Commit message (Collapse) | Author | Age | Files | Lines | |
---|---|---|---|---|---|
* | Fix a subtle iterator invalidation bug in a recursive algorithm. | Owen Anderson | 2007-08-03 | 1 | -5/+7 |
| | | | | llvm-svn: 40776 | ||||
* | Prepare for "core" website. | Reid Spencer | 2007-08-03 | 1 | -0/+26 |
| | | | | llvm-svn: 40775 | ||||
* | Long double, part 1 of N. Support in IR. | Dale Johannesen | 2007-08-03 | 15 | -4515/+3750 |
| | | | | llvm-svn: 40774 | ||||
* | add an observation | Chris Lattner | 2007-08-03 | 1 | -0/+27 |
| | | | | llvm-svn: 40772 | ||||
* | Fix an accidental commit. | Chris Lattner | 2007-08-02 | 1 | -1/+1 |
| | | | | llvm-svn: 40758 | ||||
* | More explicit keywords. | Dan Gohman | 2007-08-02 | 8 | -8/+8 |
| | | | | llvm-svn: 40757 | ||||
* | Fix the alignment requirements of several unpck and shuf instructions. | Dan Gohman | 2007-08-02 | 3 | -13/+71 |
| | | | | | | | | Generalize isPSHUFDMask and add a unary SHUFPD pattern so that SHUFPD's memory operand alignment can be tested as well, with a fix to avoid breaking MMX's use of isPSHUFDMask. llvm-svn: 40756 | ||||
* | Fix pastos in vector arithmetic intrinsics. | Dan Gohman | 2007-08-02 | 1 | -4/+4 |
| | | | | llvm-svn: 40754 | ||||
* | Fix 80 col. violations. | Owen Anderson | 2007-08-02 | 1 | -8/+14 |
| | | | | llvm-svn: 40751 | ||||
* | Fix 80 col. violations. | Owen Anderson | 2007-08-02 | 1 | -7/+15 |
| | | | | llvm-svn: 40750 | ||||
* | Fix 80 col. violations. | Owen Anderson | 2007-08-02 | 1 | -14/+23 |
| | | | | llvm-svn: 40749 | ||||
* | don't redefine a parameter | Chris Lattner | 2007-08-02 | 1 | -1/+1 |
| | | | | llvm-svn: 40748 | ||||
* | Fix a bug that was causing several miscompilations on SPEC. | Owen Anderson | 2007-08-02 | 2 | -7/+21 |
| | | | | llvm-svn: 40746 | ||||
* | Implement review feedback. | Christopher Lamb | 2007-08-02 | 1 | -4/+4 |
| | | | | llvm-svn: 40745 | ||||
* | I don't have time to restore this functionality right now. | Chris Lattner | 2007-08-02 | 1 | -0/+1 |
| | | | | llvm-svn: 40743 | ||||
* | Replacing a cast with another one does not reduce the number of | Chris Lattner | 2007-08-02 | 1 | -5/+3 |
| | | | | | | casts in the input. llvm-svn: 40741 | ||||
* | Reduced testcase for PR1594 | Chris Lattner | 2007-08-02 | 1 | -0/+10 |
| | | | | llvm-svn: 40740 | ||||
* | Disable an xform that causes an infinite loop. This fixes PR1594 | Chris Lattner | 2007-08-02 | 1 | -0/+1 |
| | | | | llvm-svn: 40739 | ||||
* | wrap some long lines. Major offenders that are left include | Chris Lattner | 2007-08-02 | 7 | -11/+16 |
| | | | | | | | | gvn, gvnpre, dse, and predsimplify. To see these, use: make check-line-length llvm-svn: 40738 | ||||
* | Update dominator info for the middle blocks created while spliting | Devang Patel | 2007-08-02 | 3 | -22/+141 |
| | | | | | | | | exit edge to preserve LCSSA. Fix dominance frontier update during loop unswitch. This fixes PR 1589, again llvm-svn: 40737 | ||||
* | Mark the SSE and MMX load instructions that | Dan Gohman | 2007-08-02 | 3 | -0/+16 |
| | | | | | | | | | | X86InstrInfo::isReallyTriviallyReMaterializable knows how to handle with the isReMaterializable flag so that it is given a chance to handle them. Without hoisting constant-pool loads from loops this isn't very visible, though it does keep CodeGen/X86/constant-pool-remat-0.ll from making a copy of the constant pool on the stack. llvm-svn: 40736 | ||||
* | Enhance instcombine to be more aggressive about folding casts of | Chris Lattner | 2007-08-02 | 2 | -32/+54 |
| | | | | | | operations of casts. This implements InstCombine/zext-fold.ll llvm-svn: 40726 | ||||
* | Fix bug spotted by Chris. | Anders Carlsson | 2007-08-02 | 1 | -4/+4 |
| | | | | llvm-svn: 40725 | ||||
* | Add extend and extOrTrunc methods that do sign or zero extension depending ↵ | Anders Carlsson | 2007-08-02 | 1 | -0/+16 |
| | | | | | | on whether the integer is signed or not llvm-svn: 40724 | ||||
