Hi,
On Tue, Mar 05, 2013 at 09:42:36AM +0800, Chen Gang wrote:
> 于 2013年03月04日 22:35, Felipe Balbi 写道:
> > this is the wrong fix, I believe. Looks like when I wrote the commits
> > you mention, I deleted more code then I should. Looks like the real fix
> > would be to add back:
> >
> > /* report disconnect; the driver is already quiesced */
> > if (driver) {
> > spin_unlock(&dev->lock);
> > driver->disconnect(&dev->gadget);
> > spin_lock(&dev->lock);
> > }
> >
>
> ok, thanks.
>
>
> > since stop_activity() also gets called from RESET interrupt, and in that
> > case we need to call driver->disconnect(). Can you make a simple test
> > that would take current code and issue a device reset to see if that
> > would break, then apply my suggestion above and run the same test again?
> >
>
> ok, thanks, that is what I should do.
>
> but excuse me for my time limitation:
> in these days, my father had a serious heart disease in hospital.
> within this week end, most of my time has to be in hospital.
> (God bless, and thank Jesus Christ, my father is safe now).
> within my company (Asianux), also has something to do.
I'm glad to hear your father is better now :-) Hope he recovers soon.
> so I try to finish it in next week end (Mar-15-2013).
>
> if we can not bear the time point, please reply (or prefer another
> member to finish it)
no problem, take your time. Family is more important than a bugfix. I
don't have that HW anyway and can't test fixes to it.
> > If you want, I have a simple libusb-1.0-based little app which can issue
> > resets to any device you ask. Currently it will reset a device 10K
> > times but you can remove the for loop if you want:
>
> thank you for your information, they are valuable to me.
no problem, any time.
--
balbi
Attachment:
signature.asc
Description: Digital signature
