From mboxrd@z Thu Jan  1 00:00:00 1970
From: Al Viro <viro@ZenIV.linux.org.uk>
Subject: Re: scsi: use-after-free in bio_copy_from_iter
Date: Mon, 5 Dec 2016 19:03:39 +0000
Message-ID: <20161205190339.GK1555@ZenIV.linux.org.uk>
References: <CACT4Y+aBd+1pssnsVx2x145n7Q1nDMU=xc7NAiOCtpy-MA-EDg@mail.gmail.com>
 <CACT4Y+auCpJbXhmxbr4OmSHbiVkatyWmDxGGShs=WvO=HCXRKA@mail.gmail.com>
 <20161203103802.dmvqzzquthwa7kd7@linux-x5ow.site>
 <CACT4Y+bB+oxjPoC0WVCUV_Bq9ePqdV3m_V4snsfUrZ2BSWub2Q@mail.gmail.com>
 <20161203181948.GA3322@linux-x5ow.site>
 <CACT4Y+Z4V=9nm04H-uyy0tA4yXLp7Ld7dqgjLiA5_opeQC_m6A@mail.gmail.com>
 <20161205151753.GI8481@linux-x5ow.site>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Return-path: <linux-kernel-owner@vger.kernel.org>
Content-Disposition: inline
In-Reply-To: <20161205151753.GI8481@linux-x5ow.site>
Sender: linux-kernel-owner@vger.kernel.org
To: Johannes Thumshirn <jthumshirn@suse.de>
Cc: Dmitry Vyukov <dvyukov@google.com>, syzkaller <syzkaller@googlegroups.com>, Doug Gilbert <dgilbert@interlog.com>, jejb@linux.vnet.ibm.com, "Martin K. Petersen" <martin.petersen@oracle.com>, linux-scsi <linux-scsi@vger.kernel.org>, LKML <linux-kernel@vger.kernel.org>, axboe@kernel.dk, linux-block@vger.kernel.org, David Rientjes <rientjes@google.com>, Hannes Reinecke <hare@suse.de>, mhocko@suse.cz
List-Id: linux-scsi@vger.kernel.org

On Mon, Dec 05, 2016 at 04:17:53PM +0100, Johannes Thumshirn wrote:
> 633         hp = &srp->header;
> [...]
> 646                 hp->dxferp = (char __user *)buf + cmd_size;

> So the memory for hp->dxferp comes from:
> 633         hp = &srp->header;

????

> >From my debug instrumentation I see that the dxferp ends up in the
> iovec_iter's kvec->iov_base and the faulting address is always dxferp + n *
> 4k with n in [1, 16] (and we're copying 16 4k pages from the iovec into the
> bio).

_Address_ of hp->dxferp comes from that assignment; the value is 'buf'
argument of sg_write() + small offset.  In this case, it should point
inside a pipe buffer, which is, indeed, at a kernel address.  Who'd
allocated srp is irrelevant.

And if you end up dereferencing more than one page worth there, you do have
a problem - pipe buffers are not going to be that large.  Could you slap
	WARN_ON((size_t)input_size > count);
right after the calculation of input_size in sg_write() and see if it triggers
on your reproducer?