59727c0496
Revert r61158 for now, as it caused some test failures.
...
llvm-svn: 61159
2008-12-17 22:17:27 +00:00
9389176009
Fix miscompilations caused by renumbering, and enable it as part of prealloc splitting.
...
llvm-svn: 61158
2008-12-17 22:06:59 +00:00
5121ec3821
Add code to renumber split intervals into new vregs. This is disabled for now until I finish working out some iterator invalidation issues.
...
llvm-svn: 61104
2008-12-16 21:35:08 +00:00
12679470bf
Switch to top-down mode and fix a crasher this exposed caused by an error in the
...
live interval updating.
llvm-svn: 60652
2008-12-07 05:33:18 +00:00
0bcbe8f6a8
Factor out some common code.
...
llvm-svn: 60553
2008-12-04 21:20:30 +00:00
ecd5b5f2b2
Add support for folding spills into preceding defs when doing pre-alloc splitting.
...
llvm-svn: 60433
2008-12-02 18:53:47 +00:00
1af37c2fca
Add support for rematerialization in pre-alloc-splitting.
...
llvm-svn: 59587
2008-11-19 04:28:29 +00:00
896512af94
Fix a bug introduced by my previous patch. With this change, SPEC is now clean with prealloc splitting enabled.
...
llvm-svn: 59544
2008-11-18 20:53:59 +00:00
d36b4c1bf8
Don't allow the restore point to be placed after terminators. With this change,
...
MultiSource/Applications is clean with the prealloc splitter. Some failures
remain in SPEC.
llvm-svn: 59267
2008-11-13 21:53:14 +00:00
fd395eb801
Don't walk into predecessors in which the vreg is not live when doing shrinkwrapping.
...
This lets several failing tests get farther along, but doesn't completely fix any of them.
llvm-svn: 59086
2008-11-11 22:11:27 +00:00
26c10f1b4a
Use the new predicate to control when we do prealloc splitting. Fix a small bug.
...
llvm-svn: 58738
2008-11-05 00:32:13 +00:00
a926ed4b50
First pass at checking for the creation of a new join point when doing pre-alloc splitting. This is not turned on yet.
...
llvm-svn: 58726
2008-11-04 22:22:41 +00:00
b067843863
Revert my last patch until I consult with Evan about it.
...
llvm-svn: 58591
2008-11-03 02:33:28 +00:00
2bb2aaa775
Don't do pre-splitting if doing so would create a value join that did not
...
exist before. Updating the live intervals in that care is tricky in the general
case.
Evan, if you see a tighter guard condition for this, let me know.
llvm-svn: 58560
2008-11-02 08:08:18 +00:00
d1f8c9f4a2
Add a fixme.
...
llvm-svn: 58514
2008-10-31 16:41:59 +00:00
ce3ccc1ea0
- More pre-split fixes: spill slot live interval computation bug; restore point bug.
...
- If a def is spilt, remember its spill index to allow its reuse.
llvm-svn: 58375
2008-10-29 08:39:34 +00:00
938e201528
- Rewrite code that update register live interval that's split.
...
- Create and update spill slot live intervals.
- Lots of bug fixes.
llvm-svn: 58367
2008-10-29 05:06:14 +00:00
5e88d90379
If def is in the same mbb as the barrier, spilt the value after the last use before the barrier.
...
llvm-svn: 58314
2008-10-28 05:28:21 +00:00
9f854697e4
Add command line option to limit the number splits to help debugging.
...
llvm-svn: 58312
2008-10-28 01:48:24 +00:00
fab31680e1
Avoid putting a split past the end of the live range; always shrink wrap live interval in the barrier mbb.
...
llvm-svn: 58309
2008-10-28 00:47:49 +00:00
e67ad47a00
Silence a bogus compile time warning.
...
llvm-svn: 58297
2008-10-27 23:29:28 +00:00
f46642ada6
Remove val# defined by a remat'ed def that is now dead.
...
llvm-svn: 58294
2008-10-27 23:21:01 +00:00
f713722975
For now, don't split live intervals around x87 stack register barriers. FpGET_ST0_80 must be right after a call instruction (and ADJCALLSTACKUP) so we need to find a way to prevent reload of x87 registers between them.
...
llvm-svn: 58230
2008-10-27 07:14:50 +00:00
ed033ede22
Do not shrink wrap live interval in a mbb if it's livein any of its successor blocks. The mbb can be revisited again after all of the successors are processed.
...
llvm-svn: 58184
2008-10-26 07:49:03 +00:00
f48367b8e9
Handle cases where there aren't uses in the barrier mbb.
...
llvm-svn: 58174
2008-10-25 23:49:39 +00:00
85d71d4588
If val# def is ~0U, meaning it's defined by a PHI, and it's previously split, spill before the barrier because it's impossible to determine if all the defs are spilled in the same spill slot.
...
llvm-svn: 58129
2008-10-25 00:52:41 +00:00
cd160f9859
Fix a pasto.
...
llvm-svn: 58102
2008-10-24 18:46:44 +00:00
f8b09d424d
Fix a end() dereference; remove an abort() that wasn't meant to be left in.
...
llvm-svn: 58072
2008-10-24 05:53:44 +00:00
4bac4d0a16
Avoid splitting an interval multiple times; avoid splitting re-materializable val# (for now).
...
llvm-svn: 58068
2008-10-24 02:05:00 +00:00
d0d0317b9a
Committing a good chunk of the pre-register allocation live interval splitting pass. It's handling simple cases and appear to do good things. Next: avoid splitting an interval multiple times; renumber registers when possible; record stack slot live intervals for coloring; rematerialize defs when possible.
...
llvm-svn: 58044
2008-10-23 20:43:13 +00:00
7e721ecd21
Add skeleton for the pre-register allocation live interval splitting pass.
...
llvm-svn: 57847
2008-10-20 21:44:59 +00:00
Owen Anderson