From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: X-Spam-Checker-Version: SpamAssassin 3.4.0 (2014-02-07) on aws-us-west-2-korg-lkml-1.web.codeaurora.org X-Spam-Level: X-Spam-Status: No, score=-5.4 required=3.0 tests=DKIM_SIGNED,DKIM_VALID, DKIM_VALID_AU,HEADER_FROM_DIFFERENT_DOMAINS,MAILING_LIST_MULTI,SIGNED_OFF_BY, SPF_HELO_NONE,SPF_PASS,USER_AGENT_SANE_1 autolearn=ham autolearn_force=no version=3.4.0 Received: from mail.kernel.org (mail.kernel.org [198.145.29.99]) by smtp.lore.kernel.org (Postfix) with ESMTP id 26D9FC35669 for ; Sat, 22 Feb 2020 03:08:51 +0000 (UTC) Received: from vger.kernel.org (vger.kernel.org [209.132.180.67]) by mail.kernel.org (Postfix) with ESMTP id E5D062071E for ; Sat, 22 Feb 2020 03:08:50 +0000 (UTC) Authentication-Results: mail.kernel.org; dkim=pass (1024-bit key) header.d=joelfernandes.org header.i=@joelfernandes.org header.b="b8sXekKg" Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1727096AbgBVDIt (ORCPT ); Fri, 21 Feb 2020 22:08:49 -0500 Received: from mail-qt1-f195.google.com ([209.85.160.195]:38938 "EHLO mail-qt1-f195.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1726032AbgBVDIt (ORCPT ); Fri, 21 Feb 2020 22:08:49 -0500 Received: by mail-qt1-f195.google.com with SMTP id p34so2795984qtb.6 for ; Fri, 21 Feb 2020 19:08:47 -0800 (PST) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=joelfernandes.org; s=google; h=date:from:to:cc:subject:message-id:references:mime-version :content-disposition:in-reply-to:user-agent; bh=CP6atjRwhdDbnoorKkqcf5kaCRKftzd9xcwwbnPKUYc=; b=b8sXekKgEQH6+jxePbWcLlA5uslQdfxzNedSgSBwsWEcqggyq7jiipBUxz4n7cl9KC hye7uA7MM7Hgq92KAt5L90XEnBO5zTQWW2oUrcCagbPu2Zr7yM+n6xGEZIC0l/IJ90sy mKurXKiNnfZDtJ5onWavJwQs1jcVzU51pUB9Q= X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20161025; h=x-gm-message-state:date:from:to:cc:subject:message-id:references :mime-version:content-disposition:in-reply-to:user-agent; bh=CP6atjRwhdDbnoorKkqcf5kaCRKftzd9xcwwbnPKUYc=; b=EMdeqexo7BvHh2ktf3XRwEjHDIEaikeQxTaIgOPuJf33fSxku1jSop7zob+GkIphP8 sawe1J9N3wnW00Q7ICkgB/arWFfURu7YVcFXDC1kkk5E2E/hU0Qpr0nUN3oh+FIGKUE+ jxl/qT8/eLcU3PziDEhSpBRF21YhKS27QcWs3QIJzkPdKH+6FACy9PN0oZQ4/DA8qY1O TCEzIcDeNY6evQ//OTda/1lUrzmSG0+b3lx/85KkeWNZALgWvQYaNMfRy4yp/9SZrI44 UQzNfvyhHq2zBv687DytpwWbLETN0RrIVz81CLaxDx5VLQ3cPrxLVa7ispOrQXItTYcG AJLA== X-Gm-Message-State: APjAAAW+BCRtuahPNeYdTYE7UqphnCVhuXT4tbekGuUQ4TtGOb98DPsI ZRZAgYMQMOj1TUnpb+ALgMyhQw== X-Google-Smtp-Source: APXvYqxsnqWMcaf7bIfKSBiNID3M78w4lFmVk2MV1D93TRWZ63pRswgtXzpuBhYRx4HIWbI9QLbTLQ== X-Received: by 2002:ac8:6f73:: with SMTP id u19mr34598485qtv.326.1582340926493; Fri, 21 Feb 2020 19:08:46 -0800 (PST) Received: from localhost ([2620:15c:6:12:9c46:e0da:efbf:69cc]) by smtp.gmail.com with ESMTPSA id w18sm304561qki.40.2020.02.21.19.08.43 (version=TLS1_3 cipher=TLS_AES_256_GCM_SHA384 bits=256/256); Fri, 21 Feb 2020 19:08:45 -0800 (PST) Date: Fri, 21 Feb 2020 22:08:43 -0500 From: Joel Fernandes To: Peter Zijlstra Cc: linux-kernel@vger.kernel.org, linux-arch@vger.kernel.org, rostedt@goodmis.org, mingo@kernel.org, gregkh@linuxfoundation.org, gustavo@embeddedor.com, tglx@linutronix.de, paulmck@kernel.org, josh@joshtriplett.org, mathieu.desnoyers@efficios.com, jiangshanlai@gmail.com, luto@kernel.org, tony.luck@intel.com, frederic@kernel.org, dan.carpenter@oracle.com, mhiramat@kernel.org Subject: Re: [PATCH v4 01/27] lockdep: Teach lockdep about "USED" <- "IN-NMI" inversions Message-ID: <20200222030843.GA191380@google.com> References: <20200221133416.777099322@infradead.org> <20200221134215.090538203@infradead.org> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <20200221134215.090538203@infradead.org> User-Agent: