55d738e2e1
Don't track kills in VNInfo. Use interval ends instead.
...
The VNInfo.kills vector was almost unused except for all the code keeping it
updated. The few places using it were easily rewritten to check for interval
ends instead.
The two new methods LiveInterval::killedAt and killedInRange are replacements.
This brings us down to 3 independent data structures tracking kills.
llvm-svn: 106905
2010-06-25 22:53:05 +00:00
779c69bbc5
Add a DebugLoc argument to TargetInstrInfo::copyRegToReg, so that it
...
doesn't have to guess.
llvm-svn: 103194
2010-05-06 20:33:48 +00:00
4a618827de
Fix "the the" and similar typos.
...
llvm-svn: 95781
2010-02-10 16:03:48 +00:00
b06015aa69
move target-independent opcodes out of TargetInstrInfo
...
into TargetOpcodes.h. #include the new TargetOpcodes.h
into MachineInstr. Add new inline accessors (like isPHI())
to MachineInstr, and start using them throughout the
codebase.
llvm-svn: 95687
2010-02-09 19:54:29 +00:00
d18966c79f
Change errs() to dbgs().
...
llvm-svn: 92596
2010-01-05 01:26:09 +00:00
3c13667d14
Revert accidental commit.
...
llvm-svn: 91635
2009-12-17 23:45:18 +00:00
819c356a09
Turn off critical edge splitting for landing pads. The introduction of a
...
non-landing pad basic block as the successor to a block that ends in an
unconditional jump will cause block folding to remove the added block as a
successor. Thus eventually removing it AND the landing pad entirely. Critical
edge splitting is an optimization, so we can safely turn it off when dealing
with landing pads.
llvm-svn: 91634
2009-12-17 23:42:32 +00:00
05fb9637f6
The Indexes Patch.
...
This introduces a new pass, SlotIndexes, which is responsible for numbering
instructions for register allocation (and other clients). SlotIndexes numbering
is designed to match the existing scheme, so this patch should not cause any
changes in the generated code.
For consistency, and to avoid naming confusion, LiveIndex has been renamed
SlotIndex.
The processImplicitDefs method of the LiveIntervals analysis has been moved
into its own pass so that it can be run prior to SlotIndexes. This was
necessary to match the existing numbering scheme.
llvm-svn: 85979
2009-11-03 23:52:08 +00:00
974e12b2d3
Remove includes of Support/Compiler.h that are no longer needed after the
...
VISIBILITY_HIDDEN removal.
llvm-svn: 85043
2009-10-25 06:57:41 +00:00
02d5f77d26
Remove VISIBILITY_HIDDEN from class/struct found inside anonymous namespaces.
...
Chris claims we should never have visibility_hidden inside any .cpp file but
that's still not true even after this commit.
llvm-svn: 85042
2009-10-25 06:33:48 +00:00
a7780905e7
Oops. Renamed remaining MachineInstrIndex references.
...
llvm-svn: 83255
2009-10-03 04:31:31 +00:00
920301ecc5
Renamed MachineInstrIndex to LiveIndex.
...
llvm-svn: 83254
2009-10-03 04:21:37 +00:00
6b81770040
Removed static qualifier from a few index related methods. These methods may require a LiveIntervals instance in future.
...
llvm-svn: 81374
2009-09-09 20:14:17 +00:00
3fffe62177
Replaces uses of unsigned for indexes in LiveInterval and VNInfo with
...
a new class, MachineInstrIndex, which hides arithmetic details from
most clients. This is a step towards allowing the register allocator
to update/insert code during allocation.
llvm-svn: 81040
2009-09-04 20:41:11 +00:00
4dc3edde9f
remove a few DOUTs here and there.
...
llvm-svn: 79832
2009-08-23 06:35:02 +00:00
3b90d973b0
Modified VNInfo. The "copy" member is now a union which holds the copy for a register interval, or the defining register for a stack interval. Access is via getCopy/setCopy and getReg/setReg.
...
llvm-svn: 78620
2009-08-10 23:43:28 +00:00
0402315d41
Use setPreservesAll and setPreservesCFG in CodeGen passes.
...
llvm-svn: 77754
2009-07-31 23:37:33 +00:00
dab7b06de9
Improved tracking of value number kills. VN kills are now represented
...
as an (index,bool) pair. The bool flag records whether the kill is a
PHI kill or not. This code will be used to enable splitting of live
intervals containing PHI-kills.
A slight change to live interval weights introduced an extra spill
into lsr-code-insertion (outside the critical sections). The test
condition has been updated to reflect this.
llvm-svn: 75097
2009-07-09 03:57:02 +00:00
16cab1935e
VNInfo cleanup.
...
llvm-svn: 73634
2009-06-17 21:01:20 +00:00
fc968ef687
Update to in-place spilling framework. Includes live interval scaling and trivial rewriter.
...
llvm-svn: 72729
2009-06-02 16:53:25 +00:00
fe1f3d6e08
Fix a bug in live-in detection that caused lost-copy problems to show up.
...
llvm-svn: 57424
2008-10-12 20:39:30 +00:00
68fec886ed
Fix compilation error on MSVC.
