netdev - Re: [PATCH net-next 3/3] net: airoha: Add L2 hw acceleration support

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <Z/V/ip/dA+aw+dCW@localhost.localdomain>
Date: Tue, 8 Apr 2025 21:56:58 +0200
From: Michal Kubiak <michal.kubiak@...el.com>
To: Lorenzo Bianconi <lorenzo@...nel.org>
CC: Andrew Lunn <andrew+netdev@...n.ch>, "David S. Miller"
	<davem@...emloft.net>, Eric Dumazet <edumazet@...gle.com>, Jakub Kicinski
	<kuba@...nel.org>, Paolo Abeni <pabeni@...hat.com>,
	<linux-arm-kernel@...ts.infradead.org>, <linux-mediatek@...ts.infradead.org>,
	<netdev@...r.kernel.org>
Subject: Re: [PATCH net-next 3/3] net: airoha: Add L2 hw acceleration support

On Mon, Apr 07, 2025 at 04:18:32PM +0200, Lorenzo Bianconi wrote:
> Similar to mtk driver, introduce the capability to offload L2 traffic
> defining flower rules in the PSE/PPE engine available on EN7581 SoC.
> Since the hw always reports L2/L3/L4 flower rules, link all L2 rules
> sharing the same L2 info (with different L3/L4 info) in the L2 subflows
> list of a given L2 PPE entry.
> 
> Signed-off-by: Lorenzo Bianconi <lorenzo@...nel.org>
> ---
>  drivers/net/ethernet/airoha/airoha_eth.c |   2 +-
>  drivers/net/ethernet/airoha/airoha_eth.h |  11 ++-
>  drivers/net/ethernet/airoha/airoha_ppe.c | 162 +++++++++++++++++++++++++------
>  3 files changed, 144 insertions(+), 31 deletions(-)
> 

[...]

> +static void airoha_ppe_foe_remove_flow(struct airoha_ppe *ppe,
> +				       struct airoha_flow_table_entry *e)
> +{
> +	lockdep_assert_held(&ppe_lock);
> +
> +	hlist_del_init(&e->list);
> +	if (e->hash != 0xffff) {
> +		e->data.ib1 &= ~AIROHA_FOE_IB1_BIND_STATE;
> +		e->data.ib1 |= FIELD_PREP(AIROHA_FOE_IB1_BIND_STATE,
> +					  AIROHA_FOE_STATE_INVALID);
> +		airoha_ppe_foe_commit_entry(ppe, &e->data, e->hash);
> +		e->hash = 0xffff;
> +	}
> +	if (e->type == FLOW_TYPE_L2_SUBFLOW) {
> +		hlist_del_init(&e->l2_subflow_node);
> +		kfree(e);
> +	}
> +}
> +
> +static void airoha_ppe_foe_remove_l2_flow(struct airoha_ppe *ppe,
> +					  struct airoha_flow_table_entry *e)
> +{
> +	struct hlist_head *head = &e->l2_flows;
> +	struct hlist_node *n;
> +
> +	lockdep_assert_held(&ppe_lock);
> +
> +	rhashtable_remove_fast(&ppe->l2_flows, &e->l2_node,
> +			       airoha_l2_flow_table_params);
> +	hlist_for_each_entry_safe(e, n, head, l2_subflow_node)
> +		airoha_ppe_foe_remove_flow(ppe, e);
> +}
> +
>  static void airoha_ppe_foe_flow_remove_entry(struct airoha_ppe *ppe,
>  					     struct airoha_flow_table_entry *e)
>  {
>  	lockdep_assert_held(&ppe_lock);
>  
> -	if (e->type == FLOW_TYPE_L2) {
> -		rhashtable_remove_fast(&ppe->l2_flows, &e->l2_node,
> -				       airoha_l2_flow_table_params);
> -	} else {
> -		hlist_del_init(&e->list);
> -		if (e->hash != 0xffff) {
> -			e->data.ib1 &= ~AIROHA_FOE_IB1_BIND_STATE;
> -			e->data.ib1 |= FIELD_PREP(AIROHA_FOE_IB1_BIND_STATE,
> -						  AIROHA_FOE_STATE_INVALID);
> -			airoha_ppe_foe_commit_entry(ppe, &e->data, e->hash);
> -			e->hash = 0xffff;
> -		}
> -	}
> +	if (e->type == FLOW_TYPE_L2)
> +		airoha_ppe_foe_remove_l2_flow(ppe, e);
> +	else
> +		airoha_ppe_foe_remove_flow(ppe, e);

It's not a hard request, more of a question: wouldn't it be better to
introduce "airoha_ppe_foe_remove_l2_flow()" and
"airoha_ppe_foe_remove_flow()" in the patch #2?
It looks like reorganizing the code can be part of the preliminary
patch and the current patch can just add the feature, e.g. L2_SUBFLOW.

Thanks,
Michal