From mboxrd@z Thu Jan  1 00:00:00 1970
From: Christoph Hellwig <hch@lst.de>
Subject: Re: [PATCH 3/4] sg: check for valid direction before starting the
        request
Date: Fri, 3 Feb 2017 11:28:13 +0100
Message-ID: <20170203102813.GC29163@lst.de>
References: <1486112091-68470-1-git-send-email-hare@suse.de> <1486112091-68470-4-git-send-email-hare@suse.de>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Return-path: <linux-scsi-owner@vger.kernel.org>
Received: from verein.lst.de ([213.95.11.211]:47243 "EHLO newverein.lst.de"
        rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP
        id S1753028AbdBCK2P (ORCPT <rfc822;linux-scsi@vger.kernel.org>);
        Fri, 3 Feb 2017 05:28:15 -0500
Content-Disposition: inline
In-Reply-To: <1486112091-68470-4-git-send-email-hare@suse.de>
Sender: linux-scsi-owner@vger.kernel.org
List-Id: linux-scsi@vger.kernel.org
To: Hannes Reinecke <hare@suse.de>
Cc: "Martin K. Petersen" <martin.petersen@oracle.com>, Christoph Hellwig <hch@lst.de>, James Bottomley <james.bottomley@hansenpartnership.com>, Johannes Thumshirn <jth@kernel.org>, Doug Gilberg <dgilbert@interlog.com>, linux-scsi@vger.kernel.org, Johannes Thumshirn <jthumshirn@suse.de>

On Fri, Feb 03, 2017 at 09:54:50AM +0100, Hannes Reinecke wrote:
> From: Johannes Thumshirn <jthumshirn@suse.de>
> 
> Check for a valid direction before starting the request, otherwise we risk
> running into an assertion in the scsi midlayer checking for vaild requests.

Good idea, but..

> +static bool sg_is_valid_direction(int dxfer_direction)
> +{
> +	switch (dxfer_direction) {
> +	case SG_DXFER_NONE:
> +	case SG_DXFER_TO_DEV:
> +	case SG_DXFER_FROM_DEV:
> +	case SG_DXFER_TO_FROM_DEV:

This isn't strictly valid as sg doesn't actually handle real bidi
commands, but we work around it.

It might be a good idea to move the warning about it from sg_write
to here and use printk_ratelimited instead of the hand-rolled
per-thread warning there.

> +	case SG_DXFER_UNKNOWN:
> +		return true;

And how valid is this one (Question to Doug I guess): for a 0-sized
transfer we should be fine, but otherwise it should probably be
rejected.