From mboxrd@z Thu Jan 1 00:00:00 1970 From: Henrique de Moraes Holschuh Subject: Re: [PATCH] md: don't use flush_signals in userspace processes Date: Fri, 9 Jun 2017 00:27:28 -0300 Message-ID: <20170609032728.GA25899@khazad-dum.debian.net> References: <87h8zrart4.fsf@notabene.neil.brown.name> <87shja9b7s.fsf@notabene.neil.brown.name> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Return-path: Content-Disposition: inline In-Reply-To: <87shja9b7s.fsf@notabene.neil.brown.name> Sender: linux-kernel-owner@vger.kernel.org To: NeilBrown Cc: Mikulas Patocka , Shaohua Li , linux-raid@vger.kernel.org, linux-kernel@vger.kernel.org, Ingo Molnar , Peter Zijlstra List-Id: linux-raid.ids On Fri, 09 Jun 2017, NeilBrown wrote: > Or maybe it could be discarded - the md_check_recovery() thing. > The idea was that if you alt-sysrq-K to kill all processes, md arrays > would go into immediate-safe-mode where the metadata is marked clean > immediately after writes finish, rather than waiting a few seconds. The > chance of having a clean array after shutdown is hopefully improved. > > I've never actually used this though, and I doubt many people know about > it. And bitmaps make it fairly pointless. Hmm, I have, although I had no idea this was why my arrays were getting far less frazzled than expected... It is really useful behavior, now that I know it can do that. If you can teach SysRq+S, and especially SysRq+U, to force all arrays into safe-mode *after* they carried their current meanings (sync/umount), that would be more useful though, and it would help a lot of people to avoid dirty arrays without them even knowing why... -- Henrique Holschuh