Am Donnerstag 11 Oktober 2007 schrieb Alan Stern:
> On Thu, 11 Oct 2007, Oliver Neukum wrote:
>
> > > > > One way around the problem would be to unbind these bad drivers before
> > > > > the suspend begins, say in a notifier routine. That would leave a
> > > >
> > > > Problematic. You'd have to get the order of doing stuff exactly right.
> > >
> > > What order are you concerned about? It will be impossible to have
> >
> > Blocking new device additions vs. calling disconnect.
>
> If we disconnect the bad drivers after khubd has gone to sleep then no
> new devices will be added. There's still the possibility of userspace
> causing a bind via sysfs, though.
Driver loading, change of configuration, ...
Did you check the other code paths?
> > > Of course there's always the other approach -- solve the problem by
> > > adding suspend and resume methods to every driver!
> >
> > Are you volunteering?
>
> How many drivers are we talking about?
Dozens, which will require testing.
> > Resumption may involve IO, which therefore may fail. I cannot see the reason
> > to not count the lack of a resume method as a failure. We'll have to deal with
> > them anyway, especially for reset_resume.
>
> If a driver doesn't have a suspend or resume method, then resuming it
> can't involve I/O. :-)
You still can have IO to a hub to wake the device.
> You appear to be saying that if a driver doesn't have a resume method,
> or the resume method returns an error, then the driver should be
> unbound as soon as possible. For non-system non-auto resumes this is
Yes.
> right away, but for system resumes it isn't until the driver core is
> stable again. Likewise, non-system suspends should fail if a driver
Yes.
> doesn't have a suspend method (I'm not so sure about this part).
Yes.
Regards
Oliver
-------------------------------------------------------------------------
This SF.net email is sponsored by: Splunk Inc.
Still grepping through log files to find problems? Stop.
Now Search log events and configuration files using AJAX and a browser.
Download your FREE copy of Splunk now >> http://get.splunk.com/
_______________________________________________
linux-usb-devel@xxxxxxxxxxxxxxxxxxxxx
To unsubscribe, use the last form field at:
https://lists.sourceforge.net/lists/listinfo/linux-usb-devel
[Home]
[Video for Linux]
[Photo]
[Yosemite Forum]
[Yosemite Photos]
[Video Projectors]
[PDAs]
[Hacking TiVo]
[Linux Kernel]
[Linux SCSI]
[XFree86]
[Devices]
[Big List of Linux Books]
[Free Dating]