Am Freitag, 15. Juni 2007 schrieb Alan Stern:
> [Added Matt to CC: list]
>
> On Fri, 15 Jun 2007, Oliver Neukum wrote:
>
> > Hi,
> >
> > I've been looking at memory management in ehci. IMHO to cleanly
> > solve the deadlock, callers need a way to pass down memory reserves
> > to the HCD. As they are unique to the bus, this needs a method to allocate
> > such reserves from the HCD.
> >
> > So a few questions arise.
> >
> > 1. Can every HCD compute how much reserves it would need in the
> > worst case to transfer one buffer of a given size over a given endpoint?
>
> I believe so. uhci-hcd certainly can.
Good.
> > 2. Can the SCSI layer deal with partially executed sg requests?
>
> Yes for READs and WRITEs; I'm not sure about other sorts of requests.
>
> But you're much better off avoiding partially executed requests if you
> can avoid it. It would be preferable to transmit the data one packet
> at a time: slower but more likely to succeed.
Why is that better? The problem is unlikely to be unique to USB
and should therefore be handled in common code if possible.
Regards
Oliver
-------------------------------------------------------------------------
This SF.net email is sponsored by DB2 Express
Download DB2 Express C - the FREE version of DB2 express and take
control of your XML. No limits. Just data. Click to get it now.
http://sourceforge.net/powerbar/db2/
_______________________________________________
linux-usb-devel@lists.sourceforge.net
To unsubscribe, use the last form field at:
https://lists.sourceforge.net/lists/listinfo/linux-usb-devel
