Chris Lattner
2c2062cdba
Build the SelectionDAG library
...
llvm-svn: 7718
2003-08-11 14:57:48 +00:00
Chris Lattner
600d308853
Initial checkin of SelectionDAG implementation. This is still rough and
...
unfinished
llvm-svn: 7717
2003-08-11 14:57:33 +00:00
Vikram S. Adve
da92f366c7
Fix assertion in MachineInstr::substituteValue().
...
llvm-svn: 7675
2003-08-07 15:01:48 +00:00
Chris Lattner
30e987470f
Use a new local data structure instead of the MachineInstr::regsUsed set
...
llvm-svn: 7621
2003-08-05 22:11:13 +00:00
Chris Lattner
92f5fb598c
Add a map
...
llvm-svn: 7620
2003-08-05 22:09:31 +00:00
Chris Lattner
e62a2a76e5
Minor cleanups
...
llvm-svn: 7619
2003-08-05 22:03:27 +00:00
Chris Lattner
b05d3508d6
Physical registers no longer live in the regsUsed set for each machine instr
...
llvm-svn: 7618
2003-08-05 21:55:58 +00:00
Chris Lattner
020b544cbc
Do not insert physical regsiters into the regsUsed set
...
llvm-svn: 7617
2003-08-05 21:55:20 +00:00
Chris Lattner
cd4f4320b0
All callers of these methods actually wanted them to preserve the flags,
...
so get rid of the def/use parameters that were getting passed in.
**** This now changes the semantics of these methods to preserve the flags,
not clobber them!
llvm-svn: 7602
2003-08-05 16:58:46 +00:00
Chris Lattner
92a199dca7
Fix bugs handling ESP in alloca references
...
llvm-svn: 7591
2003-08-05 04:13:58 +00:00
Chris Lattner
506fa6843e
Revert previous change, and be really anal about what physical registers can do.
...
llvm-svn: 7588
2003-08-05 00:49:09 +00:00
Chris Lattner
4b637cc666
Don't bother calculating info unless its needed. May reduce number of stack slots created.
...
llvm-svn: 7584
2003-08-04 23:42:37 +00:00
Chris Lattner
815b85e3ba
* Fix spelling of 'necessary'
...
* Add a lot more DEBUG output, which is better structured than before
* Fix bug: Jello/2003-08-04-PhysRegLiveFailure.llx
llvm-svn: 7583
2003-08-04 23:36:39 +00:00
Chris Lattner
41a894de77
Simplify code, eliminating the need for the X86 isVoid target instr flag
...
llvm-svn: 7534
2003-08-03 21:51:45 +00:00
Chris Lattner
74e4e9b44f
Set debug types
...
llvm-svn: 7533
2003-08-03 21:47:31 +00:00
Chris Lattner
1789d23990
Remove using decl
...
llvm-svn: 7531
2003-08-03 20:24:29 +00:00
Chris Lattner
6cce665888
Wrap at 80 columns
...
llvm-svn: 7503
2003-08-03 13:49:03 +00:00
Chris Lattner
1007f033d4
Move DEBUG to Debug.h
...
llvm-svn: 7497
2003-08-01 22:21:34 +00:00
Vikram S. Adve
9143f1962e
All constant-evaluation code now unified into
...
TargetInstrInfo::ConvertConstantToIntType().
llvm-svn: 7390
2003-07-29 19:50:12 +00:00
Vikram S. Adve
c9cb319bcb
1. Bug fix: Don't use branch operand reg. as temp. reg. when
...
spilling values used by an instruction in the delay slot of the branch
(which will eventually be moved before the branch).
2. Bug fix: Delete the delay slot instr, not the branch instr, when
moving delay slot instr. out!!!!
3. Move code to insert caller-saves moved here from SparcRegInfo:
it is now machine-independent.
llvm-svn: 7389
2003-07-29 19:49:21 +00:00
Vikram S. Adve
e2ef34a818
Add const version of getLiveRangeForValue().
...
llvm-svn: 7386
2003-07-29 19:38:22 +00:00
Vikram S. Adve
91e75d87d2
Code to insert caller-saves moved here from SparcRegInfo: it is now
...
machine-independent.
Fix problem with using branch operand reg. as temp. reg. when
spilling values used by an instruction in the delay slot of the branch
(which will eventually be moved before the branch).
llvm-svn: 7385
2003-07-29 19:37:41 +00:00
Chris Lattner
0c8de4bfca
Making this code const-correct would be a pain, so I'll hack it.
...
llvm-svn: 7350
2003-07-26 23:29:51 +00:00
Chris Lattner
414832fea6
Const correctness fixes
...
llvm-svn: 7349
2003-07-26 23:24:56 +00:00
Chris Lattner
0ed86aca71
Be const correct
...
llvm-svn: 7348
2003-07-26 23:23:41 +00:00
Chris Lattner
1316383267
Hrm, another necesary one :(
...
llvm-svn: 7347
2003-07-26 23:22:19 +00:00
Chris Lattner
824398ec6d
Remove #include
...
llvm-svn: 7342
2003-07-26 23:01:04 +00:00
Chris Lattner
981e585021
Remove extraneous #includes
...
llvm-svn: 7341
2003-07-26 23:00:29 +00:00
Vikram S. Adve
45766ab682
(1) Change the way unused regs. are marked and found to consider regType
...
info (since multiple reg types may share the same reg class).
