From mboxrd@z Thu Jan 1 00:00:00 1970 From: Greg Thelen Subject: Re: [PATCH v6 4/9] memcg: add kernel calls for memcg dirty page stats Date: Mon, 14 Mar 2011 23:32:38 -0700 Message-ID: References: <1299869011-26152-1-git-send-email-gthelen@google.com> <1299869011-26152-5-git-send-email-gthelen@google.com> <20110314151023.GF11699@barrios-desktop> Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: QUOTED-PRINTABLE Cc: Andrew Morton , linux-kernel@vger.kernel.org, linux-mm@kvack.org, containers@lists.osdl.org, linux-fsdevel@vger.kernel.org, Andrea Righi , Balbir Singh , KAMEZAWA Hiroyuki , Daisuke Nishimura , Johannes Weiner , Ciju Rajan K , David Rientjes , Wu Fengguang , Chad Talbott , Justin TerAvest , Vivek Goyal , KONISHI Ryusuke To: Minchan Kim Return-path: Received: from smtp-out.google.com ([74.125.121.67]:16114 "EHLO smtp-out.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1750937Ab1COGdD convert rfc822-to-8bit (ORCPT ); Tue, 15 Mar 2011 02:33:03 -0400 Received: from kpbe19.cbf.corp.google.com (kpbe19.cbf.corp.google.com [172.25.105.83]) by smtp-out.google.com with ESMTP id p2F6X0Pc018112 for ; Mon, 14 Mar 2011 23:33:00 -0700 Received: from qyk2 (qyk2.prod.google.com [10.241.83.130]) by kpbe19.cbf.corp.google.com with ESMTP id p2F6Wwci018394 (version=TLSv1/SSLv3 cipher=RC4-SHA bits=128 verify=NOT) for ; Mon, 14 Mar 2011 23:32:58 -0700 Received: by qyk2 with SMTP id 2so2093042qyk.7 for ; Mon, 14 Mar 2011 23:32:58 -0700 (PDT) In-Reply-To: <20110314151023.GF11699@barrios-desktop> Sender: linux-fsdevel-owner@vger.kernel.org List-ID: On Mon, Mar 14, 2011 at 8:10 AM, Minchan Kim wr= ote: > On Fri, Mar 11, 2011 at 10:43:26AM -0800, Greg Thelen wrote: >> Add calls into memcg dirty page accounting. =A0Notify memcg when pag= es >> transition between clean, file dirty, writeback, and unstable nfs. >> This allows the memory controller to maintain an accurate view of >> the amount of its memory that is dirty. >> >> Signed-off-by: Greg Thelen >> Signed-off-by: Andrea Righi >> Acked-by: KAMEZAWA Hiroyuki >> Reviewed-by: Daisuke Nishimura >> --- >> Changelog since v5: >> - moved accounting site in test_clear_page_writeback() and >> =A0 test_set_page_writeback(). >> >> =A0fs/nfs/write.c =A0 =A0 =A0| =A0 =A04 ++++ >> =A0mm/filemap.c =A0 =A0 =A0 =A0| =A0 =A01 + >> =A0mm/page-writeback.c | =A0 10 ++++++++-- >> =A0mm/truncate.c =A0 =A0 =A0 | =A0 =A01 + >> =A04 files changed, 14 insertions(+), 2 deletions(-) >> >> diff --git a/fs/nfs/write.c b/fs/nfs/write.c >> index 42b92d7..7863777 100644 >> --- a/fs/nfs/write.c >> +++ b/fs/nfs/write.c >> @@ -451,6 +451,7 @@ nfs_mark_request_commit(struct nfs_page *req) >> =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 NFS_PAGE_TAG_COMMIT); >> =A0 =A0 =A0 nfsi->ncommit++; >> =A0 =A0 =A0 spin_unlock(&inode->i_lock); >> + =A0 =A0 mem_cgroup_inc_page_stat(req->wb_page, MEMCG_NR_FILE_UNSTA= BLE_NFS); >> =A0 =A0 =A0 inc_zone_page_state(req->wb_page, NR_UNSTABLE_NFS); >> =A0 =A0 =A0 inc_bdi_stat(req->wb_page->mapping->backing_dev_info, BD= I_RECLAIMABLE); >> =A0 =A0 =A0 __mark_inode_dirty(inode, I_DIRTY_DATASYNC); >> @@ -462,6 +463,7 @@ nfs_clear_request_commit(struct nfs_page *req) >> =A0 =A0 =A0 struct page *page =3D req->wb_page; >> >> =A0 =A0 =A0 if (test_and_clear_bit(PG_CLEAN, &(req)->wb_flags)) { >> + =A0 =A0 =A0 =A0 =A0 =A0 mem_cgroup_dec_page_stat(page, MEMCG_NR_FI= LE_UNSTABLE_NFS); >> =A0 =A0 =A0 =A0 =A0 =A0 =A0 dec_zone_page_state(page, NR_UNSTABLE_NF= S); >> =A0 =A0 =A0 =A0 =A0 =A0 =A0 dec_bdi_stat(page->mapping->backing_dev_= info, BDI_RECLAIMABLE); >> =A0 =A0 =A0 =A0 =A0 =A0 =A0 return 1; >> @@ -1319,6 +1321,8 @@ nfs_commit_list(struct inode *inode, struct li= st_head *head, int how) >> =A0 =A0 =A0 =A0 =A0 =A0 =A0 req =3D nfs_list_entry(head->next); >> =A0 =A0 =A0 =A0 =A0 =A0 =A0 nfs_list_remove_request(req); >> =A0 =A0 =A0 =A0 =A0 =A0 =A0 nfs_mark_request_commit(req); >> + =A0 =A0 =A0 =A0 =A0 =A0 mem_cgroup_dec_page_stat(req->wb_page, >> + =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0= =A0 =A0MEMCG_NR_FILE_UNSTABLE_NFS); >> =A0 =A0 =A0 =A0 =A0 =A0 =A0 dec_zone_page_state(req->wb_page, NR_UNS= TABLE_NFS); >> =A0 =A0 =A0 =A0 =A0 =A0 =A0 dec_bdi_stat(req->wb_page->mapping->back= ing_dev_info, >> =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 BDI_RECL= AIMABLE); >> diff --git a/mm/filemap.c b/mm/filemap.c >> index a6cfecf..7e751fe 100644 >> --- a/mm/filemap.c >> +++ b/mm/filemap.c >> @@ -143,6 +143,7 @@ void __delete_from_page_cache(struct page *page) >> =A0 =A0 =A0 =A0* having removed the page entirely. >> =A0 =A0 =A0 =A0*/ >> =A0 =A0 =A0 if (PageDirty(page) && mapping_cap_account_dirty(mapping= )) { >> + =A0 =A0 =A0 =A0 =A0 =A0 mem_cgroup_dec_page_stat(page, MEMCG_NR_FI= LE_DIRTY); >> =A0 =A0 =A0 =A0 =A0 =A0 =A0 dec_zone_page_state(page, NR_FILE_DIRTY)= ; >> =A0 =A0 =A0 =A0 =A0 =A0 =A0 dec_bdi_stat(mapping->backing_dev_info, = BDI_RECLAIMABLE); >> =A0 =A0 =A0 } >> diff --git a/mm/page-writeback.c b/mm/page-writeback.c >> index 632b464..d8005b0 100644 >> --- a/mm/page-writeback.c >> +++ b/mm/page-writeback.c >> @@ -1118,6 +1118,7 @@ int __set_page_dirty_no_writeback(struct page = *page) >> =A0void account_page_dirtied(struct page *page, struct address_space= *mapping) >> =A0{ >> =A0 =A0 =A0 if (mapping_cap_account_dirty(mapping)) { >> + =A0 =A0 =A0 =A0 =A0 =A0 mem_cgroup_inc_page_stat(page, MEMCG_NR_FI= LE_DIRTY); >> =A0 =A0 =A0 =A0 =A0 =A0 =A0 __inc_zone_page_state(page, NR_FILE_DIRT= Y); >> =A0 =A0 =A0 =A0 =A0 =A0 =A0 __inc_zone_page_state(page, NR_DIRTIED); >> =A0 =A0 =A0 =A0 =A0 =A0 =A0 __inc_bdi_stat(mapping->backing_dev_info= , BDI_RECLAIMABLE); >> @@ -1317,6 +1318,7 @@ int clear_page_dirty_for_io(struct page *page) >> =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0* for more