[Bug middle-end/48441] [4.7 Regression] ICE in mark_oprs_set

[danglin at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=48441

--- Comment #2 from John David Anglin <danglin at gcc dot gnu.org> 2011-04-04 
22:37:38 UTC ---
(gdb) p debug_rtx (insn)
(jump_insn/v 247 246 338 29 (set (pc)
        (if_then_else (eq (reg/v:SI 116 [ rval ])
                (const_int 7 [0x7]))
            (label_ref:SI 338)
            (pc))) ../../gcc/libcpp/charset.c:481 25 {*pa.md:1330}
     (expr_list:REG_BR_PROB (const_int 9550 [0x254e])
        (nil))
 -> 338)
$1 = void
(gdb) p/x $pc
$2 = 0x343fb28
(gdb) disass 0x343fb18,0x343fb38
Dump of assembler code from 0x343fb18 to 0x343fb38:
   0x0343fb18 <mark_oprs_set+48>:    add,l r19,ret0,ret0
   0x0343fb1c <mark_oprs_set+52>:    ldw 0(ret0),ret0
   0x0343fb20 <mark_oprs_set+56>:    stw ret0,c(r3)
   0x0343fb24 <mark_oprs_set+60>:    ldw c(r3),ret0
=> 0x0343fb28 <mark_oprs_set+64>:    ldw 4(ret0),ret0
   0x0343fb2c <mark_oprs_set+68>:    stw ret0,8(r3)
   0x0343fb30 <mark_oprs_set+72>:    b,l,n 0x343fb68 <mark_oprs_set+128>,r0
   0x0343fb34 <mark_oprs_set+76>:    addil L%8f800,dp,r1
End of assembler dump.
(gdb) p/x $ret0
$3 = 0x0