gcc-4.0: a couple oopses
Hi...
I have just built 2.6.12-rc2-mm3 with gcc4 final. I got this couple oopses on boot. This is almost plain rc3-mm3, the only patches applied are the syscall traps changes by Stas Sergeev and the segment loading patch from Ingo.
Apart from this, things _seem_ to work...
Oopses follow:
ksymoops 2.4.9 on i686 2.6.11-jam14. Options used
-V (default)
-k /proc/ksyms (default)
-l /proc/modules (default)
-o /lib/modules/2.6.11-jam14/ (default) -m /boot/System.map-2.6.11-jam14 (default)
Warning: You did not tell me where to find symbol information. I will assume that the log matches the kernel and modules that are running right now and I'll use the default options above for symbol resolution. If the current kernel and/or modules do not match the log, you can get more accurate output by telling me the kernel version and where to find map, modules, ksyms etc. ksymoops -h explains the options.
Error (regular_file): read_ksyms stat /proc/ksyms failed No modules in ksyms, skipping objects
No ksyms, skipping lsmod
Apr 27 15:52:44 werewolf kernel: Unable to handle kernel paging request at virtual address 10a316b4 Apr 27 15:52:44 werewolf kernel: b011f5d5 Apr 27 15:52:44 werewolf kernel: *pde = 00000000 Apr 27 15:52:44 werewolf kernel: Oops: 0002 [#1] Apr 27 15:52:44 werewolf kernel: CPU: 0 Apr 27 15:52:44 werewolf kernel: EIP: 0060:[do_proc_dointvec_conv+15/58] Not tainted VLI Apr 27 15:52:44 werewolf kernel: EIP: 0060:[<b011f5d5>] Not tainted VLI Using defaults from ksymoops -t elf32-i386 -a i386 Apr 27 15:52:44 werewolf kernel: EFLAGS: 00010246 (2.6.11-jam14) Apr 27 15:52:44 werewolf kernel: eax: 00000000 ebx: 00000001 ecx: 10a316b4 edx: ef4eef1c Apr 27 15:52:44 werewolf kernel: esi: a7f5f001 edi: ef4eef03 ebp: 00000001 esp: ef4eeedc Apr 27 15:52:44 werewolf kernel: ds: 007b es: 007b ss: 0068 Apr 27 15:52:44 werewolf kernel: Stack: b011f8b2 00000001 00000000 00000001 eed6332c 10a316b4 00000001 00000001 Apr 27 15:52:44 werewolf kernel: 00000000 30000000 0000000a ef4eefbc ef507550 a7f5f000 00000001 ef4eef04 Apr 27 15:52:44 werewolf kernel: 00000000 00000000 a7f5f000 a7f5f000 00000001 efd9e8e0 b011f95c a7f5f000 Apr 27 15:52:44 werewolf kernel: Call Trace: Apr 27 15:52:44 werewolf kernel: [<b011f8b2>] do_proc_dointvec+0x2b2/0x32c Apr 27 15:52:44 werewolf kernel: [<b011f95c>] proc_dointvec+0x30/0x35 Apr 27 15:52:44 werewolf kernel: [<b011f5c6>] do_proc_dointvec_conv+0x0/0x3a Apr 27 15:52:44 werewolf kernel: [<b011f311>] do_rw_proc+0x78/0x84 Apr 27 15:52:44 werewolf kernel: [<b011f354>] proc_writesys+0x0/0x24 Apr 27 15:52:44 werewolf kernel: [<b011f373>] proc_writesys+0x1f/0x24 Apr 27 15:52:44 werewolf kernel: [<b0152032>] vfs_write+0x89/0x12a Apr 27 15:52:44 werewolf kernel: [<b015217e>] sys_write+0x41/0x6a Apr 27 15:52:44 werewolf kernel: [<b0102993>] sysenter_past_esp+0x54/0x75 Apr 27 15:52:44 werewolf kernel: Code: c1 89 d8 ba ff ff 00 00 f0 0f c1 10 0f 85 9a 12 00 00 89 c8 83 c4 0c 5b 5e 5f 5d c3 83 7c 24 04 00 74 0d 8b 00 85 c0 75 18 8b 02 <89> 01 31 c0 c3 8b 09 85 c9 78 13 c7 00 00 00 00 00 89 0a 31 c0
Quoted 7 lines
Quoted 7 lines
EIP;
b011f5d5<do_proc_dointvec_conv+f/3a> <=====ecx;
10a316b4<phys_startup_32+109316b4/b0000000>
edx;ef4eef1c<pg0+3f0ddf1c/4fbed400>
esi;a7f5f001<phys_startup_32+a7e5f001/b0000000>
edi;ef4eef03<pg0+3f0ddf03/4fbed400>
esp;ef4eeedc<pg0+3f0ddedc/4fbed400>
Trace; b011f8b2 <do_proc_dointvec+2b2/32c>
Trace; b011f95c <proc_dointvec+30/35>
Trace; b011f5c6 <do_proc_dointvec_conv+0/3a>
Trace; b011f311 <do_rw_proc+78/84>
Trace; b011f354 <proc_writesys+0/24>
Trace; b011f373 <proc_writesys+1f/24>
Trace; b0152032 <vfs_write+89/12a>
Trace; b015217e <sys_write+41/6a>
Trace; b0102993 <sysenter_past_esp+54/75>
This architecture has variable length instructions, decoding before eip is unreliable, take these instructions with a pinch of salt.
