netdev - Re: [PATCH v26 01/20] net: Introduce direct data placement tcp offload

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: <9d381da6-cef7-431e-be82-fd2888fc480a@redhat.com>
Date: Tue, 25 Feb 2025 11:23:28 +0100
From: Paolo Abeni <pabeni@...hat.com>
To: Aurelien Aptel <aaptel@...dia.com>, linux-nvme@...ts.infradead.org,
 netdev@...r.kernel.org, sagi@...mberg.me, hch@....de, kbusch@...nel.org,
 axboe@...com, chaitanyak@...dia.com, davem@...emloft.net, kuba@...nel.org
Cc: Boris Pismenny <borisp@...dia.com>, aurelien.aptel@...il.com,
 smalin@...dia.com, malin1024@...il.com, ogerlitz@...dia.com,
 yorayz@...dia.com, galshalom@...dia.com, mgurtovoy@...dia.com,
 edumazet@...gle.com, dsahern@...nel.org, ast@...nel.org,
 jacob.e.keller@...el.com
Subject: Re: [PATCH v26 01/20] net: Introduce direct data placement tcp
 offload

On 2/21/25 10:52 AM, Aurelien Aptel wrote:
> +int ulp_ddp_sk_add(struct net_device *netdev,
> +		   struct sock *sk,
> +		   struct ulp_ddp_config *config,
> +		   const struct ulp_ddp_ulp_ops *ops)
> +{
> +	int ret;
> +
> +	/* put in ulp_ddp_sk_del() */
> +	dev_hold(netdev);

You should use netdev_hold()/netdev_put() instead, with a paired reftracker.

> +
> +	ret = netdev->netdev_ops->ulp_ddp_ops->sk_add(netdev, sk, config);
> +	if (ret) {
> +		dev_put(netdev);
> +		return ret;
> +	}
> +
> +	inet_csk(sk)->icsk_ulp_ddp_ops = ops;
> +
> +	return 0;
> +}
> +EXPORT_SYMBOL_GPL(ulp_ddp_sk_add);
> +
> +void ulp_ddp_sk_del(struct net_device *netdev,
> +		    struct sock *sk)
> +{
> +	netdev->netdev_ops->ulp_ddp_ops->sk_del(netdev, sk);
> +	inet_csk(sk)->icsk_ulp_ddp_ops = NULL;
> +	dev_put(netdev);
> +}
> +EXPORT_SYMBOL_GPL(ulp_ddp_sk_del);
> +
> +bool ulp_ddp_is_cap_active(struct net_device *netdev, int cap_bit_nr)
> +{
> +	struct ulp_ddp_dev_caps caps;
> +
> +	if (!netdev->netdev_ops->ulp_ddp_ops)
> +		return false;
> +	netdev->netdev_ops->ulp_ddp_ops->get_caps(netdev, &caps);
> +	return test_bit(cap_bit_nr, caps.active);
> +}
> +EXPORT_SYMBOL_GPL(ulp_ddp_is_cap_active);
> diff --git a/net/ipv4/tcp_input.c b/net/ipv4/tcp_input.c
> index fbb67a098543..771720c6e0da 100644
> --- a/net/ipv4/tcp_input.c
> +++ b/net/ipv4/tcp_input.c
> @@ -5474,6 +5474,8 @@ tcp_collapse(struct sock *sk, struct sk_buff_head *list, struct rb_root *root,
>  
>  		memcpy(nskb->cb, skb->cb, sizeof(skb->cb));
>  		skb_copy_decrypted(nskb, skb);
> +		skb_copy_no_condense(nskb, skb);
> +		skb_copy_ulp_crc(nskb, skb);
>  		TCP_SKB_CB(nskb)->seq = TCP_SKB_CB(nskb)->end_seq = start;
>  		if (list)
>  			__skb_queue_before(list, skb, nskb);
> diff --git a/net/ipv4/tcp_offload.c b/net/ipv4/tcp_offload.c
> index 2308665b51c5..bcb8055bbb0f 100644
> --- a/net/ipv4/tcp_offload.c
> +++ b/net/ipv4/tcp_offload.c
> @@ -346,6 +346,7 @@ struct sk_buff *tcp_gro_receive(struct list_head *head, struct sk_buff *skb,
>  
>  	flush |= (ntohl(th2->seq) + skb_gro_len(p)) ^ ntohl(th->seq);
>  	flush |= skb_cmp_decrypted(p, skb);
> +	flush |= skb_is_ulp_crc(p) != skb_is_ulp_crc(skb);

Possibly a `skb_cmp_ulp_crc()` helper would be cleaner.

Thanks,

Paolo