On Thu, 3 Nov 2005, David Brownell wrote: > On Thursday 03 November 2005 9:33 am, Alan Stern wrote: > > On Thu, 3 Nov 2005, David Brownell wrote: > > > > > > One way to test this theory is to use the undocumented > > > > "old_scheme_first=y" parameter for the usbcore module. > > > > > > Actually that doesn't get rid of the fault mode (short reads), it just > > > moves it to a different place. > > > > It gets moved to a place where it won't execute at all if the old-style > > code still works. > > There'll still be a descriptor read where the requested and actual > lengths disagree, though.
No. There'll be a descriptor read where the requested and _total_ lengths disagree. That's quite a different thing. And it's exactly how pre-2.6.9 kernels worked. > > > Then there's the other fault mode, where > > > if SET_ADDRESS fails it's not clear whether the address change actually > > > took effect ... > > > > But that was present in pre-2.6.9 as well, so it shouldn't matter for > > Charles. > > Well, given there can be multiple faults in those paths, it _might_ not > matter. I suspect there could still be bad combinations, though they > may well show up less regularly. Remember that Charles said it worked fine in earlier kernels. I'd be surprised to learn that since then we had introduced fault paths that got activated randomly, with higher and higher probability in each new release. Alan Stern ------------------------------------------------------- SF.Net email is sponsored by: Tame your development challenges with Apache's Geronimo App Server. Download it for free - -and be entered to win a 42" plasma tv or your very own Sony(tm)PSP. Click here to play: http://sourceforge.net/geronimo.php _______________________________________________ linux-usb-devel@lists.sourceforge.net To unsubscribe, use the last form field at: https://lists.sourceforge.net/lists/listinfo/linux-usb-devel
