Duncan Sands wrote:
In later steps we can (1) turn dev->serialize into a rwsem
That doesn't bother me (unlike some folk!) but I do agree it should wait until the rest shakes out, and until we can observe benefits. Hardly any code paths need complete mutual exclusion (writelock).
(2) push the acquisition of dev->serialize down to the lower levels as they are fixed up.
Actually at least some of those lower levels are getting fixed by pushing the lock acquisition up.
The reason usbfs needs to get involved in that stuff is that it's going around the standard usb driver structure. One of the other things that "usbfs2" should do is act a lot more like a "normal" device driver, so it plays better with more of the normal usbcore mechanisms.
- Dave
------------------------------------------------------- This SF.Net email is sponsored by: IBM Linux Tutorials Free Linux tutorial presented by Daniel Robbins, President and CEO of GenToo technologies. Learn everything from fundamentals to system administration.http://ads.osdn.com/?ad_id=1470&alloc_id=3638&op=click _______________________________________________ [EMAIL PROTECTED] To unsubscribe, use the last form field at: https://lists.sourceforge.net/lists/listinfo/linux-usb-devel