Code; b011f5aa <proc_doutsstring+8c/a8>
00000000 <_EIP>:
Code; b011f5aa <proc_doutsstring+8c/a8>
0: c1 89 d8 ba ff ff 00 rorl $0x0,0xffffbad8(%ecx)
Code; b011f5b1 <proc_doutsstring+93/a8>
7: 00 f0 add %dh,%al
Code; b011f5b3 <proc_doutsstring+95/a8>
9: 0f c1 10 xadd %edx,(%eax)
Code; b011f5b6 <proc_doutsstring+98/a8>
c: 0f 85 9a 12 00 00 jne 12ac <_EIP+0x12ac>
Code; b011f5bc <proc_doutsstring+9e/a8>
12: 89 c8 mov %ecx,%eax
Code; b011f5be <proc_doutsstring+a0/a8>
14: 83 c4 0c add $0xc,%esp
Code; b011f5c1 <proc_doutsstring+a3/a8>
17: 5b pop %ebx
Code; b011f5c2 <proc_doutsstring+a4/a8>
18: 5e pop %esi
Code; b011f5c3 <proc_doutsstring+a5/a8>
19: 5f pop %edi
Code; b011f5c4 <proc_doutsstring+a6/a8>
1a: 5d pop %ebp
Code; b011f5c5 <proc_doutsstring+a7/a8>
1b: c3 ret
Code; b011f5c6 <do_proc_dointvec_conv+0/3a>
1c: 83 7c 24 04 00 cmpl $0x0,0x4(%esp)
Code; b011f5cb <do_proc_dointvec_conv+5/3a>
21: 74 0d je 30 <_EIP+0x30>
Code; b011f5cd <do_proc_dointvec_conv+7/3a>
23: 8b 00 mov (%eax),%eax
Code; b011f5cf <do_proc_dointvec_conv+9/3a>
25: 85 c0 test %eax,%eax
Code; b011f5d1 <do_proc_dointvec_conv+b/3a>
27: 75 18 jne 41 <_EIP+0x41>
Code; b011f5d3 <do_proc_dointvec_conv+d/3a>
29: 8b 02 mov (%edx),%eax
This decode from eip onwards should be reliable
Code; b011f5d5 <do_proc_dointvec_conv+f/3a>
00000000 <_EIP>:
Code; b011f5d5 <do_proc_dointvec_conv+f/3a> <=====
0: 89 01 mov %eax,(%ecx) <=====
Code; b011f5d7 <do_proc_dointvec_conv+11/3a>
2: 31 c0 xor %eax,%eax
Code; b011f5d9 <do_proc_dointvec_conv+13/3a>
4: c3 ret
Code; b011f5da <do_proc_dointvec_conv+14/3a>
5: 8b 09 mov (%ecx),%ecx
Code; b011f5dc <do_proc_dointvec_conv+16/3a>
7: 85 c9 test %ecx,%ecx
Code; b011f5de <do_proc_dointvec_conv+18/3a>
9: 78 13 js 1e <_EIP+0x1e>
Code; b011f5e0 <do_proc_dointvec_conv+1a/3a>
b: c7 00 00 00 00 00 movl $0x0,(%eax)
Code; b011f5e6 <do_proc_dointvec_conv+20/3a>
11: 89 0a mov %ecx,(%edx)
Code; b011f5e8 <do_proc_dointvec_conv+22/3a>
13: 31 c0 xor %eax,%eax
1 warning and 1 error issued. Results may not be reliable.
