| Commit message (Collapse) | Author | Age | Files | Lines | |
|---|---|---|---|---|---|
| * | Change the NEON VDUPfdf and VDUPfqf pseudo-instructions to actually be | Jim Grosbach | 2010-10-06 | 2 | -7/+28 |
| | | | | | | | pseudo instructions. llvm-svn: 115840 | ||||
| * | Add a 'pattern' arg to the ARM PseudoNeonI class. | Jim Grosbach | 2010-10-06 | 2 | -7/+9 |
| | | | | | llvm-svn: 115831 | ||||
| * | target operand flag values aren't a bitmask | Jim Grosbach | 2010-10-06 | 1 | -2/+2 |
| | | | | | llvm-svn: 115798 | ||||
| * | - Add TargetInstrInfo::getOperandLatency() to compute operand latencies. This | Evan Cheng | 2010-10-06 | 9 | -40/+298 |
| | | | | | | | | | | | | | | allow target to correctly compute latency for cases where static scheduling itineraries isn't sufficient. e.g. variable_ops instructions such as ARM::ldm. This also allows target without scheduling itineraries to compute operand latencies. e.g. X86 can return (approximated) latencies for high latency instructions such as division. - Compute operand latencies for those defined by load multiple instructions, e.g. ldm and those used by store multiple instructions, e.g. stm. llvm-svn: 115755 | ||||
| * | replace stuff like: | Chris Lattner | 2010-10-06 | 1 | -15/+11 |
| | | | | | | | | | | | | | | | let AsmString = !strconcat( !strconcat(!strconcat(opc, "${p}"), !strconcat(".", dt)), !strconcat("\t", asm)); with: let AsmString = !strconcat(opc, "${p}", ".", dt, "\t", asm); :) llvm-svn: 115720 | ||||
| * | Comment out fastisel debugging message. | Eric Christopher | 2010-10-05 | 1 | -1/+1 |
| | | | | | llvm-svn: 115717 | ||||
| * | Random cleanup and make the intermediate register in fptosi a | Eric Christopher | 2010-10-05 | 1 | -5/+5 |
| | | | | | | | | | 32-bit fp reg, not 64-bit. Fixes SingleSource. llvm-svn: 115711 | ||||
| * | Increase the number of bits used internally by the ARM target to represent the | Jim Grosbach | 2010-10-05 | 2 | -16/+16 |
| | | | | | | | addressing mode from four to five. llvm-svn: 115645 | ||||
| * | fix MSVC 2010 build. | Michael J. Spencer | 2010-10-05 | 1 | -1/+2 |
| | | | | | llvm-svn: 115594 | ||||
| * | Cleanup Whitespace. | Michael J. Spencer | 2010-10-05 | 1 | -11/+11 |
| | | | | | llvm-svn: 115593 | ||||
| * | Jim Asked us to move DataLayout on ARM back to the most specialized classes. Do | Rafael Espindola | 2010-10-03 | 4 | -38/+28 |
| | | | | | | | | | so and also change X86 for consistency. Investigating if this can be improved a bit. llvm-svn: 115469 | ||||
| * | Major changes to Cortex-A9 itinerary. | Evan Cheng | 2010-10-03 | 1 | -211/+251 |
| | | | | | | | | | | | | 1. Model dual issues as two FUs. 2. Model the pipelines correctly: two symmetric ALUs, the multiplier is a dependent pipeline on ALU0. The changes do not have much impact on codegen right now. But I plan to make pre-RA scheduler multi-issue aware which should take good advantage of the changes. llvm-svn: 115457 | ||||
| * | Start on lowering global addresses. | Eric Christopher | 2010-10-02 | 1 | -2/+43 |
| | | | | | llvm-svn: 115390 | ||||
| * | PrintSpecial() can go away now. | Jim Grosbach | 2010-10-01 | 1 | -1/+0 |
