Commit message (Collapse) | Author | Age | Files | Lines | ||
---|---|---|---|---|---|---|
... | ||||||
* | Friendly names | Andrew Lenharth | 2006-01-16 | 2 | -1/+19 | |
| | | | | llvm-svn: 25364 | |||||
* | bswap implementation | Nate Begeman | 2006-01-14 | 1 | -0/+1 | |
| | | | | llvm-svn: 25312 | |||||
* | make DAG isel the default | Andrew Lenharth | 2006-01-13 | 1 | -3/+3 | |
| | | | | llvm-svn: 25282 | |||||
* | expand unsupported stacksave/stackrestore nodes | Chris Lattner | 2006-01-13 | 1 | -0/+4 | |
| | | | | llvm-svn: 25272 | |||||
* | Add bswap, rotl, and rotr nodes | Nate Begeman | 2006-01-11 | 1 | -0/+2 | |
| | | | | | | | | | Add dag combiner code to recognize rotl, rotr Add ppc code to match rotl Targets should add rotl/rotr patterns if they have them llvm-svn: 25222 | |||||
* | this pattern was bogus | Andrew Lenharth | 2006-01-11 | 1 | -2/+1 | |
| | | | | llvm-svn: 25197 | |||||
* | Int immediate loading fix | Andrew Lenharth | 2006-01-10 | 1 | -1/+1 | |
| | | | | llvm-svn: 25182 | |||||
* | proper branch not equal sequence | Andrew Lenharth | 2006-01-09 | 1 | -0/+4 | |
| | | | | llvm-svn: 25159 | |||||
* | make 0 codegen much better | Andrew Lenharth | 2006-01-06 | 1 | -0/+4 | |
| | | | | llvm-svn: 25131 | |||||
* | unbreak the build, these are now in TargetSelectionDAG.td | Chris Lattner | 2006-01-05 | 1 | -3/+0 | |
| | | | | llvm-svn: 25109 | |||||
* | Had expand logic backward. | Jim Laskey | 2006-01-05 | 1 | -0/+2 | |
| | | | | llvm-svn: 25105 | |||||
* | Added initial support for DEBUG_LABEL allowing debug specific labels to be | Jim Laskey | 2006-01-05 | 1 | -1/+0 | |
| | | | | | | inserted in the code. llvm-svn: 25104 | |||||
* | typeo | Andrew Lenharth | 2006-01-02 | 1 | -1/+1 | |
| | | | | llvm-svn: 25060 | |||||
* | Remove a 'using namespace std'. | Chris Lattner | 2006-01-01 | 1 | -5/+4 | |
| | | | | llvm-svn: 25059 | |||||
* | Move brcond over and fix some imm patterns. This may be the last change ↵ | Andrew Lenharth | 2006-01-01 | 3 | -160/+166 | |
| | | | | | | before changing the default alpha isel. llvm-svn: 25057 | |||||
* | clean this function up some | Andrew Lenharth | 2006-01-01 | 1 | -37/+26 | |
| | | | | llvm-svn: 25055 | |||||
* | improve constant loading. Still sucks, but oh well | Andrew Lenharth | 2005-12-30 | 2 | -17/+53 | |
| | | | | llvm-svn: 25047 | |||||
* | let us get some do what I meant not what I said stuff checked in. You would ↵ | Andrew Lenharth | 2005-12-29 | 1 | -2/+2 | |
| | | | | | | think the alpha backend would be 64bit clean llvm-svn: 25040 | |||||
* | Fix up immediate handling | Andrew Lenharth | 2005-12-29 | 1 | -1/+1 | |
| | | | | llvm-svn: 25039 | |||||
* | Restore some happiness to the JIT | Andrew Lenharth | 2005-12-27 | 1 | -2/+3 | |
| | | | | llvm-svn: 25026 | |||||
* | Fix alpha regressions. | Andrew Lenharth | 2005-12-27 | 1 | -1/+1 | |
| | | | | llvm-svn: 25025 | |||||
* | Added field noResults to Instruction. | Evan Cheng | 2005-12-26 | 1 | -5/+5 | |
| | | | | | | | | | | | | Currently tblgen cannot tell which operands in the operand list are results so it assumes the first one is a result. This is bad. Ideally we would fix this by separating results from inputs, e.g. (res R32:$dst), (ops R32:$src1, R32:$src2). But that's a more distruptive change. Adding 'let noResults = 1' is the workaround to tell tblgen that the instruction does not produces a result. It works for now since tblgen does not support instructions which produce multiple results. llvm-svn: 25017 | |||||
* | add br pattern, unify JSR and BSR ISel instrs, and add BSR support for DAG | Andrew Lenharth | 2005-12-25 | 5 | -32/+38 | |
| | | | | llvm-svn: 25011 | |||||
* | All that just to lower div and rem | Andrew Lenharth | 2005-12-25 | 5 | -131/+78 | |
| | | | | llvm-svn: 25008 | |||||
* | All addressing modes are now exposed. The only remaining relocated forms | Andrew Lenharth | 2005-12-24 | 5 | -27/+48 | |