...
llvm-svn: 57046
2008-10-04 08:11:49 +00:00
0d1e9a8e04
Switch the MachineOperand accessors back to the short names like
...
isReg, etc., from isRegister, etc.
llvm-svn: 57006
2008-10-03 15:45:36 +00:00
c5b45fec6b
Use a multimap rather than a map for holding the list of copies to insert, so we don't lose copies when two of them have
...
the same source. I don't know what I was thinking when I wrote this originally.
Note: There's probably a more efficient way to do this, but I need to think about it some more, and about what determinism
guarantees need to be present.
llvm-svn: 56964
2008-10-02 19:40:33 +00:00
15e6be8cc5
Mark merged-in VNInfo's as being PHIKilled.
...
llvm-svn: 56893
2008-09-30 23:58:47 +00:00
a381313d98
We don't need to insert copies for implicit_def's.
...
llvm-svn: 56674
2008-09-26 18:50:46 +00:00
4ef4b1a7c1
Fix off-by-one error when updating live intervals.
...
llvm-svn: 56567
2008-09-24 21:30:38 +00:00
4cdc18ad80
Add initial support for inserting last minute copies.
...
llvm-svn: 56485
2008-09-23 04:37:10 +00:00
97364655dc
Significant improvements to the logic for merging live intervals. This code can't
...
just use LI::MergeValueAsValue, as its behavior in the presence of overlapping ranges
isn't what StrongPHIElimination wants.
llvm-svn: 56472
2008-09-22 21:58:58 +00:00
a79db30d28
Tidy up several unbeseeming casts from pointer to intptr_t.
...
llvm-svn: 55779
2008-09-04 17:05:41 +00:00
8de6d22392
Use empty() instead of begin() == end().
...
llvm-svn: 54780
2008-08-14 18:13:49 +00:00
ef96ac4f95
Get rid of unused variable.
...
llvm-svn: 54742
2008-08-13 17:44:52 +00:00
65fce4d813
1) Merge entire live intervals instead of parts of them.
...
2) Conditionalize temporary insertion if we don't need it.
llvm-svn: 54741
2008-08-13 17:25:42 +00:00
d172c15ab0
Do a dominator walk when scheduling copies, rather than a DFS on the CFG. Also, fix a few problems when creating
...
live intervals for temporaries created by phi elimination.
llvm-svn: 54483
2008-08-07 18:28:07 +00:00
8a8d6f0a78
Correct handle cases where two phis are coalesced together, and correct break up the case where two different
...
phis want to coalesce with the same vreg.
llvm-svn: 54426
2008-08-06 22:08:58 +00:00
987b5057d3
We don't need to try to coalesce input vregs that are the same as the output vreg.
...
llvm-svn: 54422
2008-08-06 20:29:20 +00:00
f9fca2f2dc
Only trim a live interval if the register is not used after the PHI node.
...
llvm-svn: 54421
2008-08-06 18:36:17 +00:00
d9b88a85f2
Oops, we were already checking for dead phis. Handle this the proper way, then.
...
llvm-svn: 54371
2008-08-05 21:18:51 +00:00
d4ffa4eb57
We don't need to update live intervals for dead PHIs.
...
llvm-svn: 54369
2008-08-05 20:51:26 +00:00
c818c01539
Use existing LiveInterval methods to simplify live interval merging. Thanks to Evan for pointing these out.
...
llvm-svn: 54225
2008-07-30 18:27:35 +00:00
1aebe49ae7
When merging live intervals, we also need to merge in any live ranges that are inputs to two-address instructions
...
that themselves define a range we already care about.
llvm-svn: 54185
2008-07-30 00:21:16 +00:00
6b1cc46fee
When merging a PHI operand's live interval into the PHI's live interval, we need to merge over all liveranges in
...
the operand's interval that share the relevant value number, not just the range that immediately precedes the PHI.
llvm-svn: 54174
2008-07-29 21:17:08 +00:00
54912b3e8d
Fix the issues originally addressed in r54070. After thinking about it some more, I realized that the right thing to do
...
is to have StrongPHIElimination use its knowledge of the PHIs before they're erased to update the intervals appropriate. This is
both simpler and more accurate than the alternative, which was having LIA figure it out when it renumbered things, plus it's just
the right thing to do!
llvm-svn: 54077
2008-07-25 23:38:08 +00:00
0346aba5c2
In order to avoid reprocessing a register more than once, we need to add it
...
to the handled set so it will get filtered out in future iterations.
llvm-svn: 54065
2008-07-25 21:35:43 +00:00
d9c8711d70
Remove live interval entries for an interval if we're eliminating its only VN.
...
llvm-svn: 54062
2008-07-25 21:08:41 +00:00
79b66966b8
Store the predecessor MBB in the PHIUnion, rather than an index, since the indices will change after renumbering.
...
llvm-svn: 53985
2008-07-24 17:12:16 +00:00
05d1cd99b3
Remove debugging code.
...
llvm-svn: 52016
2008-06-05 18:43:34 +00:00
65dc2b2e06
Use the newly created helper on LiveIntervals.
...
llvm-svn: 52013
2008-06-05 17:22:53 +00:00
3a697779fe
Correctly construct live intervals for the copies we inserted into the predecessors of a block containing a PHI.
...
llvm-svn: 51950
2008-06-04 17:55:58 +00:00
c777d9e6fe
We need to subtract one from this index because live ranges are open at the end.
...
llvm-svn: 51922
2008-06-04 00:38:56 +00:00
Jakob Stoklund Olesen