e366c29fa0
Adjust the sizes for a few SmallVectors to reflect their usage.
...
llvm-svn: 64381
2009-02-12 17:29:01 +00:00
90880e2598
make fast isel fall back to selectiondags for VLA llvm.declare intrinsics.
...
llvm-svn: 64379
2009-02-12 17:23:20 +00:00
cf5cd6ecfe
It's (currently) not safe to keep certain physical registers live across basic blocks, e.g. x86 fp stack registers.
...
llvm-svn: 64374
2009-02-12 10:32:17 +00:00
b570499c25
Oops. Last second clean up messed things up.
...
llvm-svn: 64373
2009-02-12 09:52:13 +00:00
8b39b545a9
If availability info is kept when fallthrough into a bb, add the available registers to live-in set.
...
llvm-svn: 64372
2009-02-12 09:43:23 +00:00
3a14efacb6
Replace one of burr scheduling heuristic with something more sensible. Now calcMaxScratches simply compute the number of true data dependencies. This actually improve a couple of tests in dejagnu suite as many tests in llvm nightly test suite.
...
llvm-svn: 64369
2009-02-12 08:59:45 +00:00
6f6abfd94b
Apparently some MachineBasicBlock's don't have corresponding llvm basic blocks.
...
llvm-svn: 64340
2009-02-11 23:42:39 +00:00
eb5ec4a0db
Remove a bogus assertion. It's possible a live-in available value is used by a previous instruction.
...
llvm-svn: 64339
2009-02-11 23:41:57 +00:00
45889d24fe
Fix a comment.
...
llvm-svn: 64328
2009-02-11 21:32:08 +00:00
6571ef3577
Don't use special heuristics for nodes with no data predecessors
...
unless they actually have data successors, and likewise for nodes
with no data successors unless they actually have data precessors.
llvm-svn: 64327
2009-02-11 21:29:39 +00:00
298a2946f1
Delete the heuristic for non-livein CopyFromReg nodes. Non-liveinness
...
is determined by whether the node has a Flag operand. However, if the
node does have a Flag operand, it will be glued to its register's
def, so the heuristic would end up spuriously applying to whatever
node is the def.
llvm-svn: 64319
2009-02-11 20:25:59 +00:00
cc5fc44d02
Make a transformation added in 63266 a bit less aggressive.
...
It was transforming (x&y)==y to (x&y)!=0 in the case where
y is variable and known to have at most one bit set (e.g. z&1).
This is not correct; the expressions are not equivalent when y==0.
I believe this patch salvages what can be salvaged, including
all the cases in bt.ll. Dan, please review.
Fixes gcc.c-torture/execute/20040709-[12].c
llvm-svn: 64314
2009-02-11 19:19:41 +00:00
a1968b0fc7
Implement PR3495: local spiller optimization. The local spiller can now keep availability information over BB boundaries. It visits BB's in depth first order. After visiting a BB if it find a successor which has a single predecessor it visits the successor next without clearing the availability information. This allows the successor to omit reloads or change them into copies.
...
llvm-svn: 64298
2009-02-11 08:24:21 +00:00
dfaf646c34
When scheduling a block in parts, keep track of the overall
...
instruction index across each part. Instruction indices are used
to make live range queries, and live ranges can extend beyond
scheduling region boundaries.
Refactor the ScheduleDAGSDNodes class some more so that it
doesn't have to worry about this additional information.
llvm-svn: 64288
2009-02-11 04:27:20 +00:00
27fa408b64
Use iterators to iterate through the Preds array instead of
...
an index. This code is on the hot-path because the current
way SDep edges are uniqued has quadratic complexity.
llvm-svn: 64262
2009-02-11 00:12:28 +00:00
b3dbb21df1
Consider any instruction that modifies the stack pointer to be
...
a scheduling region boundary. This isn't necessary for
correctness; it helps with compile time, as it avoids the need
for data- and anti-dependencies from all spills and reloads on
the stack-pointer modification.
llvm-svn: 64255
2009-02-10 23:29:38 +00:00
b95434356c
Factor out more code for computing register live-range informationfor
...
scheduling, and generalize is so that preserves state across
scheduling regions. This fixes incorrect live-range information around
terminators and labels, which are effective region boundaries.
