On Fri, Oct 30, 2020 at 6:42 AM Jakub Kicinski <k...@kernel.org> wrote:
>
> On Thu, 29 Oct 2020 02:12:21 +0800 DENG Qingfang wrote:
> > MT7530/7531 has a global RX packet length register, which can be used
> > to set MTU.
> >
> > Signed-off-by: DENG Qingfang <dqf...@gmail.com>
>
> Please wrap your code at 80 chars.
>
> > diff --git a/drivers/net/dsa/mt7530.c b/drivers/net/dsa/mt7530.c
> > index de7692b763d8..7764c66a47c9 100644
> > --- a/drivers/net/dsa/mt7530.c
> > +++ b/drivers/net/dsa/mt7530.c
> > @@ -1021,6 +1021,40 @@ mt7530_port_disable(struct dsa_switch *ds, int port)
> > mutex_unlock(&priv->reg_mutex);
> > }
> >
> > +static int
> > +mt7530_port_change_mtu(struct dsa_switch *ds, int port, int new_mtu)
> > +{
> > + struct mt7530_priv *priv = ds->priv;
> > + int length;
> > +
> > + /* When a new MTU is set, DSA always set the CPU port's MTU to the
> > largest MTU
> > + * of the slave ports. Because the switch only has a global RX length
> > register,
> > + * only allowing CPU port here is enough.
> > + */
> > + if (!dsa_is_cpu_port(ds, port))
> > + return 0;
> > +
> > + /* RX length also includes Ethernet header, MTK tag, and FCS length */
> > + length = new_mtu + ETH_HLEN + MTK_HDR_LEN + ETH_FCS_LEN;
> > + if (length <= 1522)
> > + mt7530_rmw(priv, MT7530_GMACCR, MAX_RX_PKT_LEN_MASK,
> > MAX_RX_PKT_LEN_1522);
> > + else if (length <= 1536)
> > + mt7530_rmw(priv, MT7530_GMACCR, MAX_RX_PKT_LEN_MASK,
> > MAX_RX_PKT_LEN_1536);
> > + else if (length <= 1552)
> > + mt7530_rmw(priv, MT7530_GMACCR, MAX_RX_PKT_LEN_MASK,
> > MAX_RX_PKT_LEN_1552);
> > + else
> > + mt7530_rmw(priv, MT7530_GMACCR, MAX_RX_JUMBO_MASK |
> > MAX_RX_PKT_LEN_MASK,
> > + MAX_RX_JUMBO(DIV_ROUND_UP(length, 1024)) |
> > MAX_RX_PKT_LEN_JUMBO);
>
> this line should start under priv, so it aligns to the opening
> parenthesis.
>
> Besides, don't you need to reset the JUMBO bit when going from jumbo to
> non-jumbo? The mask should always include jumbo.
MAX_RX_JUMBO works only when MAX_RX_PKT_LEN is set to 0x3, so just
changing MAX_RX_PKT_LEN to non-jumbo is enough.
FYI, the default value of MAX_RX_JUMBO is 0x9.
>
> I assume you're duplicating the mt7530_rmw() for the benefit of the
> constant validation, but it seems to be counterproductive here.
As I mentioned above, MAX_RX_JUMBO does not need to be changed when
going from jumbo to non-jumbo.
Perhaps I should use mt7530_mii_read() and mt7530_mii_write() instead?
>
> > + return 0;
> > +}
>
