Re: backport a patch of ravb for v5.10 and v5.4

[Greg KH <gregkh@linuxfoundation.org>]

On Mon, Oct 16, 2023 at 03:04:36AM +0000, Yoshihiro Shimoda wrote:
> Hello Sasha,
> 
> Thank you for backporting the latest ravb patches for stable.
> I found one of patches [1] was queued into v5.10 and v5.4 [2].
> However, another patch [3] was not queued into them. I guess
> that this is because conflict happens.
> 
> The reason for the conflict is that the condition in the following
> line is diffetent:
> ---
> v5.10 or v5.4:
> 	if (priv->chip_id != RCAR_GEN2) {
> 
> mainline:
> 	if (info->multi_irqs) {
> ---
> 
> However, this difference can be ignored when backporting. For your
> reference, I wrote a sample patch at the end of this email. Would
> you backport such a patch to v5.10 and v5.4? I would appreciate it
> if you could let me know if there is an official way to request one.
> 
> [1] https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?id=e6864af61493113558c502b5cd0d754c19b93277
> [2] https://www.spinics.net/lists/stable-commits/msg319990.html
>     https://www.spinics.net/lists/stable-commits/msg320008.html
> [3] https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?id=3971442870713de527684398416970cf025b4f89
> 
> ---
>  drivers/net/ethernet/renesas/ravb_main.c | 2 ++
>  1 file changed, 2 insertions(+)
> 
> diff --git a/drivers/net/ethernet/renesas/ravb_main.c b/drivers/net/ethernet/renesas/ravb_main.c
> index a59da6a11976..f218bacec001 100644
> --- a/drivers/net/ethernet/renesas/ravb_main.c
> +++ b/drivers/net/ethernet/renesas/ravb_main.c
> @@ -1706,6 +1706,8 @@ static int ravb_close(struct net_device *ndev)
>  			of_phy_deregister_fixed_link(np);
>  	}
>  
> +	cancel_work_sync(&priv->work);
> +
>  	if (priv->chip_id != RCAR_GEN2) {
>  		free_irq(priv->tx_irqs[RAVB_NC], ndev);
>  		free_irq(priv->rx_irqs[RAVB_NC], ndev);
> -- 
> 2.25.1
> 

Can you please just send a working backport for the kernel trees you
wish to see this applied to?  Merging patches like this is not easy and
doesn't usually result in a commit that we know actually works properly.

thanks,

greg k-h