From mboxrd@z Thu Jan 1 00:00:00 1970 From: Greg Thelen Subject: Re: [PATCH -mmotm 2/5] memcg: dirty memory documentation Date: Wed, 17 Mar 2010 09:48:18 -0800 Message-ID: <49b004811003171048h5f27405oe6ea39a103bc4ee3__40179.8427374617$1268851353$gmane$org@mail.gmail.com> References: <1268609202-15581-1-git-send-email-arighi@develer.com> <1268609202-15581-3-git-send-email-arighi@develer.com> <20100316164121.024e35d8.nishimura@mxp.nes.nec.co.jp> Mime-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable Return-path: In-Reply-To: <20100316164121.024e35d8.nishimura-YQH0OdQVrdy45+QrQBaojngSJqDPrsil@public.gmane.org> List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Sender: containers-bounces-cunTk1MwBs9QetFLy7KEm3xJsTq8ys+cHZ5vskTnxNA@public.gmane.org Errors-To: containers-bounces-cunTk1MwBs9QetFLy7KEm3xJsTq8ys+cHZ5vskTnxNA@public.gmane.org To: Daisuke Nishimura Cc: linux-mm-Bw31MaZKKs3YtjvyW6yDsg@public.gmane.org, Andrea Righi , containers-cunTk1MwBs9QetFLy7KEm3xJsTq8ys+cHZ5vskTnxNA@public.gmane.org, linux-kernel-u79uwXL29TY76Z2rM5mHXA@public.gmane.org, Trond Myklebust , Suleiman Souhlal , Andrew Morton , Vivek Goyal , Balbir Singh List-Id: containers.vger.kernel.org On Mon, Mar 15, 2010 at 11:41 PM, Daisuke Nishimura wrote: > On Mon, 15 Mar 2010 00:26:39 +0100, Andrea Righi wro= te: >> Document cgroup dirty memory interfaces and statistics. >> >> Signed-off-by: Andrea Righi >> --- >> =A0Documentation/cgroups/memory.txt | =A0 36 +++++++++++++++++++++++++++= +++++++++ >> =A01 files changed, 36 insertions(+), 0 deletions(-) >> >> diff --git a/Documentation/cgroups/memory.txt b/Documentation/cgroups/me= mory.txt >> index 49f86f3..38ca499 100644 >> --- a/Documentation/cgroups/memory.txt >> +++ b/Documentation/cgroups/memory.txt >> @@ -310,6 +310,11 @@ cache =A0 =A0 =A0 =A0 =A0 =A0- # of bytes of page c= ache memory. >> =A0rss =A0 =A0 =A0 =A0 =A0- # of bytes of anonymous and swap cache memor= y. >> =A0pgpgin =A0 =A0 =A0 =A0 =A0 =A0 =A0 - # of pages paged in (equivalent = to # of charging events). >> =A0pgpgout =A0 =A0 =A0 =A0 =A0 =A0 =A0- # of pages paged out (equivalent= to # of uncharging events). >> +filedirty =A0 =A0- # of pages that are waiting to get written back to t= he disk. >> +writeback =A0 =A0- # of pages that are actively being written back to t= he disk. >> +writeback_tmp =A0 =A0 =A0 =A0- # of pages used by FUSE for temporary wr= iteback buffers. >> +nfs =A0 =A0 =A0 =A0 =A0- # of NFS pages sent to the server, but not yet= committed to >> + =A0 =A0 =A0 =A0 =A0 =A0 =A0 the actual storage. Should these new memory.stat counters (filedirty, etc) report byte counts rather than page counts? I am thinking that byte counters would make reporting more obvious depending on how heterogeneous page sizes are used. Byte counters would also agree with /proc/meminfo. Within the kernel we could still maintain page counts. The only change would be to the reporting routine, mem_cgroup_get_local_stat(), which would scale the page counts by PAGE_SIZE as it does for for cache,rss,etc. -- Greg