2009-08-11 |
Jim Grosbach | Move ~ARMConstantPoolValue() to the .cpp file to avoid... |
tree | commitdiff |
2009-08-11 |
Evan Cheng | Fix the previous accidental commit. Now shrinking commo... |
tree | commitdiff |
2009-08-11 |
Evan Cheng | Fix Thumb2 load / store addressing mode matching code... |
tree | commitdiff |
2009-08-11 |
Evan Cheng | 80 column violation. |
tree | commitdiff |
2009-08-11 |
Evan Cheng | Cosmetic changes. |
tree | commitdiff |
2009-08-11 |
Evan Cheng | Adding a blank line back. |
tree | commitdiff |
2009-08-11 |
Bob Wilson | Use vAny type to get rid of Neon intrinsics that differ... |
tree | commitdiff |
2009-08-11 |
Bob Wilson | Use new EVT::vAny type to combine Neon intrinsics for... |
tree | commitdiff |
2009-08-11 |
David Goodwin | Fix bug in NEON convert for single-precision FP. This... |
tree | commitdiff |
2009-08-11 |
Jim Grosbach | Add stdlib.h |
tree | commitdiff |
2009-08-11 |
Jim Grosbach | SjLj based exception handling unwinding support. This... |
tree | commitdiff |
2009-08-10 |
Evan Cheng | Enable Thumb2 instruction shrinking (32-bit to 16-bit... |
tree | commitdiff |
2009-08-10 |
Dan Gohman | Fix a bug where DAGCombine was producing an illegal... |
tree | commitdiff |
2009-08-10 |
Owen Anderson | Rename MVT to EVT, in preparation for splitting SimpleV... |
tree | commitdiff |
2009-08-10 |
David Goodwin | Use NEON for single-precision int<->FP conversions. |
tree | commitdiff |
2009-08-10 |
Owen Anderson | SimpleValueType-ify a few more methods on TargetLowering. |
tree | commitdiff |
2009-08-10 |
Evan Cheng | Handle the constantfp created during post-legalization... |
tree | commitdiff |
2009-08-10 |
Owen Anderson | Continue the SimpleValueType-ification. |
tree | commitdiff |
2009-08-10 |
Chris Lattner | split MachO section handling stuff out to its out ... |
tree | commitdiff |
2009-08-10 |
Chris Lattner | arm only needs to emit one .align directive for hidden... |
tree | commitdiff |
2009-08-10 |
Chris Lattner | make sure that arm nonlazypointers are aligned properly |
tree | commitdiff |
2009-08-10 |
David Goodwin | Checkpoint scheduling itinerary changes. |
tree | commitdiff |
2009-08-10 |
Evan Cheng | Watch out for empty BB. |
tree | commitdiff |
2009-08-10 |
Evan Cheng | rev, rev16, and revsh do not set CPSR. |
tree | commitdiff |
2009-08-10 |
Evan Cheng | Duh. Most 16-bit Thumb rr instructions are two-address... |
tree | commitdiff |
2009-08-10 |
Evan Cheng | CPSR can be livein; transfer predicate operands correct... |
tree | commitdiff |
2009-08-10 |
Evan Cheng | Add support for folding loads / stores into 16-bit... |
tree | commitdiff |
2009-08-10 |
Evan Cheng | 80 col violation. |
tree | commitdiff |
2009-08-10 |
Evan Cheng | Use tMOVgpr2gpr instead of t2MOVr. |
tree | commitdiff |
2009-08-10 |
Evan Cheng | Add support to reduce most of 32-bit Thumb2 arithmetic... |
tree | commitdiff |
2009-08-10 |
Evan Cheng | Always use the 16-bit tMOVgpr2gpr instead of the 32... |
tree | commitdiff |
2009-08-10 |
Chris Lattner | Make the big switch: Change MCSectionMachO to represent... |
tree | commitdiff |
2009-08-09 |
Evan Cheng | Add support to convert 32-bit instructions to 16-bit... |
tree | commitdiff |
2009-08-08 |
Anton Korobeynikov | Use subclassing to print lane-like immediates (w/o... |
