RE: [PATCH net-next 08/18] ravb: Add R-Car common features

From: Biju Das <biju.das.jz@bp.renesas.com>
To: Sergei Shtylyov <sergei.shtylyov@gmail.com>,
	"David S. Miller" <davem@davemloft.net>,
	Jakub Kicinski <kuba@kernel.org>
Cc: Geert Uytterhoeven <geert+renesas@glider.be>,
	Sergey Shtylyov <s.shtylyov@omprussia.ru>,
	Adam Ford <aford173@gmail.com>, Andrew Lunn <andrew@lunn.ch>,
	Yuusuke Ashizuka <ashiduka@fujitsu.com>,
	Yoshihiro Shimoda <yoshihiro.shimoda.uh@renesas.com>,
	"netdev@vger.kernel.org" <netdev@vger.kernel.org>,
	"linux-renesas-soc@vger.kernel.org" 
	<linux-renesas-soc@vger.kernel.org>,
	Chris Paterson <Chris.Paterson2@renesas.com>,
	Biju Das <biju.das@bp.renesas.com>,
	Prabhakar Mahadev Lad <prabhakar.mahadev-lad.rj@bp.renesas.com>
Subject: RE: [PATCH net-next 08/18] ravb: Add R-Car common features
Date: Wed, 28 Jul 2021 10:13:47 +0000	[thread overview]
Message-ID: <TYCPR01MB59337E1C3F14B0E15F0626D986EA9@TYCPR01MB5933.jpnprd01.prod.outlook.com> (raw)
In-Reply-To: <d493b1d2-6d05-9eb3-c5f5-f3828938fe56@gmail.com>

Hi Sergei,

Thanks for the feedback.

> Subject: Re: [PATCH net-next 08/18] ravb: Add R-Car common features
> 
> Hello!
> 
> On 7/22/21 5:13 PM, Biju Das wrote:
> 
> > The below features are supported by both R-Car Gen2 and Gen3.
> >
> > 1) magic packet detection
> > 2) multiple TSRQ support
> > 3) extended descriptor in rx
> 
>    I think this one should better be called timestamping...

OK. Will change it to timestamp.

> 
> > 4) No half duplex support
> 
>    Couldn't we avoid the "negative" features?

10 Mbps/Half duplex(As per the comment below) is not supported in R-Car. May be it should be a "positive" feature for
RZ/G2L. probably we need to add this as part of RZ/G2L feature.

> 
> > 5) override mtu change
> 
>    Hm, I'd vote for the individual patches covering only single feature...

OK. That will make patches again simpler.

> 
> > Add features bits to support the same.
> >
> > Signed-off-by: Biju Das <biju.das.jz@bp.renesas.com>
> > Reviewed-by: Lad Prabhakar <prabhakar.mahadev-lad.rj@bp.renesas.com>
> > ---
> >  drivers/net/ethernet/renesas/ravb_main.c | 110
> > +++++++++++++++--------
> >  1 file changed, 71 insertions(+), 39 deletions(-)
> >
> > diff --git a/drivers/net/ethernet/renesas/ravb_main.c
> > b/drivers/net/ethernet/renesas/ravb_main.c
> > index b3c99f974632..4ef2565534d2 100644
> > --- a/drivers/net/ethernet/renesas/ravb_main.c
> > +++ b/drivers/net/ethernet/renesas/ravb_main.c
> [...]
> > @@ -680,11 +694,14 @@ static void ravb_rcv_snd_enable(struct
> > net_device *ndev)
> >  /* function for waiting dma process finished */  static int
> > ravb_stop_dma(struct net_device *ndev)  {
> > +	struct ravb_private *priv = netdev_priv(ndev);
> > +	const struct ravb_drv_data *info = priv->info;
> >  	int error;
> >
> >  	/* Wait for stopping the hardware TX process */
> > -	error = ravb_wait(ndev, TCCR,
> > -			  TCCR_TSRQ0 | TCCR_TSRQ1 | TCCR_TSRQ2 | TCCR_TSRQ3, 0);
> > +	if (info->features & RAVB_MULTI_TSRQ)
> > +		error = ravb_wait(ndev, TCCR,
> > +				  TCCR_TSRQ0 | TCCR_TSRQ1 | TCCR_TSRQ2 |
> TCCR_TSRQ3, 0);
> >  	if (error)
> 
>    What if the above *if* skips the ravb_wait() call -- didn't you get a
> complaint from gcc about the unnintialized variable?

Good catch will initialize with error = 0 and later patch when we add RZ/G2L, will remove this unnecessary initialization.

> 
> [...]
> > @@ -808,11 +826,14 @@ static bool ravb_queue_interrupt(struct
> > net_device *ndev, int q)
> >
> >  static bool ravb_timestamp_interrupt(struct net_device *ndev)  {
> > +	struct ravb_private *priv = netdev_priv(ndev);
> > +	const struct ravb_drv_data *info = priv->info;
> >  	u32 tis = ravb_read(ndev, TIS);
> >
> >  	if (tis & TIS_TFUF) {
> >  		ravb_write(ndev, ~(TIS_TFUF | TIS_RESERVED), TIS);
> > -		ravb_get_tx_tstamp(ndev);
> > +		if (info->features & RAVB_EX_RX_DESC)
> 
>    Yeah, definitely a bad feature name...
OK. AS you suggested, will change the macro RAVB_EX_RX_DESC to ravb_timestamp feature bit like sh_eth.

