Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
Date:   Tue, 27 Sep 2022 13:39:13 +0200
From:   Nam Cao <namcaov@gmail.com>
To:     Dan Carpenter <dan.carpenter@oracle.com>
Cc:     forest@alittletooquiet.net, gregkh@linuxfoundation.org,
        philipp.g.hortmann@gmail.com, linux-kernel@vger.kernel.org,
        linux-staging@lists.linux.dev
Subject: Re: [RFC PATCH 3/5] staging: vt6655: split device_alloc_rx_buf
Message-ID: <20220927113913.GB10468@nam-dell>
References: <cover.1663273218.git.namcaov@gmail.com>
 <311eaa7dabe12d0baeb0af6f85c2b43b20b230a3.1663273218.git.namcaov@gmail.com>
 <Yyg4BTymed1fEzQx@kadam>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <Yyg4BTymed1fEzQx@kadam>
Precedence: bulk

On Mon, Sep 19, 2022 at 12:36:05PM +0300, Dan Carpenter wrote:
> On Thu, Sep 15, 2022 at 10:29:34PM +0200, Nam Cao wrote:
> > The function device_alloc_rx_buf does 2 things: allocating rx buffer
> > and initializing the rx descriptor's values. Split this function into
> > two, with each does one job.
> > 
> > This split is preparation for implementing correct out-of-memory error
> > handling.
> > 
> > Signed-off-by: Nam Cao <namcaov@gmail.com>
> > ---
> >  drivers/staging/vt6655/device_main.c | 16 +++++++++++++---
> >  1 file changed, 13 insertions(+), 3 deletions(-)
> > 
> > diff --git a/drivers/staging/vt6655/device_main.c b/drivers/staging/vt6655/device_main.c
> > index 79325a693857..27fe28156257 100644
> > --- a/drivers/staging/vt6655/device_main.c
> > +++ b/drivers/staging/vt6655/device_main.c
> > @@ -133,6 +133,7 @@ static int device_init_td1_ring(struct vnt_private *priv);
> >  static int  device_rx_srv(struct vnt_private *priv, unsigned int idx);
> >  static int  device_tx_srv(struct vnt_private *priv, unsigned int idx);
> >  static bool device_alloc_rx_buf(struct vnt_private *, struct vnt_rx_desc *);
> > +static void device_init_rx_desc(struct vnt_private *priv, struct vnt_rx_desc *rd);
> >  static void device_free_rx_buf(struct vnt_private *priv,
> >  			       struct vnt_rx_desc *rd);
> >  static void device_init_registers(struct vnt_private *priv);
> > @@ -615,6 +616,8 @@ static int device_init_rd0_ring(struct vnt_private *priv)
> >  			dev_err(&priv->pcid->dev, "can not alloc rx bufs\n");
> >  			ret = -ENOMEM;
> >  			goto err_free_rd;
> > +		} else {
> > +			device_init_rx_desc(priv, desc);
> >  		}
> 
> None of these else statements make sense.  It should be:
> 
> 		ret = -ENOMEM;
> 		goto err_free_rd;
> 	}
> 
> 	device_init_rx_desc(priv, desc);
> 	desc->next = &priv->aRD0Ring[(i + 1) % priv->opts.rx_descs0];

That does look better, will be changed.

> I haven't reviewed the patch totally.  I don't understand why it's doing
> this here instead of at the end.  But then I don't understand why it
> needs to be in a separate function at all.
> 
> This patch does not make sense.  The commit description says that this
> is a "preparation" patch.  Maybe fold it in with patch 5?  The rule is
> "one thing per patch" not "half a thing per patch".

I thought splitting it like this would make it easier to review. But if
these preparation patches are not welcomed, I will squash them and
resend.

Thank you for spending time reviewing the patches.

Best regards,
Nam