[Bug rtl-optimization/54472] New: ICE (spill_failure): unable to find a register to spill in class 'AREG' with -O -fschedule-insns -fselective-scheduling

[Bug rtl-optimization/54472] New: ICE (spill_failure): unable to find a register to spill in class 'AREG' with -O -fschedule-insns -fselective-scheduling

[zsojka at seznam dot cz]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=54472

             Bug #: 54472
           Summary: ICE (spill_failure): unable to find a register to
                    spill in class 'AREG' with -O -fschedule-insns
                    -fselective-scheduling
    Classification: Unclassified
           Product: gcc
           Version: 4.8.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: rtl-optimization
        AssignedTo: unassigned@gcc.gnu.org
        ReportedBy: zsojka@seznam.cz


Created attachment 28125
  --> http://gcc.gnu.org/bugzilla/attachment.cgi?id=28125
reduced testcase

Compiler output:
$ gcc -O -fschedule-insns -fselective-scheduling testcase.c
testcase.c: In function 'main':
testcase.c:6:1: error: unable to find a register to spill in class 'AREG'
 }
 ^
testcase.c:6:1: error: this is the insn:
(insn 9 25 22 2 (parallel [
            (set (reg:DI 2 cx [63])
                (const_int 0 [0]))
            (set (reg/f:DI 5 di [61])
                (plus:DI (ashift:DI (reg:DI 2 cx [63])
                        (const_int 3 [0x3]))
                    (reg/f:DI 1 dx [60])))
            (set (mem/c:BLK (reg/f:DI 1 dx [60]) [0 MEM[(void *)&a]+0 S104
A64])
                (const_int 0 [0]))
            (use (reg:DI 4 si [62]))
            (use (reg:DI 2 cx [63]))
        ]) testcase.c:4 890 {*rep_stosdi_rex64}
     (expr_list:REG_DEAD (reg:DI 4 si [62])
        (expr_list:REG_DEAD (reg/f:DI 1 dx [60])
            (expr_list:REG_UNUSED (reg:DI 2 cx [63])
                (expr_list:REG_UNUSED (reg/f:DI 5 di [61])
                    (nil))))))
testcase.c:6:1: internal compiler error: in spill_failure, at reload1.c:2127
Please submit a full bug report,
with preprocessed source if appropriate.
See <http://gcc.gnu.org/bugs.html> for instructions.

(gdb) bt
#0  0x0000000001130b50 in fancy_abort(char const*, int, char const*) ()
#1  0x000000000096d5fb in _fatal_insn(char const*, rtx_def const*, char const*,
int, char const*) ()
#2  0x000000000096c6d0 in reload(rtx_def*, int) () at
/mnt/svn/gcc-trunk/gcc/reload1.c:2127
#3  0x000000000087eba2 in rest_of_handle_reload() () at
/mnt/svn/gcc-trunk/gcc/ira.c:4328
#4  0x000000000090092f in execute_one_pass(opt_pass*) () at
/mnt/svn/gcc-trunk/gcc/passes.c:2157
#5  0x0000000000900ce5 in execute_pass_list(opt_pass*) () at
/mnt/svn/gcc-trunk/gcc/passes.c:2212
#6  0x0000000000900cf7 in execute_pass_list(opt_pass*) () at
/mnt/svn/gcc-trunk/gcc/passes.c:2213
#7  0x00000000006bdd88 in expand_function(cgraph_node*) () at
/mnt/svn/gcc-trunk/gcc/cgraphunit.c:1609
#8  0x00000000006bfc2a in compile() () at
/mnt/svn/gcc-trunk/gcc/cgraphunit.c:1714
#9  0x00000000006c0225 in finalize_compilation_unit() () at
/mnt/svn/gcc-trunk/gcc/cgraphunit.c:2089
#10 0x00000000005a1dc0 in c_write_global_declarations() () at
/mnt/svn/gcc-trunk/gcc/c/c-decl.c:10116
#11 0x00000000009e8475 in compile_file() () at
/mnt/svn/gcc-trunk/gcc/toplev.c:560
#12 0x00000000009ea03a in toplev_main(int, char**) () at
/mnt/svn/gcc-trunk/gcc/toplev.c:1863
#13 0x00007ffff6e784bd in __libc_start_main () from /lib64/libc.so.6
#14 0x0000000000584d01 in _start ()

Tested revisions:
r190830 - crash
4.7 r188682 - crash
4.6 r188682 - crash
4.5 r188682 - crash
4.4 r188682 - crash

[Bug rtl-optimization/54472] ICE (spill_failure): unable to find a register to spill in class 'AREG' with -O -fschedule-insns -fselective-scheduling

[ubizjak at gmail dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=54472

Uros Bizjak <ubizjak at gmail dot com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |izamyatin at gmail dot com

--- Comment #1 from Uros Bizjak <ubizjak at gmail dot com> 2012-09-03 18:24:59 UTC ---
This is the problem with scheduler, where insns involving hard registers are
moved around. This part should be fixed by the WIP patch at [1].