In place of looking for terminators to anchor inter-block dependencies,
introduce special entry and exit scheduling units for this purpose.
llvm-svn: 64254
2009-02-10 23:27:53 +00:00
ce3bbe515b
Fix PR3457: Ignore control successors when looking for closest scheduled successor. A control successor doesn't read result(s) produced by the scheduling unit being evaluated.
...
llvm-svn: 64210
2009-02-10 08:30:11 +00:00
9b6c1d6313
Validate file id.
...
llvm-svn: 64204
2009-02-10 06:04:08 +00:00
3af42a8a14
If the target cannot issue a copy for the given source and dest registers, abort instead of silently continue.
...
llvm-svn: 64184
2009-02-09 22:47:36 +00:00
fe174df170
Simplify code.
...
llvm-svn: 64164
2009-02-09 21:01:06 +00:00
020588cee3
Make sure constant subscript is truncated to ptr size if it may not fit.
...
llvm-svn: 64163
2009-02-09 20:54:38 +00:00
f736bd9c79
Re-enable machine sinking pass now that the coalescer bugs and the AnalyzeBrnach bug are fixed.
...
llvm-svn: 64126
2009-02-09 08:45:39 +00:00
81ae066ab6
Fix another case ShortenDeadCopySrcLiveRange is shortening too much. No test case possible since I don't know what to grep for. :-(
...
llvm-svn: 64125
2009-02-09 08:37:45 +00:00
64dfcacd5f
Turns out AnalyzeBranch can modify the mbb being analyzed. This is a nasty
...
suprise to some callers, e.g. register coalescer. For now, add an parameter
that tells AnalyzeBranch whether it's safe to modify the mbb. A better
solution is out there, but I don't have time to deal with it right now.
llvm-svn: 64124
2009-02-09 07:14:22 +00:00
b3783639cb
Fix PR3486. Fix a bug in code that manually patch physical register live interval after its sub-register is coalesced with a virtual register.
...
llvm-svn: 64082
2009-02-08 11:04:35 +00:00
942519306e
Strengthen the previous check.
...
llvm-svn: 64076
2009-02-08 08:24:28 +00:00
fe014df914
r64073 commit message is lost. Here it is:
...
Right now if the coalesced copy def is dead and its src is a kill, and that
there are now other uses within the live range, the coalescer would mark the
def of the source register as dead. But it should also check if there are
other kills which means the value has other uses not in the live range.
llvm-svn: 64075
2009-02-08 08:00:36 +00:00
e5e95f7717
(no commit message)
...
llvm-svn: 64073
2009-02-08 07:48:37 +00:00
5469ec1072
Revert r63999. It was breaking self-hosting builds.
...
llvm-svn: 64062
2009-02-08 00:58:05 +00:00
d8f0786c31
Add missing break statements, fixing PR3503.
...
llvm-svn: 64040
2009-02-07 22:37:06 +00:00
9c310711bb
Use getDebugLoc forwarder instead of getNode()->getDebugLoc.
...
No functional change.
llvm-svn: 64026
2009-02-07 19:59:05 +00:00
747e55bc9a
Constify TargetInstrInfo::EmitInstrWithCustomInserter, allowing
...
ScheduleDAG's TLI member to use const.
llvm-svn: 64018
2009-02-07 16:15:20 +00:00
8ba7132128
Make SDNode constructors take a DebugLoc always.
...
Adjust derived classes to pass UnknownLoc where
a DebugLoc does not make sense. Pick one of
DebugLoc and non-DebugLoc variants to survive
for all such classes.
llvm-svn: 64000
2009-02-07 02:15:05 +00:00
9571621665
Enable machine sinking pass in non-fast mode.
...
llvm-svn: 63999
2009-02-07 01:57:46 +00:00
a72d41a67b
Remove now-unused constructors.
...
llvm-svn: 63995
2009-02-07 01:27:09 +00:00
47a65a167d
Don't sink the instruction if TargetRegisterInfo::isSafeToMoveRegClassDefs doesn't think it's safe. This works around PR1911.
...
llvm-svn: 63994
2009-02-07 01:21:47 +00:00
62fd95d6ec
Get rid of the last non-DebugLoc versions of getNode!
...
Many targets build placeholder nodes for special operands, e.g.
