| Commit message (Collapse) | Author | Age | Files | Lines | ||
|---|---|---|---|---|---|---|
| ... | ||||||
| * | Enhance LiveValues to work on PHI operands. | Dan Gohman | 2009-03-23 | 1 | -13/+24 | |
| | | | | | llvm-svn: 67523 | |||||
| * | Make getOperandNumForIncomingValue and getOperandNumForIncomingBlock | Dan Gohman | 2009-03-23 | 1 | -2/+10 | |
| | | | | | | | | | static member functions, and add getIncomingValueNumForOperand and getIncomingBlockNumForOperand, which are the respective inverses. llvm-svn: 67522 | |||||
| * | Correct some comments. Operand numbers start at 0. | Dan Gohman | 2009-03-23 | 1 | -6/+6 | |
| | | | | | llvm-svn: 67518 | |||||
| * | Model inline asm constraint which ties an input to an output register as ↵ | Evan Cheng | 2009-03-23 | 10 | -53/+144 | |
| | | | | | | | machine operand TIED_TO constraint. This eliminated the need to pre-allocate registers for these. This also allows register allocator can eliminate the unneeded copies. llvm-svn: 67512 | |||||
| * | Do not fold away subreg_to_reg if the source register has a sub-register ↵ | Evan Cheng | 2009-03-23 | 2 | -3/+34 | |
| | | | | | | | | | | | index. That means the source register is taking a sub-register of a larger register. e.g. On x86 %RAX<def> = ... %RAX<def> = SUBREG_TO_REG 0, %EAX:3<kill>, 3 The first def is defining RAX, not EAX so the top bits were not zero-extended. llvm-svn: 67511 | |||||
| * | fix a bug Alexei Svitkine pointed out. | Chris Lattner | 2009-03-23 | 1 | -2/+2 | |
| | | | | | llvm-svn: 67510 | |||||
| * | When we restore signal handlers, restore them back to what they | Chris Lattner | 2009-03-23 | 1 | -8/+30 | |
| | | | | | | | | were when we came around, not to their default handler. This should fix PR3848 llvm-svn: 67509 | |||||
| * | factorize signal registration, part of PR3848. | Chris Lattner | 2009-03-23 | 1 | -10/+23 | |
| | | | | | llvm-svn: 67508 | |||||
| * | Fix a grammaro in a comment that Bill noticed. | Dan Gohman | 2009-03-23 | 1 | -1/+1 | |
| | | | | | llvm-svn: 67507 | |||||
| * | tt'ify. | Chris Lattner | 2009-03-23 | 1 | -2/+2 | |
| | | | | | llvm-svn: 67505 | |||||
| * | VC++ 6.0 is not future work :) | Chris Lattner | 2009-03-23 | 1 | -64/+5 | |
| | | | | | | | Do not recommend llvm::OStream anymore. Use raw_ostream or MemoryBuffer. llvm-svn: 67504 | |||||
| * | Add comments explaining why there's only one register for | Dan Gohman | 2009-03-23 | 1 | -1/+6 | |
| | | | | | | | i8 return values. llvm-svn: 67502 | |||||
| * | Missed this as part of r67451. | Evan Cheng | 2009-03-23 | 1 | -20/+22 | |
| | | | | | llvm-svn: 67500 | |||||
| * | Fix PR3860 by correcting a predicate. | Chris Lattner | 2009-03-22 | 1 | -8/+10 | |
| | | | | | llvm-svn: 67473 | |||||
| * | Add some explanations of how apint loads and stores | Duncan Sands | 2009-03-22 | 1 | -2/+15 | |
| | | | | | | | work. llvm-svn: 67471 | |||||
| * | Really should pass -dylib to the linker... | Bill Wendling | 2009-03-22 | 1 | -1/+1 | |
| | | | | | llvm-svn: 67469 | |||||
| * | A dylib should be built as a dylib and not a bundle. | Bill Wendling | 2009-03-22 | 1 | -2/+2 | |
| | | | | | llvm-svn: 67468 | |||||
| * | add method to access a template argument. | Chris Lattner | 2009-03-22 | 1 | -0/+4 | |
