summaryrefslogtreecommitdiffstats
path: root/llvm/lib
Commit message (Collapse)AuthorAgeFilesLines
* Add a default NoItinerary class for targets to use.Chris Lattner2006-01-271-0/+1
| | | | llvm-svn: 25670
* Teach the scheduler to emit the appropriate INLINEASM MachineInstr for anChris Lattner2006-01-261-0/+29
| | | | | | ISD::INLINEASM node. llvm-svn: 25668
* initial selectiondag support for new INLINEASM node. Note that inline asmsChris Lattner2006-01-263-1/+74
| | | | | | with outputs or inputs are not supported yet. :) llvm-svn: 25664
* Use global information to fill out Dwarf compile units.Jim Laskey2006-01-262-22/+18
| | | | llvm-svn: 25662
* Improve compatibility with VC2005, patch by Morten Ofstad!Jeff Cohen2006-01-263-4/+14
| | | | llvm-svn: 25661
* Implement a method for inline asm supportChris Lattner2006-01-261-0/+17
| | | | llvm-svn: 25660
* Set up MachineDebugInfo to scan for debug information form "llvm.db"g globals.Jim Laskey2006-01-264-16/+289
| | | | | | Global Variable information is now pulled from "llvm.dbg.globals" llvm-svn: 25655
* Improve compatibility with VC2005, patch by Morten Ofstad!Chris Lattner2006-01-261-3/+5
| | | | llvm-svn: 25653
* dynamically allocate plugin space as neededAndrew Lenharth2006-01-261-5/+12
| | | | llvm-svn: 25652
* Remember plugins should someone like bugpoint want to know them.Andrew Lenharth2006-01-261-0/+15
| | | | llvm-svn: 25649
* Added preliminary x86 subtarget support.Evan Cheng2006-01-266-4/+158
| | | | llvm-svn: 25645
* fix stack corruption! Previously, 16-byte whole-FP-register stores wereDuraid Madina2006-01-261-4/+7
| | | | | | | | | | | | | | | being treated as needing only 8 bytes (though they were 16 byte aligned.) This should fix a bunch of tests - anyone have any comments, though? - in Target.td , SpillSize and SpillAlignment seem dead - is this what Size and Alignment do now? - in CodeGenRegisters.h/CodeGenTarget.cpp , DeclaredSpillSize and DeclaredSpillAlignment seem dead. - there are a bunch of comments here and there that don't clearly distinguish between 'size' and 'spillsize' etc. hmm. llvm-svn: 25644
* some hooveringDuraid Madina2006-01-263-9/+13
| | | | llvm-svn: 25643
* Rest of subtarget support, remove references to ppcChris Lattner2006-01-264-14/+38
| | | | llvm-svn: 25642
* Add trivial subtarget supportChris Lattner2006-01-266-1/+86
| | | | llvm-svn: 25641
* minor renamingAndrew Lenharth2006-01-262-16/+16
| | | | llvm-svn: 25640
* allow R28 to be used for frame calculations without entirely removing it ↵Andrew Lenharth2006-01-263-21/+4
| | | | | | from circulation llvm-svn: 25639
* Work around some x86 Darwin assembler bugsEvan Cheng2006-01-261-0/+26
| | | | llvm-svn: 25638
* add method for constraint parsingChris Lattner2006-01-261-29/+50
| | | | llvm-svn: 25637
* When trying to fold X86::SETCC into a Select, make a copy if it has more thanEvan Cheng2006-01-261-6/+22
| | | | | | one use. This allows more CMOV instructions. llvm-svn: 25634
* teach the cloner to handle inline asmsChris Lattner2006-01-261-9/+1
| | | | llvm-svn: 25633
* parse and verify the constraint string.Chris Lattner2006-01-261-0/+68
| | | | llvm-svn: 25631
* Clean up some code; improve efficiency; and fixed a potential bug involvingEvan Cheng2006-01-261-150/+127