| | | | | | | | | | are for function prologue. TODO: move external symbols over to using RelLit. : have a pattern that matches constpool|globaladdr : have a pattern that matches (add x imm) -> x, imm or (...) -> ..., 0 llvm-svn: 25003 | |||||
* | Unify the patterns for loads and stores. Now offset addressing should be | Andrew Lenharth | 2005-12-24 | 2 | -169/+96 | |
| | | | | | | supported. This almost completes memory operations. llvm-svn: 25002 | |||||
* | Let's see if we can break things. | Andrew Lenharth | 2005-12-24 | 5 | -51/+172 | |
| | | | | | | | | Lower GOT relative addresses to Lo and HI. Update both ISels to select them when they can. Saves instructions here and there. llvm-svn: 25001 | |||||
* | move loads and stores over. Smart addr selection comming | Andrew Lenharth | 2005-12-24 | 3 | -58/+53 | |
| | | | | llvm-svn: 25000 | |||||
* | remove dead code | Chris Lattner | 2005-12-22 | 1 | -13/+0 | |
| | | | | llvm-svn: 24965 | |||||
* | Disengage DEBUG_LOC from non-PPC targets. | Jim Laskey | 2005-12-21 | 1 | -0/+1 | |
| | | | | llvm-svn: 24919 | |||||
* | fix FP selects | Andrew Lenharth | 2005-12-12 | 2 | -14/+14 | |
| | | | | llvm-svn: 24672 | |||||
* | restore a more restricted select | Andrew Lenharth | 2005-12-12 | 1 | -0/+32 | |
| | | | | llvm-svn: 24668 | |||||
* | FP select improvements (and likely breakage), oh and crazy people might want ↵ | Andrew Lenharth | 2005-12-11 | 2 | -42/+46 | |
| | | | | | | to *return* floating point values. Don't see why myself llvm-svn: 24658 | |||||
* | it helps if your conditionals are not reversed | Andrew Lenharth | 2005-12-09 | 1 | -9/+9 | |
| | | | | llvm-svn: 24641 | |||||
* | fix divide and remainder | Andrew Lenharth | 2005-12-06 | 2 | -13/+14 | |
| | | | | llvm-svn: 24628 | |||||
* | more decent branches for FP. I might have to make some intermediate nodes ↵ | Andrew Lenharth | 2005-12-06 | 1 | -2/+33 | |
| | | | | | | to actually be able to use the DAG for FPcmp llvm-svn: 24625 | |||||
* | OK, this does wonders for broken stuff | Andrew Lenharth | 2005-12-06 | 1 | -0/+1 | |
| | | | | llvm-svn: 24624 | |||||
* | added instructions with inverted immediates | Andrew Lenharth | 2005-12-06 | 2 | -32/+23 | |
| | | | | llvm-svn: 24614 | |||||
* | yea, it helps to have your path set right when testing | Andrew Lenharth | 2005-12-05 | 1 | -2/+4 | |
| | | | | llvm-svn: 24613 | |||||
* | These never trigger, but whatever | Andrew Lenharth | 2005-12-05 | 2 | -9/+40 | |
| | | | | llvm-svn: 24612 | |||||
* | move this over to the dag | Andrew Lenharth | 2005-12-05 | 2 | -10/+9 | |
| | | | | llvm-svn: 24609 | |||||
* | fix constant pool loads | Andrew Lenharth | 2005-12-05 | 1 | -1/+1 | |
| | | | | llvm-svn: 24607 | |||||
* | major think-o | Andrew Lenharth | 2005-12-01 | 1 | -5/+10 | |
| | | | | llvm-svn: 24564 | |||||
* | Support multiple ValueTypes per RegisterClass, needed for upcoming vector | Nate Begeman | 2005-12-01 | 1 | -3/+3 | |
| | | | | | | work. This change has no effect on generated code. llvm-svn: 24563 | |||||
* | Flags where I think I need them, quick, before the nightly tester starts | Andrew Lenharth | 2005-12-01 | 1 | -23/+42 | |
| | | | | llvm-svn: 24560 | |||||
* | SelectNodeTo now returns its result, we must pay attention to it. | Chris Lattner | 2005-11-30 | 1 | -40/+29 | |
| | | | | llvm-svn: 24552 | |||||
* | No longer track value types for asm printer operands, and remove them as | Nate Begeman | 2005-11-30 | 1 | -3/+2 | |
| | | | | | | | an argument to every operand printing function. Requires some slight tweaks to x86, the only user. llvm-svn: 24541 | |||||
* | remove redundant code | Andrew Lenharth | 2005-11-30 | 1 | -14/+2 | |
| | | | | llvm-svn: 24538 | |||||
* | Make typesafe that which isn't: FCMOVxx | Andrew Lenharth | 2005-11-30 | 4 | -26/+52 | |
| | | | | llvm-svn: 24536 | |||||
* | FPSelect and more custom lowering | Andrew Lenharth | 2005-11-30 | 5 | -25/+69 | |
| | | | | llvm-svn: 24535 |