> 
> > +			ravb_get_tx_tstamp(ndev);
> >  		return true;
> >  	}
> >  	return false;
> [...]
> > @@ -1069,15 +1091,17 @@ static int ravb_phy_init(struct net_device
> *ndev)
> >  		netdev_info(ndev, "limited PHY to 100Mbit/s\n");
> >  	}
> >
> > -	/* 10BASE, Pause and Asym Pause is not supported */
> > -	phy_remove_link_mode(phydev, ETHTOOL_LINK_MODE_10baseT_Half_BIT);
> > -	phy_remove_link_mode(phydev, ETHTOOL_LINK_MODE_10baseT_Full_BIT);
> > -	phy_remove_link_mode(phydev, ETHTOOL_LINK_MODE_Pause_BIT);
> > -	phy_remove_link_mode(phydev, ETHTOOL_LINK_MODE_Asym_Pause_BIT);
> > +	if (info->features & RAVB_NO_HALF_DUPLEX) {
> > +		/* 10BASE, Pause and Asym Pause is not supported */
> > +		phy_remove_link_mode(phydev,
> ETHTOOL_LINK_MODE_10baseT_Half_BIT);
> > +		phy_remove_link_mode(phydev,
> ETHTOOL_LINK_MODE_10baseT_Full_BIT);
> > +		phy_remove_link_mode(phydev, ETHTOOL_LINK_MODE_Pause_BIT);
> > +		phy_remove_link_mode(phydev,
> ETHTOOL_LINK_MODE_Asym_Pause_BIT);
> >
> > -	/* Half Duplex is not supported */
> > -	phy_remove_link_mode(phydev, ETHTOOL_LINK_MODE_1000baseT_Half_BIT);
> > -	phy_remove_link_mode(phydev, ETHTOOL_LINK_MODE_100baseT_Half_BIT);
> > +		/* Half Duplex is not supported */
> > +		phy_remove_link_mode(phydev,
> ETHTOOL_LINK_MODE_1000baseT_Half_BIT);
> > +		phy_remove_link_mode(phydev,
> ETHTOOL_LINK_MODE_100baseT_Half_BIT);
> 
>     Mhm? Some of the half-duplex modes sre unsupported still?
> 
> [...]
> > @@ -1314,8 +1338,9 @@ static void ravb_get_wol(struct net_device
> > *ndev, struct ethtool_wolinfo *wol)  static int ravb_set_wol(struct
> > net_device *ndev, struct ethtool_wolinfo *wol)  {
> >  	struct ravb_private *priv = netdev_priv(ndev);
> > +	const struct ravb_drv_data *info = priv->info;
> >
> > -	if (wol->wolopts & ~WAKE_MAGIC)
> > +	if ((wol->wolopts & ~WAKE_MAGIC) || (!(info->features &
> > +RAVB_MAGIC)))
> 
>    Parens about !x not needed. And I think the second operand should come
> first instead...

OK.

> 
> >  		return -EOPNOTSUPP;
> >
> >  	priv->wol_enabled = !!(wol->wolopts & WAKE_MAGIC);
> [...]
> > @@ -1595,28 +1621,30 @@ static netdev_tx_t ravb_start_xmit(struct
> sk_buff *skb, struct net_device *ndev)
> >  	desc->dptr = cpu_to_le32(dma_addr);
> >
> >  	/* TX timestamp required */
> > -	if (q == RAVB_NC) {
> > -		ts_skb = kmalloc(sizeof(*ts_skb), GFP_ATOMIC);
> > -		if (!ts_skb) {
> > -			if (num_tx_desc > 1) {
> > -				desc--;
> > -				dma_unmap_single(ndev->dev.parent, dma_addr,
> > -						 len, DMA_TO_DEVICE);
> > +	if (info->features & RAVB_EX_RX_DESC) {
> 
>    Definitely a bad name...

OK.

> 
> [...]
> > @@ -2205,8 +2235,10 @@ static int ravb_probe(struct platform_device
> *pdev)
> >  	}
> >  	clk_prepare_enable(priv->refclk);
> >
> > -	ndev->max_mtu = 2048 - (ETH_HLEN + VLAN_HLEN + ETH_FCS_LEN);
> > -	ndev->min_mtu = ETH_MIN_MTU;
> > +	if (info->features & RAVB_OVERRIDE_MTU_CHANGE) {
> 
>    Why? :-/ Could you tell me more details?

RX buff size = 2048 for R-Car where as it is 8K for RZ/G2L.
So Do you think we should use the same max_mtu for RZ/G2L as well?

Please let us know your recommendation here in this case?  The original bsp code used default "eth_mtu_change"
which is now deprecated.

Regards,
Biju

> 
> > +		ndev->max_mtu = 2048 - (ETH_HLEN + VLAN_HLEN + ETH_FCS_LEN);
> > +		ndev->min_mtu = ETH_MIN_MTU;
> > +	}
> >
> >  	priv->num_tx_desc = info->num_tx_desc;
> >
> 
> MBR, Sergei