On Fri, 2 Mar 2007, Gordon Messmer wrote:

> > Did you really mean it that the earlier kernels don't generate those "-84" 
> > lines?
> 
> Yes.  As far as I can tell, no matter how long I watch usbmon on a 
> "working" kernel, there is no output if I don't actually move the mouse.

Hmm.  So the problem does not occur in 2.6.17 and it does occur in 2.6.18?  
Under those circumstances I would recommend bisection: try the 
intermediate -rc kernels and perhaps even git -bisect to find exactly 
when the trouble started.  However that's a whole lot of extra work for 
you, and you might not want to do it.

> > the presence or absence of 
> > an I/O error shouldn't depend on the kernel version. 
> 
> I agree completely.
> 
> > Do you have any sort of CPU power saving (like cpufreq) enabled?
> 
> Not as far as I can tell.  None of the cpufreq modules are loaded.  I 
> had seen suggestions in other people's discussions of similar problems 
> that gnome-power-manager might be at fault, but killing it doesn't solve 
> the problem.  I see resets even when no user is logged in to the 
> console, anyway.

Have you tried booting into single-user mode?  That would reduce the 
number of processes.  :-)

> When running 2.6.19 I see a process named [ksuspend_usbd], but I don't 
> know what that is.

It's a kernel thread responsible for suspending idle USB devices.  Under 
2.6.19 it doesn't do very much; in 2.6.20 it's a little more active (it 
will suspend hubs with no devices attached, for instance).  In neither 
kernel will it try to suspend a USB mouse, even one that's not in use.  
That feature hasn't been added yet.

> > Another reason to think there may be a systematic cause, something that 
> > interferes with the USB controller periodically.  But it isn't anything in 
> > the USB stack.
> 
> Perhaps not, but it does seem like the kernel is somehow at fault.  No 
> other component of the system actually changes.  It's possible that 
> something doesn't run on the older kernel due to a missing interface of 
> some type, but I don't really know what that might be.  Other than a few 
> new kernel processes, I don't see any significant difference in 'ps' 
> output between the old and new kernels.

Yes, at this point it's hard to say what's going on.

Alan Stern


-------------------------------------------------------------------------
Take Surveys. Earn Cash. Influence the Future of IT
Join SourceForge.net's Techsay panel and you'll get the chance to share your
opinions on IT & business topics through brief surveys-and earn cash
http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV
_______________________________________________
Linux-usb-users@lists.sourceforge.net
To unsubscribe, use the last form field at:
https://lists.sourceforge.net/lists/listinfo/linux-usb-users

Reply via email to