RE: [PATCH net-next 05/18] ravb: Replace chip type with a structure for driver data

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 05/18] ravb: Replace chip type with a structure for driver data
Date: Fri, 23 Jul 2021 06:08:17 +0000	[thread overview]
Message-ID: <TYCPR01MB5933A2CA3DEE8620F5AFF16786E59@TYCPR01MB5933.jpnprd01.prod.outlook.com> (raw)
In-Reply-To: <09e06a00-6b9e-c8b8-e5c3-67d82d900a74@gmail.com>

Hi Sergei,

Thanks for the feedback.

> Biju Das <biju.das@bp.renesas.com>; Prabhakar Mahadev Lad
> <prabhakar.mahadev-lad.rj@bp.renesas.com>
> Subject: Re: [PATCH net-next 05/18] ravb: Replace chip type with a
> structure for driver data
> 
> Hello!
> 
> On 7/22/21 5:13 PM, Biju Das wrote:
> 
> > The DMAC and EMAC blocks of Gigabit Ethernet IP is almost similar to
> > Ethernet AVB. With few changes in driver we can support both the IP.
> >
> > This patch is in preparation for supporting the same by replacing chip
> > type by a structure with values, feature bits and function pointers.
> >
> > Currently only values is added to structure and later patches will add
> > features and function pointers.
> >
> > 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.h      | 14 +++++
> >  drivers/net/ethernet/renesas/ravb_main.c | 76
> > +++++++++++++++++-------
> >  2 files changed, 67 insertions(+), 23 deletions(-)
> >
> > diff --git a/drivers/net/ethernet/renesas/ravb.h
> > b/drivers/net/ethernet/renesas/ravb.h
> > index 80e62ca2e3d3..0ed21262f26b 100644
> > --- a/drivers/net/ethernet/renesas/ravb.h
> > +++ b/drivers/net/ethernet/renesas/ravb.h
> > @@ -988,6 +988,18 @@ enum ravb_chip_id {
> >  	RCAR_GEN3,
> >  };
> >
> > +struct ravb_drv_data {
> 
>    I'd rather suggest *struct* ravb_hw_info... This is hardly a driver
> data, more like hwrdware's one. :-)

OK.

> 
> > +	netdev_features_t net_features;
> > +	netdev_features_t net_hw_features;
> > +	const char (*gstrings_stats)[ETH_GSTRING_LEN];
> > +	size_t gstrings_size;
> > +	size_t stats_len;
> > +	u32 num_gstat_queue;
> > +	size_t skb_sz;
> > +	u8 num_tx_desc;
> > +	enum ravb_chip_id chip_id;
> 
>    Mhm, I'd expect that chip_id is no longer needed with the feature
> structs...

Yes,If you see the subsequent 3 patches, chip_id usage is completely removed from
ravb_main.c. but it is still required for ravb_ptp.c. Please let me know do you want
me take out from there as well. Then as part of [1], I can take out chipid completely.

[1]
https://patchwork.kernel.org/project/linux-renesas-soc/patch/20210722141351.13668-7-biju.das.jz@bp.renesas.com/

> 
> [...]
> > @@ -1040,6 +1052,8 @@ struct ravb_private {
> >  	unsigned txcidm:1;		/* TX Clock Internal Delay Mode */
> >  	unsigned rgmii_override:1;	/* Deprecated rgmii-*id behavior */
> >  	int num_tx_desc;		/* TX descriptors per packet */
> > +
> > +	const struct ravb_drv_data *info;
> 
>    So data or info? :-)

As we are going to use ravb_hw_info, it will be info.

> 
> [...]
> > diff --git a/drivers/net/ethernet/renesas/ravb_main.c
> > b/drivers/net/ethernet/renesas/ravb_main.c
> > index 805397088850..84ebd6fef711 100644
> > --- a/drivers/net/ethernet/renesas/ravb_main.c
> > +++ b/drivers/net/ethernet/renesas/ravb_main.c
> [...]
> > @@ -1176,9 +1179,12 @@ static void ravb_get_ethtool_stats(struct
> > net_device *ndev,
> >
> >  static void ravb_get_strings(struct net_device *ndev, u32 stringset,
> > u8 *data)  {
> > +	struct ravb_private *priv = netdev_priv(ndev);
> > +	const struct ravb_drv_data *info = priv->info;
> > +
> >  	switch (stringset) {
> >  	case ETH_SS_STATS:
> > -		memcpy(data, ravb_gstrings_stats,
> sizeof(ravb_gstrings_stats));
> > +		memcpy(data, info->gstrings_stats, info->gstrings_size);
> >  		break;
> >  	}
> >  }
> > @@ -1924,12 +1930,36 @@ static int ravb_mdio_release(struct ravb_private
> *priv)
> >  	return 0;
> >  }
> >
> > +static const struct ravb_drv_data ravb_gen3_data = {
> > +	.net_features = NETIF_F_RXCSUM,
> > +	.net_hw_features = NETIF_F_RXCSUM,
> > +	.gstrings_stats = ravb_gstrings_stats,
> > +	.gstrings_size = sizeof(ravb_gstrings_stats),
> > +	.stats_len = ARRAY_SIZE(ravb_gstrings_stats),
> > +	.num_gstat_queue = NUM_RX_QUEUE,
> > +	.skb_sz = RX_BUF_SZ + RAVB_ALIGN - 1,
> > +	.num_tx_desc = NUM_TX_DESC_GEN3,
> > +	.chip_id = RCAR_GEN3,
> > +};
> > +
> > +static const struct ravb_drv_data ravb_gen2_data = {
> > +	.net_features = NETIF_F_RXCSUM,
> > +	.net_hw_features = NETIF_F_RXCSUM,
> 
>    Mhm, why have the fields that don't change from SoC to SoC anyway?
> I do think they should be added when a new SoC support is added...