comments. >> =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0*/ >> =A0 =A0 =A0 =A0 =A0 =A0 =A0 if (TestClearPageDirty(page)) { >> + =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 mem_cgroup_dec_page_stat(p= age, MEMCG_NR_FILE_DIRTY); >> =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 dec_zone_page_state(page= , NR_FILE_DIRTY); >> =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 dec_bdi_stat(mapping->ba= cking_dev_info, >> =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 = =A0 =A0 BDI_RECLAIMABLE); >> @@ -1352,8 +1354,10 @@ int test_clear_page_writeback(struct page *pa= ge) >> =A0 =A0 =A0 } else { >> =A0 =A0 =A0 =A0 =A0 =A0 =A0 ret =3D TestClearPageWriteback(page); >> =A0 =A0 =A0 } >> - =A0 =A0 if (ret) >> + =A0 =A0 if (ret) { >> + =A0 =A0 =A0 =A0 =A0 =A0 mem_cgroup_dec_page_stat(page, MEMCG_NR_FI= LE_WRITEBACK); >> =A0 =A0 =A0 =A0 =A0 =A0 =A0 dec_zone_page_state(page, NR_WRITEBACK); >> + =A0 =A0 } >> =A0 =A0 =A0 return ret; >> =A0} >> >> @@ -1386,8 +1390,10 @@ int test_set_page_writeback(struct page *page= ) >> =A0 =A0 =A0 } else { >> =A0 =A0 =A0 =A0 =A0 =A0 =A0 ret =3D TestSetPageWriteback(page); >> =A0 =A0 =A0 } >> - =A0 =A0 if (!ret) >> + =A0 =A0 if (!ret) { >> + =A0 =A0 =A0 =A0 =A0 =A0 mem_cgroup_inc_page_stat(page, MEMCG_NR_FI= LE_WRITEBACK); >> =A0 =A0 =A0 =A0 =A0 =A0 =A0 account_page_writeback(page); >> + =A0 =A0 } >> =A0 =A0 =A0 return ret; >> >> =A0} > > At least in mainline, NR_WRITEBACK handling codes are following as. > > 1) increase > > =A0* account_page_writeback > > 2) decrease > > =A0* test_clear_page_writeback > =A0* __nilfs_end_page_io > > I think account_page_writeback name is good to add your account funct= ion into that. > The problem is decreasement. Normall we can handle decreasement in te= st_clear_page_writeback. > But I am not sure it's okay in __nilfs_end_page_io. > I think if __nilfs_end_page_io is right, __nilfs_end_page_io should c= all > mem_cgroup_dec_page_stat(page, MEMCG_NR_FILE_WRITEBACK). > > What do you think about it? > > > > -- > Kind regards, > Minchan Kim > I would like to not have any special cases that avoid certain memory. So I think your suggestion is good. However, nilfs memcg dirty page accounting was skipped in a previous memcg dirty limit effort due to complexity. See 'clone_page' reference in: http://lkml.indiana.edu/hypermail/linux/kernel/1003.0/02997.html I admit that I don't follow all of the nilfs code path, but it looks like some of the nilfs pages are allocated but not charged to memcg. There is code in mem_cgroup_update_page_stat() to gracefully handle pages not associated with a memcg. So perhaps nilfs clone pages dirty [un]charge could be attempted. I have not succeeded in testing in exercising these code paths in nilfs. -- To unsubscribe from this list: send the line "unsubscribe linux-fsdevel= " in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html