GlobalBaseReg on X86 and PPC for the PIC base. There's no
sensible way to associate debug info with these. I've left
them built with getNode calls with explicit DebugLoc::getUnknownLoc operands.
I'm not too happy about this but don't see a good improvement;
I considered adding a getPseudoOperand or something, but it
seems to me that'll just make it harder to read.
llvm-svn: 63992
2009-02-07 00:55:49 +00:00
84935759d5
Remove more non-DebugLoc getNode variants. Use
...
getCALLSEQ_{END,START} to permit passing no DebugLoc
there. UNDEF doesn't logically have DebugLoc; add
getUNDEF to encapsulate this.
llvm-svn: 63978
2009-02-06 23:05:02 +00:00
dc93bbc4b0
And one more file.
...
llvm-svn: 63971
2009-02-06 21:55:48 +00:00
400dc2e2e4
Remove more non-DebugLoc versions of getNode.
...
llvm-svn: 63969
2009-02-06 21:50:26 +00:00
03c34d0d3c
Clear out the CurDebugLoc info when doing a 'clear' on the SDL object.
...
llvm-svn: 63967
2009-02-06 21:36:23 +00:00
ab8e4425a3
Eliminate remaining non-DebugLoc version of getTargetNode.
...
llvm-svn: 63951
2009-02-06 19:16:40 +00:00
817a24f8e9
Rename SelectionDAGISel::Schedule to
...
SelectionDAGISel::CreateScheduler, and make it just create the
scheduler. Leave running the scheduler to the higher-level code.
This makes the higher-level code a little more explicit and
easier to follow, and will help enable some future refactoring.
llvm-svn: 63944
2009-02-06 18:26:51 +00:00
cd2cd9f5d7
Delete an unused member function.
...
llvm-svn: 63941
2009-02-06 18:19:52 +00:00
066757eea1
Move getPointerRegClass from TargetInstrInfo to TargetRegisterInfo.
...
llvm-svn: 63938
2009-02-06 17:43:24 +00:00
483377c639
Move ScheduleDAGSDNodes.h to be a private header. Front-ends
...
that previously included this header should include
SchedulerRegistry.h instead.
llvm-svn: 63937
2009-02-06 17:22:58 +00:00
b5f0ec3eb7
Add TargetInstrInfo::isSafeToMoveRegisterClassDefs. It returns true if it's safe to move an instruction which defines a value in the register class. Replace pre-splitting specific IgnoreRegisterClassBarriers with this new hook.
...
llvm-svn: 63936
2009-02-06 17:17:30 +00:00
f4b08b4f6c
Move ScheduleDAGInstrs.h to be a private header. Front-ends
...
that used this header to select a scheduling policy should
use SchedulerRegistry.h instead (llvm-gcc and clang were
updated a while ago).
llvm-svn: 63934
2009-02-06 17:12:10 +00:00
2c4cf2752d
get rid of some non-DebugLoc getTargetNode variants.
...
llvm-svn: 63909
2009-02-06 02:08:06 +00:00
9f3f72f144
Get rid of one more non-DebugLoc getNode and
...
its corresponding getTargetNode. Lots of
caller changes.
llvm-svn: 63904
2009-02-06 01:31:28 +00:00
2599084ac5
isAsCheapAsMove instructions can have register src operands. Check if they are really re-materializable.
...
This fixes sse.expandfft and sse.stepfft.
llvm-svn: 63890
2009-02-05 22:24:17 +00:00
f80493bbfd
Remove a non-DebugLoc version of getNode.
...
llvm-svn: 63889
2009-02-05 22:07:54 +00:00
3eb373f5ce
Remove 3 non-DebugLoc variants of getNode.
...
llvm-svn: 63886
2009-02-05 21:20:44 +00:00
9d8bfbfdef
Machine LICM increases register pressure and it almost always increase code size. For now, disable it for optimizing for size.
...
llvm-svn: 63856
2009-02-05 08:51:13 +00:00
409c25f78d
Turn on machine LICM in non-fast mode.
...
llvm-svn: 63855
2009-02-05 08:46:33 +00:00
399660c384
Teach machine licm to CSE hoisted instructions.
