| Commit message (Collapse) | Author | Age | Files | Lines | |
|---|---|---|---|---|---|
| * | Fix some bugs in HMEditor's moveAllOperandsInto logic. | Lang Hames | 2012-02-21 | 1 | -9/+10 |
| | | | | | llvm-svn: 151006 | ||||
| * | Check for the correct size in the invariant marker. | Nick Lewycky | 2012-02-20 | 1 | -4/+7 |
| | | | | | llvm-svn: 151003 | ||||
| * | Fix machine-cp by having it to check sub-register indicies. e.g. | Evan Cheng | 2012-02-20 | 2 | -3/+27 |
| | | | | | | | | | | | | | ecx = mov eax al = mov ch The second copy is not a nop because the sub-indices of ecx,ch is not the same of that of eax/al. Re-enabled machine-cp. PR11940 llvm-svn: 151002 | ||||
| * | Fix 80-column violation. | Chad Rosier | 2012-02-20 | 1 | -1/+2 |
| | | | | | llvm-svn: 150998 | ||||
| * | InstCombine: Removing the base from the address calculation is only safe ↵ | Benjamin Kramer | 2012-02-20 | 1 | -1/+1 |
| | | | | | | | when the GEPs are inbounds. llvm-svn: 150978 | ||||
| * | Make this a bit prettier and more obvious when a derived type isn't | Eric Christopher | 2012-02-20 | 1 | -1/+3 |
| | | | | | | | derived from anything. llvm-svn: 150975 | ||||
| * | If a derived type is also a composite type, print that information | Eric Christopher | 2012-02-20 | 1 | -2/+7 |
| | | | | | | | too. llvm-svn: 150974 | ||||
| * | Add support for runtime languages on our forward declarations. | Eric Christopher | 2012-02-20 | 1 | -2/+5 |
| | | | | | llvm-svn: 150973 | ||||
| * | InstCombine: When comparing two GEPs that were derived from the same base ↵ | Benjamin Kramer | 2012-02-20 | 1 | -0/+14 |
| | | | | | | | | | pointer but use different types, expand the offset calculation and to the compare on the offset if profitable. This came up in SmallVector code. llvm-svn: 150962 | ||||
| * | InstCombine: Make OptimizePointerDifference more aggressive. | Benjamin Kramer | 2012-02-20 | 1 | -27/+27 |
| | | | | | | | | | | - Ignore pointer casts. - Also expand GEPs that aren't constantexprs when they have one use or only constant indices. - We now compile "&foo[i] - &foo[j]" into "i - j". llvm-svn: 150961 | ||||
| * | Teach the DAGCombiner that certain loadext nodes followed by ANDs can be ↵ | James Molloy | 2012-02-20 | 1 | -0/+82 |
| | | | | | | | converted to zeroexts. llvm-svn: 150957 | ||||
| * | Improve generated code for extending loads and some trunc stores on ARM. | James Molloy | 2012-02-20 | 2 | -7/+140 |
| | | | | | | | Teach TargetSelectionDAG about lengthening loads for vector types and set v4i8 as legal. Allow FP_TO_UINT for v4i16 from v4i32. llvm-svn: 150956 | ||||
| * | Make post-ra tail duplication bundle safe. No test case as recent codegen | Evan Cheng | 2012-02-20 | 1 | -3/+6 |
| | | | | | | | flow changes have already hidden the bug. rdar://10893812 llvm-svn: 150949 | ||||
| * | Rename class Evaluate to Evaluator and put it in an anonymous namespace. | Nick Lewycky | 2012-02-20 | 1 | -12/+17 |
| | | | | | llvm-svn: 150947 | ||||
| * | fold comparisons of gep'd alloca points with null to false, | Chris Lattner | 2012-02-20 | 1 | -4/+21 |
