| Commit message (Collapse) | Author | Age | Files | Lines | ||
|---|---|---|---|---|---|---|
| ... | ||||||
| * | Fix PR2748. Avoid coalescing physical register with virtual register which ↵ | Evan Cheng | 2008-09-11 | 2 | -0/+70 | |
| | | | | | | | | | | | | would create illegal extract_subreg. e.g. vr1024 = extract_subreg vr1025, 1 ... vr1024 = mov8rr AH If vr1024 is coalesced with AH, the extract_subreg is now illegal since AH does not have a super-reg whose sub-register 1 is AH. llvm-svn: 56118 | |||||
| * | Fix a bug in ANY_EXTEND handling that was breaking 403.gcc on X86-64 in fast ↵ | Owen Anderson | 2008-09-11 | 1 | -2/+7 | |
| | | | | | | | isel. llvm-svn: 56117 | |||||
| * | Fix comment typo. | Duncan Sands | 2008-09-11 | 1 | -1/+1 | |
| | | | | | llvm-svn: 56116 | |||||
| * | Intrinsics don't touch internal global variables | Duncan Sands | 2008-09-11 | 1 | -2/+4 | |
| | | | | | | | | (unless passed one via a parameter), even if they are IntrWriteMem. llvm-svn: 56115 | |||||
| * | Fix a copy+paste bug that Duncan spotted. For several | Dan Gohman | 2008-09-11 | 1 | -1/+1 | |
| | | | | | | | | cases it was still getting lucky and detecting overflow but it was clearly incorrect. llvm-svn: 56113 | |||||
| * | Fix PR2783 - coalescer bug. Missing a TargetRegisterInfo::isVirtualRegister ↵ | Evan Cheng | 2008-09-11 | 1 | -1/+2 | |
| | | | | | | | check. llvm-svn: 56112 | |||||
| * | Intrinsics don't read these kinds of global | Duncan Sands | 2008-09-11 | 1 | -5/+7 | |
| | | | | | | | variables. llvm-svn: 56105 | |||||
| * | Fix a 80 column violation. | Evan Cheng | 2008-09-11 | 1 | -1/+2 | |
| | | | | | llvm-svn: 56097 | |||||
| * | The version of AtomicSDNode::AtomicSDNode used (only) for | Dale Johannesen | 2008-09-11 | 1 | -6/+6 | |
| | | | | | | | | | | | | | | cmp-and-swap reversed the Cmp and Swap arguments; comments make it clear this is unintentional. Unfortunately, the x86 BE had a compensating reversal, which is removed here. PPC is OK. From inspection of the Alpha code I think it is OK, but if somebody has that platform please check it out. I cannot test on that platform. llvm-svn: 56091 | |||||
| * | If ISD::ANY_EXTEND fails, try ISD::ZERO_EXTEND and ISD::SIGN_EXTEND before ↵ | Owen Anderson | 2008-09-11 | 1 | -0/+7 | |
| | | | | | | | | | giving up. This fixes 445.gobmk on X86-64 in fast isel. llvm-svn: 56088 | |||||
| * | Succumb utterly to compatibility and implement | Dale Johannesen | 2008-09-11 | 1 | -4/+4 | |
| | | | | | | | | __sync_fetch_and_nand as ANDC, even though that's not what nand means. llvm-svn: 56087 | |||||
| * | Propagate subreg index when promoting a load to a copy. | Evan Cheng | 2008-09-11 | 1 | -0/+7 | |
| | | | | | llvm-svn: 56085 | |||||
| * | In my analysis for r56076 I missed the case where the original | Dan Gohman | 2008-09-11 | 1 | -1/+2 | |
| | | | | | | | multiplication overflows. llvm-svn: 56082 | |||||
| * | Fix an icmp+sdiv optimization to check for and handle an overflow | Dan Gohman | 2008-09-10 | 1 | -1/+16 | |
| | | | | | | | condition. This fixes PR2740. llvm-svn: 56076 | |||||
| * | X86FastISel support for double->float and float->double casts. | Dan Gohman | 2008-09-10 | 1 | -0/+43 | |