ksymoops 2.4.9 on i686 2.6.11-jam14. Options used
-V (default)
-k /proc/ksyms (default)
-l /proc/modules (default)
-o /lib/modules/2.6.11-jam14/ (default) -m /boot/System.map-2.6.11-jam14 (default)
Warning: You did not tell me where to find symbol information. I will assume that the log matches the kernel and modules that are running right now and I'll use the default options above for symbol resolution. If the current kernel and/or modules do not match the log, you can get more accurate output by telling me the kernel version and where to find map, modules, ksyms etc. ksymoops -h explains the options.
Error (regular_file): read_ksyms stat /proc/ksyms failed No modules in ksyms, skipping objects
No ksyms, skipping lsmod
Apr 27 15:52:44 werewolf kernel: Unable to handle kernel paging request at virtual address 10a316b4 Apr 27 15:52:44 werewolf kernel: b011f5d5 Apr 27 15:52:44 werewolf kernel: *pde = 00000000 Apr 27 15:52:44 werewolf kernel: Oops: 0002 [#2] Apr 27 15:52:44 werewolf kernel: CPU: 3 Apr 27 15:52:44 werewolf kernel: EIP: 0060:[do_proc_dointvec_conv+15/58] Not tainted VLI Apr 27 15:52:44 werewolf kernel: EIP: 0060:[<b011f5d5>] Not tainted VLI Using defaults from ksymoops -t elf32-i386 -a i386 Apr 27 15:52:44 werewolf kernel: EFLAGS: 00010246 (2.6.11-jam14) Apr 27 15:52:44 werewolf kernel: eax: 00000000 ebx: 00000001 ecx: 10a316b4 edx: eeee0f1c Apr 27 15:52:45 werewolf kernel: esi: a7fa3001 edi: eeee0f03 ebp: 00000001 esp: eeee0edc Apr 27 15:52:45 werewolf kernel: ds: 007b es: 007b ss: 0068 Apr 27 15:52:45 werewolf kernel: Stack: b011f8b2 00000001 00000000 00000001 ef3410ac 10a316b4 00000001 00000001 Apr 27 15:52:45 werewolf kernel: 00000000 30000000 0000000a eeee0fbc ef245a70 a7fa3000 00000001 eeee0f04 Apr 27 15:52:45 werewolf kernel: 00000000 00000000 a7fa3000 a7fa3000 00000001 efd9e8e0 b011f95c a7fa3000 Apr 27 15:52:45 werewolf kernel: Call Trace: Apr 27 15:52:45 werewolf kernel: [<b011f8b2>] do_proc_dointvec+0x2b2/0x32c Apr 27 15:52:45 werewolf kernel: [<b011f95c>] proc_dointvec+0x30/0x35 Apr 27 15:52:45 werewolf kernel: [<b011f5c6>] do_proc_dointvec_conv+0x0/0x3a Apr 27 15:52:45 werewolf kernel: [<b011f311>] do_rw_proc+0x78/0x84 Apr 27 15:52:45 werewolf kernel: [<b011f354>] proc_writesys+0x0/0x24 Apr 27 15:52:45 werewolf kernel: [<b011f373>] proc_writesys+0x1f/0x24 Apr 27 15:52:45 werewolf kernel: [<b0152032>] vfs_write+0x89/0x12a Apr 27 15:52:45 werewolf kernel: [<b015217e>] sys_write+0x41/0x6a Apr 27 15:52:45 werewolf kernel: [<b0102993>] sysenter_past_esp+0x54/0x75 Apr 27 15:52:45 werewolf kernel: Code: c1 89 d8 ba ff ff 00 00 f0 0f c1 10 0f 85 9a 12 00 00 89 c8 83 c4 0c 5b 5e 5f 5d c3 83 7c 24 04 00 74 0d 8b 00 85 c0 75 18 8b 02 <89> 01 31 c0 c3 8b 09 85 c9 78 13 c7 00 00 00 00 00 89 0a 31 c0
Quoted 7 lines
Quoted 7 lines
EIP;
b011f5d5<do_proc_dointvec_conv+f/3a> <=====ecx;
10a316b4<phys_startup_32+109316b4/b0000000>
edx;eeee0f1c<pg0+3eacff1c/4fbed400>
esi;a7fa3001<phys_startup_32+a7ea3001/b0000000>
edi;eeee0f03<pg0+3eacff03/4fbed400>
esp;eeee0edc<pg0+3eacfedc/4fbed400>
Trace; b011f8b2 <do_proc_dointvec+2b2/32c>
Trace; b011f95c <proc_dointvec+30/35>
Trace; b011f5c6 <do_proc_dointvec_conv+0/3a>
Trace; b011f311 <do_rw_proc+78/84>
Trace; b011f354 <proc_writesys+0/24>
Trace; b011f373 <proc_writesys+1f/24>
Trace; b0152032 <vfs_write+89/12a>
Trace; b015217e <sys_write+41/6a>
Trace; b0102993 <sysenter_past_esp+54/75>
This architecture has variable length instructions, decoding before eip is unreliable, take these instructions with a pinch of salt.
