From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from mail-io0-f197.google.com (mail-io0-f197.google.com [209.85.223.197]) by kanga.kvack.org (Postfix) with ESMTP id 24A3B6B0038 for ; Mon, 6 Nov 2017 08:36:06 -0500 (EST) Received: by mail-io0-f197.google.com with SMTP id h70so21871224ioi.5 for ; Mon, 06 Nov 2017 05:36:06 -0800 (PST) Received: from mail-sor-f41.google.com (mail-sor-f41.google.com. [209.85.220.41]) by mx.google.com with SMTPS id f78sor4882595ita.133.2017.11.06.05.36.05 for (Google Transport Security); Mon, 06 Nov 2017 05:36:05 -0800 (PST) MIME-Version: 1.0 In-Reply-To: <20171106133304.GS21978@ZenIV.linux.org.uk> References: <94eb2c05f6a018dc21055d39c05b@google.com> <20171106032941.GR21978@ZenIV.linux.org.uk> <20171106131544.GB4359@quack2.suse.cz> <20171106133304.GS21978@ZenIV.linux.org.uk> From: Dmitry Vyukov Date: Mon, 6 Nov 2017 14:35:44 +0100 Message-ID: Subject: Re: possible deadlock in generic_file_write_iter Content-Type: text/plain; charset="UTF-8" Sender: owner-linux-mm@kvack.org List-ID: To: Al Viro Cc: Jan Kara , syzbot , Andrew Morton , Johannes Weiner , jlayton@redhat.com, LKML , linux-mm@kvack.org, npiggin@gmail.com, rgoldwyn@suse.com, ross.zwisler@linux.intel.com, syzkaller-bugs@googlegroups.com, Ingo Molnar , Peter Zijlstra On Mon, Nov 6, 2017 at 2:33 PM, Al Viro wrote: > On Mon, Nov 06, 2017 at 02:15:44PM +0100, Jan Kara wrote: > >> > Should we annotate these inodes with different lock types? Or use >> > nesting annotations? >> >> Well, you'd need to have a completely separate set of locking classes for >> each filesystem to avoid false positives like these. And that would >> increase number of classes lockdep has to handle significantly. So I'm not >> sure it's really worth it... > > Especially when you consider that backing file might be on a filesystem > that lives on another loop device. *All* per-{device,fs} locks involved > would need classes split that way... This crashes our test machines left and right. We've seen 100000+ of these crashes. We need to do at least something. Can we disable all checking of these mutexes if they inherently have positives? +Ingo, Peter, maybe you have some suggestions of how to fight this lockdep false positives. Full thread is here: https://groups.google.com/forum/#!msg/syzkaller-bugs/NJ_4llH84XI/c7M9jNLTAgAJ -- 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