From: Claudio Jeker Subject: Re: bgpd: make imsg_rde void and fail hard To: Theo Buehler Cc: tech@openbsd.org Date: Wed, 26 Feb 2025 17:37:42 +0100 On Wed, Feb 26, 2025 at 05:30:38PM +0100, Theo Buehler wrote: > On Wed, Feb 26, 2025 at 05:18:58PM +0100, Claudio Jeker wrote: > > While working on all the other bits in the session engine I came to the > > conclusion that if imsg_rde() fails the game is over. So just admit to > > that and fatal in that case. > > > > This simplifies all the callers since there is no need to check the return > > value. This also allows some other functions to be come void. > > ok > > Should session_handle_{update,rrefresh}() also become void? Wouldn't > it be more consistent if the return (0) was pushed to parse_update() > and parse_rrefresh()? Maybe we should make them all void including parse_update() and parse_rrefresh(). The return value of parse_rrefresh is not checked and the code around parse_update() is most probably wrong for the error case. In the end when a parse error happens we need to generate a notification and properly throw an error in the FSM but there is no need to bubble up the error. -- :wq Claudio