From: Pan Bian <bianpan201602@163.com>
To: Lino Sanfilippo <LinoSanfilippo@gmx.de>,
        David Dillow <dave@thedillows.org>, netdev@vger.kernel.org
Cc: PanBian <bianpan2016@163.com>, linux-kernel@vger.kernel.org
Subject: Re: [PATCH 1/1] net: ethernet: 3com: set error code on failures
Date: Sat,  3 Dec 2016 23:23:12 +0800
Message-Id: <20161203141512.GA7143@bp>
Reply-To: PanBian <bianpan2016@163.com>
References: <1480771470-6404-1-git-send-email-bianpan201602@163.com> <06dd8fc8-0c39-e79d-7e34-22bfab17d8a3@gmx.de>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <06dd8fc8-0c39-e79d-7e34-22bfab17d8a3@gmx.de>
User-Agent: Mutt/1.5.21 (2010-09-15)
Sender: linux-kernel-owner@vger.kernel.org
Content-Length: 1432
Lines: 50

From: PanBian <bianpan2016@163.com>

Hi, Lino,

There is no special reason to map different errors to -EIO. In the
original source code, err is set to -EIO on most error paths. I copied 
that. I guess it has no difference with the error value provided by the 
called functions in the view of callers of typhoon_init_one().

Thanks!

Best regards,
Pan

On Sat, Dec 03, 2016 at 02:53:07PM +0100, Lino Sanfilippo wrote:
> Hi,
> 
> On 03.12.2016 14:24, Pan Bian wrote:
> > From: Pan Bian <bianpan2016@163.com>
> > 
> > In function typhoon_init_one(), returns the value of variable err on
> > errors. However, on some error paths, variable err is not set to a
> > negative errno. This patch assigns "-EIO" to err on those paths.
> > 
> > Signed-off-by: Pan Bian <bianpan2016@163.com>
> 
> >  
> > @@ -2409,6 +2410,7 @@ enum state_values {
> >  	INIT_COMMAND_WITH_RESPONSE(&xp_cmd, TYPHOON_CMD_READ_VERSIONS);
> >  	if(typhoon_issue_command(tp, 1, &xp_cmd, 3, xp_resp) < 0) {
> >  		err_msg = "Could not get Sleep Image version";
> > +		err = -EIO;
> >  		goto error_out_reset;
> >  	}
> >  
> > @@ -2451,6 +2453,7 @@ enum state_values {
> >  
> >  	if(register_netdev(dev) < 0) {
> >  		err_msg = "unable to register netdev";
> > +		err = -EIO;
> >  		goto error_out_reset;
> >  	}
> >  
> > 
> 
> Why not return the error value provided by the called functions? Is there a reason
> to map different errors to -EIO?
> 
> Regards,
> Lino