On Thu, 30 Mar 2017, Oliver Neukum wrote: > > Btw, I'm a lot more concerned about USB storage drivers. When I was > > discussing about this issue at the #raspberrypi-devel IRC channel, > > someone complained that, after switching from the RPi downstream Kernel > > to upstream, his USB data storage got corrupted. Well, if the USB > > storage drivers also assume that the buffer can be continuous, > > that can corrupt data. > > They do assume that. Wait a minute. Where does that assumption occur? And exactly what is the assumption? Mauro wrote "the buffer can be continuous", but that is certainly not what he meant. Alan Stern
