From: Andrew Morton <akpm@linux-foundation.org> To: balbir@in.ibm.com Cc: linux-kernel@vger.kernel.org, vatsa@in.ibm.com, ckrm-tech@lists.sourceforge.net, xemul@sw.ru, linux-mm@kvack.org, menage@google.com, svaidy@linux.vnet.ibm.com, devel@openvz.org Subject: Re: [RFC][PATCH][3/4] Add reclaim support Date: Mon, 19 Feb 2007 03:10:17 -0800 [thread overview] Message-ID: <20070219031017.c6e180e9.akpm@linux-foundation.org> (raw) In-Reply-To: <45D9810D.3040508@in.ibm.com> On Mon, 19 Feb 2007 16:20:53 +0530 Balbir Singh <balbir@in.ibm.com> wrote: > >> + * so, is the container over it's limit. Returns 1 if the container is above > >> + * its limit. > >> + */ > >> +int memctlr_mm_overlimit(struct mm_struct *mm, void *sc_cont) > >> +{ > >> + struct container *cont; > >> + struct memctlr *mem; > >> + long usage, limit; > >> + int ret = 1; > >> + > >> + if (!sc_cont) > >> + goto out; > >> + > >> + read_lock(&mm->container_lock); > >> + cont = mm->container; > >> + > >> + /* > >> + * Regular reclaim, let it proceed as usual > >> + */ > >> + if (!sc_cont) > >> + goto out; > >> + > >> + ret = 0; > >> + if (cont != sc_cont) > >> + goto out; > >> + > >> + mem = memctlr_from_cont(cont); > >> + usage = atomic_long_read(&mem->counter.usage); > >> + limit = atomic_long_read(&mem->counter.limit); > >> + if (limit && (usage > limit)) > >> + ret = 1; > >> +out: > >> + read_unlock(&mm->container_lock); > >> + return ret; > >> +} > > > > hm, I wonder how much additional lock traffic all this adds. > > > > It's a read_lock() and most of the locks are read_locks > which allow for concurrent access, until the container > changes or goes away read_lock isn't free, and I suspect we're calling this function pretty often (every pagefault?) It'll be measurable on some workloads, on some hardware. It probably won't be terribly bad because each lock-taking is associated with a clear_page(). But still, if there's any possibility of lightening the locking up, now is the time to think about it. > >> @@ -66,6 +67,9 @@ struct scan_control { > >> int swappiness; > >> > >> int all_unreclaimable; > >> + > >> + void *container; /* Used by containers for reclaiming */ > >> + /* pages when the limit is exceeded */ > >> }; > > > > eww. Why void*? > > > > I did not want to expose struct container in mm/vmscan.c. It's already there, via rmap.h > An additional > thought was that no matter what container goes in the field would be > useful for reclaim. Am having trouble parsing that sentence ;)
WARNING: multiple messages have this Message-ID (diff)
From: Andrew Morton <akpm@linux-foundation.org> To: balbir@in.ibm.com Cc: linux-kernel@vger.kernel.org, vatsa@in.ibm.com, ckrm-tech@lists.sourceforge.net, xemul@sw.ru, linux-mm@kvack.org, menage@google.com, svaidy@linux.vnet.ibm.com, devel@openvz.org Subject: Re: [RFC][PATCH][3/4] Add reclaim support Date: Mon, 19 Feb 2007 03:10:17 -0800 [thread overview] Message-ID: <20070219031017.c6e180e9.akpm@linux-foundation.org> (raw) In-Reply-To: <45D9810D.3040508@in.ibm.com> On Mon, 19 Feb 2007 16:20:53 +0530 Balbir Singh <balbir@in.ibm.com> wrote: > >> + * so, is the container over it's limit. Returns 1 if the container is above > >> + * its limit. > >> + */ > >> +int memctlr_mm_overlimit(struct mm_struct *mm, void *sc_cont) > >> +{ > >> + struct container *cont; > >> + struct memctlr *mem; > >> + long usage, limit; > >> + int ret = 1; > >> + > >> + if (!sc_cont) > >> + goto out; > >> + > >> + read_lock(&mm->container_lock); > >> + cont = mm->container; > >> + > >> + /* > >> + * Regular reclaim, let it proceed as usual > >> + */ > >> + if (!sc_cont) > >> + goto out; > >> + > >> + ret = 0; > >> + if (cont != sc_cont) > >> + goto out; > >> + > >> + mem = memctlr_from_cont(cont); > >> + usage = atomic_long_read(&mem->counter.usage); > >> + limit = atomic_long_read(&mem->counter.limit); > >> + if (limit && (usage > limit)) > >> + ret = 1; > >> +out: > >> + read_unlock(&mm->container_lock); > >> + return ret; > >> +} > > > > hm, I wonder how much additional lock traffic all this adds. > > > > It's a read_lock() and most of the locks are read_locks > which allow for concurrent access, until the container > changes or goes away read_lock isn't free, and I suspect we're calling this function pretty often (every pagefault?) It'll be measurable on some workloads, on some hardware. It probably won't be terribly bad because each lock-taking is associated with a clear_page(). But still, if there's any possibility of lightening the locking up, now is the time to think about it. > >> @@ -66,6 +67,9 @@ struct scan_control { > >> int swappiness; > >> > >> int all_unreclaimable; > >> + > >> + void *container; /* Used by containers for reclaiming */ > >> + /* pages when the limit is exceeded */ > >> }; > > > > eww. Why void*? > > > > I did not want to expose struct container in mm/vmscan.c. It's already there, via rmap.h > An additional > thought was that no matter what container goes in the field would be > useful for reclaim. Am having trouble parsing that sentence ;) -- 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/ . Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>
next prev parent reply other threads:[~2007-02-19 11:11 UTC|newest] Thread overview: 76+ messages / expand[flat|nested] mbox.gz Atom feed top 2007-02-19 6:50 [RFC][PATCH][0/4] Memory controller (RSS Control) Balbir Singh 2007-02-19 6:50 ` Balbir Singh 2007-02-19 6:50 ` [RFC][PATCH][1/4] RSS controller setup Balbir Singh 2007-02-19 6:50 ` Balbir Singh 2007-02-19 8:57 ` Andrew Morton 2007-02-19 8:57 ` Andrew Morton 2007-02-19 9:18 ` Paul Menage 2007-02-19 9:18 ` Paul Menage 2007-02-19 11:13 ` Balbir Singh 2007-02-19 11:13 ` Balbir Singh 2007-02-19 19:43 ` Matthew Helsley 2007-02-19 19:43 ` Matthew Helsley 2007-02-19 10:06 ` Balbir Singh 2007-02-19 10:06 ` Balbir Singh 2007-02-19 6:50 ` [RFC][PATCH][2/4] Add RSS accounting and control Balbir Singh 2007-02-19 6:50 ` Balbir Singh 2007-02-19 8:58 ` Andrew Morton 2007-02-19 8:58 ` Andrew Morton 2007-02-19 10:37 ` [ckrm-tech] " Balbir Singh 2007-02-19 10:37 ` Balbir Singh 2007-02-19 11:01 ` Andrew Morton 2007-02-19 11:01 ` Andrew Morton 2007-02-19 11:09 ` Balbir Singh 2007-02-19 11:09 ` Balbir Singh 2007-02-19 11:23 ` Andrew Morton 2007-02-19 11:23 ` Andrew Morton 2007-02-19 11:56 ` Balbir Singh 2007-02-19 11:56 ` Balbir Singh 2007-02-19 12:09 ` Paul Menage 2007-02-19 12:09 ` Paul Menage 2007-02-19 14:10 ` Balbir Singh 2007-02-19 14:10 ` Balbir Singh 2007-02-19 16:07 ` Vaidyanathan Srinivasan 2007-02-19 16:07 ` Vaidyanathan Srinivasan 2007-02-19 16:17 ` Balbir Singh 2007-02-19 16:17 ` Balbir Singh 2007-02-20 6:40 ` Vaidyanathan Srinivasan 2007-02-20 6:40 ` Vaidyanathan Srinivasan 2007-02-19 6:50 ` [RFC][PATCH][3/4] Add reclaim support Balbir Singh 2007-02-19 6:50 ` Balbir Singh 2007-02-19 8:59 ` Andrew Morton 2007-02-19 8:59 ` Andrew Morton 2007-02-19 10:50 ` Balbir Singh 2007-02-19 10:50 ` Balbir Singh 2007-02-19 11:10 ` Andrew Morton [this message] 2007-02-19 11:10 ` Andrew Morton 2007-02-19 11:16 ` Balbir Singh 2007-02-19 11:16 ` Balbir Singh 2007-02-19 9:48 ` KAMEZAWA Hiroyuki 2007-02-19 9:48 ` KAMEZAWA Hiroyuki 2007-02-19 10:52 ` Balbir Singh 2007-02-19 10:52 ` Balbir Singh 2007-02-19 6:50 ` [RFC][PATCH][4/4] RSS controller documentation Balbir Singh 2007-02-19 6:50 ` Balbir Singh 2007-02-19 8:54 ` [RFC][PATCH][0/4] Memory controller (RSS Control) Andrew Morton 2007-02-19 8:54 ` Andrew Morton 2007-02-19 9:06 ` Paul Menage 2007-02-19 9:06 ` Paul Menage 2007-02-19 9:50 ` [ckrm-tech] " Kirill Korotaev 2007-02-19 9:50 ` Kirill Korotaev 2007-02-19 9:50 ` Paul Menage 2007-02-19 9:50 ` Paul Menage 2007-02-19 10:24 ` Balbir Singh 2007-02-19 10:24 ` Balbir Singh 2007-02-19 10:39 ` Balbir Singh 2007-02-19 10:39 ` Balbir Singh 2007-02-19 9:16 ` Magnus Damm 2007-02-19 9:16 ` Magnus Damm 2007-02-19 10:45 ` Balbir Singh 2007-02-19 10:45 ` Balbir Singh 2007-02-19 11:56 ` Magnus Damm 2007-02-19 11:56 ` Magnus Damm 2007-02-19 14:07 ` Balbir Singh 2007-02-19 14:07 ` Balbir Singh 2007-02-19 10:00 ` Balbir Singh 2007-02-19 10:00 ` Balbir Singh
Reply instructions: You may reply publicly to this message via plain-text email using any one of the following methods: * Save the following mbox file, import it into your mail client, and reply-to-all from there: mbox Avoid top-posting and favor interleaved quoting: https://en.wikipedia.org/wiki/Posting_style#Interleaved_style * Reply using the --to, --cc, and --in-reply-to switches of git-send-email(1): git send-email \ --in-reply-to=20070219031017.c6e180e9.akpm@linux-foundation.org \ --to=akpm@linux-foundation.org \ --cc=balbir@in.ibm.com \ --cc=ckrm-tech@lists.sourceforge.net \ --cc=devel@openvz.org \ --cc=linux-kernel@vger.kernel.org \ --cc=linux-mm@kvack.org \ --cc=menage@google.com \ --cc=svaidy@linux.vnet.ibm.com \ --cc=vatsa@in.ibm.com \ --cc=xemul@sw.ru \ /path/to/YOUR_REPLY https://kernel.org/pub/software/scm/git/docs/git-send-email.html * If your mail client supports setting the In-Reply-To header via mailto: links, try the mailto: linkBe sure your reply has a Subject: header at the top and a blank line before the message body.
This is an external index of several public inboxes, see mirroring instructions on how to clone and mirror all data and code used by this external index.