| | | | | | llvm-svn: 67458 | |||||
| * | Factorize out a concept - no functionality change. | Duncan Sands | 2009-03-21 | 4 | -8/+15 | |
| | | | | | llvm-svn: 67454 | |||||
| * | Re-commit r67334 and r67349 with fix. | Evan Cheng | 2009-03-21 | 4 | -9/+17 | |
| | | | | | llvm-svn: 67451 | |||||
| * | add a fastpath to ConstantExpr::getBitCast to handle the case when an obviously | Chris Lattner | 2009-03-21 | 1 | -0/+5 | |
| | | | | | | | | | | unneeded bitcast is requested. This is common for frontends who just unconditionally cast even if the target is often the right type already. THis prevents going into getFoldedCast which switches on the opcode and does a bunch of other stuff before doing the same opzn. llvm-svn: 67435 | |||||
| * | add some inline methods for infix operators on sparse vectors, | Chris Lattner | 2009-03-21 | 2 | -13/+41 | |
| | | | | | | | tidy some df iteration stuff, patch by John Mosby! llvm-svn: 67428 | |||||
| * | Removed AFGR32 register class | Bruno Cardoso Lopes | 2009-03-21 | 5 | -176/+110 | |
| | | | | | | | Handle odd registers allocation in FGR32. llvm-svn: 67422 | |||||
| * | Fix a few more indentation problems and an 80-column violation. | Bob Wilson | 2009-03-20 | 1 | -8/+8 | |
| | | | | | llvm-svn: 67416 | |||||
| * | No functional changes. Fix indentation and whitespace only. | Bob Wilson | 2009-03-20 | 1 | -101/+91 | |
| | | | | | llvm-svn: 67412 | |||||
| * | Fix instcombine to not introduce undefined shifts when merging two | Chris Lattner | 2009-03-20 | 2 | -2/+23 | |
| | | | | | | | shifts together. This fixes PR3851. llvm-svn: 67411 | |||||
| * | Don't load values out of global constants with weak | Duncan Sands | 2009-03-20 | 5 | -6/+58 | |
| | | | | | | | | | | linkage: the value may be replaced with something different at link time. (Frontends that want to allow values to be loaded out of weak constants can give their constants weak_odr linkage). llvm-svn: 67407 | |||||
| * | Simplify this code; use a while instead of an if and a do-while. | Dan Gohman | 2009-03-20 | 1 | -9/+6 | |
| | | | | | llvm-svn: 67400 | |||||
| * | For inline asm output operand that matches an input. Encode the input ↵ | Evan Cheng | 2009-03-20 | 7 | -29/+62 | |
| | | | | | | | operand index in the high bits. llvm-svn: 67387 | |||||
| * | Fixed comment for libcalls. | Sanjiv Gupta | 2009-03-20 | 1 | -20/+23 | |
| | | | | | llvm-svn: 67373 | |||||
| * | Fixed build warnings for unused variables. | Sanjiv Gupta | 2009-03-20 | 1 | -3/+0 | |
| | | | | | llvm-svn: 67372 | |||||
| * | Reformatting. Inserted code comments. Cleaned interfaces. | Sanjiv Gupta | 2009-03-20 | 2 | -112/+68 | |
| | | | | | | | Removed unncessary code. No functionality change. llvm-svn: 67371 | |||||
| * | Fixed the comment. No functionality change. | Sanjiv Gupta | 2009-03-20 | 1 | -4/+4 | |
| | | | | | llvm-svn: 67370 | |||||
| * | Revert r67334 and r37349 which break "make check" on Linux. | Nick Lewycky | 2009-03-20 | 5 | -37/+29 | |
| | | | | | llvm-svn: 67368 | |||||
| * | Apply the patch requested in PR3846. | Chris Lattner | 2009-03-20 | 1 | -1/+1 | |
| | | | | | llvm-svn: 67364 | |||||