| | | | | | llvm-svn: 56070 | |||||
| * | FastISel support for i1 PHI nodes. | Dan Gohman | 2008-09-10 | 1 | -2/+7 | |
| | | | | | llvm-svn: 56069 | |||||
| * | FastISel support for i1 constants. | Dan Gohman | 2008-09-10 | 1 | -2/+10 | |
| | | | | | llvm-svn: 56068 | |||||
| * | Fix a bug in the coalescer where it didn't check if a live interval existed ↵ | Owen Anderson | 2008-09-10 | 1 | -1/+1 | |
| | | | | | | | | | before trying to manipulate it. This was exposed by fast isel's handling of shifts on X86-64. With this, FreeBench/pcompress2 passes on X86-64 in fast isel. llvm-svn: 56067 | |||||
| * | Add X86FastISel support for static allocas, and refences | Dan Gohman | 2008-09-10 | 5 | -77/+105 | |
| | | | | | | | | to static allocas. As part of this change, refactor the address mode code for laods and stores. llvm-svn: 56066 | |||||
| * | Fix PR2664 - spiller GetRegForReload wasn't respecting sub-register indices ↵ | Evan Cheng | 2008-09-10 | 1 | -2/+4 | |
| | | | | | | | on machine operands. llvm-svn: 56065 | |||||
| * | Fix a fastcc + sret bug. If fastcc and sret, callee doesn't need to pop the ↵ | Evan Cheng | 2008-09-10 | 1 | -2/+4 | |
| | | | | | | | hidden struct ptr; Re-enable fastcc. llvm-svn: 56061 | |||||
| * | Handle new intrinsics with vector arguments. | Dale Johannesen | 2008-09-10 | 1 | -0/+5 | |
| | | | | | | | Patch by Paul Redmond. llvm-svn: 56059 | |||||
| * | Add a break statement that I accidentally deleted when | Dan Gohman | 2008-09-10 | 1 | -0/+1 | |
| | | | | | | | | I shuffled the fast-isel command-line options around. This fixes a bunch of fast-isel failures. llvm-svn: 56057 | |||||
| * | fix white spaces. | Devang Patel | 2008-09-10 | 1 | -5/+5 | |
| | | | | | llvm-svn: 56056 | |||||
| * | Fix name. | Duncan Sands | 2008-09-10 | 1 | -1/+1 | |
| | | | | | llvm-svn: 56055 | |||||
| * | Add trampoline support for the new FastCC calling | Duncan Sands | 2008-09-10 | 1 | -0/+1 | |
| | | | | | | | | convention (not related to recent Ada testsuite failures). llvm-svn: 56054 | |||||
| * | Turn off the new FastCC for the moment. It causes | Duncan Sands | 2008-09-10 | 1 | -2/+0 | |
| | | | | | | | | a slew of Ada testsuite failures on x86-32 linux. Seems to be related to the use of float. llvm-svn: 56053 | |||||
| * | Remove unnecessary bit-wise AND from the limited precision work. | Bill Wendling | 2008-09-10 | 1 | -6/+4 | |
| | | | | | llvm-svn: 56049 | |||||
| * | Fix 80 col violation. | Daniel Dunbar | 2008-09-10 | 1 | -1/+2 | |
| | | | | | llvm-svn: 56048 | |||||
| * | Fix a warning about comparing signed and unsigned values. | Dan Gohman | 2008-09-10 | 1 | -1/+1 | |
| | | | | | llvm-svn: 56040 | |||||
| * | Fix typo. | Evan Cheng | 2008-09-10 | 1 | -2/+2 | |
| | | | | | llvm-svn: 56037 | |||||
| * | Check that both operands are f32 before attempting to lower. | Bill Wendling | 2008-09-10 | 1 | -0/+1 | |
| | | | | | llvm-svn: 56036 | |||||
| * | Implement "visitPow". This is mainly used to see if we have a pow() call of this | Bill Wendling | 2008-09-10 | 2 | -4/+146 | |
| | | | | | | | | | | | | form: powf(10.0f, x); If this is the case, and also we want limited precision floating-point calculations, then lower to do the limited-precision stuff. llvm-svn: 56035 | |||||