| | | | | | llvm-svn: 115376 | ||||
| * | Stub out constant GV handling, fixes C++ eh tests. | Eric Christopher | 2010-10-01 | 1 | -2/+14 |
| | | | | | llvm-svn: 115375 | ||||
| * | Nuke the rest of the :comment references | Jim Grosbach | 2010-10-01 | 4 | -14/+10 |
| | | | | | llvm-svn: 115373 | ||||
| * | Nuke a bunch of no-longer-needed comment-only asm strings. | Jim Grosbach | 2010-10-01 | 1 | -57/+28 |
| | | | | | llvm-svn: 115370 | ||||
| * | Fix r115332: correctly model AGU / NEON mux. | Evan Cheng | 2010-10-01 | 1 | -133/+266 |
| | | | | | llvm-svn: 115365 | ||||
| * | Thread the determination of branch prediction hit rates back through the ↵ | Owen Anderson | 2010-10-01 | 4 | -13/+19 |
| | | | | | | | | | | if-conversion heuristic APIs. For now, stick with a constant estimate of 90% (branch predictors are good!), but we might find that we want to provide more nuanced estimates in the future. llvm-svn: 115364 | ||||
| * | Rename the AsmPrinter directory to InstPrinter for those targets that have | Jim Grosbach | 2010-10-01 | 6 | -2/+2 |
| | | | | | | | | been MC-ized for assembly printing. MSP430 is mostly so, but still has the asm printer and lowering code in the printer subdir for the moment. llvm-svn: 115360 | ||||
| * | Fix scheduling infor for vmovn and vshrn which I broke accidentially. | Evan Cheng | 2010-10-01 | 2 | -2/+2 |
| | | | | | llvm-svn: 115354 | ||||
| * | Add operand cycles for vldr / vstr. | Evan Cheng | 2010-10-01 | 2 | -8/+17 |
| | | | | | llvm-svn: 115353 | ||||
| * | Direct calls only for arm fast isel for now. | Eric Christopher | 2010-10-01 | 1 | -2/+3 |
| | | | | | llvm-svn: 115350 | ||||
| * | NEON scheduling info fix. vmov reg, reg are single cycle instructions. | Evan Cheng | 2010-10-01 | 4 | -15/+39 |
| | | | | | llvm-svn: 115344 | ||||
| * | Fix thinko on store instructions. Fixes test_indvars failure. | Eric Christopher | 2010-10-01 | 1 | -6/+6 |
| | | | | | llvm-svn: 115342 | ||||
| * | Make the spelling of the flags for old-style if-conversion heuristics ↵ | Owen Anderson | 2010-10-01 | 1 | -4/+4 |
| | | | | | | | consistent between ARM and Thumb2. llvm-svn: 115341 | ||||
| * | Provide an option to restore old-style if-conversion heuristics for Thumb2. | Owen Anderson | 2010-10-01 | 2 | -0/+35 |
| | | | | | llvm-svn: 115339 | ||||
| * | Per Cortex-A9 pipeline diagram. AGU (core load / store issue) and NEON/FP ↵ | Evan Cheng | 2010-10-01 | 1 | -151/+232 |
| | | | | | | | issue are multiplexed. Model it correctly. llvm-svn: 115332 | ||||
| * | grammar | Jim Grosbach | 2010-10-01 | 1 | -2/+2 |
| | | | | | llvm-svn: 115314 | ||||
| * | Implement double return values in calls. Fixes | Eric Christopher | 2010-10-01 | 1 | -9/+35 |
| | | | | | | | SingleSource/Regression/C/casts.c. llvm-svn: 115246 | ||||
| * | Temporarily add a flag to make it easier to compare the new-style ARM if | Owen Anderson | 2010-09-30 | 1 | -0/+19 |
| | | | | | | | conversion heuristics to the old-style ones. llvm-svn: 115239 | ||||
| * | Movement and cleanup. | Eric Christopher | 2010-09-30 | 1 | -26/+26 |
| | | | | | llvm-svn: 115225 | ||||
| * | Start of generalized call support for ARM fast isel. | Eric Christopher | 2010-09-30 | 1 | -4/+146 |
