Commit Graph

109 Commits

Author SHA1 Message Date
Evan Cheng
9193514e2e Re-materialize all loads from fixed stack slots.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@35660 91177308-0d34-0410-b5e6-96231b3b80d8
2007-04-04 07:40:01 +00:00
Evan Cheng
7cb33c8652 Don't add the same MI to register reuse "last def/use" twice if it reads the
register more than once.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@35513 91177308-0d34-0410-b5e6-96231b3b80d8
2007-03-30 20:21:35 +00:00
Evan Cheng
018d6e1537 Don't call getOperandConstraint() if operand index is greater than
TID->numOperands.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@35375 91177308-0d34-0410-b5e6-96231b3b80d8
2007-03-27 00:48:28 +00:00
Evan Cheng
ad7ccf34b5 Fix for PR1266. Don't mark a two address operand IsKill.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@35365 91177308-0d34-0410-b5e6-96231b3b80d8
2007-03-26 22:40:42 +00:00
Evan Cheng
2638e1a6b9 First cut trivial re-materialization support.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@35208 91177308-0d34-0410-b5e6-96231b3b80d8
2007-03-20 08:13:50 +00:00
Evan Cheng
a7288df62d Only propagate IsKill if the last use is a kill.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@34878 91177308-0d34-0410-b5e6-96231b3b80d8
2007-03-03 06:32:37 +00:00
Evan Cheng
6b448092bf - Keep track all def and uses of stack slot available in register.
- Available value use may be deleted (e.g. noop move).


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@34841 91177308-0d34-0410-b5e6-96231b3b80d8
2007-03-02 08:52:00 +00:00
Evan Cheng
c3dd314b78 Invalidate last use of a reused register if the use is a deleted noop copy.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@34839 91177308-0d34-0410-b5e6-96231b3b80d8
2007-03-02 05:41:42 +00:00
Evan Cheng
c0ba1bc49b A restore is promoted to copy (or deleted entirely), remove the kill from the last use of the targetted register.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@34773 91177308-0d34-0410-b5e6-96231b3b80d8
2007-03-01 02:27:30 +00:00
Evan Cheng
de4e942faa A couple of more places where a register liveness has been extended and its last kill should be updated accordingly.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@34597 91177308-0d34-0410-b5e6-96231b3b80d8
2007-02-25 09:51:27 +00:00
Evan Cheng
50d25d7ff7 Reuse extends the liveness of a register. Transfer the kill to the operand that reuse it.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@34536 91177308-0d34-0410-b5e6-96231b3b80d8
2007-02-23 21:47:50 +00:00
Evan Cheng
91e2390818 A spill kills the register being stored. But it is later being reused by spiller, its live range has to be extended.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@34517 91177308-0d34-0410-b5e6-96231b3b80d8
2007-02-23 01:13:26 +00:00
Evan Cheng
957840b3e1 Use BitVector instead. No functionality change.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@34460 91177308-0d34-0410-b5e6-96231b3b80d8
2007-02-21 02:22:03 +00:00
Evan Cheng
667089de04 Dead code.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@34435 91177308-0d34-0410-b5e6-96231b3b80d8
2007-02-20 01:29:10 +00:00
Evan Cheng
f50d09ad21 Fixed a long standing spiller bug that's exposed by Thumb:
The code sequence before the spiller is something like:
                 = tMOVrr
        %reg1117 = tMOVrr
        %reg1078 = tLSLri %reg1117, 2

The it starts spilling:
        %r0 = tRestore <fi#5>, 0
        %r1 = tRestore <fi#7>, 0
        %r1 = tMOVrr %r1<kill>
        tSpill %r1, <fi#5>, 0
        %reg1078 = tLSLri %reg1117, 2

It restores the value while processing the first tMOVrr. At this point, the
spiller remembers fi#5 is available in %r0. Next it processes the second move.
It restores the source before the move and spills the result afterwards. The
move becomes a noop and is deleted. However, a spill has been inserted and that
should invalidate reuse of %r0 for fi#5 and add reuse of %r1 for fi#5.
Therefore, %reg1117 (which is also assigned fi#5) should get %r1, not %r0.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@34039 91177308-0d34-0410-b5e6-96231b3b80d8
2007-02-08 06:04:54 +00:00
Chris Lattner
08a4d5a343 Switch this to use SmallSet to avoid mallocs in the common case.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@33457 91177308-0d34-0410-b5e6-96231b3b80d8
2007-01-23 00:59:48 +00:00
Evan Cheng
3c82cab9aa GetRegForReload() now keeps track which registers have been considered and rejected during its quest to find a suitable reload register. This avoids an infinite loop in case like this:
t1 := op t2, t3
  t2 <- assigned r0 for use by the reload but ended up reuse r1
  t3 <- assigned r1 for use by the reload but ended up reuse r0
  t1 <- desires r1
        sees r1 is taken by t2, tries t2's reload register r0
        sees r0 is taken by t3, tries t3's reload register r1
        sees r1 is taken by t2, tries t2's reload register r0 ...