| | | | | | chain successors. llvm-svn: 25630
* Remove the uses of STATUS flag register. Rely on node property SDNPInFlag,Evan Cheng2006-01-262-258/+221
| | | | | | SDNPOutFlag, and SDNPOptInFlag instead. llvm-svn: 25629
* Make sure the only user of InlineAsm's are direct calls.Chris Lattner2006-01-261-0/+4
| | | | llvm-svn: 25626
* oopsAndrew Lenharth2006-01-251-1/+1
| | | | llvm-svn: 25623
* add bc reader/writer support for inline asmChris Lattner2006-01-256-34/+74
| | | | llvm-svn: 25621
* forgot oneAndrew Lenharth2006-01-251-0/+1
| | | | llvm-svn: 25620
* regenerateChris Lattner2006-01-253-1302/+1339
| | | | llvm-svn: 25619
* Parse inline asm objectsChris Lattner2006-01-253-3/+56
| | | | llvm-svn: 25618
* Print InlineAsm objectsChris Lattner2006-01-252-9/+27
| | | | llvm-svn: 25617
* make things compile againAndrew Lenharth2006-01-251-0/+4
| | | | llvm-svn: 25614
* Don't break the optimized build (by incorrect placement of #endif)Reid Spencer2006-01-251-1/+1
| | | | llvm-svn: 25613
* Change inline asms to be uniqued like constants, not embedded in a Module.Chris Lattner2006-01-253-43/+6
| | | | llvm-svn: 25610
* initialize an instance var, apparently I forgot to commit this long agoChris Lattner2006-01-251-0/+1
| | | | llvm-svn: 25609
* No need to keep track of top and bottom nodes in a group since the vector isEvan Cheng2006-01-251-5/+0
| | | | | | already in order. Thanks Jim for pointing it out. llvm-svn: 25608
* Set SchedulingForLatency to be the default scheduling preference for all.Evan Cheng2006-01-254-3/+1
| | | | llvm-svn: 25607
* First part of bug 680:Nate Begeman2006-01-2512-311/+337
| | | | | | | Remove TLI.LowerVA* and replace it with SDNodes that are lowered the same way as everything else. llvm-svn: 25606
* Make it even more portable.Jeff Cohen2006-01-252-2/+2
| | | | llvm-svn: 25605
* Fix VC++ compilation error.Jeff Cohen2006-01-251-1/+1
| | | | llvm-svn: 25604
* Default scheduling preference is SchedulingForLatency.Evan Cheng2006-01-253-0/+3
| | | | llvm-svn: 25603
* X86 prefer scheduling for reduced register pressure.Evan Cheng2006-01-251-0/+1
| | | | llvm-svn: 25602
* Bottom up register usage reducing list scheduler.Evan Cheng2006-01-251-21/+451
| | | | llvm-svn: 25601
* Keep track of bottom / top element of a set of flagged nodes.Evan Cheng2006-01-251-1/+6
| | | | llvm-svn: 25600
* If scheduler choice is the default (-sched=default), use target schedulingEvan Cheng2006-01-251-2/+10
| | | | | | | | preference to determine which scheduler to use. SchedulingForLatency == Breadth first; SchedulingForRegPressure == bottom up register reduction list scheduler. llvm-svn: 25599
* Fix a selectcc lowering bug. Make a copy of X86ISD::CMP when folding it.Evan Cheng2006-01-251-2/+10
| | | | llvm-svn: 25596
* Loosen up these checks to allow direct uses of ESPChris Lattner2006-01-251-4/+12
| | | | llvm-svn: 25595
* Portably cast a pointer to an integer.Jeff Cohen2006-01-252-2/+2
| | | | llvm-svn: 25594
* add bundling! well not really, for now it's just stop-insertion.Duraid Madina2006-01-256-154/+277
| | | | llvm-svn: 25593
* maintaining stackpointer alignment. Perhaps it doesn't matterAndrew Lenharth2006-01-251-0/+4
| | | | llvm-svn: 25592
OpenPOWER on IntegriCloud