summaryrefslogtreecommitdiffstats
path: root/llvm
Commit message (Collapse)AuthorAgeFilesLines
* Instruction formats added used to generate multiply instructions of V5TE.Raul Herbster2007-08-301-39/+52
| | | | llvm-svn: 41629
* Unused relocation type reloc_arm_absolute removed.Raul Herbster2007-08-301-2/+0
| | | | llvm-svn: 41628
* Comments added. It now generates V5TE multiply instructions. However, it is ↵Raul Herbster2007-08-301-58/+175
| | | | | | still necessary to model PUWLSH bits more clearly. llvm-svn: 41627
* ARM instruction table was modified by adding information to generate ↵Raul Herbster2007-08-301-42/+57
| | | | | | multiply instruction of V5TE. llvm-svn: 41626
* JITInfo now resolves function addrs and also relocations. It always emits a ↵Raul Herbster2007-08-301-30/+43
| | | | | | stub. llvm-svn: 41625
* Added Graphviz escaping for the '|' character.Ted Kremenek2007-08-301-1/+1
| | | | llvm-svn: 41610
* Do not run on darwin.Tanya Lattner2007-08-301-2/+2
| | | | llvm-svn: 41608
* Added support to fold X86 load / store instructions. This allow ↵Evan Cheng2007-08-303-50/+101
| | | | | | rematerialized loads to be folded into their uses. llvm-svn: 41599
* Try fold re-materialized load instructions into its uses.Evan Cheng2007-08-301-11/+22
| | | | llvm-svn: 41598
* Add a variant of foldMemoryOperand to fold any load / store, not just load / ↵Evan Cheng2007-08-307-5/+42
| | | | | | store from / to stack slots. llvm-svn: 41597
* Added CreateFrameIndex to create a FrameIndex MachineOperand without a ↵Evan Cheng2007-08-301-0/+12
| | | | | | MachineInstr. llvm-svn: 41596
* Mark load instructions with isLoad = 1.Evan Cheng2007-08-305-6/+21
| | | | llvm-svn: 41595
* Use i64 on a PPC64 machineBill Wendling2007-08-301-1/+1
| | | | llvm-svn: 41590
* Change LegalFPImmediates to use APFloat.Dale Johannesen2007-08-3011-35/+141
| | | | | | | | | Add APFloat interfaces to ConstantFP, SelectionDAG. Fix integer bit in double->APFloat conversion. Convert LegalizeDAG to use APFloat interface in ConstantFPSDNode uses. llvm-svn: 41587
* Fix use of declaration inside case blockAnton Korobeynikov2007-08-291-1/+2
| | | | llvm-svn: 41584
* Change LiveRange so it keeps a pointer to the VNInfo rather than an index.Evan Cheng2007-08-295-352/+316
| | | | | | | Changes related modules so VNInfo's are not copied. This decrease copy coalescing time by 45% and overall compilation time by 10% on siod. llvm-svn: 41579
* Lower FRAME_TO_ADDR_OFFSET to zero by default (if not custom lowered)Anton Korobeynikov2007-08-291-1/+13
| | | | llvm-svn: 41578
* Move getX86RegNum into X86RegisterInfo and use itDuncan Sands2007-08-296-88/+80
| | | | | | | | in the trampoline lowering. Lookup the jump and mov opcodes for the trampoline rather than hard coding them. llvm-svn: 41577
* Alphabetize Raul's entry.Raul Herbster2007-08-291-3/+4
| | | | llvm-svn: 41576
* New contributor addedRaul Herbster2007-08-291-1/+3
| | | | llvm-svn: 41569
* Updating configure script to enable MIPS.Tanya Lattner2007-08-291-25/+30
| | | | llvm-svn: 41568
* Silence implicit 64->32-bit conversion warnings.Chris Lattner2007-08-291-20/+6
| | | | llvm-svn: 41567
* getMinSignedBits needs to take into consider the sign bit when the value is ↵Chris Lattner2007-08-291-1/+1
| | | | | | positive. llvm-svn: 41566
* update theseChris Lattner2007-08-292-2/+2
| | | | llvm-svn: 41565
* Add an option, -view-sunit-dags, for viewing the actual SUnit DAGs used byDan Gohman2007-08-283-1/+145
| | | | | | scheduling. llvm-svn: 41556
* Update win32 VStudio project files to keep pace with recently added and ↵Chuck Rose III2007-08-284-16/+27
| | | | | | deleted files llvm-svn: 41553
* Recover most of the compile time regression due to recent live interval changes.Evan Cheng2007-08-284-98/+100
| | | | | | | | 1. Eliminate the costly live interval "swapping". 2. Change ValueNumberInfo container from SmallVector to std::vector. The former performs slowly when the vector size is very large. llvm-svn: 41536
* Added method to get Mips register numbersBruno Cardoso Lopes2007-08-283-61/+113
| | | | | | | Changed the stack frame layout, StackGrowsUp fits better to Mips strange stack. Stack offset calculation bug fixed! llvm-svn: 41529
* Changed stack allocation On LowerFORMAL_ARGUMENTS.Bruno Cardoso Lopes2007-08-281-20/+29
| | | | | | | Added comments about new stack allocation. Expand SelectCC for i32 results llvm-svn: 41527
* Mask directive completed with CalleeSave infoBruno Cardoso Lopes2007-08-282-73/+111
| | | | | | Comments for Mips directives added. llvm-svn: 41526