...
llvm-svn: 63854
2009-02-05 08:45:46 +00:00
0f734469eb
ReMaterializeTrivialDef need to trim the live interval to the last kill if the copy kills the source register. This fixes uint64tof64.ll after ARM::MOVi is marked as isAsCheapAsAMove.
...
llvm-svn: 63853
2009-02-05 08:45:04 +00:00
7e1d2862ca
if we have a large GEP offset on a 32-bit or other target, make
...
sure to print the value properly sext'd to the right pointer size.
This fixes PR3481.
llvm-svn: 63843
2009-02-05 06:55:21 +00:00
7411523ee6
Pre-alloc splitting needs to be more careful to avoid inserting spills/restores
...
between call frame setup/restore points. Unfortunately, this regresses
code size a bit, but at least it's correct now!
llvm-svn: 63837
2009-02-05 05:58:41 +00:00
3f0e0a6dea
Fix a bug where we were not emitting a cvt rnd sat node for converting
...
between a unsigned integer and signed integer.
llvm-svn: 63831
2009-02-05 04:47:42 +00:00
b842d529a3
Reapply 63765. Patches for clang and llvm-gcc to follow.
...
llvm-svn: 63812
2009-02-05 01:49:45 +00:00
12c572b6fa
Get rid of 3 non-DebugLoc getNode variants.
...
llvm-svn: 63808
2009-02-05 01:01:16 +00:00
7ae8c8b108
Remove non-DebugLoc versions of getMergeValues, ZeroExtendInReg.
...
llvm-svn: 63800
2009-02-05 00:20:09 +00:00
f08a47bb70
Remove non-DebugLoc forms of CopyToReg and CopyFromReg.
...
Adjust callers.
llvm-svn: 63789
2009-02-04 23:02:30 +00:00
ae616c2c61
Reverting 63765. This broke the build of both clang
...
and llvm-gcc.
llvm-svn: 63786
2009-02-04 22:47:25 +00:00
556bd92698
80 column rule.
...
llvm-svn: 63768
2009-02-04 20:30:10 +00:00
021052a705
Remove non-DebugLoc versions of getLoad and getStore.
...
Adjust the many callers of those versions.
llvm-svn: 63767
2009-02-04 20:06:27 +00:00
6ae3aa83d0
New feature: add support for target intrinsics being defined in the
...
target directories themselves. This also means that VMCore no longer
needs to know about every target's list of intrinsics. Future work
will include converting the PowerPC target to this interface as an
example implementation.
llvm-svn: 63765
2009-02-04 19:47:21 +00:00
34650735d0
Avoids generating a legalization assert for the case where a vector type is legal
...
but when legalizing the operation, we split the vector type and generate a library
call whose type needs to be promoted. For example, X86 with SSE on but MMX off,
a divide v2i64 will be scalarized to 2 calls to a library using i64.
llvm-svn: 63760
2009-02-04 19:38:14 +00:00
9ec370f748
Skip over zero registers.
...
llvm-svn: 63748
2009-02-04 18:18:58 +00:00
ffee3d831a
Since I'm obliged to work with a development OS that currently doesn't
...
support GraphViz, I've been using the foo->dump() facility. This
patch is a minor rewrite to the SelectionDAG dump() stuff to make it a
little more helpful. The existing foo->dump() functionality does not
change; this patch adds foo->dumpr(). All of this is only useful when
running LLVM under a debugger.
llvm-svn: 63736
2009-02-04 16:46:19 +00:00
4ded601445
Don't call isInvariantLoad twice.
...
llvm-svn: 63729
2009-02-04 09:21:58 +00:00
1d9f7ac8cc
Don't bother hoisting out a "cheap" instruction if all of its uses are PHIs. LICM "cheap" instructions are not particularly beneficial to start with. This will just end up making the copies harder to coalesce.
...
llvm-svn: 63728
2009-02-04 09:19:56 +00:00
0649abdfe2
For now, only hoist re-materilizable instructions. LICM will increase register pressure. We want to avoid spilling more instructions if it's possible.
...
llvm-svn: 63725
2009-02-04 07:17:49 +00:00
679073b420
Remove non-DebugLoc forms of the exotic forms
...
of Lod and Sto; patch uses.
llvm-svn: 63716
2009-02-04 02:34:38 +00:00
f2bb6f09a3
Remove some more non-DebugLoc versions of construction
...
functions, with callers adjusted to fit.
llvm-svn: 63705
2009-02-04 01:48:28 +00:00
efb82cfbf2
Check in file I forgot.
