summaryrefslogtreecommitdiffstats
path: root/llvm/lib
Commit message (Collapse)AuthorAgeFilesLines
* For inline asm output operand that matches an input. Encode the input ↵Evan Cheng2009-03-206-29/+44
| | | | | | operand index in the high bits. llvm-svn: 67387
* Fixed comment for libcalls.Sanjiv Gupta2009-03-201-20/+23
| | | | llvm-svn: 67373
* Fixed build warnings for unused variables.Sanjiv Gupta2009-03-201-3/+0
| | | | llvm-svn: 67372
* Reformatting. Inserted code comments. Cleaned interfaces.Sanjiv Gupta2009-03-202-112/+68
| | | | | | Removed unncessary code. No functionality change. llvm-svn: 67371
* Fixed the comment. No functionality change.Sanjiv Gupta2009-03-201-4/+4
| | | | llvm-svn: 67370
* Apply the patch requested in PR3846.Chris Lattner2009-03-201-1/+1
| | | | llvm-svn: 67364
* Added option to enable generating less precise mad (multiply addition)Mon P Wang2009-03-201-0/+12
| | | | | | for those architectures that support the instruction. llvm-svn: 67363
* The last use in a block that doesn't have successorsDan Gohman2009-03-201-0/+5
| | | | | | (return or unreachable) is a kill. llvm-svn: 67357
* Fix the Win32 VS2008 build:Sebastian Redl2009-03-194-12/+15
| | | | | | | | | | - 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 Cheng2009-03-195-24/+37
| | | | llvm-svn: 67335
* Clear the cached cost when removing a function inDale Johannesen2009-03-193-11/+19
| | | | | | | | | 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 computesDan Gohman2009-03-191-0/+177
| | | | | | | | | 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 Johannesen2009-03-191-1/+1
| | | | llvm-svn: 67307
* This pass keeps a map of Instructions to Rank numbers,Dale Johannesen2009-03-191-8/+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 PEI to not walk off the start of a block when an updated instructionChris Lattner2009-03-191-2/+6
| | | | | | is the first in its block. This is PR3842. llvm-svn: 67304
* Fix comment typo.Duncan Sands2009-03-191-1/+1
| | | | llvm-svn: 67302
* Fix a couple glaring whitespace issues. This file isn't internally consistentNick Lewycky2009-03-191-2/+2
| | | | | | either. llvm-svn: 67288
* Remove strange extra semicolons.Nick Lewycky2009-03-194-4/+4
| | | | llvm-svn: 67287
* Add support to tablegen for naming the nodes themselves, not just the operands, Nate Begeman2009-03-191-1/+1
| | | | | | | in selectiondag patterns. This is required for the upcoming shuffle_vector rewrite, and as it turns out, cleans up a hack in the Alpha instruction info. llvm-svn: 67286
* Added support for Mips O32 Calling ConventionBruno Cardoso Lopes2009-03-192-33/+133
| | | | llvm-svn: 67280
* aha, DAE does have to think about PHI nodes. Many thanks to "Dr Evil" (aka ↵Chris Lattner2009-03-181-3/+3
| | | | | | | | Duncan) for pointing this out :) llvm-svn: 67212
* Fix PR3826 - InstComb assert with vector shift, by not calling ↵Chris Lattner2009-03-181-7/+8
| | | | | | ComputeNumSignBits on a vector. llvm-svn: 67211
* add an assertion to make it clear that PHI nodes are not allowed.Chris Lattner2009-03-181-1/+4
| | | | llvm-svn: 67210
* Explicitly check for StoreInst, do not lose the chance to delete Zhou Sheng2009-03-181-1/+1
| | | | | | unused loads or bitcasts. llvm-svn: 67202
* Revert my previous change on Local.cpp, instead, fix the bug on scalarrepl.Zhou Sheng2009-03-182-4/+1
| | | | | | | If the instruction has no users, it is also not only used by debug info and should not be deleted. llvm-svn: 67194
* Fix a bug.Zhou Sheng2009-03-181-0/+3
| | | | | | If I->use_empty(), this method should return false. llvm-svn: 67180
* Added missing support for widening when splitting an unary op (PR3683)Mon P Wang2009-03-184-4/+91
| | | | | | | and expanding a bit convert (PR3711). In both cases, we extract the valid part of the widen vector and then do the conversion. llvm-svn: 67175
* Disable the "call to immediate" optimization on x86-64. It isChris Lattner2009-03-181-1/+5
| | | | | | | | | | | not safe in general because the immediate could be an arbitrary value that does not fit in a 32-bit pcrel displacement. Conservatively fall back to loading the value into a register and calling through it. We still do the optzn on X86-32. llvm-svn: 67142
* Fix PR3807 by inserting 'insertelement' instructions in the normal dest of Chris Lattner2009-03-181-5/+10
| | | | | | an invoke instead of after the invoke (in its block), which is invalid. llvm-svn: 67139
* LSR shouldn't ever try to hack on integer IV's larger than 64-bits. Right nowChris Lattner2009-03-171-0/+6
