David Brownell wrote:
>
> > > > http://cvs.sourceforge.net/cgi-bin/viewcvs.cgi/linux-usb/hcd/
> > >
> > > Yep, I think I got that problem that'd been so annoying. With
> > > current CVS and a prototype USB 2.0 storage device, this now
> > > hotplugs usb-storage ... usb-storage gets past its first bulk request,
> > > hangs on the second, and then gets unhappy when I unload the
> > > ehci-hcd module while it's still trying to probe that device! :)
> >
> > Do you think it is:
>
> "it" being ...? usb-storage shouldn't ignore the error reported during
> probe(); oopsing is bad. The hang is under investigation; feels most
> like a device problem, or some driver issue. (The driver bulk code
> is the same as the control code except in a couple QH setup paths,
> but that second bulk QH should act the same as the first...)
I was less concerned about the oops, and more about the failing bulk request.
I'll be leaving this one for a while though...
> > > - should behave on some systems with 64bit dma_addr_t
> > Is this the Annex A behaviour?
>
> Per spec, yes.
No testing? I will try to get an EHCI card to Paulus or Anton Blanchard in the
near future. That should get us some 64 bit testing :)
> > We still need -ac patches, right? Maybe that hub code for USB 2.0 needs to go
> > into 2.4.5?
>
> I'd like to see that (and the usbdevfs update too) go in before long,
> but one can't quite argue it as a bugfix ... :)
It is a clear bug against the USB 2.0 (currently approved version) spec.
Johannes: are you going to feed this?
Brad
_______________________________________________
[EMAIL PROTECTED]
To unsubscribe, use the last form field at:
http://lists.sourceforge.net/lists/listinfo/linux-usb-devel
