Luiz Capitulino wrote:
> The call of readline_show_prompt() in CHR_EVENT_RESET's body will
> trig a segfault if readline is not being used, because 'mon->rs'
> will be NULL.
> 
> This fixes the problem by adding the proper check.
> 
> I've trigged this while playing with an off-tree code that disables
> readline support, I'm not sure whether in-tree code can trig this.
> 
> Signed-off-by: Luiz Capitulino <lcapitulino@redhat.com>
> ---
>  monitor.c |    2 +-
>  1 files changed, 1 insertions(+), 1 deletions(-)
> 
> diff --git a/monitor.c b/monitor.c
> index 6b45f6c..787101d 100644
> --- a/monitor.c
> +++ b/monitor.c
> @@ -3002,7 +3002,7 @@ static void monitor_event(void *opaque, int event)
>      case CHR_EVENT_RESET:
>          monitor_printf(mon, "QEMU %s monitor - type 'help' for more "
>                         "information\n", QEMU_VERSION);
> -        if (mon->chr->focus == 0)
> +        if (mon->rs && mon->chr->focus == 0)
>              readline_show_prompt(mon->rs);
>          break;
>      }

In-tree code is not affected as no monitor user will call qemu_chr_reset
for the associated char device if there is no readline active as well.
Yeah, secret de-facto rule.

The patch is not incorrect, but I would like to understand the
out-of-tree use case behind it. So you do want the info line printed,
but provide your own readline processor? What kind of terminal is this?

Jan