This is a preparation patch for supporting RZ/G2L and there is a difference w.r.to RZ/G2L
See [2]. That is the reason to isolate the new SoC changes w.r.to existing one it is added
Here. Andrew also suggested to make smaller changes.

[2]
https://patchwork.kernel.org/project/linux-renesas-soc/patch/20210722141351.13668-18-biju.das.jz@bp.renesas.com/

> 
> > +	.gstrings_stats = ravb_gstrings_stats,
> > +	.gstrings_size = sizeof(ravb_gstrings_stats),
> > +	.stats_len = ARRAY_SIZE(ravb_gstrings_stats),
> 
>    Same question...

Same as above.

> > +	.num_gstat_queue = NUM_RX_QUEUE,
> > +	.skb_sz = RX_BUF_SZ + RAVB_ALIGN - 1,
> 
>    Again why?

Same as above.

> 
> > +	.num_tx_desc = NUM_TX_DESC_GEN2,
> > +	.chip_id = RCAR_GEN2,
> > +};
> > +
> [...]
> > @@ -2052,15 +2082,15 @@ static int ravb_probe(struct platform_device
> *pdev)
> >  	if (!ndev)
> >  		return -ENOMEM;
> >
> > -	ndev->features = NETIF_F_RXCSUM;
> > -	ndev->hw_features = NETIF_F_RXCSUM;
> > +	info = of_device_get_match_data(&pdev->dev);
> > +
> > +	ndev->features = info->net_features;
> > +	ndev->hw_features = info->net_hw_features;
> >
> >  	pm_runtime_enable(&pdev->dev);
> >  	pm_runtime_get_sync(&pdev->dev);
> >
> > -	chip_id = (enum ravb_chip_id)of_device_get_match_data(&pdev->dev);
> > -
> > -	if (chip_id == RCAR_GEN3)
> > +	if (info->chip_id == RCAR_GEN3)
> 
>    Ugh...
This chip id is replaced in subsequent patch [3]

[3]
https://patchwork.kernel.org/project/linux-renesas-soc/patch/20210722141351.13668-8-biju.das.jz@bp.renesas.com/

> 
> >  		irq = platform_get_irq_byname(pdev, "ch22");
> >  	else
> >  		irq = platform_get_irq(pdev, 0);
> [...]
> > @@ -2099,7 +2130,7 @@ static int ravb_probe(struct platform_device
> *pdev)
> >  	priv->avb_link_active_low =
> >  		of_property_read_bool(np, "renesas,ether-link-active-low");
> >
> > -	if (chip_id == RCAR_GEN3) {
> > +	if (info->chip_id == RCAR_GEN3) {
> 
>    Ugh...
Same as above.

> 
> >  		irq = platform_get_irq_byname(pdev, "ch24");
> >  		if (irq < 0) {
> >  			error = irq;
> [...]
> > @@ -2184,7 +2214,7 @@ static int ravb_probe(struct platform_device
> *pdev)
> >  	INIT_LIST_HEAD(&priv->ts_skb_list);
> >
> >  	/* Initialise PTP Clock driver */
> > -	if (chip_id != RCAR_GEN2)
> > +	if (info->chip_id != RCAR_GEN2)
> >  		ravb_ptp_init(ndev, pdev);
> 
>    Ugh...
Same as above.
> 
> >
> >  	/* Debug message level */
> > @@ -2232,7 +2262,7 @@ static int ravb_probe(struct platform_device
> *pdev)
> >  			  priv->desc_bat_dma);
> >
> >  	/* Stop PTP Clock driver */
> > -	if (chip_id != RCAR_GEN2)
> > +	if (info->chip_id != RCAR_GEN2)
> >  		ravb_ptp_stop(ndev);
> 
>   Ugh...

Same as above.

Regards,
Biju

> 
> 
> >  out_disable_refclk:
> >  	clk_disable_unprepare(priv->refclk);
> 
> MBR, Sergei