summaryrefslogtreecommitdiffstats
path: root/llvm/lib
Commit message (Collapse)AuthorAgeFilesLines
* llvm-mc: Fix thinko in emitting values.Daniel Dunbar2009-08-281-1/+1
| | | | llvm-svn: 80340
* Fix -Asserts warning.Daniel Dunbar2009-08-281-2/+1
| | | | llvm-svn: 80338
* finish a half formed thought :)Chris Lattner2009-08-281-1/+4
| | | | llvm-svn: 80334
* Mark Andersen's as experimental.Chris Lattner2009-08-281-1/+2
| | | | llvm-svn: 80328
* Fix PR3913, patch by Jakub Staszak!Chris Lattner2009-08-281-1/+1
| | | | llvm-svn: 80327
* v4, v5 does not support sxtb / sxth.Evan Cheng2009-08-281-3/+9
| | | | llvm-svn: 80322
* Closure is a very generic name. Use AppleBlock instead.Devang Patel2009-08-271-1/+1
| | | | llvm-svn: 80307
* Revert 80278 for now, it caused a lot of MIPS tests to failBruno Cardoso Lopes2009-08-274-61/+40
| | | | llvm-svn: 80280
* Revamp our friend Mips :)Bruno Cardoso Lopes2009-08-274-40/+61
| | | | | | | | Add MO flags to simplify the printing of relocations. Remove the support for printing large code model relocs (which aren't supported anyway). llvm-svn: 80278
* Don't mark CMOV_GR8 as two-address, or commutable, since it's a pseudo.Dan Gohman2009-08-271-2/+2
| | | | llvm-svn: 80271
* Adjust the MachineBasicBlock verifier rules to be moreDan Gohman2009-08-271-11/+14
| | | | | | tolerant of blocks that end with "unreachable". llvm-svn: 80270
* Add FIXME for when we support more specific XMM registers.Eric Christopher2009-08-271-0/+2
| | | | llvm-svn: 80269
* Nuke trailing whitespace.Eric Christopher2009-08-271-150/+150
| | | | llvm-svn: 80268
* Use stripPointerCasts instead of doing the same manually.Dan Gohman2009-08-271-14/+3
| | | | llvm-svn: 80267
* Teach getUnderlyingObject and skipPointerCasts about GlobalAliases.Dan Gohman2009-08-271-0/+8
| | | | llvm-svn: 80265
* Global Aliases are not identifiable objects.Dan Gohman2009-08-271-2/+4
| | | | llvm-svn: 80263
* Handle TargetData with const.Dan Gohman2009-08-271-3/+4
| | | | llvm-svn: 80262
* Hopefully the final missing part :(Anton Korobeynikov2009-08-271-12/+0
| | | | | | scalar_to_vector is fully legal now llvm-svn: 80251
* Forgot about actual change :)Anton Korobeynikov2009-08-271-0/+20
| | | | llvm-svn: 80250
* Remove an entry from ValueMap before removing elements.Devang Patel2009-08-271-1/+1
| | | | llvm-svn: 80247
* Transform float scalar_to_vector into subreg accesses.Anton Korobeynikov2009-08-272-1/+14
| | | | | | No idea whether this is profitable or not. llvm-svn: 80245
* STRD and LDRD require ARMv5TE, not just ARMv5T.Misha Brukman2009-08-271-2/+2
| | | | | | See http://llvm.org/PR4687 for more info and links. llvm-svn: 80244
* llvm-mc/Mach-O: Add MCCodeEmitter support, for encoding instructions.Daniel Dunbar2009-08-271-5/+24
| | | | | | - No relocations yet, of course. llvm-svn: 80235
* llvm-mc/X86: Implement single instruction encoding interface for MC.Daniel Dunbar2009-08-273-0/+258
| | | | | | | | | | | | - Note, this is a gigantic hack, with the sole purpose of unblocking further work on the assembler (its also possible to test the mathcer more completely now). - Despite being a hack, its actually good enough to work over all of 403.gcc (although some encodings are probably incorrect). This is a testament to the beauty of X86's MachineInstr, no doubt! ;) llvm-svn: 80234
* llvm-mc: Print encodings after the instruction, and only when we have an asmDaniel Dunbar2009-08-271-17/+18
| | | | | | printer. llvm-svn: 80233
* X86: Mark EH_RETURN as code-gen-only.Daniel Dunbar2009-08-271-1/+1
| | | | llvm-svn: 80232
* Add {MCInst,MCOperand}::{print,dump}Daniel Dunbar2009-08-272-0/+53
| | | | llvm-svn: 80231
* Implement a new optimization in the inliner: if inlining multipleChris Lattner2009-08-271-25/+125
| | | | | | | | | | | | | | | | | | | calls into a function and if the calls bring in arrays, try to merge them together to reduce stack size. For example, in the testcase we'd previously end up with 4 allocas, now we end up with 2 allocas. As described in the comments, this is not really the ideal solution to this problem, but it is surprisingly effective. For example, on 176.gcc, we end up eliminating 67 arrays at "gccas" time and another 24 at "llvm-ld" time. One piece of concern that I didn't look into: at -O0 -g with forced inlining this will almost certainly result in worse debug info. I think this is acceptable though given that this is a case of "debugging optimized code", and we don't want debug info to prevent the optimizer from doing things anyway. llvm-svn: 80215
