From mboxrd@z Thu Jan 1 00:00:00 1970 Message-ID: <445D6CFA.2080102@cyberone.com.au> Date: Sun, 07 May 2006 13:43:54 +1000 From: Nick Piggin MIME-Version: 1.0 Subject: Re: [RFC][PATCH] tracking dirty pages in shared mappings References: <1146861313.3561.13.camel@lappy> <445CA22B.8030807@cyberone.com.au> <1146922446.3561.20.camel@lappy> <445CA907.9060002@cyberone.com.au> <1146929357.3561.28.camel@lappy> <445D41F0.800@cyberone.com.au> In-Reply-To: <445D41F0.800@cyberone.com.au> Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit Sender: owner-linux-mm@kvack.org Return-Path: To: Peter Zijlstra Cc: Linus Torvalds , Andi Kleen , Rohit Seth , Andrew Morton , clameter@sgi.com, mbligh@google.com, hugh@veritas.com, riel@redhat.com, andrea@suse.de, arjan@infradead.org, apw@shadowen.org, mel@csn.ul.ie, marcelo@kvack.org, anton@samba.org, paulmck@us.ibm.com, linux-mm List-ID: Nick Piggin wrote: > Peter Zijlstra wrote: > >> On Sat, 2006-05-06 at 23:47 +1000, Nick Piggin wrote: >> >> >>> Yep. Let's not distract from getting the basic mechanism working though. >>> balance_dirty_pages would be patch 2..n ;) >>> >> >> Attached are both a new version of the shared_mapping_dirty patch, and >> balance_dirty_pages; to be applied in that order. >> It makes my testcase survive and not OOM like it used to. >> > > Looks OK. I wonder if test_clear_page_dirty could skip the page_wrprotect > entirely? It would speed up cases like truncate that don't care. OTOH, it > looks like several filesystems do not use clear_page_dirty_for_io where > they possibly should be... > > Perhaps you could consolidate both checks into test_set_page_writeback()? Actually, you'll skip the slow path of taking the lock most of the times in situations like truncate because it will bail out after the page_mapped check. So doing anything further is probably a premature optimisation at this stage. Nick -- 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: email@kvack.org