Re: [PATCH] irqchip/irq-bcm7120-l2: add put_device() after of_find_device_by_node()

[Florian Fainelli]

On 11/8/2021 7:29 PM, cgel.zte@xxxxxxxxx wrote:
> From: Ye Guojin <ye.guojin@xxxxxxxxxx>
>
> This was found by coccicheck:
> ./drivers/irqchip/irq-bcm7120-l2.c,328,1-7,ERROR  missing put_device;
> call of_find_device_by_node on line 234, but without a corresponding
> object release within this function.
> ./drivers/irqchip/irq-bcm7120-l2.c,341,1-7,ERROR  missing put_device;
> call of_find_device_by_node on line 234, but without a corresponding
> object release within this function.
>
> Addtionally, fixup the potential problem that memory is not released
> before return.
>
> Reported-by: Zeal Robot <zealci@xxxxxxxxxx>
> Signed-off-by: Ye Guojin <ye.guojin@xxxxxxxxxx>
> ---
>   drivers/irqchip/irq-bcm7120-l2.c | 3 ++-
>   1 file changed, 2 insertions(+), 1 deletion(-)
>
> diff --git a/drivers/irqchip/irq-bcm7120-l2.c b/drivers/irqchip/irq-bcm7120-l2.c
> index d80e67a6aad2..aaa70619f3c0 100644
> --- a/drivers/irqchip/irq-bcm7120-l2.c
> +++ b/drivers/irqchip/irq-bcm7120-l2.c
> @@ -325,7 +325,7 @@ static int __init bcm7120_l2_intc_probe(struct device_node *dn,
>   	pr_info("registered %s intc (%pOF, parent IRQ(s): %d)\n",
>   		intc_name, dn, data->num_parent_irqs);
>   
> -	return 0;
> +	ret = 0;

That cannot be right, we would take the irq_free_domain label and 
unregister the interrupt domain. The put_device() ought to be done when 
we stop making use of the platform_device, that is at line 246 after we 
call platform_irq_count().

Thanks!
--
Florian