> > On Sat, May 9, 2009 at 3:09 PM, Hans Petter Selasky <[email protected]> wrote: > > > Hi, > > > > If the strings are corrupt I would guess at a busdma issue. > > thank you, adding a "flush all" instruction makes the USB controller > works properly. > > Now i should check the implementation of the FA526 invalidate/flush > instruction, because I shouldn't need the extra flush instruction if > bus_dmamap_sync works properly, right?. > > > Are you running stock 8-current ? > > yes
Sounds like the multiple kernel mapping problem in ARM. It was mentioned when the new USB stack was comitted. Muliple kernel mappings can also occur in other situations. There is a patch that has been around for a month and is about to be committed. I just put three counters in the patch to determine if PG_UNMANAGED pages could also be shared. --Mark Tinguely _______________________________________________ [email protected] mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-usb To unsubscribe, send any mail to "[email protected]"
