round libcalls: | Bruno Cardoso Lopes | 2008-08-07 | 3 | -2/+41 | |
| | | | | | | | | | | | | | FPROUND_F80_F32, FPROUND_PPCF128_F32, FPROUND_F80_F64, FPROUND_PPCF128_F64 Support for soften float fp_round operands is added, Mips needs this to round f64->f32. Also added support to soften float FABS result, Mips doesn't support double fabs results while in 'single float only' mode. llvm-svn: 54484 | |||||
* | Do a dominator walk when scheduling copies, rather than a DFS on the CFG. ↵ | Owen Anderson | 2008-08-07 | 1 | -14/+44 | |
| | | | | | | | | Also, fix a few problems when creating live intervals for temporaries created by phi elimination. llvm-svn: 54483 | |||||
* | Let SRETPromotion properly preserve the function name instead of (implicitly) | Matthijs Kooijman | 2008-08-07 | 1 | -1/+2 | |
| | | | | | | postfixing it with a number. llvm-svn: 54468 | |||||
* | Fix SRETPromotion, it was generating functions without returns statements since | Matthijs Kooijman | 2008-08-07 | 1 | -1/+1 | |
| | | | | | | | r53941 (but this was not noticed due to the lack of a basic test for SRETPromotion). llvm-svn: 54467 | |||||
* | Add some debug output to SRETPromotion. | Matthijs Kooijman | 2008-08-07 | 1 | -0/+4 | |
| | | | | llvm-svn: 54464 | |||||
* | Remove dead forward decl | Anton Korobeynikov | 2008-08-07 | 1 | -1/+0 | |
| | | | | llvm-svn: 54461 | |||||
* | Print section flags ok on platforms, which use '@' as comment string. Fix test. | Anton Korobeynikov | 2008-08-07 | 1 | -3/+9 | |
| | | | | llvm-svn: 54460 | |||||
* | Add assertion for easy debugging of missing stuff | Anton Korobeynikov | 2008-08-07 | 1 | -0/+2 | |
| | | | | llvm-svn: 54459 | |||||
* | Switch ARM to new section handling stuff | Anton Korobeynikov | 2008-08-07 | 5 | -236/+227 | |
| | | | | llvm-svn: 54458 | |||||
* | Switch Alpha to new section handling stuff | Anton Korobeynikov | 2008-08-07 | 1 | -63/+60 | |
| | | | | llvm-svn: 54457 | |||||
* | Use EmitAlignment consistently | Anton Korobeynikov | 2008-08-07 | 3 | -6/+4 | |
| | | | | llvm-svn: 54456 | |||||
* | Cleanup | Anton Korobeynikov | 2008-08-07 | 1 | -6/+6 | |
| | | | | llvm-svn: 54455 | |||||
* | Cleanup | Anton Korobeynikov | 2008-08-07 | 1 | -9/+9 | |
| | | | | llvm-svn: 54454 | |||||
* | Switch IA64 to new section-handling stuff | Anton Korobeynikov | 2008-08-07 | 3 | -71/+90 | |
| | | | | llvm-svn: 54453 | |||||
* | Cleanup | Anton Korobeynikov | 2008-08-07 | 1 | -18/+16 | |
| | | | | llvm-svn: 54452 | |||||
* | Provide convenient helpers | Anton Korobeynikov | 2008-08-07 | 2 | -10/+4 | |
| | | | | llvm-svn: 54451 | |||||
* | Switch Sparc to new section handling stuff. Refactor printing of ↵ | Anton Korobeynikov | 2008-08-07 | 4 | -77/+120 | |
| | | | | | | module-level GVs significantly. llvm-svn: 54450 | |||||
* | Add hook for constant pool section selection for darwin. | Anton Korobeynikov | 2008-08-07 | 2 | -5/+21 | |
| | | | | llvm-svn: 54449 | |||||
* | Select section for constant pool entries | Anton Korobeynikov | 2008-08-07 | 2 | -3/+23 | |
| | | | | llvm-svn: 54448 | |||||
* | Don't verify passes when assertions are disabled. | Chris Lattner | 2008-08-07 | 1 | -4/+7 | |
