Re: [PATCH v3 bpf-next 2/9] bpf: tcp: Add a new per-socket flag and kfunc for bpf_tcp_ops.

[sashiko-bot@kernel.org]

Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
- [Low] tcp: missing documentation update for tcp_sock cacheline
--

commit a581d0e37015bf601147a1857ac83871b8f4eae9
Author: Kuniyuki Iwashima <kuniyu@google.com>

bpf: tcp: Add a new per-socket flag and kfunc for bpf_tcp_ops.

This commit introduces a new per-socket flag field, bpf_tcp_ops_flags, to
struct tcp_sock to avoid the overhead of the legacy SOCK_OPS approach and
support fast-path callbacks. It also adds a new kfunc,
bpf_tcp_ops_set_flags, allowing BPF programs to atomically enable or disable
these flags. The new field is placed within the tcp_sock_read_txrx cacheline
group.

Link: https://lore.kernel.org/netdev/CAMB2axMwBuz3X4Uwn5uzZUqm91EbiMnbQX832wUFOQ44fOwDnQ@mail.gmail.com/

> diff --git a/include/linux/tcp.h b/include/linux/tcp.h
> --- a/include/linux/tcp.h
> +++ b/include/linux/tcp.h

[ ... ]

> @@ -233,6 +233,13 @@ struct tcp_sock {
>  		is_sack_reneg:1,    /* in recovery from loss with SACK reneg? */
>  		is_cwnd_limited:1,/* forward progress limited by snd_cwnd? */
>  		recvmsg_inq : 1;/* Indicate # of bytes in queue upon recvmsg */
> +#ifdef CONFIG_BPF
> +	u32	bpf_tcp_ops_flags;

[Severity: Low]
This isn't a functional bug, but should
Documentation/networking/net_cachelines/tcp_sock.rst be updated to reflect
the addition of the bpf_tcp_ops_flags field to the tcp_sock_read_txrx
cacheline group?

The tcp_sock structure header explicitly mentions that the document should be
updated when adding new fields.

> +#define BPF_TCP_OPS_TEST_FLAG(TP, ARG) \
> +	(READ_ONCE((TP)->bpf_tcp_ops_flags) & BPF_TCP_OPS_FLAG_ ## ARG)
> +#else
> +#define BPF_TCP_OPS_TEST_FLAG(TP, ARG) (0)
> +#endif
>  	__cacheline_group_end(tcp_sock_read_txrx);

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20261005154533.4147685-1-kuniyu@google.com?part=2