Bug 8491 - OOPS triggered by ip(8) deconfiguring a network interface
Summary: OOPS triggered by ip(8) deconfiguring a network interface
Status: CLOSED PATCH_ALREADY_AVAILABLE
Alias: None
Product: Power Management
Classification: Unclassified
Component: Other (show other bugs)
Hardware: i386 Linux
: P2 high
Assignee: power-management_other
URL:
Keywords:
Depends on:
Blocks:
 
Reported: 2007-05-17 07:00 UTC by Ben Collins
Modified: 2011-07-30 05:26 UTC (History)
1 user (show)

See Also:
Kernel Version: 2.6.22-rc1
Subsystem:
Regression: No
Bisected commit-id:

Attachments
Dmesg showing oops (121.12 KB, text/plain)
2007-05-17 07:03 UTC, Ben Collins 		Details
Add an attachment (proposed patch, testcase, etc.)

Description Ben Collins 2007-05-17 07:00:37 UTC 
Most recent kernel where this bug did *NOT* occur:
Distribution: Ubuntu Feisty/Gutsy
Hardware Environment: Pentium M, e1000, ipw2200
Software Environment: network-manager
Problem Description: Oops removing proc entry during ifdown

Steps to reproduce: Suspend/resume with interfaces alive
Comment 1 Ben Collins 2007-05-17 07:03:40 UTC 
Created attachment 11525 [details]
Dmesg showing oops
Comment 2 Ben Collins 2007-05-17 07:07:21 UTC 
Original bug report: https://launchpad.net/bugs/115184
Comment 3 Herbert Xu 2007-05-20 03:30:08 UTC 
Andrew Morton <akpm@linux-foundation.org> wrote:
>
> [43092.388000] Stack: e30f433c cab60000 c028608e 00000000 e2e8ae40 00000000 e2e8a5c0 dfe0aa58 
> [43092.388000]        c826ca00 c826ca00 f10a72ff f1088989 00a33d71 000080fe 00000000 00000001 
> [43092.388000]        cab60000 e2e8a5c0 dfe0aa58 00000000 c826ca00 f1089594 c826ca54 00000040 
> [43092.388000] Call Trace:
> [43092.388000]  [<c028608e>] pneigh_queue_purge+0x1e/0x30
> [43092.388000]  [<f10a72ff>] snmp6_unregister_dev+0x2f/0x40 [ipv6]

Looks like a problem with your resume.  The second argument to
remove_proc_entry, parent, has the value 00a33d71 which is clearly
bogus.  That value in turn comes from proc_net_devsnmp6.  So somehow
during the suspend/resume that pointer has been corrupted.

Cheers,
Comment 4 Ben Collins 2007-05-29 05:42:44 UTC 
This appears to be fixed, at least in 2.6.22-rc2.
Note You need to log in before you can comment on or make changes to this bug.