Re: [PATCH net v2 1/1] can: bcm: limit RX operation resource usage per socket

[sashiko-bot@kernel.org]

Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
- [Low] can: misplaced kernel-doc comment above can_rx_register_internal()
--

commit 6e185f899c0a564d22495ef1e055cc4b9c9b27f0
Author: Zihan Xi <zihanx@nebusec.ai>

can: bcm: limit RX operation resource usage per socket

This patch bounds the number and total allocation size of RX operations per
BCM socket to prevent kernel memory exhaustion. It also introduces
can_rx_register_with_release() so the socket's receiver count remains
charged until the RCU grace period completes.

Link: https://bugtracker.nebusec.ai/f/11134

> diff --git a/net/can/af_can.c b/net/can/af_can.c
> index 34fe3b28d576f..a3f0ed35fbb52 100644
> --- a/net/can/af_can.c
> +++ b/net/can/af_can.c
> @@ -441,15 +441,20 @@ static struct hlist_head *can_rcv_list_find(canid_t *can_id, canid_t *mask,
>   *  -ENOMEM on missing cache mem to create subscription entry
>   *  -ENODEV unknown device
>   */
> -int can_rx_register(struct net *net, struct net_device *dev, canid_t can_id,
> -		    canid_t mask, void (*func)(struct sk_buff *, void *),
> -		    void *data, char *ident, struct sock *sk)
> +static int can_rx_register_internal(struct net *net, struct net_device *dev,
> +				    canid_t can_id, canid_t mask,
> +				    void (*func)(struct sk_buff *, void *),
> +				    void *data, char *ident, struct sock *sk,
> +				    void (*release)(struct sock *sk))

[Severity: Low]
When splitting can_rx_register() into can_rx_register_internal() and the
exported wrapper in net/can/af_can.c, can_rx_register_internal() is placed
directly beneath the existing kernel-doc comment for can_rx_register():

/**
 * can_rx_register - subscribe CAN frames from a specific interface
...
 */
static int can_rx_register_internal(struct net *net, struct net_device *dev,
				    canid_t can_id, canid_t mask,
				    void (*func)(struct sk_buff *, void *),
				    void *data, char *ident, struct sock *sk,
				    void (*release)(struct sock *sk))

Doesn't this cause scripts/kernel-doc (under W=1 builds) to warn that the
prototype is for can_rx_register_internal() instead of can_rx_register()
and that the release parameter is not described?

Should the kernel-doc comment be moved down to the exported
can_rx_register() wrapper (or updated to match the function it precedes)?

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/cover.1793109154.git.zihanx@nebusec.ai?part=1