On Wed, Mar 09, 2011 at 02:33:04AM +0000, Al Viro wrote: > On Tue, Mar 08, 2011 at 09:15:25PM -0500, Stephen Wilson wrote: > > > I think we could also remove the intermediate copy in both mem_read() and > > mem_write() as well, but I think such optimizations could be left for > > follow on patches. > > How? We do copy_.._user() in there; it can trigger page faults and > that's not something you want while holding mmap_sem on some mm. Ah, OK. I did not think thru that subtlety. Was merely mentioning "things we might do afterwords" as opposed to a genuine proposal. > Looks like a deadlock country... So we can't do that from inside > access_process_vm() or its analogs, which means buffering in caller. Thanks for the feed back -- I am certainly (relatively speaking) new to the code so your insights are most valuable. Thanks again! -- steve -- To unsubscribe, send a message with 'unsubscribe linux-mm' in the body to majordomo@xxxxxxxxxx For more info on Linux MM, see: http://www.linux-mm.org/ . Fight unfair telecom internet charges in Canada: sign http://stopthemeter.ca/ Don't email: <a href=mailto:"dont@xxxxxxxxx";> email@xxxxxxxxx </a>