From: Michal Hocko Subject: Re: [patch 00/13] preempt: Make preempt count unconditional Date: Tue, 29 Sep 2020 10:19:38 +0200 Message-ID: <20200929081938.GC22035@dhcp22.suse.cz> References: <871rj4owfn.fsf@nanos.tec.linutronix.de> <87bli75t7v.fsf@nanos.tec.linutronix.de> <20200916152956.GV29330@paulmck-ThinkPad-P72> <20200916205840.GD29330@paulmck-ThinkPad-P72> Mime-Version: 1.0 Return-path: DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=suse.com; s=susede1; t=1601367582; h=from:from:reply-to:subject:subject:date:date:message-id:message-id: to:to:cc:cc:mime-version:mime-version:content-type:content-type: in-reply-to:in-reply-to:references:references; bh=KFyfyEQCS1dcCR4BhfL+zwiQ3CuzyAy4x23tr+VkzjE=; b=SSpMjosSL8o4OPVa55QL9wIlM8PHwJpdBy+0T8Zs2+GxBxG6YA9l00PJ9Y66qcHmMfGLv9 u+g+tcEvJ7RxAvAP57LRStfEelxtPsdHOGOgFzbQCNh9qQZDiKdnSarvXzcr1YnPqWmMjI wdLRjalKNBbwgyGfH1Q3bT+vgMo4axA= Content-Disposition: inline In-Reply-To: List-ID: Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit To: Daniel Vetter Cc: "Paul E. McKenney" , Juri Lelli , Peter Zijlstra , Sebastian Andrzej Siewior , Lai Jiangshan , dri-devel , Ben Segall , Linux-MM , "open list:KERNEL SELFTEST FRAMEWORK" , linux-hexagon@vger.kernel.org, Will Deacon , Ingo Molnar , Anton Ivanov , linux-arch , Vincent Guittot , Herbert Xu , Brian Cain , Richard Weinberger , Russell King On Wed 16-09-20 23:43:02, Daniel Vetter wrote: > I can > then figure out whether it's better to risk not spotting issues with > call_rcu vs slapping a memalloc_noio_save/restore around all these > critical section which force-degrades any allocation to GFP_ATOMIC at did you mean memalloc_noreclaim_* here? > most, but has the risk that we run into code that assumes "GFP_KERNEL > never fails for small stuff" and has a decidedly less tested fallback > path than rcu code. Even if the above then please note that memalloc_noreclaim_* or PF_MEMALLOC should be used with an extreme care. Essentially only for internal memory reclaimers. It grants access to _all_ the available memory so any abuse can be detrimental to the overall system operation. Allocation failure in this mode means that we are out of memory and any code relying on such an allocation has to carefuly consider failure. This is not a random allocation mode. -- Michal Hocko SUSE Labs