[Bug rtl-optimization/42295] New: ICE: 'error: unable to find a register to spill in class "AREG"' with -fschedule-insns

[Bug rtl-optimization/42295] New: ICE: 'error: unable to find a register to spill in class "AREG"' with -fschedule-insns

[zsojka at seznam dot cz]

Command line:
for 4.5/trunk:
g++-4.5 -O1 -fschedule-insns -fselective-scheduling -o tmp.o testcase-manual.ii
-c
for 4.4 can be used as well:
g++-4.4 -O1 -fschedule-insns -o tmp.o testcase-manual.ii -c

Gives (trunk r154953):
testcase-manual.ii: In function ‘void testsum(ms*)’:
testcase-manual.ii:20:1: error: unable to find a register to spill in class
‘AREG’
testcase-manual.ii:20:1: error: this is the insn:
(insn 26 24 27 2 testcase-manual.ii:16 (parallel [
            (set (reg:SI 5 di [83])
                (div:SI (reg:SI 5 di [orig:82 a_2(D)->n ] [82])
                    (reg:SI 2 cx [85])))
            (set (reg:SI 1 dx [84])
                (mod:SI (reg:SI 5 di [orig:82 a_2(D)->n ] [82])
                    (reg:SI 2 cx [85])))
            (clobber (reg:CC 17 flags))
        ]) 352 {*divmodsi4} (expr_list:REG_DEAD (reg:SI 2 cx [85])
        (expr_list:REG_DEAD (reg:SI 5 di [orig:82 a_2(D)->n ] [82])
            (expr_list:REG_UNUSED (reg:SI 1 dx [84])
                (expr_list:REG_UNUSED (reg:CC 17 flags)
                    (nil))))))
testcase-manual.ii:20:1: internal compiler error: in spill_failure, at
reload1.c:2141
Please submit a full bug report,
with preprocessed source if appropriate.
See <http://gcc.gnu.org/bugs.html> for instructions.

Tested versions:
trunk r154953 crash
trunk r154886 crash
trunk r154830 crash
trunk r153685 crash
4.4 r154975 crash
4.4 r154724 crash
4.4 r154216 crash
4.4.2 (gentoo) crash
4.3.4 4.2.4 4.1.2 3.4.6 3.3.6 (all gentoo) OK


Full output:
$ /mnt/svn/gcc-trunk/binary-154953-lto/bin/g++ -O1 -fschedule-insns
-fselective-scheduling -o tmp.o testcase-manual.ii -c -v
Using built-in specs.
COLLECT_GCC=/mnt/svn/gcc-trunk/binary-154953-lto/bin/g++
COLLECT_LTO_WRAPPER=/mnt/svn/gcc-trunk/binary-154953-lto/libexec/gcc/x86_64-unknown-linux-gnu/4.5.0/lto-wrapper
Target: x86_64-unknown-linux-gnu
Configured with: ../configure --enable-languages=c,c++,lto
--prefix=/mnt/svn/gcc-trunk/binary-154953-lto
Thread model: posix
gcc version 4.5.0 20091203 (experimental) (GCC)
COLLECT_GCC_OPTIONS='-O1' '-fschedule-insns' '-fselective-scheduling' '-o'
'tmp.o' '-c' '-v' '-shared-libgcc' '-mtune=generic'

/mnt/svn/gcc-trunk/binary-154953-lto/libexec/gcc/x86_64-unknown-linux-gnu/4.5.0/cc1plus
-fpreprocessed testcase-manual.ii -quiet -dumpbase testcase-manual.ii
-mtune=generic -auxbase-strip tmp.o -O1 -version -fschedule-insns
-fselective-scheduling -o /tmp/ccpvVv5N.s
GNU C++ (GCC) version 4.5.0 20091203 (experimental) (x86_64-unknown-linux-gnu)
        compiled by GNU C version 4.5.0 20091203 (experimental), GMP version
4.3.1, MPFR version 2.4.1-p5, MPC version 0.8
GGC heuristics: --param ggc-min-expand=30 --param ggc-min-heapsize=4096
GNU C++ (GCC) version 4.5.0 20091203 (experimental) (x86_64-unknown-linux-gnu)
        compiled by GNU C version 4.5.0 20091203 (experimental), GMP version
4.3.1, MPFR version 2.4.1-p5, MPC version 0.8
GGC heuristics: --param ggc-min-expand=30 --param ggc-min-heapsize=4096
Compiler executable checksum: 4b09d9924550b4f9676aafda3b16134c
testcase-manual.ii: In function &#8216;void testsum(ms*)&#8217;:
testcase-manual.ii:20:1: error: unable to find a register to spill in class
&#8216;AREG&#8217;
testcase-manual.ii:20:1: error: this is the insn:
(insn 26 24 27 2 testcase-manual.ii:16 (parallel [
            (set (reg:SI 5 di [83])
                (div:SI (reg:SI 5 di [orig:82 a_2(D)->n ] [82])
                    (reg:SI 2 cx [85])))
            (set (reg:SI 1 dx [84])
                (mod:SI (reg:SI 5 di [orig:82 a_2(D)->n ] [82])
                    (reg:SI 2 cx [85])))
            (clobber (reg:CC 17 flags))
        ]) 352 {*divmodsi4} (expr_list:REG_DEAD (reg:SI 2 cx [85])
        (expr_list:REG_DEAD (reg:SI 5 di [orig:82 a_2(D)->n ] [82])
            (expr_list:REG_UNUSED (reg:SI 1 dx [84])
                (expr_list:REG_UNUSED (reg:CC 17 flags)
                    (nil))))))
testcase-manual.ii:20:1: internal compiler error: in spill_failure, at
reload1.c:2141
Please submit a full bug report,
with preprocessed source if appropriate.
See <http://gcc.gnu.org/bugs.html> for instructions.


