Re: [PATCH net-next v24 09/23] ovpn: implement packet processing

[Sabrina Dubroca <sd@queasysnail.net>]

2025-03-18, 02:40:44 +0100, Antonio Quartulli wrote:
> +int ovpn_crypto_state_reset(struct ovpn_crypto_state *cs,
> +			    const struct ovpn_peer_key_reset *pkr)
> +{
> +	struct ovpn_crypto_key_slot *old = NULL, *new;
> +	u8 idx;
> +
> +	if (pkr->slot != OVPN_KEY_SLOT_PRIMARY &&
> +	    pkr->slot != OVPN_KEY_SLOT_SECONDARY)
> +		return -EINVAL;
> +
> +	new = ovpn_aead_crypto_key_slot_new(&pkr->key);
> +	if (IS_ERR(new))
> +		return PTR_ERR(new);
> +
> +	spin_lock_bh(&cs->lock);

At this point, should there be a check that we're not installing 2
keys with the same key_id at the same time? I expect a well-behaved
userspace never does that, but it would confuse
ovpn_crypto_key_id_to_slot if it ever happened.

["well, then the tunnel is broken. if userspace sets up a broken
config that's not the kernel's problem." is an acceptable answer]

-- 
Sabrina