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@33382 91177308-0d34-0410-b5e6-96231b3b80d8
2007-01-19 22:40:14 +00:00
Chris Lattner
cd3245ac45 Eliminate static ctors from Statistics
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@32698 91177308-0d34-0410-b5e6-96231b3b80d8
2006-12-19 22:41:21 +00:00
Bill Wendling
5c7e326585 Added an automatic cast to "std::ostream*" etc. from OStream. We then can
rework the hacks that had us passing OStream in. We pass in std::ostream*
instead, check for null, and then dispatch to the correct print() method.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@32636 91177308-0d34-0410-b5e6-96231b3b80d8
2006-12-17 05:15:13 +00:00
Evan Cheng
86facc2828 Minor clean up.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@32593 91177308-0d34-0410-b5e6-96231b3b80d8
2006-12-15 06:41:01 +00:00
Evan Cheng
7a0d51c8e3 Fix a long-standing spiller bug:
If a spillslot value is available in a register, and there is a noop copy that
targets that register, the spiller correctly decide not to invalidate the
spillslot register.

However, even though the noop copy does not clobbers the value. It does start a
new intersecting live range. That means the spillslot register is available for
use but should not be reused for a two-address instruction modref operand which
would clobber the new live range.

When we remove the noop copy, update the available information by clearing the
canClobber bit.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@32576 91177308-0d34-0410-b5e6-96231b3b80d8
2006-12-14 07:54:05 +00:00
Evan Cheng
cc22a7a2ad Move findTiedToSrcOperand to TargetInstrDescriptor.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@32366 91177308-0d34-0410-b5e6-96231b3b80d8
2006-12-08 18:45:48 +00:00
Evan Cheng
5c2a46052b Proper fix for PR1037: to determine is a VR is a modref, check 1) whether it is
tied to another oeprand, 2) whether is is being tied to by another operand. So
the destination operand of a two-address MI can be correctly identified.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@32354 91177308-0d34-0410-b5e6-96231b3b80d8
2006-12-08 08:02:34 +00:00
Reid Spencer
fe46361719 Revision 1.83 causes PR1037.
Reverted.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@32305 91177308-0d34-0410-b5e6-96231b3b80d8
2006-12-07 16:21:19 +00:00
Bill Wendling
e81561909d Changed llvm_ostream et all to OStream. llvm_cerr, llvm_cout, llvm_null, are
now cerr, cout, and NullStream resp.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@32298 91177308-0d34-0410-b5e6-96231b3b80d8
2006-12-07 01:30:32 +00:00
Evan Cheng
51cdcd1972 MI keeps a ptr of TargetInstrDescriptor, use it.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@32296 91177308-0d34-0410-b5e6-96231b3b80d8
2006-12-07 01:21:59 +00:00
Evan Cheng
6065f66e7c getOperandConstraint returns -1 if the operand does have the specific constraint. This bug was causing excessive spills.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@32295 91177308-0d34-0410-b5e6-96231b3b80d8
2006-12-07 00:46:04 +00:00
Chris Lattner
ac0b6ae358 Detemplatize the Statistic class. The only type it is instantiated with
is 'unsigned'.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@32279 91177308-0d34-0410-b5e6-96231b3b80d8
2006-12-06 17:46:33 +00:00
Evan Cheng
ba59a1e453 Match TargetInstrInfo changes.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@32098 91177308-0d34-0410-b5e6-96231b3b80d8
2006-12-01 21:52:58 +00:00
Bill Wendling
b2b9c20b61 More removal of std::cerr and DEBUG, replacing with DOUT instead.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@31806 91177308-0d34-0410-b5e6-96231b3b80d8
2006-11-17 02:09:07 +00:00
Evan Cheng
e077ef6e85 Fixed some spiller bugs exposed by the recent two-address code changes. Now
there may be other def(s) apart from the use&def two-address operand. We need
to check if the register reuse for a use&def operand may conflicts with another
def. Provide a mean to recover from the conflict if it is detected when the
defs are processed later.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@31439 91177308-0d34-0410-b5e6-96231b3b80d8
2006-11-04 00:21:55 +00:00
Evan Cheng
e6ae14e1f4 Rename
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@31364 91177308-0d34-0410-b5e6-96231b3b80d8
2006-11-01 23:18:32 +00:00
Evan Cheng
360c2dd25a Two-address instructions no longer have to be A := A op C. Now any pair of dest / src operands can be tied together.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@31363 91177308-0d34-0410-b5e6-96231b3b80d8
2006-11-01 23:06:55 +00:00
Chris Lattner
6ec3626be4 restore my previous patch, now that the X86 backend bug has been fixed:
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20061009/038518.html


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@30906 91177308-0d34-0410-b5e6-96231b3b80d8
2006-10-12 17:45:38 +00:00
Evan Cheng
b870100f2a Backing out Chris' last commit. It's breaking llvm-gcc bootstrapping.
It's turning:
        movl -24(%ebp), %esp
        subl $16, %esp
        movl -24(%ebp), %ecx
into
        movl -24(%ebp), %esp
        subl $16, %esp
        movl %esp, (%esp)