...
llvm-svn: 63704
2009-02-04 01:33:20 +00:00
85263882aa
Remove a few non-DebugLoc versions of node creation
...
functions.
llvm-svn: 63703
2009-02-04 01:17:06 +00:00
9888edee10
Fill in more omissions in DebugLog propagation.
...
I think that's it for this directory.
llvm-svn: 63690
2009-02-04 00:13:36 +00:00
0b4403f9bd
Constify and don't return a reference.
...
llvm-svn: 63689
2009-02-04 00:05:34 +00:00
3a09f5589d
DebugLoc propagation; adjustment to things omitted
...
from SelectionDagBuild.
llvm-svn: 63680
2009-02-03 23:04:43 +00:00
77b3be8ea3
Erm. Would be nice to initialize this variable. . .
...
llvm-svn: 63677
2009-02-03 22:55:54 +00:00
2c0a4fe4e2
Add getDebugLocTuple to retrieve the DebugLocTuple for a given DebugLoc object.
...
llvm-svn: 63676
2009-02-03 22:49:58 +00:00
abf66b8343
Add some DL propagation to places that didn't
...
have it yet. More coming.
llvm-svn: 63673
2009-02-03 22:26:09 +00:00
dd7fa33679
More whitespace fixin'.
...
llvm-svn: 63663
2009-02-03 21:38:21 +00:00
eaf02df7b4
Whitespace and comment changes. No functionality change.
...
llvm-svn: 63660
2009-02-03 21:17:20 +00:00
9c7c0db37a
Remove dead code.
...
llvm-svn: 63653
2009-02-03 19:46:28 +00:00
d564353193
Change the post-RA scheduler to iterate through the
...
basic-block segments bottom-up instead of top down. This
is the first step in a general restructuring of the way
register liveness is tracked in the post-RA scheduler.
llvm-svn: 63643
2009-02-03 18:57:45 +00:00
70da8e8425
First initialize DAG otherwise dwarf writer is used uninitialized.
...
Duncan spotted this. Thanks!
llvm-svn: 63641
2009-02-03 18:46:32 +00:00
a77c5f758c
Fix PR3411. When replacing values, nodes are analyzed
...
in any old order. Since analyzing a node analyzes its
operands also, this can mean that when we pop a node
off the list of nodes to be analyzed, it may already
have been analyzed.
llvm-svn: 63632
2009-02-03 10:23:33 +00:00
67cd395d45
Explicitly pass in debug location information to BuildMI.
...
llvm-svn: 63599
2009-02-03 02:29:34 +00:00
135227a060
Pass in something sensible for the debug location information when creating the
...
initial PHI nodes of the machine function.
llvm-svn: 63598
2009-02-03 02:20:52 +00:00
db39362c90
Fill in some missing DL propagation in getNode()s.
...
llvm-svn: 63595
2009-02-03 01:55:44 +00:00
143a2c3470
Use SDL->getCurDebugLoc() instead of unknown loc for landing pads.
...
llvm-svn: 63594
2009-02-03 01:55:42 +00:00
fa50a23f8a
Explicitly pass in the "unknown" debug location. This is probably not
...
correct. We need more infrastructure before we can get the DebugLoc info for
these instructions.
llvm-svn: 63593
2009-02-03 01:33:28 +00:00
9862a64419
Alphabetize includes.
...
llvm-svn: 63591
2009-02-03 01:32:22 +00:00
17450acc3b
Propagate debug loc info during SDNode -> machine instr creation.
...
llvm-svn: 63585
2009-02-03 01:02:39 +00:00
e3c78361d3
Create DebugLoc information in FastISel. Several temporary methods were
...
created. Specifically, those BuildMIs which use
"DebugLoc::getUnknownLoc()". I'll remove them soon.
llvm-svn: 63584
2009-02-03 00:55:04 +00:00
f1163e9a4d
Propagation in TargetLowering. Includes passing a DL
...
into SimplifySetCC which gets called elsewhere.
llvm-svn: 63583
2009-02-03 00:47:48 +00:00
Dan Gohman