Date:   Wed, 15 Feb 2023 11:45:03 +0100
From:   Peter Zijlstra <peterz@infradead.org>
To:     Boqun Feng <boqun.feng@gmail.com>
Cc:     Alan Stern <stern@rowland.harvard.edu>,
        Kent Overstreet <kent.overstreet@linux.dev>,
        Kent Overstreet <kent.overstreet@gmail.com>,
        Linus Torvalds <torvalds@linux-foundation.org>,
        Coly Li <colyli@suse.de>,
        Tetsuo Handa <penguin-kernel@i-love.sakura.ne.jp>,
        syzkaller <syzkaller@googlegroups.com>,
        Dmitry Vyukov <dvyukov@google.com>,
        Greg Kroah-Hartman <gregkh@linuxfoundation.org>,
        "Rafael J. Wysocki" <rafael@kernel.org>,
        Ingo Molnar <mingo@redhat.com>,
        LKML <linux-kernel@vger.kernel.org>,
        USB list <linux-usb@vger.kernel.org>,
        Hillf Danton <hdanton@sina.com>
Subject: Re: [PATCH RFC] drivers/core: Replace lockdep_set_novalidate_class()
 with unique class keys
Message-ID: <Y+y3r8Q5GT+oJsvd@hirez.programming.kicks-ass.net>
References: <Y+hTEtCKPuO0zGIt@moria.home.lan>
 <Y+hW74TAVzCpSv7c@rowland.harvard.edu>
 <Y+hYn6uzIUBaxDdV@moria.home.lan>
 <Y+kEgDLSRwdODRdD@rowland.harvard.edu>
 <Y+oBveWO2z6xdTW/@hirez.programming.kicks-ass.net>
 <Y+pWhyFJeE93nlWd@rowland.harvard.edu>
 <Y+plfZnEqw6mG+XH@hirez.programming.kicks-ass.net>
 <Y+rpD7QPheQQ8Lxj@boqun-archlinux>
 <Y+tm59SmBEY1Ywq7@hirez.programming.kicks-ass.net>
 <Y+u1RBMMcCkvKISZ@boqun-archlinux>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <Y+u1RBMMcCkvKISZ@boqun-archlinux>
Precedence: bulk

On Tue, Feb 14, 2023 at 08:22:28AM -0800, Boqun Feng wrote:

> Ah, right, I was missing the fact that it works with 2 classes...
> 
> But I think with only one class, the nest_lock() still works, right?
> In other words, if P and Cn are the same lock class in your example.

I don't think so, but I don't think I've carefully considered that case.

> Also seems I gave a wrong answer to Alan, just to clarify, the following
> is not a deadlock to lockdep:
> 
> T1:
> 	mutex_lock(P)
> 	mutex_lock_next_lock(C1, P)
> 	mutex_lock_next_lock(C2, P)
> 	mutex_lock(B)
> 
> T2:
> 	mutex_lock(P)
> 	mutex_lock(B)
> 	mutex_lock_next_lock(C1, P)
> 	mutex_lock_next_lock(C2, P)
> 

This should in fact complain about a CB-BC deadlock, (but I've not
tested it, just going on memories of how I implemented it).

> Because of any pair of
> 
> 	mutex_lock(L);
> 	... // other locks maybe
> 	mutex_lock_nest_lock(M, L);
> 
> lockdep will not add M into the dependency graph, since it's nested and
> should be serialized by L.

We do enter M into the dependency graph, but instead ignore M-M
recursion. Specifically so that we might catch the above deadlock vs B.