[1] http://gcc.gnu.org/ml/gcc-patches/2012-08/msg00810.html

[Bug rtl-optimization/54472] ICE (spill_failure): unable to find a register to spill in class 'AREG' with -O -fschedule-insns -fselective-scheduling

[ubizjak at gmail dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=54472

--- Comment #2 from Uros Bizjak <ubizjak at gmail dot com> 2012-09-04 14:48:02 UTC ---
For some reason, -fselective-scheduling is moving

(insn 19 16 22 2 (use (reg/i:SI 0 ax)) testcase.c:6 -1
     (nil))

around. This insn marks function return, so IMO should be at the function exit
all the time. -fschedule-insns (without -fselective-scheduling) leaves the insn
at the function exit. Is it OK to move this marker around?

[Bug rtl-optimization/54472] ICE (spill_failure): unable to find a register to spill in class 'AREG' with -O -fschedule-insns -fselective-scheduling

[abel at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=54472

--- Comment #3 from Andrey Belevantsev <abel at gcc dot gnu.org> 2012-09-04 17:47:24 UTC ---
(In reply to comment #2)
> For some reason, -fselective-scheduling is moving
> 
> (insn 19 16 22 2 (use (reg/i:SI 0 ax)) testcase.c:6 -1
>      (nil))
> 
> around. This insn marks function return, so IMO should be at the function exit
> all the time. -fschedule-insns (without -fselective-scheduling) leaves the insn
> at the function exit. Is it OK to move this marker around?

I will take a look but I'm out of office until Friday at least (maybe Monday).

[Bug rtl-optimization/54472] ICE (spill_failure): unable to find a register to spill in class 'AREG' with -O -fschedule-insns -fselective-scheduling

[ubizjak at gmail dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=54472

--- Comment #4 from Uros Bizjak <ubizjak at gmail dot com> 2012-10-04 09:29:15 UTC ---
(In reply to comment #3)
> > For some reason, -fselective-scheduling is moving
> > 
> > (insn 19 16 22 2 (use (reg/i:SI 0 ax)) testcase.c:6 -1
> >      (nil))
> > 
> > around. This insn marks function return, so IMO should be at the function exit
> > all the time. -fschedule-insns (without -fselective-scheduling) leaves the insn
> > at the function exit. Is it OK to move this marker around?
> 
> I will take a look but I'm out of office until Friday at least (maybe Monday).

Any progress on this?

[Bug rtl-optimization/54472] ICE (spill_failure): unable to find a register to spill in class 'AREG' with -O -fschedule-insns -fselective-scheduling

[abel at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=54472

Andrey Belevantsev <abel at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |ASSIGNED
   Last reconfirmed|                            |2012-10-04
         AssignedTo|unassigned at gcc dot       |abel at gcc dot gnu.org
                   |gnu.org                     |
     Ever Confirmed|0                           |1

--- Comment #5 from Andrey Belevantsev <abel at gcc dot gnu.org> 2012-10-04 12:43:36 UTC ---
(In reply to comment #4)
> (In reply to comment #3)
> > > For some reason, -fselective-scheduling is moving
> > > 
> > > (insn 19 16 22 2 (use (reg/i:SI 0 ax)) testcase.c:6 -1
> > >      (nil))
> > > 
> > > around. This insn marks function return, so IMO should be at the function exit
> > > all the time. -fschedule-insns (without -fselective-scheduling) leaves the insn
> > > at the function exit. Is it OK to move this marker around?
> > 
> > I will take a look but I'm out of office until Friday at least (maybe Monday).
> 
> Any progress on this?

Sorry Uros, I forgot about this.  I took a look, the reason is actually in
selective scheduler not handling implicit_sets (we miss dependencies between
implicit_sets register sets and other sets/clobbers).  In this example, we miss
a dependency between insns 9 and 16 and thus schedule insn #16 earlier than
insn #9 and then its dependent insn #19 earlier than insn #9.  Thus we are
moving the return mark before the actual function code.

