[Bug middle-end/52997] New: [4.8 Regression] FAIL: gcc.dg/c99-intconst-1.c (internal compiler error)
danglin at gcc dot gnu.org
gcc-bugzilla@gcc.gnu.org
Sun Apr 15 17:37:00 GMT 2012
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=52997
Bug #: 52997
Summary: [4.8 Regression] FAIL: gcc.dg/c99-intconst-1.c
(internal compiler error)
Classification: Unclassified
Product: gcc
Version: 4.8.0
Status: UNCONFIRMED
Severity: normal
Priority: P3
Component: middle-end
AssignedTo: unassigned@gcc.gnu.org
ReportedBy: danglin@gcc.gnu.org
CC: bernds@gcc.gnu.org
Host: hppa-unknown-linux-gnu
Target: hppa-unknown-linux-gnu
Build: hppa-unknown-linux-gnu
Compiler generates a segmentation fault in reload1.c:
Program received signal SIGSEGV, Segmentation fault.
0x004ac200 in calculate_elim_costs_all_insns () at ../../gcc/gcc/reload1.c:1636
1636 && reg_renumber[REGNO (SET_DEST (set))] < 0
(gdb) bt
#0 0x004ac200 in calculate_elim_costs_all_insns ()
at ../../gcc/gcc/reload1.c:1636
#1 0x003eaca0 in ira_costs () at ../../gcc/gcc/ira-costs.c:2061
#2 0x003e5604 in ira_build () at ../../gcc/gcc/ira-build.c:3076
#3 0x003df6a8 in ira (f=<optimized out>) at ../../gcc/gcc/ira.c:4142
#4 rest_of_handle_ira () at ../../gcc/gcc/ira.c:4315
#5 0x00459764 in execute_one_pass (pass=0x15cccd8)
at ../../gcc/gcc/passes.c:2079
#6 0x004599e8 in execute_pass_list (pass=0x0) at ../../gcc/gcc/passes.c:2134
#7 0x004599fc in execute_pass_list (pass=0xa25410)
at ../../gcc/gcc/passes.c:2135
#8 0x001ca47c in tree_rest_of_compilation (node=<optimized out>)
at ../../gcc/gcc/cgraphunit.c:1806
#9 0x001ca888 in cgraph_expand_function (node=0x4076b500)
at ../../gcc/gcc/cgraphunit.c:1876
#10 0x001ccd28 in cgraph_output_in_order () at ../../gcc/gcc/cgraphunit.c:2041
#11 cgraph_optimize () at ../../gcc/gcc/cgraphunit.c:2629
#12 0x001cd08c in cgraph_finalize_compilation_unit ()
at ../../gcc/gcc/cgraphunit.c:2720
#13 0x001cd08c in cgraph_finalize_compilation_unit ()
at ../../gcc/gcc/cgraphunit.c:2720
(gdb) disass $pc-16,$pc+16
Dump of assembler code from 0x4ac1f0 to 0x4ac210:
0x004ac1f0 <calculate_elim_costs_all_insns+360>: cmpb,<>,n
r20,r21,0x4ac388 <calculate_elim_costs_all_insns+768>
0x004ac1f4 <calculate_elim_costs_all_insns+364>: ldw 4(r19),r4
0x004ac1f8 <calculate_elim_costs_all_insns+368>: addil L%2a800,dp,r1
0x004ac1fc <calculate_elim_costs_all_insns+372>: ldw 308(r1),r19
=> 0x004ac200 <calculate_elim_costs_all_insns+376>: ldh,s r4(r19),r19
0x004ac204 <calculate_elim_costs_all_insns+380>: extrw,s r19,31,16,r19
0x004ac208 <calculate_elim_costs_all_insns+384>: cmpib,<= 0,r19,0x4ac388
<calculate_elim_costs_all_insns+768>
0x004ac20c <calculate_elim_costs_all_insns+388>: addil L%7c000,dp,r1
End of assembler dump.
(gdb) p/x $r19
$1 = 0x4061a008
(gdb) p/x $r4
$2 = 0x20ffc
(gdb) p/x reg_renumber
$4 = 0x4061a008
(gdb) p debug_rtx (set)
(set (reg:SI 135164 [3982])
(const_int 16384 [0x4000]))
$5 = void
(gdb) p/x 135164
$6 = 0x20ffc
(gdb) p/x ®_renumber[135164]
$7 = 0x4065c000
(gdb) p reg_renumber[135164]
Cannot access memory at address 0x4065c000
Apr 15 13:21:44 mx3210 kernel: do_page_fault() pid=14445 command='cc1' type=15
a
ddress=0x4065c000Apr 15 13:21:44 mx3210 kernel: vm_start = 0x4051a000, vm_end =
0x4065c000
The register number causes an access beyond the region allocated for
reg_renumber.
The bug was introduced in r186378.
More information about the Gcc-bugs
mailing list