From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1757515AbcEEPYL (ORCPT ); Thu, 5 May 2016 11:24:11 -0400 Received: from mx1.redhat.com ([209.132.183.28]:33244 "EHLO mx1.redhat.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1756718AbcEEPYJ (ORCPT ); Thu, 5 May 2016 11:24:09 -0400 Date: Thu, 5 May 2016 17:24:06 +0200 From: Andrea Arcangeli To: "Kirill A. Shutemov" Cc: Alex Williamson , kirill.shutemov@linux.intel.com, linux-kernel@vger.kernel.org, "linux-mm@kvack.org" Subject: Re: [BUG] vfio device assignment regression with THP ref counting redesign Message-ID: <20160505152406.GH28755@redhat.com> References: <20160428204542.5f2053f7@ul30vt.home> <20160429070611.GA4990@node.shutemov.name> <20160429163444.GM11700@redhat.com> <20160502104119.GA23305@node.shutemov.name> <20160502152307.GA12310@redhat.com> <20160502160042.GC24419@node.shutemov.name> <20160502180307.GB12310@redhat.com> <20160504191927.095cdd90@t450s.home> <20160505143924.GC28755@redhat.com> <20160505151110.GA13972@node.shutemov.name> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <20160505151110.GA13972@node.shutemov.name> User-Agent: Mutt/1.6.1 (2016-04-27) Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org On Thu, May 05, 2016 at 06:11:10PM +0300, Kirill A. Shutemov wrote: > Hm. How total_mapcount equal to NULL wouldn't lead to NULL-pointer > dereference inside page_trans_huge_mapcount()? Sorry for the confusion, this was still work in progress and then I've seen the email from Alex and I sent the last version I had committed right away. An earlier version of course had the proper checks for NULL but they got wiped as I transitioned from one model to another and back. > > + page_move_anon_rmap(old_page, vma, address); > > compound_head() is missing, I believe. Oh yes, fixed that too. if (total_mapcount == 1) { /* * The page is all ours. Move it to * our anon_vma so the rmap code will * not search our parent or siblings. * Protected against the rmap code by * the page lock. */ page_move_anon_rmap(compound_head(old_page), vma, address); } If there's no other issue I can git send-email. Then we should look into calling page_move_anon_rmap from THP COWs too, hugetlbfs calls it too. I think we probably need to make page_move_anon_rmap smarter and optionally let it take the lock for us after reading page->mapping first to be sure it's really moving it. The question is then if trylock or lock_page should be used, my preference would be just trylock.