Bug 10989
| Summary: | kernel oopses when wiggling the mouse to make it known to hidd | ||
|---|---|---|---|
| Product: | Drivers | Reporter: | Daniel Vetter (daniel) |
| Component: | Bluetooth | Assignee: | drivers_bluetooth (drivers_bluetooth) |
| Status: | CLOSED CODE_FIX | ||
| Severity: | normal | CC: | akpm, bunk, daniel, dmitry.torokhov, hidave.darkstar, jikos, marcel |
| Priority: | P1 | ||
| Hardware: | All | ||
| OS: | Linux | ||
| Kernel Version: | 2.6.26-rc7-00051-g62786b9 | Subsystem: | |
| Regression: | Yes | Bisected commit-id: | |
| Bug Depends on: | |||
| Bug Blocks: | 10492 | ||
| Attachments: |
Use a more unique bus name for connections
dmesg output for comment #10 |
||
|
Description
Daniel Vetter
2008-06-26 10:32:07 UTC
You sort-of imply that this is a regression, but that isn't clear? Hmm, I think we had a similar issue earlier when we tear down connection that is not completely set up. Maybe Marcel has an idea. Created attachment 16642 [details]
Use a more unique bus name for connections
You can test this patch if it works. Also you can pull in the bluetooth-2.6 tree since it contains something like 22 pending patches. You said 2.6.25-rc5 worked. Does 2.6.25 also work? Two things: 1. Why kobject EXIST? I remember some csr dongles need to reset firstly before being used or will cause something like this. 2. How can we retreive the work function return value? any idea? If the device_add fails we ignore the return value now, so further bug is triggered. > ------- Comment #5 from bunk@kernel.org 2008-06-26 14:05 -------
> You said 2.6.25-rc5 worked.
>
> Does 2.6.25 also work?
Debian flavour 2.6.25-2 works. I only started tracking -linus very
recently, but if you want I could recheck mainline 2.6.25. The problem is
that it's sometimes very hard to hit: Yesterday I rebooted consecutively at least 20
times without hitting it, right this morning I've hitted it on the first
boot-up.
I just saw that I screwed up the version numbers. I meant 2.6.26-rc7-git doesn't work and 2.6.26-rc5 _does_ work. So s/25-rc/26-rc/ in all my comments. Sorry for that mess-up. > ------- Comment #4 from marcel@holtmann.org 2008-06-26 13:44 -------
> You can test this patch if it works. Also you can pull in the bluetooth-2.6
> tree since it contains something like 22 pending patches.
I tested:
- 2.6.26-rc8 vanilla, which crashed,
- 2.6.26-rc8 + bluetooth-2.6 and your patched reverted (it did not apply
to rc8), which eventually crashed and
- 2.6.26-rc8 + bluetooth-2.6, which survived quite a lot of test-runs.
So I think your patch does prevent the oops from happening.
While retesting mainline I capture an oops for the second case (i.e. beside the mouse everything works 'till shutdown, were presumably a process is stuck somewhere odd). This is on 2.6.26-rc8-00076-g1702b52 from -linus. Created attachment 16670 [details] dmesg output for comment #10 Handled-By : Marcel Holtmann <marcel@holtmann.org> On Thursday, 10 of July 2008, Daniel Vetter wrote:
> Hi
>
> After about a dozen rebots I managed to crash 2.6.26-rc9-00005-g1b40a89 in
> the same fashion.
>
> Yours Daniel
> On Sun, Jul 06, 2008 at 01:45:53PM +0200, Rafael J. Wysocki wrote:
> > This message has been generated automatically as a part of a report
> > of recent regressions.
> >
> > The following bug entry is on the current list of known regressions
> > from 2.6.25. Please verify if it still should be listed.
> >
> >
> > Bug-Entry : http://bugzilla.kernel.org/show_bug.cgi?id=10989
> > Subject : kernel oopses when wiggling the mouse to make it
> known to hidd
> > Submitter : Daniel Vetter <daniel@ffwll.ch>
> > Date : 2008-06-26 10:32 (11 days old)
On Monday, 14 of July 2008, Daniel Vetter wrote:
> 2.6.26 crashed right away on the first boot-up.
>
> On Sun, Jul 13, 2008 at 08:00:46PM +0200, Rafael J. Wysocki wrote:
> > This message has been generated automatically as a part of a report
> > of recent regressions.
> >
> > The following bug entry is on the current list of known regressions
> > from 2.6.25. Please verify if it still should be listed.
> >
> >
> > Bug-Entry : http://bugzilla.kernel.org/show_bug.cgi?id=10989
> > Subject : kernel oopses when wiggling the mouse to make it
> known to hidd
> > Submitter : Daniel Vetter <daniel@ffwll.ch>
> > Date : 2008-06-26 10:32 (18 days old)
> > Handled-By : Marcel Holtmann <marcel@holtmann.org>
I'm closing this (fix arrived in -linus in time for 2.6.27). btw: is it the reporters duty to close a bug in a case like this? |