On Sun, 21 Jul 2013, Greg Kroah-Hartman wrote:
> On Mon, Jul 22, 2013 at 01:00:41AM +0200, Marcin Ślusarz wrote:
> > If you want me to test some patches, just shout ;)
>
> Interesting patch, it makes sense to have this applied. Can you resend
> it with a "Signed-off-by:" line so that I can add it to the tree?
Wait a minute. The patch catches cases where io->dev is NULL in
sg_complete(). But how does that happen?
The only place io->dev gets set to NULL is in sg_clean(), and
sg_clean() gets called in only two places:
if an URB could not be allocated in usb_sg_init(),
or after all the URBs have completed in usb_sg_wait().
In either case, no more URBs will complete after sg_clean() is called,
so sg_complete() won't run.
A patch like this merely covers up the symptom without fixing the
underlying cause. How did this manage to go wrong in the first place?
There must be a bug in the logic of sg_complete() or usb_sg_wait().
_That_ logic bug is what needs to be fixed.
Alan Stern
--
To unsubscribe from this list: send the line "unsubscribe linux-usb" in
the body of a message to [email protected]
More majordomo info at http://vger.kernel.org/majordomo-info.html