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@30902 91177308-0d34-0410-b5e6-96231b3b80d8
2006-10-12 08:00:47 +00:00
Chris Lattner
f183cabba8 If we see a load from a stack slot into a physreg, consider it as providing
the stack slot.  This fixes PR943.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@30898 91177308-0d34-0410-b5e6-96231b3b80d8
2006-10-12 02:34:07 +00:00
Chris Lattner
2926869b4a Fix a long-standing wart in the code generator: two-address instruction lowering
actually *removes* one of the operands, instead of just assigning both operands
the same register.  This make reasoning about instructions unnecessarily complex,
because you need to know if you are before or after register allocation to match
up operand #'s with the target description file.

Changing this also gets rid of a bunch of hacky code in various places.

This patch also includes changes to fold loads into cmp/test instructions in
the X86 backend, along with a significant simplification to the X86 spill
folding code.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@30108 91177308-0d34-0410-b5e6-96231b3b80d8
2006-09-05 02:12:02 +00:00
Chris Lattner
a4f0b3a084 s|llvm/Support/Visibility.h|llvm/Support/Compiler.h|
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@29911 91177308-0d34-0410-b5e6-96231b3b80d8
2006-08-27 12:54:02 +00:00
Chris Lattner
f7da2c7b0c Take advantage of the recent improvements to the liveintervals set (tracking
instructions which define each value#) to simplify and improve the coallescer.
In particular, this patch:

1. Implements iterative coallescing.
2. Reverts an unsafe hack from handlePhysRegDef, superceeding it with a
   better solution.
3. Implements PR865, "coallescing" away the second copy in code like:

   A = B
   ...
   B = A

This also includes changes to symbolically print registers in intervals
when possible.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@29862 91177308-0d34-0410-b5e6-96231b3b80d8
2006-08-24 22:43:55 +00:00
Bill Wendling
d36d03bf2d Added a check so that if we have two machine instructions in this form
MOV R0, R1
    MOV R1, R0

the second machine instruction is removed. Added a regression test.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@29792 91177308-0d34-0410-b5e6-96231b3b80d8
2006-08-21 07:33:33 +00:00
Jim Laskey
cd4317efcf Eliminate data relocations by using NULL instead of global empty list.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@29250 91177308-0d34-0410-b5e6-96231b3b80d8
2006-07-21 21:15:20 +00:00
Andrew Lenharth
ed41f1bb19 Reduce number of exported symbols
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@29220 91177308-0d34-0410-b5e6-96231b3b80d8
2006-07-20 17:28:38 +00:00
Chris Lattner
f8c68f694c Shave another 27K off libllvmgcc.dylib with visibility hidden
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@28973 91177308-0d34-0410-b5e6-96231b3b80d8
2006-06-28 22:17:39 +00:00
Chris Lattner
e53f4a055f Move some methods out of MachineInstr into MachineOperand
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@28102 91177308-0d34-0410-b5e6-96231b3b80d8
2006-05-04 17:52:23 +00:00
Chris Lattner
229924a79f Fix a latent bug that my spiller patch last week exposed: we were leaving
instructions in the virtregfolded map that were deleted.  Because they
were deleted, newly allocated instructions could end up at the same address,
magically finding themselves in the map.  The solution is to remove entries
from the map when we delete the instructions.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@28041 91177308-0d34-0410-b5e6-96231b3b80d8
2006-05-01 22:03:24 +00:00
Chris Lattner
8a18c13fa5 When promoting a load to a reg-reg copy, where the load was a previous
instruction folded with spill code, make sure the remove the load from
the virt reg folded map.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@28040 91177308-0d34-0410-b5e6-96231b3b80d8
2006-05-01 21:17:10 +00:00
Chris Lattner
35f2705e3d Remove previous patch, which wasn't quite right.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@28039 91177308-0d34-0410-b5e6-96231b3b80d8
2006-05-01 21:16:03 +00:00
Evan Cheng
10dbd3ead8 Remove temp. option -spiller-check-liveout, it didn't cause any failure nor performance regressions.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@28029 91177308-0d34-0410-b5e6-96231b3b80d8
2006-05-01 08:54:57 +00:00
Evan Cheng
200370fb56 Local spiller kills a store if the folded restore is turned into a copy.
But this is incorrect if the spilled value live range extends beyond the
current BB.
It is currently controlled by a temporary option -spiller-check-liveout.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@28024 91177308-0d34-0410-b5e6-96231b3b80d8
2006-04-30 08:41:47 +00:00
Chris Lattner
ba1fc3daf7 Mapping of physregs can make it so that the designated and input physregs are
the same.  In this case, don't emit a noop copy.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@28008 91177308-0d34-0410-b5e6-96231b3b80d8
2006-04-28 04:43:18 +00:00