About this list Date view Thread view Subject view Author view Attachment view

From: Eje Gustafsson (Macahan_at_fament.com)
Date: Thu 13 Feb 2003 - 21:26:52 GMT


Thursday, February 13, 2003, 1:58:13 AM, you wrote:

HP> On Wed, Feb 12, 2003 at 03:06:48PM -0600, John Goerzen wrote:
>> Hello,
>>
>> A little while ago, I reported a crash with 2.4.20ctx16. Now, it has
>> happened again, and I have captured much more debug info:

HP> this one was very useful/informational ...

HP> I would suggest to focus on NIC driver and sending
HP> udp/(tcp) messages ...

Hmm. interesting.. The box in question is a web server and crash
happen during prime time (site pushes around 35Gig/mo).
Earlier that day httpd had crashed (I think it was due to it it had a
kernel panic was in to much of a hurry to really look the message
over).

HP> It should be possible to trigger this issue by
HP> sending a lot of udp/(tcp) packets via hping2
HP> (or something similar) on systems that crashed
HP> at least once with "kernel BUG at sched.c:570"

that was the same one I got. Will do some tests if I get a chance.

HP> my guess would be that some NIC driver has some
HP> issues with schedule in softirq (shared code)

I got a Gigabyte GA-7ZXE mobo with via chipset on it.
IRQ 9 is shared between the USB Controller: Via Technologies and my
network card identified in /proc/pci as
Ethernet Controller: VIA Technologies, Inc VT86C100A (Rhine)
also on Irq 9.

HP> but maybe I'm totally wrong, so do not blame me
HP> if it is something completely different ...
HP> #include <std_disclaimer.h>

Sounds like a very plausible explanation that fits right into my case.
Would be interesting to hear John Goerzen's system setup since we both
are running the 2.4.20ctx-16 kernel and both had the same type of
crash