Code; b011f5aa <proc_doutsstring+8c/a8>
00000000 <_EIP>:
Code; b011f5aa <proc_doutsstring+8c/a8>
0: c1 89 d8 ba ff ff 00 rorl $0x0,0xffffbad8(%ecx)
Code; b011f5b1 <proc_doutsstring+93/a8>
7: 00 f0 add %dh,%al
Code; b011f5b3 <proc_doutsstring+95/a8>
9: 0f c1 10 xadd %edx,(%eax)
Code; b011f5b6 <proc_doutsstring+98/a8>
c: 0f 85 9a 12 00 00 jne 12ac <_EIP+0x12ac>
Code; b011f5bc <proc_doutsstring+9e/a8>
12: 89 c8 mov %ecx,%eax
Code; b011f5be <proc_doutsstring+a0/a8>
14: 83 c4 0c add $0xc,%esp
Code; b011f5c1 <proc_doutsstring+a3/a8>
17: 5b pop %ebx
Code; b011f5c2 <proc_doutsstring+a4/a8>
18: 5e pop %esi
Code; b011f5c3 <proc_doutsstring+a5/a8>
19: 5f pop %edi
Code; b011f5c4 <proc_doutsstring+a6/a8>
1a: 5d pop %ebp
Code; b011f5c5 <proc_doutsstring+a7/a8>
1b: c3 ret
Code; b011f5c6 <do_proc_dointvec_conv+0/3a>
1c: 83 7c 24 04 00 cmpl $0x0,0x4(%esp)
Code; b011f5cb <do_proc_dointvec_conv+5/3a>
21: 74 0d je 30 <_EIP+0x30>
Code; b011f5cd <do_proc_dointvec_conv+7/3a>
23: 8b 00 mov (%eax),%eax
Code; b011f5cf <do_proc_dointvec_conv+9/3a>
25: 85 c0 test %eax,%eax
Code; b011f5d1 <do_proc_dointvec_conv+b/3a>
27: 75 18 jne 41 <_EIP+0x41>
Code; b011f5d3 <do_proc_dointvec_conv+d/3a>
29: 8b 02 mov (%edx),%eax
This decode from eip onwards should be reliable
Code; b011f5d5 <do_proc_dointvec_conv+f/3a>
00000000 <_EIP>:
Code; b011f5d5 <do_proc_dointvec_conv+f/3a> <=====
0: 89 01 mov %eax,(%ecx) <=====
Code; b011f5d7 <do_proc_dointvec_conv+11/3a>
2: 31 c0 xor %eax,%eax
Code; b011f5d9 <do_proc_dointvec_conv+13/3a>
4: c3 ret
Code; b011f5da <do_proc_dointvec_conv+14/3a>
5: 8b 09 mov (%ecx),%ecx
Code; b011f5dc <do_proc_dointvec_conv+16/3a>
7: 85 c9 test %ecx,%ecx
Code; b011f5de <do_proc_dointvec_conv+18/3a>
9: 78 13 js 1e <_EIP+0x1e>
Code; b011f5e0 <do_proc_dointvec_conv+1a/3a>
b: c7 00 00 00 00 00 movl $0x0,(%eax)
Code; b011f5e6 <do_proc_dointvec_conv+20/3a>
11: 89 0a mov %ecx,(%edx)
Code; b011f5e8 <do_proc_dointvec_conv+22/3a>
13: 31 c0 xor %eax,%eax
1 warning and 1 error issued. Results may not be reliable.