tree | commitdiff |
2009-08-08 |
Chris Lattner | 1. Make MCSection an abstract class. |
tree | commitdiff |
2009-08-08 |
Chris Lattner | now that getOrCreateSection is all object-file specific, |
tree | commitdiff |
2009-08-08 |
Daniel Dunbar | Update CMake |
tree | commitdiff |
2009-08-08 |
Anton Korobeynikov | Add insert_elt / extract_elt patterns for v4f32 stuff. |
tree | commitdiff |
2009-08-08 |
Anton Korobeynikov | Lane number should be printed w/o hash |
tree | commitdiff |
2009-08-08 |
Anton Korobeynikov | Use VLDM / VSTM to spill/reload 128-bit Neon registers |
tree | commitdiff |
2009-08-08 |
Bob Wilson | Implement Neon VZIP and VUZP instructions. These are... |
tree | commitdiff |
2009-08-08 |
Bob Wilson | Implement Neon VTRN instructions. For now, anyway... |
tree | commitdiff |
2009-08-08 |
Evan Cheng | Add a skeleton Thumb2 instruction size reduction pass. |
tree | commitdiff |
2009-08-08 |
Evan Cheng | Code refactoring. No functionality change. |
tree | commitdiff |
2009-08-08 |
Evan Cheng | tADDhirr should target GPR, not tGPR. |
tree | commitdiff |
2009-08-08 |
Evan Cheng | I can type. |
tree | commitdiff |
2009-08-08 |
Chris Lattner | make printInstruction return void since its result... |
tree | commitdiff |
2009-08-07 |
David Goodwin | Make NEON single-precision FP support the default for... |
tree | commitdiff |
2009-08-07 |
Anton Korobeynikov | Unbreak the stuff |
tree | commitdiff |
2009-08-07 |
Anton Korobeynikov | 2 more vdup.32 cases |
tree | commitdiff |
2009-08-07 |
Evan Cheng | A big oops. Thumb1 default CC is a def of CPSR, not... |
tree | commitdiff |
2009-08-07 |
Evan Cheng | Thumb2 32-bit ldm / stm needs .w suffix if submode... |
tree | commitdiff |
2009-08-07 |
Evan Cheng | This is done. |
tree | commitdiff |
2009-08-07 |
Evan Cheng | Use 16-bit tMOVgpr2gpr instead of tMOVr to copy GPR... |
tree | commitdiff |
2009-08-07 |
Evan Cheng | Fix support to use NEON for single precision fp math. |
tree | commitdiff |
2009-08-07 |
Evan Cheng | Error out, rather than infinite looping, if constant... |
tree | commitdiff |
2009-08-07 |
Evan Cheng | tBfar is bl, which clobbers LR. |
tree | commitdiff |
2009-08-07 |
Dan Gohman | Fix a bunch of namespace pollution. |
tree | commitdiff |
2009-08-07 |
Evan Cheng | It turns out most of the thumb2 instructions are not... |
tree | commitdiff |
2009-08-06 |
Bob Wilson | Implement Neon VST[234] operations. |
tree | commitdiff |
2009-08-06 |
David Goodwin | Add parameter to pattern classes to enable an itinerary... |
tree | commitdiff |
2009-08-06 |
Bob Wilson | Neon does not actually have VLD{234}.64 instructions. |
tree | commitdiff |
2009-08-05 |
Bob Wilson | Add a new pre-allocation pass to assign adjacent regist... |
tree | commitdiff |
2009-08-05 |
David Goodwin | When using NEON for single-precision FP, the NEON resul... |
tree | commitdiff |
2009-08-05 |
Anton Korobeynikov | Remove redundand checks: the only way to have, e.g... |
tree | commitdiff |
2009-08-05 |
Anton Korobeynikov | Unbreak the stuff, this is ugly, but we cannot do bette... |
tree | commitdiff |
2009-08-05 |
Anton Korobeynikov | Missed pieces for ARM HardFP ABI. |
tree | commitdiff |
2009-08-05 |
Daniel Dunbar | Remove some dead code. |