(2) Remove machine-specific regalloc. methods that are no longer needed.
In particular, arguments and return value from a call do not need
machine-specific code for allocation.
(3) Rename TargetRegInfo::getRegType variants to avoid unintentional
overloading when an include file is omitted.
llvm-svn: 7329
2003-07-25 21:06:09 +00:00
Vikram S. Adve
abcd8d75ff
Remove unused field. Change the way unused regs. are marked and
...
found to consider regType info (since multiple reg types may share
the same reg class, e.g., single and double on Sparc).
llvm-svn: 7327
2003-07-25 21:00:13 +00:00
Chris Lattner
79f22fe02f
Remove unnecessary casts
...
llvm-svn: 7250
2003-07-23 15:17:51 +00:00
Chris Lattner
b976da9ecf
Random cleanups
...
llvm-svn: 7247
2003-07-23 14:59:40 +00:00
Chris Lattner
0470412961
Remove using decl
...
llvm-svn: 7246
2003-07-23 14:55:59 +00:00
Vikram S. Adve
e2e50f2880
Values stored in CallArgsDescriptor cannot be const.
...
llvm-svn: 7156
2003-07-10 19:46:15 +00:00
Vikram S. Adve
abf331db03
isMarkedForSpill() should be const.
...
llvm-svn: 7155
2003-07-10 19:45:28 +00:00
Vikram S. Adve
689adb19b0
Change interface to MachineInstr::substituteValue to specify more precisely
...
which args can be substituted: defsOnly, defsAndUses or usesOnly.
llvm-svn: 7154
2003-07-10 19:45:07 +00:00
Vikram S. Adve
39587d6264
Print out all neighbors in interference graph, not just higher-numbered ones.
...
llvm-svn: 7153
2003-07-10 19:43:33 +00:00
Vikram S. Adve
5224b19deb
Several fixes to handling of int CC register:
...
(1) An int CC live range must be spilled if there are any interferences,
even if no other "neighbour" in the interf. graph has been allocated
that reg. yet. This is actually true of any class with only one reg!
(2) SparcIntCCRegClass::colorIGNode sets the color even if the LR must
be spilled so that the machine-independent spill code doesn't have to
make the machine-dependent decision of which CC name to use based on
operand type: %xcc or %icc. (These are two halves of the same
register.)
(3) LR->isMarkedForSpill() is no longer the same as LR->hasColor().
These should never have been the same, and this is necessary now for #2 .
(4) All RDCCR and WRCCR instructions are directly generated with the
phony number for %ccr so that EmitAssembly/EmitBinary doesn't have to
deal with this.
llvm-svn: 7152
2003-07-10 19:42:55 +00:00
Vikram S. Adve
3bb921e8df
Choose register instead of immediate for ConstantExpr in ChooseRegOrImmed.
...
llvm-svn: 7112
2003-07-06 20:33:21 +00:00
Vikram S. Adve
1fce4cfa95
Minor beautification: fold a couple of lines of code.
...
llvm-svn: 7054
2003-07-02 01:24:00 +00:00
Vikram S. Adve
fa1dde06aa
A def. operand of a machine instruction may be an ordinary Value*,
...
not just an Instruction*, at least in one unfortunate case:
the first operand to the va_arg instruction.
Modify ValueToDefVecMap to map from Value*, not Instruction*.
llvm-svn: 7052
2003-07-02 01:16:01 +00:00
John Criswell
3ef61afb76
Merged in autoconf branch. This provides configuration via the autoconf
...
system.
llvm-svn: 7014
2003-06-30 21:59:07 +00:00
Chris Lattner
52de05c6b3
Remove a ton of extraneous #includes
...
llvm-svn: 6842
2003-06-22 03:08:05 +00:00
Chris Lattner
55491774bb
Actually, change it to use explicit new/delete, which is more likely to be
...
optimized INTO an alloca
llvm-svn: 6727
2003-06-16 22:29:09 +00:00
Chris Lattner
5b7e3cae42
Remove two using decls
...
Remove usage of alloca
llvm-svn: 6725
2003-06-16 22:18:28 +00:00
Brian Gaeke
7a719ed20c
Isolate machine-dependent use of <alloca.h> in "Support/Alloca.h",
...
so that we can easily change its use to be conditional on the result of
an autoconf test later.
llvm-svn: 6723
2003-06-16 21:54:01 +00:00
John Criswell
cfac736363
Included assert.h so that the code compiles under newer versions of GCC.
...
llvm-svn: 6682
2003-06-11 14:01:36 +00:00
Guochun Shi
c40f66c928
add some comments
...
add a function ModuloScheduling::dumpFinalSchedule() to print out final schedule
llvm-svn: 6677
2003-06-10 20:04:30 +00:00
Guochun Shi
f6b88e55ed
add an brief instruction what this pass is
...
llvm-svn: 6676
2003-06-10 20:03:39 +00:00
Guochun Shi
d2656536f5
a simple introduction to this pass
...
llvm-svn: 6675
2003-06-10 20:02:16 +00:00