On Tue, 14 Jan 2014, Mike Snitzer wrote: > On Tue, Jan 14 2014 at 3:43pm -0500, > Mikulas Patocka <mpatocka@xxxxxxxxxx> wrote: > > > > > > > On Tue, 14 Jan 2014, Mike Snitzer wrote: > > > > > On Mon, Jan 13 2014 at 7:14pm -0500, > > > Mikulas Patocka <mpatocka@xxxxxxxxxx> wrote: > > > > > > > + dm_bufio_set_minimum_buffers(client, DM_PREFETCH_CHUNKS + 1); > > > > + > > > > > > Why the +1? I'd prefer that be managed in dm-bufio rather than relying > > > on all callers to know to add +1. > > > > > > Can you supply a followup patch that cleans this up in bufio and remove > > > the +1 in the above dm_bufio_set_minimum_buffers call? > > > > If DM_PREFETCH_CHUNKS is 12, we need 13 buffers in dm-bufio (one for the > > current buffer that is being read and 12 read ahead buffers). > > > > It is correct to have +1 in snapshot code, not in dm-bufio code. > > OK, thanks for clarifying. > > I've folded the dm bufio changes needed for the respective dm snapshot > changes -- so there are now 4 patches instead of your otiginal 6. > I added a comment for "1 + DM_PREFETCH_CHUNKS" above the > dm_bufio_set_minimum_buffers call and tweaked the headers a little. > > I also staged your non-modular kobject release patch (tweaked the header > on that a little bit too). > > Please see the 5 topmost commits here: > http://git.kernel.org/cgit/linux/kernel/git/device-mapper/linux-dm.git/log/?h=for-next OK. The patches look good. Mikulas -- dm-devel mailing list dm-devel@xxxxxxxxxx https://www.redhat.com/mailman/listinfo/dm-devel