Re: [BlueZ v2 1/2] mpris-proxy: Fix possible crash

[Bastien Nocera <hadess@hadess.net>]

On Tue, 2026-05-05 at 12:48 +0200, Bastien Nocera wrote:
> find_player_by_obex() doesn't check whether session->obex is a valid
> pointer before dereferecing it, but all code paths that assign it use

There's a small typo here, any other comments on those 2 patches?

Any OBEX experts?

> create_obex_session() to assign it, a function that can fail.
> 
> Check whether session->obex is null before dereferencing it.
> 
>  #0 find_player_by_obex at tools/mpris-proxy.c:2819
>  #1 obex_property_changed at tools/mpris-proxy.c:2929
>  #2 add_property at gdbus/client.c:373
>  #3 update_properties at gdbus/client.c:399
>  #5 properties_changed at gdbus/client.c:537
>  #6 signal_filter at gdbus/watch.c:416
>  #7 message_filter at gdbus/watch.c:566
>  #10 message_dispatch at gdbus/mainloop.c:59
>  #13 g_main_context_dispatch_unlocked at ../glib/gmain.c:4451
>  #14 g_main_context_iterate_unlocked at ../glib/gmain.c:4516
> 
> Closes: https://bugzilla.redhat.com/show_bug.cgi?id=2466640
> ---
> Changes since v1:
> - Fix missing space before = sign
> 
>  tools/mpris-proxy.c | 6 +++++-
>  1 file changed, 5 insertions(+), 1 deletion(-)
> 
> diff --git a/tools/mpris-proxy.c b/tools/mpris-proxy.c
> index 1d7a421e9278..eb607347aa32 100644
> --- a/tools/mpris-proxy.c
> +++ b/tools/mpris-proxy.c
> @@ -2816,8 +2816,12 @@ static struct player
> *find_player_by_obex(const char *path)
>  	for (l = players; l; l = l->next) {
>  		struct player *player = l->data;
>  		struct obex_session *session = player->obex;
> -		const char *obex_path =
> g_dbus_proxy_get_path(session->obex);
> +		const char *obex_path = NULL;
>  
> +		if (session == NULL)
> +			continue;
> +
> +		obex_path = g_dbus_proxy_get_path(session->obex);
>  		if (g_str_has_prefix(path, obex_path))
>  			return player;
>  	}