On Fri, Jul 26, 2013 at 4:55 AM, Greg Kroah-Hartman <gregkh@xxxxxxxxxxxxxxxxxxx> wrote: > On Fri, Jul 26, 2013 at 04:52:36AM +0800, Ming Lei wrote: >> Hi Greg, >> >> On Fri, Jul 26, 2013 at 3:48 AM, Alan Stern <stern@xxxxxxxxxxxxxxxxxxx> wrote: >> > On Thu, 25 Jul 2013, Greg Kroah-Hartman wrote: >> > >> >> On Wed, Jul 03, 2013 at 10:53:09PM +0800, Ming Lei wrote: >> >> > The patch does the below improvement: >> >> > >> >> > - think QH_STATE_COMPLETING as unlinking state since all URBs on the >> >> > endpoint should be in unlinking or unlinked when doing endpoint_disable() >> >> > >> >> > - add "WARN_ON(!list_empty(&qh->qtd_list));" if qh->qh_state is >> >> > QH_STATE_LINKED because there shouldn't be any active transfer in qh >> >> >> >> If this can "never" happen, why spit a nasty message out to the error >> >> log? What can a user do with this? >> > >> > In this context, "never" means the only way it can happen is if there's >> > a serious bug in the way usbcore handles altsetting and config changes. >> > That code is sufficiently complicated that a little extra error >> > checking won't hurt, IMO. >> > >> > If it ever happens, a user could post the log message on the mailing >> > list, giving us a chance to find and fix the bug. >> > >> >> I really don't like adding stuff like this to the kernel if at all >> >> possible. >> > >> > Up to you. Still, I think bringing situations like this to people's >> > attention is a large part of what WARN_ON is intended for. >> >> Yes, I agree with Alan's comment. >> >> Could you merge this patchset since I have some following work >> which depends on this set? Or please let me know if resend is needed. > > Give me some time, this isn't trivial stuff, and I'm not quite sure I > agree that it should even be done yet... OK, any comments are appreciated, :-) I have played this stuff on my laptop and arm boxes for more than one month, and looks it is working well. Now the feature is only enabled on EHCI, I hope it can be merged and let more guys give more tests via -next tree. Suppose regressions caused and they aren't fixed, looks we still can revert the patchset, at least now, I don't find any, :-) Thanks, -- Ming Lei -- To unsubscribe from this list: send the line "unsubscribe linux-usb" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
