On Friday 21 August 2009, Felipe Balbi wrote: > On Fri, Aug 21, 2009 at 08:50:40PM +0200, ext David Brownell wrote: > > On Friday 14 August 2009, Felipe Balbi wrote: > > > The problem is then we we try to transfer files of over 31k over an obex > > > interface (probably acm and g_serial suffer the same) we get a short > > > transfer (511) where we shouldn't get. > > > > Why "should" that not happen? ... > > > > Which says to me the issue is your host side code. It > > should be working with a byte stream model, where packet > > boundaries do not matter. But it's treating them as > > significant, and thus is wrongly getting upset ... > > still the 511-byte packet is considered short packet by the host side > and thus end of transfer, no ? OBEX transfers aren't supposed to use that particular policy ... don't they include explicit lengths? > What should host do in that case ? kick another bulk in transfer ? Yes; if you've got a partial transfer then try to read the rest. - Dave -- 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