199 Commits

Author SHA1 Message Date
Evan Cheng
916802a78e Start of a series of patches related to implicit_def.
There is no point in creating a long live range defined by an implicit_def. Scheduler now duplicates implicit_def instruction for each of its uses. Therefore, if an implicit_def node has multiple uses, it will become a number of very short live ranges, rather than a long one. This will make coalescer's job easier.

llvm-svn: 49164
2008-04-03 16:36:07 +00:00
Evan Cheng
025cea1126 Backing out 48222 temporarily.
llvm-svn: 49124
2008-04-03 03:13:16 +00:00
Roman Levenstein
358e04a185 Use a linked data structure for the uses lists of an SDNode, just like
LLVM Value/Use does and MachineRegisterInfo/MachineOperand does.
This allows constant time for all uses list maintenance operations.

The idea was suggested by Chris. Reviewed by Evan and Dan.
Patch is tested and approved by Dan.

On normal use-cases compilation speed is not affected. On very big basic
blocks there are compilation speedups in the range of 15-20% or even better. 

llvm-svn: 48822
2008-03-26 12:39:26 +00:00
Dan Gohman
bdc24adaaf A quick nm audit turned up several fixed tables and objects that were
marked read-write. Use const so that they can be allocated in a
read-only segment.

llvm-svn: 48800
2008-03-25 21:45:14 +00:00
Christopher Lamb
d3d0ad3f58 Make insert_subreg a two-address instruction, vastly simplifying LowerSubregs pass. Add a new TII, subreg_to_reg, which is like insert_subreg except that it takes an immediate implicit value to insert into rather than a register.
llvm-svn: 48412
2008-03-16 03:12:01 +00:00
Evan Cheng
0e7b00d79f Replace all target specific implicit def instructions with a target independent one: TargetInstrInfo::IMPLICIT_DEF.
llvm-svn: 48380
2008-03-15 00:03:38 +00:00
Evan Cheng
db443ca377 Livein copy scheduling fixes: do not coalesce physical register copies, correctly determine the safe location to insert the copies.
llvm-svn: 48348
2008-03-14 00:14:55 +00:00
Evan Cheng
65e9d5f1a8 Experimental scheduler change to schedule / coalesce the copies added for function livein's. Take 2008-03-10-RegAllocInfLoop.ll, the schedule looks like this after these copies are inserted:
entry: 0x12049d0, LLVM BB @0x1201fd0, ID#0:
Live Ins: %EAX %EDX %ECX
        %reg1031<def> = MOVPC32r 0
        %reg1032<def> = ADD32ri %reg1031, <es:_GLOBAL_OFFSET_TABLE_>, %EFLAGS<imp-def>
        %reg1028<def> = MOV32rr %EAX
        %reg1029<def> = MOV32rr %EDX
        %reg1030<def> = MOV32rr %ECX
        %reg1027<def> = MOV8rm %reg0, 1, %reg0, 0, Mem:LD(1,1) [0x1201910 + 0]
        %reg1025<def> = MOV32rr %reg1029
        %reg1026<def> = MOV32rr %reg1030
        %reg1024<def> = MOV32rr %reg1028

The copies unnecessarily increase register pressure and it will end up requiring a physical register to be spilled.

With -schedule-livein-copies:
entry: 0x12049d0, LLVM BB @0x1201fa0, ID#0:
Live Ins: %EAX %EDX %ECX
        %reg1031<def> = MOVPC32r 0
        %reg1032<def> = ADD32ri %reg1031, <es:_GLOBAL_OFFSET_TABLE_>, %EFLAGS<imp-def>
        %reg1024<def> = MOV32rr %EAX
        %reg1025<def> = MOV32rr %EDX
        %reg1026<def> = MOV32rr %ECX
        %reg1027<def> = MOV8rm %reg0, 1, %reg0, 0, Mem:LD(1,1) [0x12018e0 + 0]

Much better!

