Re: [PATCH v2] selftests: mptcp: Fix incorrect file descriptor check in main_loop

[Matthieu Baerts <matttbe@kernel.org>]

Hi Cong Liu,

(-Cc everybody, except the MPTCP ML)

Thank you for the v2.

On 16/01/2025 09:54, Cong Liu wrote:
> Fix a bug where the code was checking the wrong file descriptor
> when opening the input file. The code was checking 'fd' instead
> of 'fd_in', which could lead to incorrect error handling.
> 
> Fixes: ca7ae8916043 ("selftests: mptcp: mptfo Initiator/Listener")

I think the first issue has been introduced in 05be5e273c84 ("selftests:
mptcp: add disconnect tests"), then in the one you mentioned. Because it
is just in the selftests, feel free to add the older one, or both.

> Signed-off-by: Cong Liu <liucong2@kylinos.cn>
> ---
>  tools/testing/selftests/net/mptcp/mptcp_connect.c | 4 ++--
>  1 file changed, 2 insertions(+), 2 deletions(-)
> 
> diff --git a/tools/testing/selftests/net/mptcp/mptcp_connect.c b/tools/testing/selftests/net/mptcp/mptcp_connect.c
> index 4209b9569039..31f4c5618569 100644
> --- a/tools/testing/selftests/net/mptcp/mptcp_connect.c
> +++ b/tools/testing/selftests/net/mptcp/mptcp_connect.c
> @@ -1249,7 +1249,7 @@ int main_loop(void)
>  
>  	if (cfg_input && cfg_sockopt_types.mptfo) {
>  		fd_in = open(cfg_input, O_RDONLY);
> -		if (fd < 0)
> +		if (fd_in < 0)
>  			xerror("can't open %s:%d", cfg_input, errno);
>  	}
>  
> @@ -1272,7 +1272,7 @@ int main_loop(void)
>  
>  	if (cfg_input && !cfg_sockopt_types.mptfo) {
>  		fd_in = open(cfg_input, O_RDONLY);
> -		if (fd < 0)
> +		if (fd_in < 0)
>  			xerror("can't open %s:%d", cfg_input, errno);
>  	}

The CI is not able to apply your patch. On which base are you?
Do you mind using our 'export' branch from our GitHub repository?

> When I tried to solve the problem where fd_in was closed in the again
> code block but not reopened in certain scenarios, I encountered some 
> issues. Here's my code, but it causes the disconnect test to fail and I'm
> not sure how to fix this. 

Some fixes have been recently applied around the 'disconnect' tests,
maybe this issue has already been fixed?

Cheers,
Matt
-- 
Sponsored by the NGI0 Core fund.