--
J.A. Magallon <jamagallon()able!es> \ Software is like sex:
werewolf!able!es \ It's better when it's free
Mandriva Linux release 2006.0 (Cooker) for i586 Linux 2.6.11-jam14 (gcc 4.0.0 (4.0.0-1mdk for Mandriva Linux
Quoted 3 lines · J.A. Magallon
Hi...
I have just built 2.6.12-rc2-mm3 with gcc4 final. I got this couple oopses on boot.
...
Quoted 18 lines
Apr 27 15:52:44 werewolf kernel: Unable to handle kernel paging request at virtual address 10a316b4 Apr 27 15:52:44 werewolf kernel: b011f5d5 Apr 27 15:52:44 werewolf kernel: *pde = 00000000 Apr 27 15:52:44 werewolf kernel: Oops: 0002 [#1] Apr 27 15:52:44 werewolf kernel: CPU: 0 Apr 27 15:52:44 werewolf kernel: EIP: 0060:[do_proc_dointvec_conv+15/58] Not tainted VLI Apr 27 15:52:44 werewolf kernel: EIP: 0060:[<b011f5d5>] Not tainted VLI Using defaults from ksymoops -t elf32-i386 -a i386 Apr 27 15:52:44 werewolf kernel: EFLAGS: 00010246 (2.6.11-jam14) Apr 27 15:52:44 werewolf kernel: eax: 00000000 ebx: 00000001 ecx: 10a316b4 edx: ef4eef1c Apr 27 15:52:44 werewolf kernel: esi: a7f5f001 edi: ef4eef03 ebp: 00000001 esp: ef4eeedc Apr 27 15:52:44 werewolf kernel: ds: 007b es: 007b ss: 0068 Apr 27 15:52:44 werewolf kernel: Stack: b011f8b2 00000001 00000000 00000001 eed6332c 10a316b4 00000001 00000001 Apr 27 15:52:44 werewolf kernel: 00000000 30000000 0000000a ef4eefbc ef507550 a7f5f000 00000001 ef4eef04 Apr 27 15:52:44 werewolf kernel: 00000000 00000000 a7f5f000 a7f5f000 00000001 efd9e8e0 b011f95c a7f5f000 Apr 27 15:52:44 werewolf kernel: Call Trace: Apr 27 15:52:44 werewolf kernel: [<b011f8b2>] do_proc_dointvec+0x2b2/0x32c Apr 27 15:52:44 werewolf kernel: [<b011f95c>] proc_dointvec+0x30/0x35
Quoted 8 lines
Apr 27 15:52:44 werewolf kernel: [<b011f5c6>] do_proc_dointvec_conv+0x0/0x3a Apr 27 15:52:44 werewolf kernel: [<b011f311>] do_rw_proc+0x78/0x84 Apr 27 15:52:44 werewolf kernel: [<b011f354>] proc_writesys+0x0/0x24 Apr 27 15:52:44 werewolf kernel: [<b011f373>] proc_writesys+0x1f/0x24 Apr 27 15:52:44 werewolf kernel: [<b0152032>] vfs_write+0x89/0x12a Apr 27 15:52:44 werewolf kernel: [<b015217e>] sys_write+0x41/0x6a Apr 27 15:52:44 werewolf kernel: [<b0102993>] sysenter_past_esp+0x54/0x75 Apr 27 15:52:44 werewolf kernel: Code: c1 89 d8 ba ff ff 00 00 f0 0f c1 10 0f 85 9a 12 00 00 89 c8 83 c4 0c 5b 5e 5f 5d c3 83 7c 24 04 00 74 0d 8b 00 85 c0 75 18 8b 02 <89> 01 31 c0 c3 8b 09 85 c9 78 13 c7 00 00 00 00 00 89 0a 31 c0
Old news :-) See:
http://marc.theaimsgroup.com/?l=linux-kernel&m=111425013208085&w=2
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=21173
To fix the bug, apply this patch to gcc-4.0.0 final:
http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/tree-ssa-pre.c.diff?cvsroot=gcc&only_with_tag=gcc-4_0-branch&r1=2.65.4.2&r2=2.65.4.3
/Mikael