[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20240424075624.GB60057@ragnatech.se>
Date: Wed, 24 Apr 2024 09:56:24 +0200
From: Niklas Söderlund <niklas.soderlund+renesas@...natech.se>
To: Geert Uytterhoeven <geert+renesas@...der.be>
Cc: Claudiu Beznea <claudiu.beznea.uj@...renesas.com>,
Sergey Shtylyov <s.shtylyov@....ru>,
Paul Barker <paul.barker.ct@...renesas.com>,
"David S . Miller" <davem@...emloft.net>,
Eric Dumazet <edumazet@...gle.com>,
Jakub Kicinski <kuba@...nel.org>, Paolo Abeni <pabeni@...hat.com>,
netdev@...r.kernel.org, linux-renesas-soc@...r.kernel.org
Subject: Re: [PATCH] net: ravb: Fix registered interrupt names
Hi Geert,
Thanks for fixing this! I noticed this as well but was dead-set needing
to finding a way to preserve ehtX names, and wrote down I needed to to a
large rework here to allow for this. Using the device name is more
clever!
On 2024-04-24 09:45:21 +0200, Geert Uytterhoeven wrote:
> As interrupts are now requested from ravb_probe(), before calling
> register_netdev(), ndev->name still contains the template "eth%d",
> leading to funny names in /proc/interrupts. E.g. on R-Car E3:
>
> 89: 0 0 GICv2 93 Level eth%d:ch22:multi
> 90: 0 3 GICv2 95 Level eth%d:ch24:emac
> 91: 0 23484 GICv2 71 Level eth%d:ch0:rx_be
> 92: 0 0 GICv2 72 Level eth%d:ch1:rx_nc
> 93: 0 13735 GICv2 89 Level eth%d:ch18:tx_be
> 94: 0 0 GICv2 90 Level eth%d:ch19:tx_nc
>
> Worse, on platforms with multiple RAVB instances (e.g. R-Car V4H), all
> interrupts have similar names.
>
> Fix this by using the device name instead, like is done in several other
> drivers:
>
> 89: 0 0 GICv2 93 Level e6800000.ethernet:ch22:multi
> 90: 0 1 GICv2 95 Level e6800000.ethernet:ch24:emac
> 91: 0 28578 GICv2 71 Level e6800000.ethernet:ch0:rx_be
> 92: 0 0 GICv2 72 Level e6800000.ethernet:ch1:rx_nc
> 93: 0 14044 GICv2 89 Level e6800000.ethernet:ch18:tx_be
> 94: 0 0 GICv2 90 Level e6800000.ethernet:ch19:tx_nc
>
> Rename the local variable dev_name, as it shadows the dev_name()
> function, and pre-initialize it, to simplify the code.
>
> Fixes: 32f012b8c01ca9fd ("net: ravb: Move getting/requesting IRQs in the probe() method")
> Signed-off-by: Geert Uytterhoeven <geert+renesas@...der.be>
Reviewed-by: Niklas Söderlund <niklas.soderlund+renesas@...natech.se>
> ---
> drivers/net/ethernet/renesas/ravb_main.c | 11 +++++------
> 1 file changed, 5 insertions(+), 6 deletions(-)
>
> diff --git a/drivers/net/ethernet/renesas/ravb_main.c b/drivers/net/ethernet/renesas/ravb_main.c
> index b621ddd4539cf517..384ddad65aaf641a 100644
> --- a/drivers/net/ethernet/renesas/ravb_main.c
> +++ b/drivers/net/ethernet/renesas/ravb_main.c
> @@ -2729,19 +2729,18 @@ static int ravb_setup_irq(struct ravb_private *priv, const char *irq_name,
> struct platform_device *pdev = priv->pdev;
> struct net_device *ndev = priv->ndev;
> struct device *dev = &pdev->dev;
> - const char *dev_name;
> + const char *devname = dev_name(dev);
> unsigned long flags;
> int error, irq_num;
>
> if (irq_name) {
> - dev_name = devm_kasprintf(dev, GFP_KERNEL, "%s:%s", ndev->name, ch);
> - if (!dev_name)
> + devname = devm_kasprintf(dev, GFP_KERNEL, "%s:%s", devname, ch);
> + if (!devname)
> return -ENOMEM;
>
> irq_num = platform_get_irq_byname(pdev, irq_name);
> flags = 0;
> } else {
> - dev_name = ndev->name;
> irq_num = platform_get_irq(pdev, 0);
> flags = IRQF_SHARED;
> }
> @@ -2751,9 +2750,9 @@ static int ravb_setup_irq(struct ravb_private *priv, const char *irq_name,
> if (irq)
> *irq = irq_num;
>
> - error = devm_request_irq(dev, irq_num, handler, flags, dev_name, ndev);
> + error = devm_request_irq(dev, irq_num, handler, flags, devname, ndev);
> if (error)
> - netdev_err(ndev, "cannot request IRQ %s\n", dev_name);
> + netdev_err(ndev, "cannot request IRQ %s\n", devname);
>
> return error;
> }
> --
> 2.34.1
>
--
Kind Regards,
Niklas Söderlund
Powered by blists - more mailing lists