[2.4] Nforce2 oops and occasional hang (tried the lockups patch, no difference)
I've posted this a couple of times, with no response.
So long as memory pressure is kept to a minimum, the system is stable. Running without swap and without serious work kept it going for a couple of weeks.
Running (currently with the nforce2-lockups patches and HZ=1000 but no APIC/IO-APIC) results in the same oopses as every kernel I've tried, including 2.4.22.
Suggestions? I'd really love to be able to use this thing reliably under Linux :(
Unable to handle kernel NULL pointer dereference at virtual address 00000089 c012dff7 *pde = 00000000 Oops: 0000 CPU: 0 EIP: 0010:[<c012dff7>] Tainted: P
Using defaults from ksymoops -t elf32-i386 -a i386
EFLAGS: 00010206
eax: 00000081 ebx: cff72b58 ecx: cee3a000 edx: 0000cad6
esi: 00001000 edi: cbabd9b4 ebp: 00000081 esp: cee3bf08
ds: 0018 es: 0018 ss: 0018
Process giftd (pid: 29738, stackpage=cee3b000)
Stack: c03155c0 c238b1c0 c44671c0 00000038 cee3a000 00000eea 00000000 00000000
00000116 cbabd900 c012e6a0 cee3bf74 c2974660 c2974640 00001000 00001000
00000000 00000000 c012e7f2 c2974640 c2974660 cee3bf74 c012e6a0 c0222e67
Call Trace: [<c012e6a0>] [<c012e7f2>] [<c012e6a0>] [<c0222e67>] [<c013d413>]
[<c0108fdf>]
Code: 39 78 08 74 05 8b 40 10 eb f2 39 68 0c 75 f6 85 c0 89 c6 0fQuoted 6 lines
Quoted 6 lines
EIP;
c012dff7<do_generic_file_read+157/4e0> <=====ebx;
cff72b58<_end+fc02f6c/104a5494>
ecx;cee3a000<_end+eaca414/104a5494>
edi;cbabd9b4<_end+b74ddc8/104a5494>
esp;cee3bf08<_end+eacc31c/104a5494>
Trace; c012e6a0 <file_read_actor+0/a0>
Trace; c012e7f2 <generic_file_read+b2/1a0>
Trace; c012e6a0 <file_read_actor+0/a0>
Trace; c0222e67 <sys_send+37/40>
Trace; c013d413 <sys_read+a3/110>
Trace; c0108fdf <system_call+33/38>
Code; c012dff7 <do_generic_file_read+157/4e0>
00000000 <_EIP>:
Code; c012dff7 <do_generic_file_read+157/4e0> <=====
0: 39 78 08 cmp %edi,0x8(%eax) <=====
Code; c012dffa <do_generic_file_read+15a/4e0>
3: 74 05 je a <_EIP+0xa>
Code; c012dffc <do_generic_file_read+15c/4e0>
5: 8b 40 10 mov 0x10(%eax),%eax
Code; c012dfff <do_generic_file_read+15f/4e0>
8: eb f2 jmp fffffffc <_EIP+0xfffffffc>
Code; c012e001 <do_generic_file_read+161/4e0>
a: 39 68 0c cmp %ebp,0xc(%eax)
Code; c012e004 <do_generic_file_read+164/4e0>
d: 75 f6 jne 5 <_EIP+0x5>
Code; c012e006 <do_generic_file_read+166/4e0>
f: 85 c0 test %eax,%eax
Code; c012e008 <do_generic_file_read+168/4e0>
11: 89 c6 mov %eax,%esi
Code; c012e00a <do_generic_file_read+16a/4e0>
13: 0f 00 00 sldtl (%eax)
Signature · 1 line
Disconnect <[email protected]>
Quoted 6 lines · Disconnect
I've posted this a couple of times, with no response.
So long as memory pressure is kept to a minimum, the system is stable. Running without swap and without serious work kept it going for a couple of weeks.
Quoted 7 lines
Running (currently with the nforce2-lockups patches and HZ=1000 but no APIC/IO-APIC) results in the same oopses as every kernel I've tried, including 2.4.22.
Suggestions? I'd really love to be able to use this thing reliably under Linux :(
Do you see hard locks with APIC/IO-APIC enabled? Do you see this oops with APIC/IO-APIC enabled? Just curious because I can reproduce the hard locks eventually, but never get an oops. Also, what patches are you running and what kernel version?