* | Switch some multiplication instructions over to the new scheme for testing. | Evan Cheng | 2007-08-02 | 2 | -12/+53 |
| | | | | llvm-svn: 40723 | ||||
* | Do not emit copies for physical register output if it's not used. | Evan Cheng | 2007-08-02 | 3 | -2/+32 |
| | | | | llvm-svn: 40722 | ||||
* | Fix test. | Evan Cheng | 2007-08-02 | 1 | -1/+1 |
| | | | | llvm-svn: 40721 | ||||
* | Fix PR1575 and test/Transforms/CondProp/2007-08-01-InvalidRead.ll | Chris Lattner | 2007-08-02 | 2 | -8/+824 |
| | | | | llvm-svn: 40720 | ||||
* | Adjust for new CallInst constructor interface. | Reid Spencer | 2007-08-02 | 1 | -6/+3 |
| | | | | | | This fixes test/Feature/llvm2cpp.ll llvm-svn: 40714 | ||||
* | Style police: Expand the tabs to spaces! | Scott Michel | 2007-08-02 | 1 | -5/+5 |
| | | | | llvm-svn: 40712 | ||||
* | Teach BasicAA about noalias parameter attributes, but do it correctly this time. | Christopher Lamb | 2007-08-02 | 3 | -1/+58 |
| | | | | llvm-svn: 40711 | ||||
* | Instead of adding copyfromreg's to handle physical definitions. Now isel can | Evan Cheng | 2007-08-02 | 2 | -68/+87 |
| | | | | | | | | | | | | | | simply specify them as results and let scheduledag handle them. That is, instead of SDOperand Flag = DAG.getTargetNode(Opc, MVT::i32, MVT::Flag, ...) SDOperand Result = DAG.getCopyFromReg(Chain, X86::EAX, MVT::i32, Flag) Just write: SDOperand Result = DAG.getTargetNode(Opc, MVT::i32, MVT::i32, ...) And let scheduledag emit the move from X86::EAX to a virtual register. llvm-svn: 40710 | ||||
* | Added TargetInstrDescriptor::numDefs - num of results. | Evan Cheng | 2007-08-02 | 2 | -1/+3 |
| | | | | llvm-svn: 40709 | ||||
* | Can't handle offset and scale if rip-relative addressing is to be used. | Evan Cheng | 2007-08-01 | 1 | -6/+10 |
| | | | | llvm-svn: 40703 | ||||
* | Mac OS X X86-64 low 4G address not available. | Evan Cheng | 2007-08-01 | 1 | -4/+4 |
| | | | | llvm-svn: 40702 | ||||
* | Mac OS X X86-64 low 4G address not available. | Evan Cheng | 2007-08-01 | 4 | -1/+21 |
| | | | | llvm-svn: 40701 | ||||
* | Undo previous check-in. | Devang Patel | 2007-08-01 | 2 | -85/+22 |
| | | | | llvm-svn: 40698 | ||||
* | Some out operands were incorrectly specified as input operands. | Evan Cheng | 2007-08-01 | 2 | -17/+17 |
| | | | | llvm-svn: 40697 | ||||
* | Make sure these tests pass for the right reasons (verifier error, rather than | Reid Spencer | 2007-08-01 | 2 | -6/+6 |
| | | | | | | failure to assemble). llvm-svn: 40696 | ||||
* | Update dominator info for the middle blocks created while spliting | Devang Patel | 2007-08-01 | 3 | -22/+115 |
| | | | | | | | | exit edge to preserve LCSSA. Fix dominance frontier update during loop unswitch. This fixes PR 1589. llvm-svn: 40695 | ||||
* | Make non-local memdep not be recursive, and fix a bug on 403.gcc that this ↵ | Owen Anderson | 2007-08-01 | 3 | -55/+62 |
| | | | | | | exposed. llvm-svn: 40692 | ||||
* | Missing Requires. | Evan Cheng | 2007-08-01 | 1 | -2/+2 |
| | | | | llvm-svn: 40691 | ||||
* | New test. Bogus implicit-def prevented a copy from being coalesced. | Evan Cheng | 2007-08-01 | 1 | -0/+8 |
| | | | | llvm-svn: 40690 | ||||
* | Be more precise. | Evan Cheng | 2007-08-01 | 1 | -1/+1 |
| | | | | llvm-svn: 40689 | ||||
* | Bugs: missing partial uses and redundant partial defs. | Evan Cheng | 2007-08-01 | 1 | -4/+20 |
| | | | | llvm-svn: 40688 | ||||
* | Add a test for the load/store alignment. | Lauro Ramos Venancio | 2007-08-01 | 1 | -0/+17 |
| | | | | llvm-svn: 40687 | ||||
* | Expand unaligned loads/stores when the target doesn't support them. (PR1548) | Lauro Ramos Venancio | 2007-08-01 | 2 | -3/+156 |
| | | | | llvm-svn: 40682 | ||||
* | Clean up C++ restrict test cases and add a test for restrict qualified methods. | Christopher Lamb | 2007-08-01 | 1 | -0/+13 |
| | | | | llvm-svn: 40681 | ||||
* | these tests aren't xfailed. | Chris Lattner | 2007-08-01 | 2 | -2/+0 |
| | | | | llvm-svn: 40680 | ||||
* | we're now handling this right :) | Chris Lattner | 2007-08-01 | 1 | -0/+12 |
| | | | | llvm-svn: 40675 |