On Wed, Jul 24, 2024 at 11:50:15PM +0200, Stefano Brivio wrote:
> In tap_sock_unix_open(), if we have a given path for the socket from
> configuration, we don't need to loop over possible paths, so we exit
> the loop on the first iteration, unconditionally.
> 
> But if we failed to bind() the socket to that explicit path, we should
> exit, instead of continuing. Otherwise we'll pretend we're up and
> running, but nobody can contact us, and this might be mildly confusing
> for users.
> 
> Link: https://bugzilla.redhat.com/show_bug.cgi?id=2299474
> Signed-off-by: Stefano Brivio <sbrivio@redhat.com>

Reviewed-by: David Gibson <david@gibson.dropbear.id.au>

> ---
>  tap.c | 7 +++++--
>  1 file changed, 5 insertions(+), 2 deletions(-)
> 
> diff --git a/tap.c b/tap.c
> index 6930ad8..44bd444 100644
> --- a/tap.c
> +++ b/tap.c
> @@ -1139,8 +1139,11 @@ int tap_sock_unix_open(char *sock_path)
>  		close(ex);
>  
>  		unlink(path);
> -		if (!bind(fd, (const struct sockaddr *)&addr, sizeof(addr)) ||
> -		    *sock_path)
> +		ret = bind(fd, (const struct sockaddr *)&addr, sizeof(addr));
> +		if (*sock_path && ret)
> +			die_perror("Failed to bind UNIX domain socket");
> +
> +		if (!ret)
>  			break;
>  	}
>  

-- 
David Gibson (he or they)	| I'll have my music baroque, and my code
david AT gibson.dropbear.id.au	| minimalist, thank you, not the other way
				| around.
http://www.ozlabs.org/~dgibson