* Added methods to record SPOffsets from LowerFORMAL_ARGUMENTSBruno Cardoso Lopes2007-08-281-3/+53
| | | | llvm-svn: 41525
* Cut off crazy computation. This helps PR1622 slightly.Chris Lattner2007-08-281-0/+4
| | | | llvm-svn: 41522
* Use simpler test to filter loops.Devang Patel2007-08-271-14/+3
| | | | llvm-svn: 41516
* Add explicit triples to avoid default behavior that varies by host.Dan Gohman2007-08-274-4/+4
| | | | llvm-svn: 41510
* Update InvokeInst to work like CallInstDavid Greene2007-08-2713-46/+76
| | | | llvm-svn: 41506
* Add a comment about using libc memset/memcpy or generating inline code.Rafael Espindola2007-08-271-2/+5
| | | | llvm-svn: 41502
* Remove this test as it is too hard to fix after theDuncan Sands2007-08-271-1621/+0
| | | | | | | latest EH changes, and in any case it is hard to imagine how the original bug could be reintroduced. llvm-svn: 41497
* Fix this test: llvm-gcc-4.2 optimizes almost everything away,Duncan Sands2007-08-271-2/+2
| | | | | | | resulting in zero matches, while llvm-gcc-4.1 manages to remove one pointless selector changing the number of matches. llvm-svn: 41494
* Make DAGCombiner's global alias analysis query more precise in the caseDan Gohman2007-08-271-2/+3
| | | | | | where both pointers have non-zero offsets. llvm-svn: 41491
* Now that we don't output cleanups by default, the actionDuncan Sands2007-08-271-1/+1
| | | | | | offset needs to be adjusted in this test. llvm-svn: 41490
* If the source and destination pointers in an llvm.memmove are knownDan Gohman2007-08-276-12/+62
| | | | | | to not alias each other, it can be translated as an llvm.memcpy. llvm-svn: 41489
* Change comments to refer to @malloc and @free instead of %malloc and %free.Dan Gohman2007-08-271-3/+3
| | | | llvm-svn: 41488
* There is an impedance matching problem between LLVM andDuncan Sands2007-08-274-31/+85
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | gcc exception handling: if an exception unwinds through an invoke, then execution must branch to the invoke's unwind target. We previously tried to enforce this by appending a cleanup action to every selector, however this does not always work correctly due to an optimization in the C++ unwinding runtime: if only cleanups would be run while unwinding an exception, then the program just terminates without actually executing the cleanups, as invoke semantics would require. I was hoping this wouldn't be a problem, but in fact it turns out to be the cause of all the remaining failures in the LLVM testsuite (these also fail with -enable-correct-eh-support, so turning on -enable-eh didn't make things worse!). Instead we need to append a full-blown catch-all to the end of each selector. The correct way of doing this depends on the personality function, i.e. it is language dependent, so can only be done by gcc. Thus this patch which generalizes the eh.selector intrinsic so that it can handle all possible kinds of action table entries (before it didn't accomodate cleanups): now 0 indicates a cleanup, and filters have to be specified using the number of type infos plus one rather than the number of type infos. Related gcc patches will cause Ada to pass a cleanup (0) to force the selector to always fire, while C++ will use a C++ catch-all (null). llvm-svn: 41484
* Add explicit keywords and remove spurious trailing semicolons.Dan Gohman2007-08-2712-26/+27
| | | | llvm-svn: 41482
* call libc memcpy/memset if array size is bigger then threshold.Rafael Espindola2007-08-272-18/+20
| | | | | | | Coping 100MB array (after a warmup) shows that glibc 2.6.1 implementation on x86-64 (core 2) is 30% faster (from 0.270917s to 0.188079s) llvm-svn: 41479
* Don't promote volatile loads/stores. This is needed (for example) to handle ↵Anton Korobeynikov2007-08-262-3/+53
| | | | | | | | setjmp/longjmp properly. This fixes PR1520. llvm-svn: 41461
* Don't DSe volatile stores.Owen Anderson2007-08-261-18/+25
| | | | llvm-svn: 41456
* Revise per review comments.Dale Johannesen2007-08-263-5/+5
| | | | llvm-svn: 41409
* Add APFloat interface to ConstantFPSDNode. ChangeDale Johannesen2007-08-253-16/+30
| | | | | | | over uses in DAGCombiner. Fix interfaces to work with APFloats. llvm-svn: 41407
* Move exit condition and exit branch from exiting block into loop header and ↵Devang Patel2007-08-251-0/+155
| | | | | | dominator info. This avoid execution of dead iteration. Loop is already filter in the beginning such that this change is safe. llvm-svn: 41394
OpenPOWER on IntegriCloud