From mboxrd@z Thu Jan 1 00:00:00 1970 Date: Mon, 1 Aug 2005 21:35:58 -0700 (PDT) From: Linus Torvalds Subject: Re: [patch 2.6.13-rc4] fix get_user_pages bug In-Reply-To: <1122956713.6338.19.camel@npiggin-nld.site> Message-ID: References: <20050801032258.A465C180EC0@magilla.sf.frob.com> <42EDDB82.1040900@yahoo.com.au> <42EECC1F.9000902@yahoo.com.au> <1122956713.6338.19.camel@npiggin-nld.site> MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-linux-mm@kvack.org Return-Path: To: Nick Piggin Cc: Robin Holt , Andrew Morton , Roland McGrath , Hugh Dickins , linux-mm@kvack.org, lkml List-ID: On Tue, 2 Aug 2005, Nick Piggin wrote: > > Any chance you can change the __follow_page test to account for > writeable clean ptes? Something like > > if (write && !pte_dirty(pte) && !pte_write(pte)) > goto out; > > And then you would re-add the set_page_dirty logic further on. Hmm.. That should be possible. I wanted to do the simplest possible code sequence, but yeah, I guess there's nothing wrong with allowing the code to dirty the page. Somebody want to send me a proper patch? Also, I haven't actually heard from whoever actually noticed the problem in the first place (Robin?) whether the fix does fix it. It "obviously does", but testing is always good ;) Linus -- 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