From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from psmtp.com (na3sys010amx108.postini.com [74.125.245.108]) by kanga.kvack.org (Postfix) with SMTP id AC8566B13F0 for ; Fri, 3 Feb 2012 15:03:39 -0500 (EST) Received: by qauh8 with SMTP id h8so2844621qau.14 for ; Fri, 03 Feb 2012 12:03:38 -0800 (PST) MIME-Version: 1.0 In-Reply-To: <20120203113822.19cf6fd2.kamezawa.hiroyu@jp.fujitsu.com> References: <1328233033-14246-1-git-send-email-yinghan@google.com> <20120203113822.19cf6fd2.kamezawa.hiroyu@jp.fujitsu.com> Date: Fri, 3 Feb 2012 12:03:38 -0800 Message-ID: Subject: Re: [PATCH] memcg: fix up documentation on global LRU. From: Ying Han Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: quoted-printable Sender: owner-linux-mm@kvack.org List-ID: To: KAMEZAWA Hiroyuki Cc: Michal Hocko , Balbir Singh , Rik van Riel , Hugh Dickins , Johannes Weiner , Mel Gorman , Pavel Emelyanov , linux-mm@kvack.org On Thu, Feb 2, 2012 at 6:38 PM, KAMEZAWA Hiroyuki wrote: > On Thu, =A02 Feb 2012 17:37:13 -0800 > Ying Han wrote: > >> In v3.3-rc1, the global LRU has been removed with commit >> "mm: make per-memcg LRU lists exclusive". The patch fixes up the memcg d= ocs. >> >> Signed-off-by: Ying Han >> --- >> =A0Documentation/cgroups/memory.txt | =A0 25 ++++++++++++------------- >> =A01 files changed, 12 insertions(+), 13 deletions(-) >> >> diff --git a/Documentation/cgroups/memory.txt b/Documentation/cgroups/me= mory.txt >> index 4c95c00..847a2a4 100644 >> --- a/Documentation/cgroups/memory.txt >> +++ b/Documentation/cgroups/memory.txt >> @@ -34,8 +34,7 @@ Current Status: linux-2.6.34-mmotm(development version= of 2010/April) >> >> =A0Features: >> =A0 - accounting anonymous pages, file caches, swap caches usage and lim= iting them. >> - - private LRU and reclaim routine. (system's global LRU and private LR= U >> - =A0 work independently from each other) >> + - pages are linked to per-memcg LRU exclusively, and there is no globa= l LRU. >> =A0 - optionally, memory+swap usage can be accounted and limited. >> =A0 - hierarchical accounting >> =A0 - soft limit >> @@ -154,7 +153,7 @@ updated. page_cgroup has its own LRU on cgroup. >> =A02.2.1 Accounting details >> >> =A0All mapped anon pages (RSS) and cache pages (Page Cache) are accounte= d. >> -Some pages which are never reclaimable and will not be on the global LR= U >> +Some pages which are never reclaimable and will not be on the LRU >> =A0are not accounted. We just account pages under usual VM management. >> >> =A0RSS pages are accounted at page_fault unless they've already been acc= ounted >> @@ -209,19 +208,19 @@ In this case, setting memsw.limit_in_bytes=3D3G wi= ll prevent bad use of swap. >> =A0By using memsw limit, you can avoid system OOM which can be caused by= swap >> =A0shortage. >> >> -* why 'memory+swap' rather than swap. >> -The global LRU(kswapd) can swap out arbitrary pages. Swap-out means >> -to move account from memory to swap...there is no change in usage of >> -memory+swap. In other words, when we want to limit the usage of swap wi= thout >> -affecting global LRU, memory+swap limit is better than just limiting sw= ap from >> -OS point of view. >> - >> =A0* What happens when a cgroup hits memory.memsw.limit_in_bytes >> =A0When a cgroup hits memory.memsw.limit_in_bytes, it's useless to do sw= ap-out >> =A0in this cgroup. Then, swap-out will not be done by cgroup routine and= file >> -caches are dropped. But as mentioned above, global LRU can do swapout m= emory >> -from it for sanity of the system's memory management state. You can't f= orbid >> -it by cgroup. >> +caches are dropped. >> + >> +TODO: >> +* use 'memory+swap' rather than swap was due to existence of global LRU= . It can >> +swap out arbitrary pages. Swap-out means to move account from memory to= swap... >> +there is no change in usage of memory+swap. In other words, when we wan= t to >> +limit the usage of swap without affecting global LRU, memory+swap limit= is >> +better than just limiting swap from OS point of view. However, the glob= al LRU >> +has been removed now and all pages are linked in private LRU. We might = want to >> +revisit this in the future. >> > > Could you devide this memory+swap discussion to otehr patch ? yes, will do that. > > Do you want to do memory locking by setting swap_limit=3D0 ? hmm, not sure what do you mean here? --Ying > > Thanks, > -Kame > > > > > > -- To unsubscribe, send a message with 'unsubscribe linux-mm' in the body to majordomo@kvack.org. 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: email@kvack.org