| | | | | llvm-svn: 54446 | |||||
* | Re-enable elimination of unnecessary SUBREG_TO_REG instructions in | Dan Gohman | 2008-08-07 | 3 | -12/+25 | |
| | | | | | | | | | | | LowerSubregs, and fix an x86-64 isel bug that this exposed. SUBREG_TO_REG for x86-64 implicit zero extension is only safe for isel to generate when the source is known to always have zeros in the high 32 bits. The EXTRACT_SUBREG instruction does not clear the high 32 bits. llvm-svn: 54444 | |||||
* | Rewrite JIT handling of GlobalVariables so they | Dale Johannesen | 2008-08-07 | 4 | -23/+179 | |
| | | | | | | | | | | | | | | are allocated in the same buffer as the code, jump tables, etc. The default JIT memory manager does not handle buffer overflow well. I didn't introduce this and I'm not attempting to fix it here, but it is more likely to be hit now since we're putting more stuff in the buffer. This affects one test that I know of so far, MultiSource/Benchmarks/NPB-serial/is. llvm-svn: 54442 | |||||
* | Factor code that finalize PHI nodes, jump tables, etc. out of ↵ | Evan Cheng | 2008-08-07 | 1 | -14/+23 | |
| | | | | | | SelectBasicBlock. No functionality changes. llvm-svn: 54438 | |||||
* | SDISel's constant branch folding can fold away self-loops, which doesn't ↵ | Owen Anderson | 2008-08-06 | 1 | -15/+40 | |
| | | | | | | | | result in any dead blocks, but rather an incorrect phi input. Add code to UnreachableMachineBlockElim to get rid of these entries. llvm-svn: 54432 | |||||
* | Correct handle cases where two phis are coalesced together, and correct ↵ | Owen Anderson | 2008-08-06 | 1 | -6/+39 | |
| | | | | | | | | break up the case where two different phis want to coalesce with the same vreg. llvm-svn: 54426 | |||||
* | Oops, didn't mean to commit this. | Owen Anderson | 2008-08-06 | 1 | -0/+2 | |
| | | | | llvm-svn: 54425 | |||||
* | We don't need to try to coalesce input vregs that are the same as the output ↵ | Owen Anderson | 2008-08-06 | 1 | -0/+7 | |
| | | | | | | vreg. llvm-svn: 54422 | |||||
* | Only trim a live interval if the register is not used after the PHI node. | Owen Anderson | 2008-08-06 | 1 | -2/+5 | |
| | | | | llvm-svn: 54421 | |||||
* | Only remap each VNInfo once when doing renumbering. | Owen Anderson | 2008-08-06 | 1 | -13/+15 | |
| | | | | llvm-svn: 54420 | |||||
* | Re-introduce the 8-bit subreg zext-inreg patterns for x86-32, | Dan Gohman | 2008-08-06 | 2 | -6/+10 | |
| | | | | | | | this time using MOV32to32_ and MOV16to16_. Thanks to Evan for suggesting this. llvm-svn: 54418 | |||||
* | Fix a shufflevector instcombine that was emitting invalid masks indices | Dan Gohman | 2008-08-06 | 1 | -3/+5 | |
| | | | | | | when it meant to be emitting undef indices. llvm-svn: 54417 | |||||
* | Fix PR2355: bug in ChangeCompareStride. When the loop termination compare is ↵ | Evan Cheng | 2008-08-06 | 1 | -2/+2 | |
| | | | | | | the only use of its iv stride, the stride can be eliminated by moving it to another stride. If the scale is negative, swap the predicate instead of using a inverse predicate. llvm-svn: 54415 | |||||
* | xchg does not modify FLAGS. | Dan Gohman | 2008-08-06 | 2 | -2/+3 | |
| | | | | llvm-svn: 54411 |