| * | A few more places where FPOW is being ignored. | Evan Cheng | 2008-09-09 | 1 | -7/+13 | |
| | | | | | llvm-svn: 56032 | |||||
| * | Change -fast-isel-no-abort to -fast-isel-abort, which now defaults | Dan Gohman | 2008-09-09 | 1 | -7/+12 | |
| | | | | | | | | | to being off by default. Also, add assertion checks to check that the various fast-isel-related command-line options are only used when -fast-isel itself is enabled. llvm-svn: 56029 | |||||
| * | Legalizer was missing code that expand fpow to a libcall. | Evan Cheng | 2008-09-09 | 1 | -10/+12 | |
| | | | | | llvm-svn: 56028 | |||||
| * | Adding 6-, 12-, and 18-bit limited-precision floating-point support for exp2 | Bill Wendling | 2008-09-09 | 1 | -26/+111 | |
| | | | | | | | function. llvm-svn: 56025 | |||||
| * | Move the uglier parts of deciding not to emit a | Dale Johannesen | 2008-09-09 | 2 | -15/+25 | |
| | | | | | | | | | | UsedDirective for some symbols in llvm.used into Darwin-specific code. I've decided LessPrivateGlobal is potentially a useful abstraction and left it in the target-independent area, with improved comment. llvm-svn: 56024 | |||||
| * | Add support for 6-, 12-, and 18-bit limited precision calculations of exp for | Bill Wendling | 2008-09-09 | 1 | -7/+134 | |
| | | | | | | | floating-point numbers. llvm-svn: 56023 | |||||
| * | Add a new option, -fast-isel-verbose, that can be used with | Dan Gohman | 2008-09-09 | 1 | -15/+21 | |
| | | | | | | | | -fast-isel-no-abort to get a dump of all unhandled instructions, without an abort. llvm-svn: 56021 | |||||
| * | Clear preference when it no longer makes sense. | Evan Cheng | 2008-09-09 | 1 | -0/+8 | |
| | | | | | llvm-svn: 56019 | |||||
| * | if loop induction variable is always sign or zero extended then | Devang Patel | 2008-09-09 | 1 | -1/+118 | |
| | | | | | | | extend the type of induction variable. llvm-svn: 56017 | |||||
| * | Add assertion check. | Devang Patel | 2008-09-09 | 1 | -1/+1 | |
| | | | | | llvm-svn: 56016 | |||||
| * | fix overflow check. | Devang Patel | 2008-09-09 | 1 | -2/+1 | |
| | | | | | llvm-svn: 56011 | |||||
| * | Clean this up, based on Evan's suggestions. | Owen Anderson | 2008-09-09 | 1 | -2/+2 | |
| | | | | | llvm-svn: 56009 | |||||
| * | - Add support for 6-, 12-, and 18-bit limited precision floating-point "log" | Bill Wendling | 2008-09-09 | 1 | -112/+216 | |
| | | | | | | | | values. - Refactored some of the code. llvm-svn: 56008 | |||||
| * | Fix PR2757. Ignore liveinterval register allocation preference if the ↵ | Evan Cheng | 2008-09-09 | 1 | -1/+1 | |
| | | | | | | | preference register is not in the right register class. This can happen due to sub-register coalescing. llvm-svn: 56006 | |||||
| * | Make safer variant of alias resolution routine to be default | Anton Korobeynikov | 2008-09-09 | 7 | -10/+9 | |
| | | | | | llvm-svn: 56005 | |||||
| * | Simplify this some more. No functionality change. | Duncan Sands | 2008-09-09 | 1 | -15/+8 | |
| | | | | | llvm-svn: 56003 | |||||
| * | Resolve aliases, when possible | Anton Korobeynikov | 2008-09-09 | 1 | -0/+22 | |
| | | | | | llvm-svn: 56001 | |||||

