netdev - Re: [PATCH net-next v4 10/11] net-timestamp: export the tskey for TCP bpf extension

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 for Android: free password hash cracker in your pocket

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

Message-ID: <9f5081bb-ed66-4171-acef-786ae02cf69c@linux.dev>
Date: Thu, 12 Dec 2024 16:28:37 -0800
From: Martin KaFai Lau <martin.lau@...ux.dev>
To: Jason Xing <kerneljasonxing@...il.com>
Cc: davem@...emloft.net, edumazet@...gle.com, kuba@...nel.org,
 pabeni@...hat.com, dsahern@...nel.org, willemdebruijn.kernel@...il.com,
 willemb@...gle.com, ast@...nel.org, daniel@...earbox.net, andrii@...nel.org,
 eddyz87@...il.com, song@...nel.org, yonghong.song@...ux.dev,
 john.fastabend@...il.com, kpsingh@...nel.org, sdf@...ichev.me,
 haoluo@...gle.com, jolsa@...nel.org, bpf@...r.kernel.org,
 netdev@...r.kernel.org, Jason Xing <kernelxing@...cent.com>
Subject: Re: [PATCH net-next v4 10/11] net-timestamp: export the tskey for TCP
 bpf extension

On 12/7/24 9:38 AM, Jason Xing wrote:
> From: Jason Xing <kernelxing@...cent.com>
> 
> For now, there are three phases where we are not able to fetch
> the right seqno from the skops->skb_data, because:
> 1) in __dev_queue_xmit(), the skb->data doesn't point to the start
> offset in tcp header.
> 2) in tcp_ack_tstamp(), the skb doesn't have the tcp header.
> 
> In the long run, we may add other trace points for bpf extension.
> And the shinfo->tskey is always the same value for both bpf and
> non-bpf cases. With that said, let's directly use shinfo->tskey
> for TCP protocol.
> 
> Signed-off-by: Jason Xing <kernelxing@...cent.com>
> ---
>   net/core/skbuff.c | 7 +++++--
>   1 file changed, 5 insertions(+), 2 deletions(-)
> 
> diff --git a/net/core/skbuff.c b/net/core/skbuff.c
> index 7c59ef501c74..2e13643f791c 100644
> --- a/net/core/skbuff.c
> +++ b/net/core/skbuff.c
> @@ -5544,7 +5544,7 @@ static void __skb_tstamp_tx_bpf(struct sock *sk, struct sk_buff *skb,
>   				int tstype)
>   {
>   	struct timespec64 tstamp;
> -	u32 args[2] = {0, 0};
> +	u32 args[3] = {0, 0, 0};
>   	int op;
>   
>   	if (!sk)
> @@ -5569,7 +5569,10 @@ static void __skb_tstamp_tx_bpf(struct sock *sk, struct sk_buff *skb,
>   		return;
>   	}
>   
> -	bpf_skops_tx_timestamping(sk, skb, op, 2, args);
> +	if (sk_is_tcp(sk))
> +		args[2] = skb_shinfo(skb)->tskey;

Instead of only passing one info "skb_shinfo(skb)->tskey" of a skb, pass the 
whole skb ptr to the bpf prog. Take a look at bpf_skops_init_skb. Lets start 
with end_offset = 0 for now so that the bpf prog won't use it to read the 
skb->data. It can be revisited later.

	bpf_skops_init_skb(&sock_ops, skb, 0);

The bpf prog can use bpf_cast_to_kern_ctx() and bpf_core_cast() to get to the 
skb_shinfo(skb). Take a look at the md_skb example in type_cast.c.

Then it needs to add a bpf_sock->op check to the existing 
bpf_sock_ops_{load,store}_hdr_opt() helpers to ensure these helpers can only be 
used by the BPF_SOCK_OPS_PARSE_HDR_OPT_CB, BPF_SOCK_OPS_HDR_OPT_LEN_CB, and 
BPF_SOCK_OPS_WRITE_HDR_OPT_CB callback.

btw, how is the ack_skb used for the SCM_TSTAMP_ACK by the user space now?