-- 
           Summary: ICE: 'error: unable to find a register to spill in class
                    "AREG"' with -fschedule-insns
           Product: gcc
           Version: 4.4.3
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: rtl-optimization
        AssignedTo: unassigned at gcc dot gnu dot org
        ReportedBy: zsojka at seznam dot cz
  GCC host triplet: x86_64-pc-linux-gnu


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=42295

[Bug rtl-optimization/42295] ICE: 'error: unable to find a register to spill in class "AREG"' with -fschedule-insns

[zsojka at seznam dot cz]

------- Comment #1 from zsojka at seznam dot cz  2009-12-05 01:55 -------
Created an attachment (id=19234)
 --> (http://gcc.gnu.org/bugzilla/attachment.cgi?id=19234&action=view)
somehow reduced testcase


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=42295

[Bug target/42295] ICE: 'error: unable to find a register to spill in class "AREG"' with -fschedule-insns

[rguenth at gcc dot gnu dot org]

------- Comment #2 from rguenth at gcc dot gnu dot org  2009-12-07 11:17 -------
-fschedule-insns is quite tricky for x86 targets, if it doesn't work do not use
it.


-- 

rguenth at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
          Component|rtl-optimization            |target
 GCC target triplet|                            |x86_64-*-*


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=42295

[Bug target/42295] ICE: 'error: unable to find a register to spill in class "AREG"' with -fschedule-insns

[zsojka at seznam dot cz]

------- Comment #3 from zsojka at seznam dot cz  2009-12-07 15:33 -------
This testcase (the call to f() isn't needed in this case) crashes with 4.4 with
these flags as well:
-O3 -fselective-scheduling2 -fsel-sched-pipelining
-fsel-sched-pipelining-outer-loops
(-O2 is enough for trunk)

The crash message is:
testcase-manual.ii: In function ‘void testsum(ms*)’:
testcase-manual.ii:20: internal compiler error: in init_seqno, at
sel-sched.c:6433 (sel-sched.c:6722 for trunk)
Please submit a full bug report,
with preprocessed source if appropriate.
See <http://gcc.gnu.org/bugs.html> for instructions.

Is this the same issue, or should I open separate bug report? (eg.
-fselective-scheduling2 is not expected to always work on x86_64 as well)


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=42295

[Bug target/42295] ICE: 'error: unable to find a register to spill in class "AREG"' with -fschedule-insns

[zsojka at seznam dot cz]

------- Comment #4 from zsojka at seznam dot cz  2009-12-16 22:01 -------
Created an attachment (id=19330)
 --> (http://gcc.gnu.org/bugzilla/attachment.cgi?id=19330&action=view)
simplified testcase

(testcase from comment#3 turned out to be a dup of pr39453)


-- 

zsojka at seznam dot cz changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
  Attachment #19234|0                           |1
        is obsolete|                            |


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=42295

[Bug target/42295] ICE: 'error: unable to find a register to spill in class "AREG"' with -fschedule-insns

[abel at gcc dot gnu dot org]

------- Comment #5 from abel at gcc dot gnu dot org  2010-01-14 11:32 -------
With the recent patches to sel-sched, testcases from comments #3 and #4 do not
longer fail for me.  I'm still seeing the spill failure of the original report
though.


-- 

abel at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |abel at gcc dot gnu dot org


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=42295

[Bug target/42295] ICE: 'error: unable to find a register to spill in class "AREG"' with -fschedule-insns

[ubizjak at gmail dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=42295

Uros Bizjak <ubizjak at gmail dot com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |RESOLVED
      Known to work|                            |4.6.3, 4.7.1
         Resolution|                            |FIXED
   Target Milestone|---                         |4.8.0

--- Comment #9 from Uros Bizjak <ubizjak at gmail dot com> 2012-09-02 07:13:53 UTC ---
Closed as fixed in 4.8.0.

[Bug target/42295] ICE: 'error: unable to find a register to spill in class "AREG"' with -fschedule-insns

[uros at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=42295

--- Comment #8 from uros at gcc dot gnu.org 2012-09-02 07:07:50 UTC ---
Author: uros
Date: Sun Sep  2 07:07:45 2012
New Revision: 190854

URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=190854
Log:
    PR target/42295
    * g++.dg/opt/pr42295.C: New test.


Added:
    trunk/gcc/testsuite/g++.dg/opt/pr42295.C
Modified:
    trunk/gcc/testsuite/ChangeLog

[Bug target/42295] ICE: 'error: unable to find a register to spill in class "AREG"' with -fschedule-insns

[dominiq at lps dot ens.fr]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=42295

--- Comment #7 from Dominique d'Humieres <dominiq at lps dot ens.fr> 2012-09-01 22:19:09 UTC ---
> All testcases compile OK for me with recent mainline.

For me too with 4.6.3, 4.7.1, and trunk. I get the ICE with 4.4.6 and 4.5.3.

[Bug target/42295] ICE: 'error: unable to find a register to spill in class "AREG"' with -fschedule-insns

[ubizjak at gmail dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=42295

--- Comment #6 from Uros Bizjak <ubizjak at gmail dot com> 2012-09-01 21:25:35 UTC ---
All testcases compile OK for me with recent mainline.