On Fri, 12 Sep 2014 17:07:32 +0300
Stratos Psomadakis <pso...@grnet.gr> wrote:
> Commit cdaa86a54 ("Add G_IO_HUP handler for socket chardev") exposed a bug in
> the way the HMP monitor handles its command buffer. When a client closes the
> connection to the monitor, tcp_chr_read() will detect the G_IO_HUP condition
> and call tcp_chr_disconnect() to close the server-side connection too. Due to
> the fact that monitor reads 1 byte at a time (for each tcp_chr_read()), the
> monitor readline state / buffers might contain junk (i.e. a half-finished
> command). Thus, without calling readline_restart() on mon->rs upon
> CHR_EVENT_CLOSED, future HMP commands will fail.
What's your reproducer? Are you using the mux feature? We also reset it
in CHR_EVENT_OPENED if the mux feature is not used, why isn't that
good enough?
>
> Signed-off-by: Stratos Psomadakis <pso...@grnet.gr>
> Signed-off-by: Dimitris Aragiorgis <dim...@grnet.gr>
> ---
> monitor.c | 1 +
> 1 file changed, 1 insertion(+)
>
> diff --git a/monitor.c b/monitor.c
> index 34cee74..7857300 100644
> --- a/monitor.c
> +++ b/monitor.c
> @@ -5252,6 +5252,7 @@ static void monitor_event(void *opaque, int event)
> break;
>
> case CHR_EVENT_CLOSED:
> + readline_restart(mon->rs);
> mon_refcount--;
> monitor_fdsets_cleanup();
> break;
