On Wed, Feb 27, 2002 at 11:30:50PM +0100, Oliver Neukum wrote: > > > In testing I could not measure any speed differences with this patch, > > and without it. I would not recommend this approach for drivers that > > really care about speed (we are doing to kmalloc() calls for every > > write), but the visor driver does not have to worry about things like > > this (the limiting factor is the device itself, not the transfer > > speeds.) > > > > Comments? > > How about abstracting the code and let usb_alloc_urb() > take urbs from a pool ? > Though I have no idea on how to find a sensible size for > the pool. How bad for the TLB are we ?
I don't know. I was thinking that this would be a good idea too (a usb subsystem wide urb pool). But more drivers would have to be changed to do what this visor driver does (right now everyone usually just creates them once, on device init.) thanks, greg k-h _______________________________________________ [EMAIL PROTECTED] To unsubscribe, use the last form field at: https://lists.sourceforge.net/lists/listinfo/linux-usb-devel