The below patch fixes the test case for me.  I need to double-check whether I
have caught all cases of implicit_sets generating dependencies in sched-deps.c
and then I will submit the patch for review.

diff --git a/gcc/sel-sched-ir.c b/gcc/sel-sched-ir.c
index e7ca3f1..0fbfdbe 100644
--- a/gcc/sel-sched-ir.c
+++ b/gcc/sel-sched-ir.c
@@ -3185,7 +3185,7 @@ has_dependence_note_reg_set (int regno)
          || reg_last->clobbers != NULL)
        *dsp = (*dsp & ~SPECULATIVE) | DEP_OUTPUT;

-      if (reg_last->uses)
+      if (reg_last->uses || reg_last->implicit_sets)
        *dsp = (*dsp & ~SPECULATIVE) | DEP_ANTI;
     }
 }
@@ -3205,7 +3205,7 @@ has_dependence_note_reg_clobber (int regno)
       if (reg_last->sets)
        *dsp = (*dsp & ~SPECULATIVE) | DEP_OUTPUT;

-      if (reg_last->uses)
+      if (reg_last->uses || reg_last->implicit_sets)
        *dsp = (*dsp & ~SPECULATIVE) | DEP_ANTI;
     }
 }

[Bug rtl-optimization/54472] ICE (spill_failure): unable to find a register to spill in class 'AREG' with -O -fschedule-insns -fselective-scheduling

[izamyatin at gmail dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=54472

--- Comment #6 from Igor Zamyatin <izamyatin at gmail dot com> 2012-10-24 11:09:49 UTC ---
Have you managed to check the patch?

[Bug rtl-optimization/54472] ICE (spill_failure): unable to find a register to spill in class 'AREG' with -O -fschedule-insns -fselective-scheduling

[abel at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=54472

--- Comment #7 from Andrey Belevantsev <abel at gcc dot gnu.org> 2012-10-26 14:04:03 UTC ---
(In reply to comment #6)
> Have you managed to check the patch?

The patch does the right thing for sets and clobbers, but the uses case should
be amended, too, judging from the sched-deps.c code.   Sorry, I was too busy to
send the patch, now, I'll try doing this on the next week.

[Bug rtl-optimization/54472] ICE (spill_failure): unable to find a register to spill in class 'AREG' with -O -fschedule-insns -fselective-scheduling

[abel at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=54472

--- Comment #8 from Andrey Belevantsev <abel at gcc dot gnu.org> 2012-11-09 12:28:30 UTC ---
Author: abel
Date: Fri Nov  9 12:28:21 2012
New Revision: 193358

URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=193358
Log:
    PR rtl-optimization/54472

    * sel-sched-ir.c (has_dependence_note_reg_set): Handle implicit sets.
    (has_dependence_note_reg_clobber,
    has_dependence_note_reg_use): Likewise.

    * gcc.dg/pr54472.c: New test.

Added:
    trunk/gcc/testsuite/gcc.dg/pr54472.c
Modified:
    trunk/gcc/ChangeLog
    trunk/gcc/sel-sched-ir.c
    trunk/gcc/testsuite/ChangeLog

[Bug rtl-optimization/54472] ICE (spill_failure): unable to find a register to spill in class 'AREG' with -O -fschedule-insns -fselective-scheduling

[abel at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=54472

--- Comment #9 from Andrey Belevantsev <abel at gcc dot gnu.org> 2012-11-09 12:33:33 UTC ---
Fixed on trunk, backports to 4.7/4.6 are needed.

[Bug rtl-optimization/54472] ICE (spill_failure): unable to find a register to spill in class 'AREG' with -O -fschedule-insns -fselective-scheduling

[mpolacek at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=54472

Marek Polacek <mpolacek at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |mpolacek at gcc dot gnu.org

--- Comment #10 from Marek Polacek <mpolacek at gcc dot gnu.org> 2013-04-03 14:53:36 UTC ---
So are we going to backport this one?

[Bug rtl-optimization/54472] ICE (spill_failure): unable to find a register to spill in class 'AREG' with -O -fschedule-insns -fselective-scheduling

[abel at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=54472

--- Comment #11 from Andrey Belevantsev <abel at gcc dot gnu.org> 2013-04-03 15:51:50 UTC ---
(In reply to comment #10)
> So are we going to backport this one?

Sorry, I've missed this one when backporting other stuff. I can do this
tomorrow, the patch is safe. However, the branch is frozen now, we need to ask
Richi for approval to squeeze it into the next 4.7 release.