* unbreak the build, yay for symlinks + makefiles. :(Chris Lattner2009-08-271-1/+1
| | | | llvm-svn: 80205
* reduce header #include'ageChris Lattner2009-08-271-1/+4
| | | | llvm-svn: 80204
* enhance InlineFunction to be able to optionally returnChris Lattner2009-08-271-6/+15
| | | | | | a the list of static allocas that it inlined. llvm-svn: 80203
* smallvectorize the list of returns built by CloneAndPruneFunctionInto.Chris Lattner2009-08-273-8/+8
| | | | llvm-svn: 80202
* remove CloneTrace, which appears to be dead since 2004.Chris Lattner2009-08-272-120/+0
| | | | llvm-svn: 80201
* reduce inlining factor some stuff out to a static helper function,Chris Lattner2009-08-272-213/+259
| | | | | | and other code cleanups. No functionality change. llvm-svn: 80199
* Revert accidental commit.Bill Wendling2009-08-272-221/+28
| | | | llvm-svn: 80198
* --- Reverse-merging r80147 into '.':Bill Wendling2009-08-272-28/+221
| | | | | | | | | | | | | | | | | | | | | | A include/llvm/ADT/iterator.cmake U autoconf/configure.ac --- Reverse-merging r80161 into '.': U cmake/config-ix.cmake --- Reverse-merging r80171 into '.': U Makefile --- Reverse-merging r80173 into '.': U configure U include/llvm/Config/config.h.in --- Reverse-merging r80180 into '.': A include/llvm/ADT/iterator.h.in Despite common miscomceptions, iterator.h is alive and well. It broke the build bots for several hours. And yet no one bothered to look at them. Gabor and Doug, please review your changes and make sure that they actually build before resubmitting them. llvm-svn: 80197
* Add some checks for MachineCFG consistency. Use AnalyzeBranch andDan Gohman2009-08-271-0/+133
| | | | | | | do extra checking when it succeeds, as those are cases where CodeGen will be doing particularly interesting CFG modifications. llvm-svn: 80196
* Update CMakeDaniel Dunbar2009-08-271-0/+1
| | | | llvm-svn: 80195
* llvm-mc: Tweak MCCodeEmitter skeleton.Daniel Dunbar2009-08-271-0/+18
| | | | llvm-svn: 80193
* Initialize the PoisonMemory member before initializingDan Gohman2009-08-271-7/+7
| | | | | | | | | members that call methods that read the PoisonMemory member. This fixes potential spurious (though probably otherwise harmless) poising of unused memory, and fixes the associated valgrind error. llvm-svn: 80192
* Fix PR4789. Teach eliminateFrameIndex how to handle VLDRQ and VSTRQ which ↵Evan Cheng2009-08-274-42/+66
| | | | | | cannot fold any immediate offset. llvm-svn: 80191
* Sketch TargetRegistry support for MCCodeEmitter abstract interface.Daniel Dunbar2009-08-271-4/+27
| | | | | | - Of course, nothing actually can provide this interface yet. llvm-svn: 80188
* Try to make MSVC just a little happier.Daniel Dunbar2009-08-271-2/+2
| | | | llvm-svn: 80187
* X86FastISel support for loading and storing values of type i1.Dan Gohman2009-08-271-3/+15
| | | | llvm-svn: 80186
* Expand i8 selects into control flow instead of 16-bit conditionalDan Gohman2009-08-272-2/+15
| | | | | | | | | | | | moves. This avoids the need to promote the operands (or implicitly extend them, a partial register update condition), and can reduce i8 register pressure. This substantially speeds up code such as write_hex in lib/Support/raw_ostream.cpp. subclass-coalesce.ll is too trivial and no longer tests what it was originally intended to test. llvm-svn: 80184
* Simplify.Daniel Dunbar2009-08-261-3/+3
| | | | llvm-svn: 80176
* Make this into a static method.Owen Anderson2009-08-261-15/+13
| | | | llvm-svn: 80170
* llvm-mc/Mach-O: Don't put assembler temporary labels in the symbol table.Daniel Dunbar2009-08-262-1/+8
| | | | | | | | - I moved section creation back into AsmParser. I think policy decisions like this should be pushed higher, not lower, when possible (in addition the assembler has flags which change this behavior, for example). llvm-svn: 80162
* Rework getPersonalityIndex slightly - 0 is now a valid and not-NULLEric Christopher2009-08-261-3/+3
| | | | | | personality function. llvm-svn: 80153
* If we're emitting additional CIEs due to personality functionsEric Christopher2009-08-262-3/+8
| | | | | | | don't emit the default one. Explicitly check for the NULL CIE later. llvm-svn: 80146
OpenPOWER on IntegriCloud