[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <734244.1580725424@warthog.procyon.org.uk>
Date: Mon, 03 Feb 2020 10:23:44 +0000
From: David Howells <dhowells@...hat.com>
To: Jakub Kicinski <kuba@...nel.org>
Cc: dhowells@...hat.com, netdev@...r.kernel.org,
linux-afs@...ts.infradead.org, linux-kernel@...r.kernel.org
Subject: Re: [PATCH net 3/4] rxrpc: Fix missing active use pinning of rxrpc_local object
Jakub Kicinski <kuba@...nel.org> wrote:
> > + BUG_ON(!conn->params.local);
> > + BUG_ON(!conn->params.local->socket);
>
> Is this really, really not possible to convert those into a WARN_ON()
> and return?
I can take those out - I actually put them in to help figure out which pointer
had become NULL - but turning them into a WARN_ON() and return doesn't
actually help that much since, without this patch, there was nothing to stop
the relevant pointer getting cleared between this point and the next access,
so there's a chance you'd end up with the same oops anyway.
David
Powered by blists - more mailing lists