Re: [PATCH 1/1] net: also flag accepted sockets supporting msghdr originated zerocopy

[Pavel Begunkov <asml.silence@gmail.com>]

On 10/25/22 14:36, Stefan Metzmacher wrote:
> Without this only the client initiated tcp sockets have SOCK_SUPPORT_ZC.
> The listening socket on the server also has it, but the accepted
> connections didn't, which meant IORING_OP_SEND[MSG]_ZC will always
> fails with -EOPNOTSUPP.

Jakub pointed out a bunch of places yesterday that we have to
consider. I'll pick it up and resend with the rest. Thanks


> Fixes: e993ffe3da4b ("net: flag sockets supporting msghdr originated zerocopy")
> Cc: <stable@vger.kernel.org> # 6.0
> CC: Pavel Begunkov <asml.silence@gmail.com>
> Cc: Jakub Kicinski <kuba@kernel.org>
> CC: Jens Axboe <axboe@kernel.dk>
> Link: https://lore.kernel.org/io-uring/20221024141503.22b4e251@kernel.org/T/#m38aa19b0b825758fb97860a38ad13122051f9dda
> Signed-off-by: Stefan Metzmacher <metze@samba.org>
> ---
>   net/ipv4/af_inet.c | 2 ++
>   1 file changed, 2 insertions(+)
> 
> diff --git a/net/ipv4/af_inet.c b/net/ipv4/af_inet.c
> index 3dd02396517d..4728087c42a5 100644
> --- a/net/ipv4/af_inet.c
> +++ b/net/ipv4/af_inet.c
> @@ -754,6 +754,8 @@ int inet_accept(struct socket *sock, struct socket *newsock, int flags,
>   		  (TCPF_ESTABLISHED | TCPF_SYN_RECV |
>   		  TCPF_CLOSE_WAIT | TCPF_CLOSE)));
>   
> +	if (test_bit(SOCK_SUPPORT_ZC, &sock->flags))
> +		set_bit(SOCK_SUPPORT_ZC, &newsock->flags);
>   	sock_graft(sk2, newsock);
>   
>   	newsock->state = SS_CONNECTED;

-- 
Pavel Begunkov