RE: [PATCH 4/6] net: stmmac: dwmac-s32cc: add basic NXP S32G/S32R glue

[Jan Petrous (OSS)]

> -----Original Message-----
> From: Simon Horman <horms@xxxxxxxxxx>
> Sent: Monday, 5 August, 2024 19:00
> To: Jan Petrous (OSS) <jan.petrous@xxxxxxxxxxx>
> Cc: Maxime Coquelin <mcoquelin.stm32@xxxxxxxxx>; Alexandre Torgue
> <alexandre.torgue@xxxxxxxxxxx>; dl-S32 <S32@xxxxxxx>; linux-
> kernel@xxxxxxxxxxxxxxx; linux-stm32@xxxxxxxxxxxxxxxxxxxxxxxxxxxx; linux-
> arm-kernel@xxxxxxxxxxxxxxxxxxx; Claudiu Manoil <claudiu.manoil@xxxxxxx>;
> netdev@xxxxxxxxxxxxxxx
> Subject: Re: [PATCH 4/6] net: stmmac: dwmac-s32cc: add basic NXP
> S32G/S32R glue
> 
> On Sun, Aug 04, 2024 at 08:50:10PM +0000, Jan Petrous (OSS) wrote:
> > NXP S32G2xx/S32G3xx and S32R45 are automotive grade SoCs
> > that integrate one or two Synopsys DWMAC 5.10/5.20 IPs.
> >
> > The basic driver supports only RGMII interface.
> >
> > Signed-off-by: Jan Petrous (OSS) <jan.petrous@xxxxxxxxxxx>
> 
> ...
> 
> > diff --git a/drivers/net/ethernet/stmicro/stmmac/dwmac-s32cc.c
> b/drivers/net/ethernet/stmicro/stmmac/dwmac-s32cc.c
> 
> ...
> 
> > +static int s32cc_gmac_init(struct platform_device *pdev, void *priv)
> > +{
> > +	struct s32cc_priv_data *gmac = priv;
> > +	int ret;
> > +
> > +	ret = clk_set_rate(gmac->tx_clk, GMAC_TX_RATE_125M);
> > +	if (!ret)
> > +		ret = clk_prepare_enable(gmac->tx_clk);
> > +
> > +	if (ret) {
> > +		dev_err(&pdev->dev, "Can't set tx clock\n");
> > +		return ret;
> > +	}
> > +
> > +	ret = clk_prepare_enable(gmac->rx_clk);
> > +	if (ret)
> > +		dev_dbg(&pdev->dev, "Can't set rx, clock source is
> disabled.\n");
> > +	else
> > +		gmac->rx_clk_enabled = true;
> > +
> > +	ret = s32cc_gmac_write_phy_intf_select(gmac);
> > +	if (ret) {
> > +		dev_err(&pdev->dev, "Can't set PHY interface mode\n");
> 
> Should operations on tx_clk and rx_clk be unwound here?
> 
> Flagged by Smatch.

Yes, it shall be there. Added to v2.

> 
> > +		return ret;
> > +	}
> > +
> > +	return 0;
> > +}
> 
> ...
> 
> > +static int s32cc_dwmac_probe(struct platform_device *pdev)
> > +{
> > +	struct device *dev = &pdev->dev;
> > +	struct plat_stmmacenet_data *plat;
> > +	struct s32cc_priv_data *gmac;
> > +	struct stmmac_resources res;
> > +	int ret;
> 
> Please consider arranging local variables in Networking code
> in reverse xmas tree order - longest line to shortest.
> 
> Flagged by: https://github.com/ecree-solarflare/xmastree
> 

Done for v2.

> > +
> > +	gmac = devm_kzalloc(&pdev->dev, sizeof(*gmac), GFP_KERNEL);
> > +	if (!gmac)
> > +		return PTR_ERR(gmac);
> 
> This will return 0, perhaps return -ENOMEM ?
> 
> Flagged by Smatch.

Changed to -ENOMEM.

Thanks for review.
/Jan