| | | | | | llvm-svn: 115203 | ||||
| * | Nuke a few more unused asm strings | Jim Grosbach | 2010-09-30 | 3 | -19/+19 |
| | | | | | llvm-svn: 115193 | ||||
| * | Move getPointerSize() to the base class since it's not dependent on MachO | Jim Grosbach | 2010-09-30 | 1 | -8/+4 |
| | | | | | | | vs. ELF llvm-svn: 115180 | ||||
| * | Remove extraneous ';' | Jim Grosbach | 2010-09-30 | 1 | -2/+2 |
| | | | | | llvm-svn: 115176 | ||||
| * | The asm strings are never used at all, so just nuke 'em entirely. | Jim Grosbach | 2010-09-30 | 3 | -14/+7 |
| | | | | | llvm-svn: 115160 | ||||
| * | Adds getPointerSize() to the AsmBackend which will be needed by the final patch | Kevin Enderby | 2010-09-30 | 1 | -0/+8 |
| | | | | | | | for the dwarf .loc support to emit dwarf line number tables. llvm-svn: 115153 | ||||
| * | 80 column fix | Jim Grosbach | 2010-09-30 | 1 | -1/+3 |
| | | | | | llvm-svn: 115149 | ||||
| * | Fix two tiny issues (ARM does not need COFF) and comment sanity. | Jason W Kim | 2010-09-30 | 1 | -2/+1 |
| | | | | | llvm-svn: 115147 | ||||
| * | trailing whitespace | Jim Grosbach | 2010-09-30 | 1 | -3/+3 |
| | | | | | llvm-svn: 115136 | ||||
| * | Remove misplaced ';'. Make buildbots happy, hopefully. | Jim Grosbach | 2010-09-30 | 1 | -1/+1 |
| | | | | | llvm-svn: 115135 | ||||
| * | Tiny patch for proof-of-concept cleanup of ARMAsmPrinter::EmitStartOfAsmFile() | Jason W Kim | 2010-09-30 | 1 | -1/+1 |
| | | | | | | | | Small test for sanity check of resulting ARM .s file. Tested against -r115129. llvm-svn: 115133 | ||||
| * | Go ahead and jump! | Jim Grosbach | 2010-09-30 | 3 | -40/+7 |
| | | | | | | | | Now that the MC lowering handles the expansion of the pseudos, kill the horrible blobs of text. llvm-svn: 115130 | ||||
| * | I added a new file ARMAsmBackend which stubs out in similar ways to | Jason W Kim | 2010-09-30 | 4 | -0/+153 |
| | | | | | | | | | | | the eqv X86 class. For now, I split the ELFARMAsmBackend from the DarwinARMAsmBackend (also mimicking X86) Tested against -r115126 llvm-svn: 115129 | ||||
| * | Now that the pseudos that needed this are all custom lowered, we can go back | Jim Grosbach | 2010-09-30 | 2 | -9/+1 |
| | | | | | | | to an empty PrintSpecial() llvm-svn: 115128 | ||||
| * | Nuke it from orbit. It's the only way to be sure. | Jim Grosbach | 2010-09-30 | 3 | -766/+22 |
| | | | | | | | (Kill the dead non-MC asm printer for the ARM target.) llvm-svn: 115127 | ||||
| * | ARM instruction itinerary fixes: | Evan Cheng | 2010-09-30 | 7 | -315/+475 |
| | | | | | | | | | 1. Cortex-a9 8-bit and 16-bit loads / stores AGU cycles are 1 cycle longer than 32-bit ones. 2. Cortex-a9 is out-of-order so model all read cycles as cycle 1. 3. Lots of other random fixes for A8 and A9. llvm-svn: 115121 | ||||
| * | Refactor arm fast isel libcall handling so that pieces can be used | Eric Christopher | 2010-09-29 | 1 | -63/+98 |
| | | | | | | | for generic call handling. llvm-svn: 115105 | ||||
| * | Model Cortex-a9 load to SUB, RSB, ADD, ADC, SBC, RSC, CMN, MVN, or CMP | Evan Cheng | 2010-09-29 | 4 | -23/+52 |
| | | | | | | | pipeline forwarding path. llvm-svn: 115098 | ||||