| | | | | | | | | it is not APInt clean, but even when it is it needs to be evaluated carefully to determine whether it is actually profitable. This fixes a crash on PR3806 llvm-svn: 67134
* Don't force promotion of return arguments on the callee.Rafael Espindola2009-03-171-9/+0
| | | | | | | Some architectures (like x86) don't require it. This fixes bug 3779. llvm-svn: 67132
* Add BUILTIN_EXPECT Support/Compiler macro.Daniel Dunbar2009-03-171-1/+2
| | | | | | | | | - Use for exceptional buffer conditions in raw_ostream:write to shave off a cycle or two. - Please rename if you have a better one. llvm-svn: 67103
* Fix codegen to compute the size of an allocation by multiplying theChris Lattner2009-03-171-8/+17
| | | | | | | | | | | size by the array amount as an i32 value instead of promoting from i32 to i64 then doing the multiply. Not doing this broke wrap-around assumptions that the optimizers (validly) made. The ultimate real fix for this is to introduce i64 version of alloca and remove mallocinst. This fixes PR3829 llvm-svn: 67093
* Remove a condition which is always true.Chris Lattner2009-03-171-3/+3
| | | | llvm-svn: 67089
* CellSPU:Scott Michel2009-03-172-16/+16
| | | | | | Revert inadvertent mis-fix of fneg. llvm-svn: 67084
* r66870 missed this out.Sanjiv Gupta2009-03-171-0/+1
| | | | llvm-svn: 67082
* typoGabor Greif2009-03-171-1/+1
| | | | llvm-svn: 67080
* Reapply r67049, with the test adjusted for darwinDuncan Sands2009-03-171-39/+61
| | | | | | (which produces "call L_f$stub" rather than "call f"). llvm-svn: 67079
* Fix a problem with DAGCombine where we were building an illegal buildMon P Wang2009-03-171-6/+11
| | | | | | | | vector shuffle mask. Forced the mask to be built using i32. Note: this will be irrelevant once vector_shuffle no longer takes a build vector for the shuffle mask. llvm-svn: 67076
* Recognize bswapl as bswap too.Dan Gohman2009-03-171-2/+5
| | | | llvm-svn: 67072
* Recognize "bswapq" as an alternate spelling for the bswap instruction.Dan Gohman2009-03-171-2/+2
| | | | llvm-svn: 67071
* raw_ostream: Put all exceptional conditions in raw_ostream::writeDaniel Dunbar2009-03-171-18/+30
| | | | | | | | under a single branch. Also, add a FIXME for formatted output. llvm-svn: 67069
* Spiller may unfold load / mod / store instructions as an optimization when ↵Evan Cheng2009-03-172-9/+30
| | | | | | the would be loaded value is available in a register. It needs to check if it's legal to clobber the register. Also, the register can contain values of multiple spill slots, make sure to check all instead of just the one being unfolded. llvm-svn: 67068
* CellSPU:Scott Michel2009-03-175-488/+417
| | | | | | | | | | | | - Fix fabs, fneg for f32 and f64. - Use BuildVectorSDNode.isConstantSplat, now that the functionality exists - Continue to improve i64 constant lowering. Lower certain special constants to the constant pool when they correspond to SPU's shufb instruction's special mask values. This avoids the overhead of performing a shuffle on a zero-filled vector just to get the special constant when the memory load suffices. llvm-svn: 67067
* raw_ostream: Rework implementation of unbuffered streams so outputtingDaniel Dunbar2009-03-171-4/+12
| | | | | | | | | a single character requires only one branch to follow slow path. - Never use a buffer when writing on an unbuffered stream. - Move default buffer size to header. llvm-svn: 67066
* Fix a debug info dependency in jump threading.Dale Johannesen2009-03-171-2/+5
| | | | llvm-svn: 67064
* raw_ostream: Replace flush_impl with write_impl, which takes data toDaniel Dunbar2009-03-161-32/+15
| | | | | | | | | write as arguments. - Add raw_ostream::GetNumBytesInBuffer. - Privatize buffer pointers. - Get rid of slow and unnecessary code for writing out large strings. llvm-svn: 67060
* raw_ostream: Lift out flush_nonempty.Daniel Dunbar2009-03-161-20/+21
| | | | | | | | | | | | | | | | | - Flush a known non-empty buffers; enforces the interface to flush_impl and kills off HandleFlush (which I saw no reason to be an inline method, Chris?). - Clarify invariant that flush_impl is only called with OutBufCur > OutBufStart. - This also cleary collects all places where we have to deal with the buffer possibly not existing. - A few more comments and fixing the unbuffered behavior remain in this commit sequence. llvm-svn: 67057
* Make raw_ostream::operator<<(const void *) fast; it doesn't matter butDaniel Dunbar2009-03-161-2/+19
| | | | | | it is easy. llvm-svn: 67054
* Add slow path for single character write, and use exclusively forDaniel Dunbar2009-03-161-8/+8
| | | | | | single characters writes outside of the fast path in raw_ostream.h llvm-svn: 67053
OpenPOWER on IntegriCloud