llvm-svn: 48307
2008-03-12 22:19:41 +00:00
Christopher Lamb
aa7c2105de Recommitting parts of r48130. These do not appear to cause the observed failures.
llvm-svn: 48223
2008-03-11 10:09:17 +00:00
Evan Cheng
e88a625ecd When the register allocator runs out of registers, spill a physical register around the def's and use's of the interval being allocated to make it possible for the interval to target a register and spill it right away and restore a register for uses. This likely generates terrible code but is before than aborting.
llvm-svn: 48218
2008-03-11 07:19:34 +00:00
Chris Lattner
3e0ec65678 variadic instructions don't have operand info for variadic arguments.
llvm-svn: 48208
2008-03-11 03:14:42 +00:00
Chris Lattner
d3090bcfc8 If a register operand comes from the variadic part of a node, don't
verify the register constraint matches what the instruction expects.

llvm-svn: 48205
2008-03-11 00:59:28 +00:00
Dan Gohman
2a3aeb1f72 Correctly clone FlaggedNodes.
llvm-svn: 48196
2008-03-10 23:48:14 +00:00
Evan Cheng
d4e1d9eeb2 Revert 48125, 48126, and 48130 for now to unbreak some x86-64 tests.
llvm-svn: 48167
2008-03-10 19:31:26 +00:00
Christopher Lamb
4ba3f0430b Allow insert_subreg into implicit, target-specific values.
Change insert/extract subreg instructions to be able to be used in TableGen patterns.
Use the above features to reimplement an x86-64 pseudo instruction as a pattern.

llvm-svn: 48130
2008-03-10 06:12:08 +00:00
Chris Lattner
86829f0ff7 teach X86InstrInfo::copyRegToReg how to copy into ST(0) from
an RFP register class.

Teach ScheduleDAG how to handle CopyToReg with different src/dst 
reg classes.

This allows us to compile trivial inline asms that expect stuff
on the top of x87-fp stack.

llvm-svn: 48107
2008-03-09 09:15:31 +00:00
Chris Lattner
9e07537e8c Add ScheduleDAG support for copytoreg where the src/dst register are
in different register classes, e.g. copy of ST(0) to RFP*.  This gets
some really trivial inline asm working that plops things on the top of
stack (PR879)

llvm-svn: 48105
2008-03-09 08:49:15 +00:00
Chris Lattner
381bbdb924 fix 80 col violation
llvm-svn: 48100
2008-03-09 07:51:01 +00:00
Roman Levenstein
c62c2bb4d0 Some improvements related to the computation of heights, depths of SUnits.
The basic idea is that all these algorithms are computing the longest paths from the root node or to the exit node. Therefore the existing implementation that uses and iterative and potentially
exponential algorithm was changed to a well-known graph algorithm based on dynamic programming. It has a linear run-time.

llvm-svn: 47884
2008-03-04 11:19:43 +00:00
Evan Cheng
567d2e5b57 Rename isOperand() to isOperandOf() (and other similar methods). It always confuses me.
llvm-svn: 47872
2008-03-04 00:41:45 +00:00
Evan Cheng
a465bfb87c Keep track how many commutes are performed by the scheduler.
llvm-svn: 47710
2008-02-28 07:40:24 +00:00
Dan Gohman
27ae573900 Rename CountMemOperands to ComputeMemOperandsEnd to reflect what
it actually does. Simplify CountOperands a little by reusing
ComputeMemOperandsEnd. And reword some comments for both.

llvm-svn: 47198
2008-02-16 00:36:48 +00:00
Nate Begeman
53e1b3f9d5 Change how FP immediates are handled.
1) ConstantFP is now expand by default
2) ConstantFP is not turned into TargetConstantFP during Legalize
   if it is legal.

This allows ConstantFP to be handled like Constant, allowing for 
targets that can encode FP immediates as MachineOperands.