Quoted 63 lines
Unable to handle kernel NULL pointer dereference at virtual address 00000089 c012dff7 *pde = 00000000 Oops: 0000 CPU: 0 EIP: 0010:[<c012dff7>] Tainted: PUsing defaults from ksymoops -t elf32-i386 -a i386
EFLAGS: 00010206
eax: 00000081 ebx:cff72b58ecx:cee3a000edx:0000cad6
esi: 00001000 edi:cbabd9b4ebp: 00000081 esp:cee3bf08
ds: 0018 es: 0018 ss: 0018
Process giftd (pid: 29738, stackpage=cee3b000)Stack: c03155c0 c238b1c0 c44671c0 00000038 cee3a000 00000eea 00000000 00000000 00000116 cbabd900 c012e6a0 cee3bf74 c2974660 c2974640 00001000 00001000 00000000 00000000 c012e7f2 c2974640 c2974660 cee3bf74 c012e6a0 c0222e67 Call Trace: [<c012e6a0>] [<c012e7f2>] [<c012e6a0>] [<c0222e67>] [<c013d413>] [<c0108fdf>] Code: 39 78 08 74 05 8b 40 10 eb f2 39 68 0c 75 f6 85 c0 89 c6 0fQuoted 6 lines
Quoted 6 lines
EIP;
c012dff7<do_generic_file_read+157/4e0> <=====ebx;
cff72b58<_end+fc02f6c/104a5494>
ecx;cee3a000<_end+eaca414/104a5494>
edi;cbabd9b4<_end+b74ddc8/104a5494>
esp;cee3bf08<_end+eacc31c/104a5494>Trace;
c012e6a0<file_read_actor+0/a0>
Trace;c012e7f2<generic_file_read+b2/1a0>
Trace;c012e6a0<file_read_actor+0/a0>
Trace;c0222e67<sys_send+37/40>
Trace;c013d413<sys_read+a3/110>
Trace;c0108fdf<system_call+33/38>Code;
c012dff7<do_generic_file_read+157/4e0>
00000000 <_EIP>:
Code;c012dff7<do_generic_file_read+157/4e0> <=====
0: 39 78 08 cmp %edi,0x8(%eax) <=====
Code;c012dffa<do_generic_file_read+15a/4e0>
3: 74 05 je a <_EIP+0xa>
Code;c012dffc<do_generic_file_read+15c/4e0>
5: 8b 40 10 mov 0x10(%eax),%eax
Code;c012dfff<do_generic_file_read+15f/4e0>
8: eb f2 jmp fffffffc <_EIP+0xfffffffc>
Code;c012e001<do_generic_file_read+161/4e0>
a: 39 68 0c cmp %ebp,0xc(%eax)
Code;c012e004<do_generic_file_read+164/4e0>
d: 75 f6 jne 5 <_EIP+0x5>
Code;c012e006<do_generic_file_read+166/4e0>
f: 85 c0 test %eax,%eax
Code;c012e008<do_generic_file_read+168/4e0>
11: 89 c6 mov %eax,%esi
Code;c012e00a<do_generic_file_read+16a/4e0>
13: 0f 00 00 sldtl (%eax)Signature · 7 lines
Disconnect <[email protected]> - To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to [email protected] More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
Signature · 6 lines
Josh McKinney | Webmaster: http://joshandangie.org
--------------------------------------------------------------------------
| They that can give up essential liberty
Linux, the choice -o) | to obtain a little temporary safety deserve
of the GNU generation /\ | neither liberty or safety.
_\_v | -Benjamin FranklinQuoted 4 lines · Josh McKinney
Do you see hard locks with APIC/IO-APIC enabled? Do you see this oops with APIC/IO-APIC enabled? Just curious because I can reproduce the hard locks eventually, but never get an oops. Also, what patches are you running and what kernel version?
With apic/io-apic enabled it crashes hard during boot (with oopses, as I recall). I haven't tested apic support with the fixup patches.
The kernel is 2.4.23 plus the patches from "Fixes for nforce2 hard lockup, apic, io-apic, udma133 covered"
(http://marc.theaimsgroup.com/?l=linux-kernel&m=107080280512734&w=2) and athcool was running. (Its currently disabled, to see if that might have something to do with it.) The machine hasn't been rebooted since the oops, since it seems to be limping along. (I did find some minorly corrupted data earlier; libbz2 had to be reinstalled. Once its straightened out I'll run debsums against it and fix up anything that got mangled..)
The same oops was occurring in stock 2.4.22.
Signature · 1 line
Disconnect <[email protected]>
Oh, and the modules list:
Module Size Used by Tainted: P
i2c-dev 4548 0 (unused)
i2c-core 13604 0 [i2c-dev]
ipt_mark 472 18 ipt_state 568 0 (unused) ipt_mac 664 3 ip_nat_irc 2288 0 (unused) ip_conntrack_irc 3120 1 ip_conntrack_ftp 4144 1 (autoclean) ip_nat_ftp 3024 0 (unused) iptable_filter 1772 1 iptable_mangle 2168 1 ipt_LOG 3480 0 (unused) ipt_TOS 1016 0 (unused) ipt_REJECT 3544 0 (unused) ipt_MARK 792 4 ipt_MASQUERADE 1592 8 ipt_REDIRECT 824 4 iptable_nat 17102 3 [ip_nat_irc ip_nat_ftp ipt_MASQUERADE ipt_REDIRECT] ip_conntrack 21892 4 [ipt_state ip_nat_irc ip_conntrack_irc ip_conntrack_ftp ip_nat_ftp ipt_MASQUERADE ipt_REDIRECT iptable_nat] ip_tables 12544 14 [ipt_mark ipt_state ipt_mac iptable_filter iptable_mangle ipt_LOG ipt_TOS ipt_REJECT ipt_MARK ipt_MASQUERADE ipt_REDIRECT iptable_nat] soundcore 4036 0 (autoclean) sd_mod 11084 0 (autoclean) (unused) sg 28540 0 (autoclean) (unused) sr_mod 15640 0 (autoclean) (unused) scsi_mod 89056 3 (autoclean) [sd_mod sg sr_mod] ide-cd 32096 0 (autoclean) cdrom 28320 0 (autoclean) [sr_mod ide-cd] nvnet 26400 1 mousedev 4276 0 (unused) hid 21572 0 (unused) usbmouse 2264 0 (unused) ehci-hcd 17868 0 (unused) ipv6 182548 -1 tulip 40960 1 crc32 2880 0 [tulip] keybdev 2084 0 (unused) usbkbd 3640 0 (unused) input 3488 0 [mousedev hid usbmouse keybdev usbkbd] usb-ohci 19656 0 (unused) usbcore 65100 1 [hid usbmouse ehci-hcd usbkbd usb-ohci]
Signature · 1 line
Disconnect <[email protected]>