/ Eje

HP> best,
HP> Herbert

>> ksymoops 2.4.8 on i686 2.4.20ctx-16. Options used
>> -V (default)
>> -k /proc/ksyms (default)
>> -l /proc/modules (default)
>> -o /lib/modules/2.4.20ctx-16/ (default)
>> -m /boot/System.map-2.4.20ctx-16 (default)
>>
>> kernel BUG at sched.c:570!
>> invalid operand: 0000
>> CPU: 1
>> EIP: 0010:[<c011a201>] Not tainted
>> Using defaults from ksymoops -t elf32-i386 -a i386
>> EFLAGS: 00010282
>> eax: 000000018 ebx: 00000080 ecx: 00000004 edx: 000000000
>> esi: d61ca000 edi: d61cbb6c ebp: d71cbb44 esp: d71cbb14
>> ds: 0018 es: 0018 ss: 0018
>> Process amavisd (pid: 5552, stackpage=d71cb000)
>> Stack: c02b910a c02fc6b4 c1030020 000000207 ffffffff d71ca00 d52b0080
>> 00000001 c0279edf c02fbef8 d71ca000 d71cbb6c e9733180 c02a510c 00000014
>> c02fbef8 e22eeb00 d1852827 c02a4f99 c02fdef8 d71cbb6c ffffffff 00000000
>> 00000000
>> Call trace: [<c0279edf>] [<c02a510c>] [<c02a4f99>] [<c012c12a>] [<c02882ef>]
>> [<c027ada9>]
>> [<c024ab25>]
>> [<c027b2f9>] [<c0260e30>] [<c0260f7e>] [<c0260e30>] [<c0258240>]
>> [<c0260e30>] [<c0260fa0>] [<c0260bb3>]
>> [<c0260e30>] [<c0261169>] [<c0257e74>] [<c0260fa0>] [<c0258240>]
>> [<c0260fa0>] [<c0260d7a>] [<c0260fa0>]
>> [<c0250c43>] [<c0250d7c>] [<c0250eb9>] [<c0122e39>] [<c02650f0>]
>> [<c02587d6>] [<c02650f0>] [<c0263ef3>]
>> [<c02650f0>] [<c0279d2d>] [<c02745d7>] [<c0275194>]
>> [<c02696cd>] [<c0159f85>] [<c012e97d>] [<c02896b2>]
>> [<c0247f65>] [<c02481fe>] [<c01429b3>] [<c01093ef>]
>> Code: 0f 0b 3a 02 02 91 2b c0 e9 3f fb ff ff 0f 0b 33 02 02 91 2b
>>
>>
>> >>EIP; c011a201 <schedule+501/530> <=====
>>
>> >>esi; d61ca000 <_end+15dfc91c/384b797c>
>> >>edi; d61cbb6c <_end+15dfe488/384b797c>
>> >>ebp; d71cbb44 <_end+16dfe460/384b797c>
>> >>esp; d71cbb14 <_end+16dfe430/384b797c>
>>
>> Trace; c0279edf <tcp_v4_send_reset+11f/190>
>> Trace; c02a510c <rwsem_down_failed_common+5c/7e>
>> Trace; c02a4f99 <rwsem_down_write_failed+29/40>
>> Trace; c012c12a <.text.lock.sys+ea/190>
>> Trace; c02882ef <inet_sock_destruct+ff/1c0>
>> Trace; c027ada9 <tcp_v4_do_rcv+139/1c0>
>> Trace; c024ab25 <sk_free+85/90>
>> Trace; c027b2f9 <tcp_v4_rcv+4c9/5a0>
>> Trace; c0260e30 <ip_local_deliver_finish+0/170>
>> Trace; c0260f7e <ip_local_deliver_finish+14e/170>
>> Trace; c0260e30 <ip_local_deliver_finish+0/170>
>> Trace; c0258240 <nf_hook_slow+100/1d0>
>> Trace; c0260e30 <ip_local_deliver_finish+0/170>
>> Trace; c0260fa0 <ip_rcv_finish+0/226>
>> Trace; c0260bb3 <ip_local_deliver+53/80>
>> Trace; c0260e30 <ip_local_deliver_finish+0/170>
>> Trace; c0261169 <ip_rcv_finish+1c9/226>
>> Trace; c0257e74 <nf_iterate+54/a0>
>> Trace; c0260fa0 <ip_rcv_finish+0/226>
>> Trace; c0258240 <nf_hook_slow+100/1d0>
>> Trace; c0260fa0 <ip_rcv_finish+0/226>
>> Trace; c0260d7a <ip_rcv+19a/250>
>> Trace; c0260fa0 <ip_rcv_finish+0/226>
>> Trace; c0250c43 <netif_receive_skb+d3/190>
>> Trace; c0250d7c <process_backlog+7c/120>
>> Trace; c0250eb9 <net_rx_action+99/140>
>> Trace; c0122e39 <do_softirq+d9/e0>
>> Trace; c02650f0 <ip_queue_xmit2+0/250>
>> Trace; c02587d6 <.text.lock.netfilter+c0/ea>
>> Trace; c02650f0 <ip_queue_xmit2+0/250>
>> Trace; c0263ef3 <ip_queue_xmit+2e3/300>
>> Trace; c02650f0 <ip_queue_xmit2+0/250>
>> Trace; c0279d2d <tcp_v4_send_check+4d/e0>
>> Trace; c02745d7 <tcp_transmit_skb+2c7/460>
>> Trace; c0275194 <tcp_write_xmit+184/270>
>> Trace; c02696cd <tcp_sendmsg+57d/10c0>
>> Trace; c0159f85 <__mark_inode_dirty+b5/c0>
>> Trace; c012e97d <do_wp_page+2bd/300>
>> Trace; c02896b2 <inet_sendmsg+42/50>
>> Trace; c0247f65 <sock_sendmsg+75/c0>
>> Trace; c02481fe <sock_write+ae/c0>
>> Trace; c01429b3 <sys_write+a3/160>
>> Trace; c01093ef <system_call+33/38>
>>
>> Code; c011a201 <schedule+501/530>
>> 00000000 <_EIP>:
>> Code; c011a201 <schedule+501/530> <=====
>> 0: 0f 0b ud2a <=====
>> Code; c011a203 <schedule+503/530>
>> 2: 3a 02 cmp (%edx),%al
>> Code; c011a205 <schedule+505/530>
>> 4: 02 91 2b c0 e9 3f add 0x3fe9c02b(%ecx),%dl
>> Code; c011a20b <schedule+50b/530>
>> a: fb sti
>> Code; c011a20c <schedule+50c/530>
>> b: ff (bad)
>> Code; c011a20d <schedule+50d/530>
>> c: ff 0f decl (%edi)
>> Code; c011a20f <schedule+50f/530>
>> e: 0b 33 or (%ebx),%esi
>> Code; c011a211 <schedule+511/530>
>> 10: 02 02 add (%edx),%al
>> Code; c011a213 <schedule+513/530>
>> 12: 91 xchg %eax,%ecx
>> Code; c011a214 <schedule+514/530>
>> 13: 2b 00 sub (%eax),%eax
>>
>> <0>Kernel panic: Aiee, killing interrupt handler!
>>
>> 1 warning issued. Results may not be reliable.
HP> ---
HP> [This E-mail scanned for viruses by Declude Virus]

Best regards,
 Eje Gustafsson mailto:MacAhan_at_fament.com

---
The Family Entertainment Network      http://www.fament.com
Phone : 620-231-7777                  Fax   : 620-231-4066
eBay UserID : macahan
          - Your Full Time Professionals -

--- [This E-mail scanned for viruses by Declude Virus]


About this list Date view Thread view Subject view Author view Attachment view
[Next/Previous Months] [Main vserver Project Homepage] [Howto Subscribe/Unsubscribe] [Paul Sladen's vserver stuff]
Generated on Thu 13 Feb 2003 - 21:44:20 GMT by hypermail 2.1.3