[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <2545100.1699264877@warthog.procyon.org.uk>
Date: Mon, 06 Nov 2023 10:01:17 +0000
From: David Howells <dhowells@...hat.com>
To: Damian Tometzki <damian@...cv-rocks.de>
Cc: dhowells@...hat.com, Eric Biggers <ebiggers@...nel.org>,
Steve French <smfrench@...il.com>,
John Sanpe <sanpeqf@...il.com>,
CIFS <linux-cifs@...r.kernel.org>,
Namjae Jeon <linkinjeon@...nel.org>,
Paulo Alcantara <pc@...guebit.com>,
LKML <linux-kernel@...r.kernel.org>
Subject: Re: smb cifs: Linux 6.7 pre rc-1 kernel dump in smb2_get_aead_req
Damian Tometzki <damian@...cv-rocks.de> wrote:
> the revert of f1b4cb650b9a0eeba206d8f069fcdc532bfbcd74 solved the issue of
> the kernel dump.
That almost certainly did not fix the problem - merely hid the wanring.
Prior to f1b4cb650b9a0eeba206d8f069fcdc532bfbcd74, ->user_backed is explicitly
set if the iov_iter is initialised to a user-backed type, now it's just
inferred from the type being 0 or 1 - so I think that the iov_iter has not
been initialised somewhere.
Somewhere being from SMB2_tcon() and cifs_send_recv() on down.
David
Powered by blists - more mailing lists