As a bonus, fix up Itanium FP constants, which now correctly match,
and match more constants!  Hooray.

llvm-svn: 47121
2008-02-14 08:57:00 +00:00
Dan Gohman
11f6212bc0 From Chris' review: use isa instead of explicitly using classof.
llvm-svn: 46964
2008-02-11 19:00:34 +00:00
Dan Gohman
991056808b From Chris' review: minor corrections in comments.
llvm-svn: 46963
2008-02-11 19:00:03 +00:00
Dan Gohman
3a4be0fdef Rename MRegisterInfo to TargetRegisterInfo.
llvm-svn: 46930
2008-02-10 18:45:23 +00:00
Dan Gohman
2d489b5081 Re-apply the memory operand changes, with a fix for the static
initializer problem, a minor tweak to the way the
DAGISelEmitter finds load/store nodes, and a renaming of the
new PseudoSourceValue objects.

llvm-svn: 46827
2008-02-06 22:27:42 +00:00
Evan Cheng
efd142a920 SDIsel processes llvm.dbg.declare by recording the variable debug information descriptor and its corresponding stack frame index in MachineModuleInfo. This only works if the local variable is "homed" in the stack frame. It does not work for byval parameter, etc.
Added ISD::DECLARE node type to represent llvm.dbg.declare intrinsic. Now the intrinsic calls are lowered into a SDNode and lives on through out the codegen passes.
For now, since all the debugging information recording is done at isel time, when a ISD::DECLARE node is selected, it has the side effect of also recording the variable. This is a short term solution that should be fixed in time.

llvm-svn: 46659
2008-02-02 04:07:54 +00:00
Evan Cheng
d6e44ab5ec Remove the nasty LABEL hack with a much less evil one. Now llvm.dbg.func.start implies a stoppoint is set. SelectionDAGISel records a new source line but does not create a ISD::LABEL node for this special stoppoint. Asm printer will magically print this label. This ensures nothing is emitted before.
llvm-svn: 46635
2008-02-01 09:10:45 +00:00
Evan Cheng
27b32b87ed Revert 46556 and 46585. Dan please fix the PseudoSourceValue problem and re-commit.
llvm-svn: 46623
2008-01-31 21:00:00 +00:00
Evan Cheng
f4f1d44779 Add a comment for a nasty short term hack.
llvm-svn: 46610
2008-01-31 10:05:13 +00:00
Evan Cheng
1c6c16ea11 Add an extra operand to LABEL nodes which distinguishes between debug, EH, or misc labels. This fixes the EH breakage. However I am not convinced this is *the* solution.
llvm-svn: 46609
2008-01-31 09:59:15 +00:00
Christopher Lamb
58ffa8c57a Add more thorough error checking for NULL register classes.
llvm-svn: 46605
2008-01-31 07:09:08 +00:00
Dan Gohman
3646fdda67 Create a new class, MemOperand, for describing memory references
in the backend. Introduce a new SDNode type, MemOperandSDNode, for
holding a MemOperand in the SelectionDAG IR, and add a MemOperand
list to MachineInstr, and code to manage them. Remove the offset
field from SrcValueSDNode; uses of SrcValueSDNode that were using
it are all all using MemOperandSDNode now.

Also, begin updating some getLoad and getStore calls to use the
PseudoSourceValue objects.

Most of this was written by Florian Brander, some
reorganization and updating to TOT by me.

llvm-svn: 46585
2008-01-31 00:25:39 +00:00
Evan Cheng
a3395a61cc Treat the label for the first @llvm.dbg.stoppoint the same way as the dbg_func_start label. Make sure nothing else is inserted before them.
Note this solution might be somewhat fragile since ISD::LABEL may be used for other
purposes. If that ends up to be an issue, we may need to introduce a different node
for debug labels.

llvm-svn: 46571
2008-01-30 20:08:35 +00:00
Evan Cheng
a3ff8e6110 A semi-gross fix for a debug info issue. When inserting the "function start" label (i.e. first label in the entry block) take care to insert it at the beginning of the block.
llvm-svn: 46568
2008-01-30 19:35:32 +00:00
Evan Cheng
29cfb67e28 Even though InsertAtEndOfBasicBlock is an ugly hack it still deserves a proper name. Rename it to EmitInstrWithCustomInserter since it does not necessarily insert
instruction at the end.

