On Wed, Sep 04, 2013 at 09:44:47PM +0200, Oliver Neukum wrote: > On Mon, 2013-08-19 at 10:23 -0700, Greg Kroah-Hartman wrote: > > > Please, use perf, and other things, to find out where real problems are > > Actually, first thing about the design. > > > in the USB stack. I'm sure there are locations that we can improve on, > > but until you get some real numbers, it's going to be hard to accept > > stuff like this. > > I suspect the actual issue is allocating memory for each > submission. That is a consequence of allocating URBs unspecific > to an endpoint. I'll not disagree with that, if that's a overall design decision that would be good to make, I'll be all for it. Odds are, we could benifit a lot, but it might take some benchmarks to prove/disprove it. thanks, greg k-h -- To unsubscribe from this list: send the line "unsubscribe linux-usb" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