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | implementing PR12013. We now compile the testcase to: __Z4testv: ## @_Z4testv ## BB#0: ## %_ZN4llvm15SmallVectorImplIiE9push_backERKi.exit pushq %rbx subq $64, %rsp leaq 32(%rsp), %rbx movq %rbx, (%rsp) leaq 64(%rsp), %rax movq %rax, 16(%rsp) movl $1, 32(%rsp) leaq 36(%rsp), %rax movq %rax, 8(%rsp) leaq (%rsp), %rdi callq __Z1gRN4llvm11SmallVectorIiLj8EEE movq (%rsp), %rdi cmpq %rbx, %rdi je LBB0_2 ## BB#1: callq _free LBB0_2: ## %_ZN4llvm11SmallVectorIiLj8EED1Ev.exit addq $64, %rsp popq %rbx ret instead of: __Z4testv: ## @_Z4testv ## BB#0: pushq %rbx subq $64, %rsp xorl %eax, %eax leaq (%rsp), %rbx addq $32, %rbx movq %rbx, (%rsp) movq %rbx, 8(%rsp) leaq 64(%rsp), %rcx movq %rcx, 16(%rsp) je LBB0_2 ## BB#1: movl $1, 32(%rsp) movq %rbx, %rax LBB0_2: ## %_ZN4llvm15SmallVectorImplIiE9push_backERKi.exit addq $4, %rax movq %rax, 8(%rsp) leaq (%rsp), %rdi callq __Z1gRN4llvm11SmallVectorIiLj8EEE movq (%rsp), %rdi cmpq %rbx, %rdi je LBB0_4 ## BB#3: callq _free LBB0_4: ## %_ZN4llvm11SmallVectorIiLj8EED1Ev.exit addq $64, %rsp popq %rbx ret This doesn't shrink clang noticably though. llvm-svn: 150944 | ||||
| * | Move constructors out-of-line and flesh out their documentation. No | Chandler Carruth | 2012-02-20 | 1 | -2/+32 |
| | | | | | | | | functionality changed. This is in preparation for some refactoring of how this class behaves. llvm-svn: 150941 | ||||
| * | Move EvaluateFunction and EvaluateBlock into a class, and make the class store | Nick Lewycky | 2012-02-19 | 1 | -122/+137 |
| | | | | | | | the information that they pass around between them. No functionality change! llvm-svn: 150939 | ||||
| * | Silence operator precedence warning. | Benjamin Kramer | 2012-02-19 | 1 | -1/+1 |
| | | | | | llvm-svn: 150921 | ||||
| * | Remove dead code. Improve llvm_unreachable text. Simplify some control flow. | Ahmed Charles | 2012-02-19 | 17 | -88/+22 |
| | | | | | llvm-svn: 150918 | ||||
| * | StringRef'ize EmitSourceFileHeader(). | Ahmed Charles | 2012-02-19 | 1 | -1/+1 |
| | | | | | llvm-svn: 150917 | ||||
| * | Remove some unneeded includes and fix ordering in X86ISelLowering.cpp. ↵ | Craig Topper | 2012-02-19 | 1 | -4/+1 |
| | | | | | | | Remove unneeded 'using namespace'. llvm-svn: 150916 | ||||
| * | Add machinery for pushing live ranges onto bundle starts while bundling. | Lang Hames | 2012-02-19 | 1 | -2/+207 |
| | | | | | llvm-svn: 150915 | ||||
| * | Simplify moveEnteringDownFrom rules. | Lang Hames | 2012-02-19 | 1 | -11/+3 |
| | | | | | llvm-svn: 150914 | ||||
| * | Unify all shuffle mask checking functions take a mask and VT instead of ↵ | Craig Topper | 2012-02-19 | 1 | -85/+68 |
| | | | | | | | VectorShuffleSDNode. llvm-svn: 150913 | ||||
| * | Skip through instructions rather than operands when looking for last use slot. | Lang Hames | 2012-02-19 | 1 | -1/+1 |
| | | | | | llvm-svn: 150912 | ||||
| * | APFloat::toString(): Fix overrun at scanning. | NAKAMURA Takumi | 2012-02-19 | 1 | -1/+1 |
| | | | | | | | FYI, clang/test/SemaTemplate/template-id-printing.cpp had been failing due to it on cygwin-clang. llvm-svn: 150911 | ||||
| * | Fix TODO and trailing whitespace. | Lang Hames | 2012-02-19 | 1 | -10/+14 |
| | | | | | llvm-svn: 150910 | ||||
| * | Defer sanity checks on live intervals until after all have been updated. ↵ | Lang Hames | 2012-02-19 | 1 | -171/+170 |
| | | | | | | | Hold (LiveInterval, LiveRange) pairs to update, rather than vregs. llvm-svn: 150909 | ||||
| * | Make a bunch of X86ISelLowering shuffle functions static now that they are ↵ | Craig Topper | 2012-02-19 | 2 | -172/+51 |
| | | | | | | | no longer needed by isel. llvm-svn: 150908 | ||||
| * | comment fix ARM.h | Jia Liu | 2012-02-19 | 1 | -1/+1 |
| | | | | | llvm-svn: 150904 | ||||
| * | some comment fix for X86 and ARM | Jia Liu | 2012-02-19 | 8 | -8/+8 |
| | | | | | llvm-svn: 150902 | ||||
