[Bug target/56455] New: [4.8 Regression on AVR] ICE in final_scan_insn

[Bug target/56455] New: [4.8 Regression on AVR] ICE in final_scan_insn

[vishnu.k_s at atmel dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=56455

             Bug #: 56455
           Summary: [4.8 Regression on AVR] ICE in final_scan_insn
    Classification: Unclassified
           Product: gcc
           Version: 4.8.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: target
        AssignedTo: unassigned@gcc.gnu.org
        ReportedBy: vishnu.k_s@atmel.com


AVR target built with specs:

COLLECT_GCC=/home/vishnu.k_s/builds/build-linux_x86_64/install/bin/avr-gcc
COLLECT_LTO_WRAPPER=/home/vishnu.k_s/builds/build-linux_x86_64/install/bin/../libexec/gcc/avr/4.8.0/lto-wrapper
Target: avr
Configured with: ../gcc-trunk/configure
--prefix=/home/vishnu.k_s/builds/build-linux_x86_64//install --target=avr
--enable-languages=c,c++ --disable-nls --with-gmp=/proj/install/gmp-4.3.2/
--with-mpfr=/proj/install/mpfr-2.4.2/ --with-

mpc=/proj/install/mpc-0.8.2/    
Thread model: single
gcc version 4.8.0 20130116 (experimental) (GCC)

Above compiler gives an ICE in final_scan_insn, at final.c for testcases 
present in gcc.dg/tree-ssa/pr36881.c, gcc.dg/pr45259.c and
gcc.c-torture/compile/981006-1.c.

Output for pr36881.c
---------------------

/home/vishnu.k_s/builds/build-linux_x86_64/gcc-trunk/gcc/testsuite/gcc.dg/tree-ssa/pr36881.c:
In function 'foo':
/home/vishnu.k_s/builds/build-linux_x86_64/gcc-trunk/gcc/testsuite/gcc.dg/tree-ssa/pr36881.c:22:1:
internal compiler error: in final_scan_insn, at final.c:2460

Output for pr45259.c 
--------------------

/home/vishnu.k_s/builds/build-linux_x86_64/gcc-trunk/gcc/testsuite/gcc.dg/pr45259.c:
In function 'foo':
/home/vishnu.k_s/builds/build-linux_x86_64/gcc-trunk/gcc/testsuite/gcc.dg/pr45259.c:43:1:
internal compiler error: in final_scan_insn, at final.c:2460

Output for 981006-1.c
---------------------

/home/vishnu.k_s/builds/build-linux_x86_64/gcc-trunk/gcc/testsuite/gcc.c-torture/compile/981006-1.c:
In function 'foo':
/home/vishnu.k_s/builds/build-linux_x86_64/gcc-trunk/gcc/testsuite/gcc.c-torture/compile/981006-1.c:49:1:
internal compiler error: in final_scan_insn, at final.c:2460

[Bug target/56455] [4.8 Regression] ICE in final_scan_insn

[rguenth at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=56455

Richard Biener <rguenth at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Target Milestone|---                         |4.8.0
            Summary|[4.8 Regression on AVR] ICE |[4.8 Regression] ICE in
                   |in final_scan_insn          |final_scan_insn

[Bug target/56455] [4.8 Regression] ICE in final_scan_insn

[jakub at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=56455

Jakub Jelinek <jakub at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |NEW
   Last reconfirmed|                            |2013-02-26
                 CC|                            |jakub at gcc dot gnu.org,
                   |                            |steven at gcc dot gnu.org
     Ever Confirmed|0                           |1

--- Comment #1 from Jakub Jelinek <jakub at gcc dot gnu.org> 2013-02-26 12:03:15 UTC ---
Seems to be broken by http://gcc.gnu.org/ml/gcc-patches/2012-05/msg00163.html
The assumption that only alpha-vms doesn't define ASM_OUTPUT_ADDR_DIFF_ELT
doesn't seem to be true.

echo `for i in config/*/; do grep -q ASM_OUTPUT_ADDR_DIFF_ELT $i/*  || echo $i;
done`
config/avr/ config/m32c/ config/mep/ config/moxie/ config/picochip/
config/stormy16/ config/vms/

2013-02-26  Jakub Jelinek  <jakub@redhat.com>

    PR target/56455
    * stmt.c (expand_switch_as_decision_tree_p): If flag_pic
    and ASM_OUTPUT_ADDR_DIFF_ELT isn't defined, return true.

--- gcc/stmt.c.jj    2013-01-11 09:03:17.000000000 +0100
+++ gcc/stmt.c    2013-02-26 12:43:16.362812625 +0100
@@ -1760,6 +1760,10 @@ expand_switch_as_decision_tree_p (tree r
     return true;
   if (!flag_jump_tables)
     return true;
+#ifndef ASM_OUTPUT_ADDR_DIFF_ELT
+  if (flag_pic)
+    return true;
+#endif

   /* If the switch is relatively small such that the cost of one
      indirect jump on the target are higher than the cost of a

fixes this, or do you want Steven to add that if (flag_pic) flag_jump_tables =
false; to all those targets?

[Bug target/56455] [4.8 Regression] ICE in final_scan_insn

[jakub at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=56455

--- Comment #2 from Jakub Jelinek <jakub at gcc dot gnu.org> 2013-02-27 07:27:02 UTC ---
Author: jakub
Date: Wed Feb 27 07:26:53 2013
New Revision: 196304

URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=196304
Log:
    PR target/56455
    * stmt.c (expand_switch_as_decision_tree_p): If flag_pic
    and ASM_OUTPUT_ADDR_DIFF_ELT isn't defined, return true.

Modified:
    trunk/gcc/ChangeLog
    trunk/gcc/stmt.c

[Bug target/56455] [4.8 Regression] ICE in final_scan_insn

[jakub at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=56455

Jakub Jelinek <jakub at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED

--- Comment #3 from Jakub Jelinek <jakub at gcc dot gnu.org> 2013-02-27 07:29:44 UTC ---
Fixed.