tree | commitdiff |
2009-08-05 |
Bob Wilson | Remove a redundant declaration. |
tree | commitdiff |
2009-08-05 |
David Goodwin | Disable NEON single-precision FP support for Cortex... |
tree | commitdiff |
2009-08-05 |
Devang Patel | Remove dead code. MDNode and MDString are not Constant... |
tree | commitdiff |
2009-08-05 |
David Goodwin | By default, for cortex-a8 use NEON for single-precision FP. |
tree | commitdiff |
2009-08-05 |
Evan Cheng | 80 col violations. |
tree | commitdiff |
2009-08-05 |
Bob Wilson | Oops. I didn't mean to commit this piece yet. |
tree | commitdiff |
2009-08-05 |
Dan Gohman | Major calling convention code refactoring. |
tree | commitdiff |
2009-08-05 |
Dan Gohman | Don't flush the raw_ostream between each MachineFunctio... |
tree | commitdiff |
2009-08-05 |
Bob Wilson | Change DAG nodes for Neon VLD2/3/4 operations to return... |
tree | commitdiff |
2009-08-04 |
Evan Cheng | Fix part 1 of pr4682. PICADD is a 16-bit instruction... |
tree | commitdiff |
2009-08-04 |
Bob Wilson | Replace dregsingle operand modifier with explicit escap... |
tree | commitdiff |
2009-08-04 |
Evan Cheng | Enable load / store multiple pass for Thumb2. It's... |
tree | commitdiff |
2009-08-04 |
David Goodwin | Add NEON single-precision FP support for fabs and fneg. |
tree | commitdiff |
2009-08-04 |
Evan Cheng | In thumb mode, r7 is used as frame register. This fixes... |
tree | commitdiff |
2009-08-04 |
David Goodwin | Match common pattern for FNMAC. Add NEON SP support. |
tree | commitdiff |
2009-08-04 |
David Goodwin | Initial support for single-precision FP using NEON... |
tree | commitdiff |
2009-08-04 |
Anton Korobeynikov | Ooops, I was too fast to commit the wrong fix :( |
tree | commitdiff |
2009-08-04 |
Anton Korobeynikov | Fix a typo - this unbreaks llvm-gcc build on arm |
tree | commitdiff |
2009-08-04 |
Evan Cheng | Thumb2 does not have ib (increment before) and da ... |
tree | commitdiff |
2009-08-04 |
Evan Cheng | Remove ARM specific getInlineAsmLength. We'll rely... |
tree | commitdiff |
2009-08-04 |
Evan Cheng | Load / store multiple pass fixes for Thumb2. Not enable... |
tree | commitdiff |
2009-08-04 |
Evan Cheng | Emit sub r, #c instead of transforming it to add r... |
tree | commitdiff |
2009-08-04 |
Bob Wilson | Lower Neon VLD* intrinsics to custom DAG nodes, and... |
tree | commitdiff |
2009-08-04 |
Bob Wilson | Minor cleanup. No functional changes intended. |
tree | commitdiff |
2009-08-03 |
Chris Lattner | use TLOF to compute the section for a function instead of |
tree | commitdiff |
2009-08-03 |
Chris Lattner | convert macho stub emission to use SwitchToSection... |
tree | commitdiff |
2009-08-03 |
Bob Wilson | Lower CONCAT_VECTOR during legalization instead of... |
tree | commitdiff |
2009-08-03 |
Benjamin Kramer | llvm_report_error already prints "LLVM ERROR:". So... |
tree | commitdiff |
2009-08-03 |
Evan Cheng | These are done. |
tree | commitdiff |
2009-08-03 |
Evan Cheng | Use the i12 variant of load / store opcodes if offset... |
tree | commitdiff |
2009-08-02 |
Daniel Dunbar | Move most targets TargetMachine constructor to only... |
tree | commitdiff |
2009-08-02 |
Daniel Dunbar | Normalize Subtarget constructors to take a target tripl... |
tree | commitdiff |
next |