| * | Add vmfunc instruction to X86 assembler and disassembler. | Craig Topper | 2012-02-19 | 4 | -20/+27 |
| | | | | | llvm-svn: 150899 | ||||
| * | Fix issue with bitwise and precedence. | Ahmed Charles | 2012-02-18 | 1 | -1/+1 |
| | | | | | llvm-svn: 150897 | ||||
| * | Hashing.h - utilities for hashing various data types. | Talin | 2012-02-18 | 2 | -0/+47 |
| | | | | | llvm-svn: 150890 | ||||
| * | White space fixes. | Rafael Espindola | 2012-02-18 | 1 | -7/+7 |
| | | | | | llvm-svn: 150886 | ||||
| * | Temporarily disable this assert. Looks like it found a similar issue when | Rafael Espindola | 2012-02-18 | 1 | -1/+3 |
| | | | | | | | building bullet. llvm-svn: 150885 | ||||
| * | Don't skip debug instructions when looking for the insertion point of | Rafael Espindola | 2012-02-18 | 1 | -5/+8 |
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | the cast. If we do, we can end up with inst1 --------------- < Insertion point dbg inst new inst instead of the desired inst1 new inst --------------- < Insertion point dbg inst Another option would be for InsertNoopCastOfTo (or its callers) to move the insertion point and we would end up with inst1 dbg inst new inst --------------- < Insertion point but that complicates the callers. This fixes PR12018 (and firefox's build). llvm-svn: 150884 | ||||
| * | Emacs-tag and some comment fix for all ARM, CellSPU, Hexagon, MBlaze, ↵ | Jia Liu | 2012-02-18 | 293 | -384/+385 |
| | | | | | | | MSP430, PPC, PTX, Sparc, X86, XCore. llvm-svn: 150878 | ||||
| * | Add X86InstrSVM.td that I forgot to add in r150873. | Craig Topper | 2012-02-18 | 1 | -0/+62 |
| | | | | | llvm-svn: 150874 | ||||
| * | Add X86 assembler and disassembler support for AMD SVM instructions. ↵ | Craig Topper | 2012-02-18 | 4 | -71/+61 |
| | | | | | | | Original patch by Kay Tiong Khoo. Few tweaks by me for code density and to reduce replication. llvm-svn: 150873 | ||||
| * | Trivial cleanup to group the generic 'armvN' cases with the 'arm' case, | Chandler Carruth | 2012-02-18 | 1 | -4/+4 |
| | | | | | | | etc. No functionality changed. llvm-svn: 150867 | ||||
| * | Fix a rather nasty regression from r150690: LHS != RHS does not imply ↵ | Eli Friedman | 2012-02-18 | 1 | -1/+2 |
| | | | | | | | LHS->stripPointerCasts() != RHS->stripPointerCasts(). llvm-svn: 150863 | ||||
| * | Bring HMEditor into line with LLVM coding standards. | Lang Hames | 2012-02-17 | 1 | -160/+163 |
| | | | | | llvm-svn: 150851 | ||||
| * | Ignore the lifetime intrinsics in fast-isel. | Eric Christopher | 2012-02-17 | 1 | -0/+4 |
| | | | | | llvm-svn: 150848 | ||||
| * | Don't print out pointer values in SUnit::dump(). | Jakob Stoklund Olesen | 2012-02-17 | 1 | -5/+3 |
| | | | | | llvm-svn: 150842 | ||||
| * | Sink variable into assert | Matt Beaumont-Gay | 2012-02-17 | 1 | -2/+2 |
| | | | | | llvm-svn: 150841 | ||||
| * | Add support for regmask slots to HMEditor. Also fixes a comment error. | Lang Hames | 2012-02-17 | 1 | -3/+17 |
| | | | | | llvm-svn: 150840 | ||||
| * | Fix typo in comment ldopen() -> dlopen(). | Kevin Enderby | 2012-02-17 | 1 | -1/+1 |
| | | | | | llvm-svn: 150836 | ||||
| * | Handle regmask operands in ARMInstrInfo. | Jakob Stoklund Olesen | 2012-02-17 | 2 | -7/+9 |
| | | | | | llvm-svn: 150833 | ||||
| * | Put back the initializing the targets in the disassembler API with a comment as | Kevin Enderby | 2012-02-17 | 1 | -0/+13 |
| | | | | | | | | to why this is needed. This broke the darwin's otool(1) program. This change was made in r144385. llvm-svn: 150832 | ||||