llvm-svn: 46562
2008-01-30 18:18:23 +00:00
Dan Gohman
70de4cb1cd Use empty() instead of comparing size() with zero.
llvm-svn: 46514
2008-01-29 13:02:09 +00:00
Chris Lattner
03ad885039 rename TargetInstrDescriptor -> TargetInstrDesc.
Make MachineInstr::getDesc return a reference instead
of a pointer, since it can never be null.

llvm-svn: 45695
2008-01-07 07:27:27 +00:00
Chris Lattner
e99a6caee4 Rename all the M_* flags to be namespace qualified enums, and switch
all clients over to using predicates instead of these flags directly.
These are now private values which are only to be used to statically
initialize the tables.

llvm-svn: 45692
2008-01-07 06:42:05 +00:00
Chris Lattner
08a69ac2f5 add more and significantly better comments to the rest of the machineinstr
flags that can be set.  Add predicates for the ones lacking it, and switch
some clients over to using the predicates instead of Flags directly.

llvm-svn: 45690
2008-01-07 06:21:53 +00:00
Chris Lattner
f376c99ea0 rename hasVariableOperands() -> isVariadic(). Add some comments.
Evan, please review the comments I added to getNumDefs to make sure
that they are accurate, thx.

llvm-svn: 45687
2008-01-07 05:19:29 +00:00
Chris Lattner
b0d06b4381 Move a bunch more accessors from TargetInstrInfo to TargetInstrDescriptor
llvm-svn: 45680
2008-01-07 03:13:06 +00:00
Chris Lattner
d34c47653e remove some uses of MachineOpCode, move getSchedClass
into TargetInstrDescriptor from TargetInstrInfo.

llvm-svn: 45678
2008-01-07 02:46:03 +00:00
Chris Lattner
e55e115616 Add predicates methods to TargetOperandInfo, and switch all clients
over to using them, instead of diddling Flags directly.  Change the
various flags from const variables to enums.

llvm-svn: 45677
2008-01-07 02:39:19 +00:00
Chris Lattner
a98c679de0 Rename MachineInstr::getInstrDescriptor -> getDesc(), which reflects
that it is cheap and efficient to get.

Move a variety of predicates from TargetInstrInfo into 
TargetInstrDescriptor, which makes it much easier to query a predicate
when you don't have TII around.  Now you can use MI->getDesc()->isBranch()
instead of going through TII, and this is much more efficient anyway. Not
all of the predicates have been moved over yet.

Update old code that used MI->getInstrDescriptor()->Flags to use the
new predicates in many places.

llvm-svn: 45674
2008-01-07 01:56:04 +00:00
Owen Anderson
7a73ae9a86 Move copyRegToReg from MRegisterInfo to TargetInstrInfo. This is part of the
Machine-level API cleanup instigated by Chris.

llvm-svn: 45470
2007-12-31 06:32:00 +00:00
Chris Lattner
21ec2b4769 update a couple of references to SSARegMap.
llvm-svn: 45468
2007-12-31 04:16:08 +00:00
Chris Lattner
a10fff51d9 Rename SSARegMap -> MachineRegisterInfo in keeping with the idea
that "machine" classes are used to represent the current state of
the code being compiled.  Given this expanded name, we can start 
moving other stuff into it.  For now, move the UsedPhysRegs and
LiveIn/LoveOuts vectors from MachineFunction into it.

Update all the clients to match.

This also reduces some needless #includes, such as MachineModuleInfo
from MachineFunction.

llvm-svn: 45467
2007-12-31 04:13:23 +00:00
Chris Lattner
bbbae8e1ce use simplified operand addition methods.
llvm-svn: 45435
2007-12-30 00:51:11 +00:00