summaryrefslogtreecommitdiffstats
path: root/llvm/lib
Commit message (Collapse)AuthorAgeFilesLines
* Remove meaningless const.Devang Patel2009-10-221-4/+4
| | | | | | Pass StringRef by value. llvm-svn: 84804
* Revert the main portion of r31856. It was causing BranchFoldingDan Gohman2009-10-221-5/+5
| | | | | | | | | | | | to break up CFG diamonds by banishing one of the blocks to the end of the function, which is bad for code density and branch size. This does pessimize MultiSource/Benchmarks/Ptrdist/yacr2, the benchmark cited as the reason for the change, however I've examined the code and it looks more like a case of gaming a particular branch than of being generally applicable. llvm-svn: 84803
* Derive metadata hierarchy from Value instead of User.Devang Patel2009-10-215-104/+7
| | | | llvm-svn: 84801
* revert r84754, it isn't the right approach. Edwin, please propose Chris Lattner2009-10-211-60/+6
| | | | | | patches for fixes like this instead of committing them directly. llvm-svn: 84799
* Missing piece of the ARM frame index post-scavenging conditionalizationJim Grosbach2009-10-213-0/+16
| | | | llvm-svn: 84798
* Fix thinko noticed by Chris.Benjamin Kramer2009-10-211-1/+1
| | | | llvm-svn: 84797
* Rename msasm to alignstack per review.Dale Johannesen2009-10-218-17/+19
| | | | llvm-svn: 84795
* Remove pointless return; at end of function.Benjamin Kramer2009-10-211-3/+0
| | | | llvm-svn: 84794
* The spill restore needs to be resolved to the SP/FP just like the spillJim Grosbach2009-10-211-0/+2
| | | | llvm-svn: 84792
* Conditionalize ARM/T2 frame index post-scavenging while working out fixesJim Grosbach2009-10-211-4/+31
| | | | | | for a few bugs. llvm-svn: 84791
* Simplify code. No intended functionality change.Benjamin Kramer2009-10-211-12/+6
| | | | llvm-svn: 84790
* Use StringRef.Devang Patel2009-10-211-4/+5
| | | | llvm-svn: 84786
* Most of the NEON shuffle instructions do not support 64-bit element types.Bob Wilson2009-10-211-3/+18
| | | | llvm-svn: 84785
* Do not use SmallVector to store MDNode elements.Devang Patel2009-10-213-20/+21
| | | | llvm-svn: 84784
* Revert r84764, it breaks mingw buildAnton Korobeynikov2009-10-211-1/+1
| | | | llvm-svn: 84783
* Improve handling of immediates by splitting 32-bit immediates into two 16-bitJim Grosbach2009-10-213-0/+99
| | | | | | immediate operands when they will fit into the using instruction. llvm-svn: 84778
* Add DAG printing for RMW stuff debuggingAnton Korobeynikov2009-10-211-0/+16
| | | | llvm-svn: 84776
* RMW preprocessing stuff was incorrect. Grab the stuff from x86 backend and ↵Anton Korobeynikov2009-10-212-54/+73
| | | | | | disable some tests until it will be clever enough to handle them. llvm-svn: 84775
* Implement branch foldingAnton Korobeynikov2009-10-213-6/+178
| | | | llvm-svn: 84774
* Cosmetic changes, no functionality changesAnton Korobeynikov2009-10-215-41/+40
| | | | llvm-svn: 84773
* Make changes to rev 84292 as requested by Chris Lattner.Victor Hernandez2009-10-214-34/+42
| | | | | | | Most changes are cleanup, but there is 1 correctness fix: I fixed InstCombine so that the icmp is removed only if the malloc call is removed (which requires explicit removal because the Worklist won't DCE any calls since they can have side-effects). llvm-svn: 84772
* Fix NEON VST2LN instruction encoding.Bob Wilson2009-10-211-7/+7
| | | | | | Patch by Johnny Chen. llvm-svn: 84767
* Revert 84732. It was the wrong fix.Bob Wilson2009-10-211-6/+6
| | | | llvm-svn: 84766
* Incorporate various suggestions Chris gave during metadata review.Devang Patel2009-10-211-30/+16
| | | | | | | | | | - i < getNumElements() instead of getNumElements() > i - Make setParent() private - Fix use of resizeOperands - Reset HasMetadata bit after removing all metadata attached to an instruction - Efficient use of iterators llvm-svn: 84765
* Build shared lib instead of an archive.Sanjiv Gupta2009-10-211-1/+1
| | | | llvm-svn: 84764
