On Mon, Jun 30, 2003 at 04:25:14PM -0700, David Brownell wrote: > Greg KH wrote: > > >Hm, in beating on this for a while, the only thing I saw "odd" was > >reports of calls failing to usb_unlink_urb() by drivers in their > >disconnect() functions. Which is what we would expect to have happen > >here :) > > In this case, "no news is good news" !! There are drivers that > don't bother looking at completion codes, and might get into > infinite resubmit loops -- evil. > > Which drivers did you beat on? I'd expect "visor" would be happier > after this patch (was it?), and that "hub" would behave (I seem to > recall teaching it about -ESHUTDOWN ages ago for other reasons).
I tested a usb trackball, some different usb-serial devices, and yes, the visor driver. No real problems there. I didn't try a hub, as I started to run into the previously mentioned UHCI problems... > >>OK, then for anyone who wants to test out such a change, try > >>this (untested) patch to see what breaks. It doesn't update > >>the API docs, or the other cases where driver disconnect() > >>methods are called, or any of the disconnect() methods, but > >>it should be a start. > > > > > >I like it, want to flush out the rest of the places this needs to be > >changed? > > OK, at least so far as usbcore is concerned; and I'll sanity > check "usbnet" but not very many other drivers. > > I suspect the basic rule of thumb should be that if the driver > has URB completion logic that understands -ESHUTDOWN, it'll > behave. We can only hope :) thanks, greg k-h ------------------------------------------------------- This SF.Net email sponsored by: Free pre-built ASP.NET sites including Data Reports, E-commerce, Portals, and Forums are available now. Download today and enter to win an XBOX or Visual Studio .NET. http://aspnet.click-url.com/go/psa00100006ave/direct;at.asp_061203_01/01 _______________________________________________ [EMAIL PROTECTED] To unsubscribe, use the last form field at: https://lists.sourceforge.net/lists/listinfo/linux-usb-devel