| * | Added option to enable generating less precise mad (multiply addition) | Mon P Wang | 2009-03-20 | 2 | -1/+21 | |
| | | | | | | | for those architectures that support the instruction. llvm-svn: 67363 | |||||
| * | The last use in a block that doesn't have successors | Dan Gohman | 2009-03-20 | 1 | -0/+5 | |
| | | | | | | | (return or unreachable) is a kill. llvm-svn: 67357 | |||||
| * | Fix typo's. | Evan Cheng | 2009-03-20 | 2 | -3/+3 | |
| | | | | | llvm-svn: 67349 | |||||
| * | Fix the Win32 VS2008 build: | Sebastian Redl | 2009-03-19 | 9 | -17/+24 | |
| | | | | | | | | | | | - Make type declarations match the struct/class keyword of the definition. - Move AddSignalHandler into the namespace where it belongs. - Correctly call functions from template base. - Some other small changes. With this patch, LLVM and Clang should build properly and with far less noise under VS2008. llvm-svn: 67347 | |||||
| * | Added MachineInstr::isRegTiedToDefOperand to check for two-addressness. | Evan Cheng | 2009-03-19 | 6 | -24/+42 | |
| | | | | | llvm-svn: 67335 | |||||
| * | More makefile changes to allow dejagnu tests to pass when system tools ↵ | Evan Cheng | 2009-03-19 | 5 | -29/+37 | |
| | | | | | | | default to a different target from the llvm configuration (e.g. 64-bit gcc and 32-bit llvm). llvm-svn: 67334 | |||||
| * | Renamed unittest files to have a consistent {Tt}est suffix. | Misha Brukman | 2009-03-19 | 2 | -0/+0 | |
| | | | | | llvm-svn: 67326 | |||||
| * | SADDO and UADDO are commutative. | Dan Gohman | 2009-03-19 | 1 | -0/+2 | |
| | | | | | llvm-svn: 67319 | |||||
| * | Clear the cached cost when removing a function in | Dale Johannesen | 2009-03-19 | 4 | -11/+26 | |
| | | | | | | | | | | the inliner; prevents nondeterministic behavior when the same address is reallocated. Don't build call graph nodes for debug intrinsic calls; they're useless, and there were typically a lot of them. llvm-svn: 67311 | |||||
| * | Add a liveness analysis pass for LLVM IR values. This computes | Dan Gohman | 2009-03-19 | 4 | -0/+287 | |
| | | | | | | | | | | the set of blocks in which values are used, the set in which values are live-through, and the set in which values are killed. For the live-through and killed sets, conservative approximations are used. llvm-svn: 67309 | |||||
| * | Fix comment typo. | Dale Johannesen | 2009-03-19 | 1 | -1/+1 | |
| | | | | | llvm-svn: 67307 | |||||
| * | This pass keeps a map of Instructions to Rank numbers, | Dale Johannesen | 2009-03-19 | 1 | -8/+14 | |
| | | | | | | | | | | and was deleting Instructions without clearing the corresponding map entry. This led to nondeterministic behavior if the same address got allocated to another Instruction within a short time. llvm-svn: 67306 | |||||
| * | Fix regression in 'tblgen -gen-clang-diags-defs': Emit the diagnostic kind ↵ | Ted Kremenek | 2009-03-19 | 1 | -1/+2 | |
| | | | | | | | instead of "DIAGNOSTICCONTROLLED". llvm-svn: 67305 | |||||
| * | Fix PEI to not walk off the start of a block when an updated instruction | Chris Lattner | 2009-03-19 | 1 | -2/+6 | |
| | | | | | | | is the first in its block. This is PR3842. llvm-svn: 67304 | |||||
| * | Fix comment typo. | Duncan Sands | 2009-03-19 | 1 | -1/+1 | |
| | | | | | llvm-svn: 67302 | |||||