Mutt/1.10.1 (2018-07-13) Sender: linux-kernel-owner@vger.kernel.org Precedence: bulk List-ID: X-Mailing-List: linux-kernel@vger.kernel.org On Fri, Feb 21, 2020 at 02:34:17PM +0100, Peter Zijlstra wrote: > nmi_enter() does lockdep_off() and hence lockdep ignores everything. > > And NMI context makes it impossible to do full IN-NMI tracking like we > do IN-HARDIRQ, that could result in graph_lock recursion. The patch makes sense to me. Reviewed-by: Joel Fernandes (Google) NOTE: Also, I was wondering if we can detect the graph_lock recursion case and avoid doing anything bad, that way we enable more of the lockdep functionality for NMI where possible. Not sure if the suggestion makes sense though! thanks, - Joel > However, since look_up_lock_class() is lockless, we can find the class > of a lock that has prior use and detect IN-NMI after USED, just not > USED after IN-NMI. > > NOTE: By shifting the lockdep_off() recursion count to bit-16, we can > easily differentiate between actual recursion and off. > > Signed-off-by: Peter Zijlstra (Intel) > --- > kernel/locking/lockdep.c | 53 ++++++++++++++++++++++++++++++++++++++++++++--- > 1 file changed, 50 insertions(+), 3 deletions(-) > > --- a/kernel/locking/lockdep.c > +++ b/kernel/locking/lockdep.c > @@ -379,13 +379,13 @@ void lockdep_init_task(struct task_struc > > void lockdep_off(void) > { > - current->lockdep_recursion++; > + current->lockdep_recursion += BIT(16); > } > EXPORT_SYMBOL(lockdep_off); > > void lockdep_on(void) > { > - current->lockdep_recursion--; > + current->lockdep_recursion -= BIT(16); > } > EXPORT_SYMBOL(lockdep_on); > > @@ -575,6 +575,7 @@ static const char *usage_str[] = > #include "lockdep_states.h" > #undef LOCKDEP_STATE > [LOCK_USED] = "INITIAL USE", > + [LOCK_USAGE_STATES] = "IN-NMI", > }; > #endif > > @@ -787,6 +788,7 @@ static int count_matching_names(struct l > return count + 1; > } > > +/* used from NMI context -- must be lockless */ > static inline struct lock_class * > look_up_lock_class(const struct lockdep_map *lock, unsigned int subclass) > { > @@ -4463,6 +4465,34 @@ void lock_downgrade(struct lockdep_map * > } > EXPORT_SYMBOL_GPL(lock_downgrade); > > +/* NMI context !!! */ > +static void verify_lock_unused(struct lockdep_map *lock, struct held_lock *hlock, int subclass) > +{ > + struct lock_class *class = look_up_lock_class(lock, subclass); > + > + /* if it doesn't have a class (yet), it certainly hasn't been used yet */ > + if (!class) > + return; > + > + if (!(class->usage_mask & LOCK_USED)) > + return; > + > + hlock->class_idx = class - lock_classes; > + > + print_usage_bug(current, hlock, LOCK_USED, LOCK_USAGE_STATES); > +} > + > +static bool lockdep_nmi(void) > +{ > + if (current->lockdep_recursion & 0xFFFF) > + return false; > + > + if (!in_nmi()) > + return false; > + > + return true; > +} > + > /* > * We are not always called with irqs disabled - do that here, > * and also avoid lockdep recursion: > @@ -4473,8 +4503,25 @@ void lock_acquire(struct lockdep_map *lo > { > unsigned long flags; > > - if (unlikely(current->lockdep_recursion)) > + if (unlikely(current->lockdep_recursion)) { > + /* XXX allow trylock from NMI ?!? */ > + if (lockdep_nmi() && !trylock) { > + struct held_lock hlock; > + > + hlock.acquire_ip = ip; > + hlock.instance = lock; > + hlock.nest_lock = nest_lock; > + hlock.irq_context = 2; // XXX > + hlock.trylock = trylock; > + hlock.read = read; > + hlock.check = check; > + hlock.hardirqs_off = true; > + hlock.references = 0; > + > + verify_lock_unused(lock, &hlock, subclass); > + } > return; > + } > > raw_local_irq_save(flags); > check_flags(flags); > >