On Thu, 8 Jan 2004, David Brownell wrote: > Oliver Neukum wrote: > > > > Yes, that is true. But, maybe I am dense, but why can't this be done > > by making state atomic or protect it by a spinlock? > > Not dense ... but you do seem to have overlooked the (earlier) part > of that note which said: > > >> Maybe having a spinlock protect dev->state transitions would be > >> workable, but it'd likely be a larger patch. > > "Atomic" might work too, except that the value is an enum not > an integer so there'd be some type punning in the atomic_set() > and atomic_read() calls.
Actually, making state atomic probably won't work. It doesn't resolve such problems as usb_disconnect() setting state to NOTATTACHED just before usb_set_configuration() sets it to CONFIGURED. Using a spinlock would be the conventional way to do it. It wouldn't be that much more work because there are only a few places where the state changes after usb_new_device() runs. Alan Stern ------------------------------------------------------- This SF.net email is sponsored by: Perforce Software. Perforce is the Fast Software Configuration Management System offering advanced branching capabilities and atomic changes on 50+ platforms. Free Eval! http://www.perforce.com/perforce/loadprog.html _______________________________________________ [EMAIL PROTECTED] To unsubscribe, use the last form field at: https://lists.sourceforge.net/lists/listinfo/linux-usb-devel