* Cleanup of frame index scavenging. Better code flow and more accuratelyJim Grosbach2009-10-211-35/+60
| | | | | | handles T2 and ARM use cases. llvm-svn: 84761
* Fix PR5262: when folding select into PHI, make sure all operands are availableTorok Edwin2009-10-211-6/+60
| | | | | | | in the PHI's Basic Block. This uses a conservative approach, because we don't have dominator info in instcombine. llvm-svn: 84754
* Add a pass to overlay pic16 data sections for function frame and automaticSanjiv Gupta2009-10-2111-6/+341
| | | | | | | | variables. This pass can be invoked by llvm-ld or opt to traverse over the call graph to detect what function frames and their automatic variables can be overlaid. Currently this builds an archive , but needs to be changed to a loadable module. llvm-svn: 84753
* Match more patterns to movt.Evan Cheng2009-10-214-1/+42
| | | | llvm-svn: 84751
* Need a comma after imp-use.Evan Cheng2009-10-211-1/+3
| | | | llvm-svn: 84749
* De-bork CMake buildDouglas Gregor2009-10-211-1/+0
| | | | llvm-svn: 84744
* make GVN work better when TD is not around:Chris Lattner2009-10-211-15/+21
| | | | | | | | | | | | | | | | | "In the existing code, if the load and the value to replace it with are of different types *and* target data is available, it tries to use the target data to coerce the replacement value to the type of the load. Otherwise, it skips all effort to handle the type mismatch and just feeds the wrongly-typed replacement value to replaceAllUsesWith, which triggers an assertion. The patch replaces it with an outer if checking for type mismatch, and an inner if-else that checks whether target data is available and, if not, returns false rather than trying to replace the load." Patch by Kenneth Uildriks! llvm-svn: 84739
* tidyChris Lattner2009-10-211-1/+1
| | | | llvm-svn: 84738
* Fix some more NEON instruction encoding problems.Bob Wilson2009-10-211-6/+6
| | | | | | Thanks to Johnny Chen for discovering the problem. llvm-svn: 84732
* Do not remove dead metadata for now.Devang Patel2009-10-211-1/+2
| | | | llvm-svn: 84731
* Leave some NEON instruction encoding bits unspecified instead of settingBob Wilson2009-10-212-198/+247
| | | | | | | a default value of zero. This is important for decoding the instructions. Patch by Johnny Chen, with some changes from me, too. llvm-svn: 84730
* Revert r84658 and r84691. They were causing llvm-gcc bootstrap to fail.Dan Gohman2009-10-211-489/+319
| | | | llvm-svn: 84727
* IPSCCP is missing stuff.Chris Lattner2009-10-211-0/+7
| | | | llvm-svn: 84725
* Delete the MacOSJITEventListener per echristo's request. It was disabled byJeffrey Yasskin2009-10-211-172/+0
| | | | | | default and didn't work anyway. llvm-svn: 84720
* Add noteAnton Korobeynikov2009-10-211-0/+2
| | | | llvm-svn: 84713
* Be crazy and assert in case of unsupported modifier passed.Anton Korobeynikov2009-10-211-0/+1
| | | | llvm-svn: 84712
* Handle external symbolsAnton Korobeynikov2009-10-212-2/+14
| | | | llvm-svn: 84711
* Distinguish between pcrel imm operands and 'normal' ones. Fix fixes gross ↵Anton Korobeynikov2009-10-214-9/+23
| | | | | | weirdness of asmprinting. llvm-svn: 84710
* Add basic block operands & jump kindsAnton Korobeynikov2009-10-215-16/+43
| | | | llvm-svn: 84709
* Ignore all implicit reg operandsAnton Korobeynikov2009-10-211-0/+2
| | | | llvm-svn: 84708
* Add a workaround for different memops prefixesAnton Korobeynikov2009-10-211-3/+6
| | | | llvm-svn: 84707
* Checkpoint MCInst printer. We (almostly) able to print global / JT / ↵Anton Korobeynikov2009-10-215-11/+106
| | | | | | constpool entries llvm-svn: 84706
* Add simple operand printing stuffAnton Korobeynikov2009-10-212-4/+19
| | | | llvm-svn: 84704
* Add experimental MSP430 MCInstLowering stuffAnton Korobeynikov2009-10-214-1/+155
| | | | llvm-svn: 84703
* Wire up MSP430 printMCInst() methodAnton Korobeynikov2009-10-211-0/+4
| | | | llvm-svn: 84702
OpenPOWER on IntegriCloud