Re: [PATCH 3/5] ppc/spapr: Add FWNMI machine check delivery warnings

[Greg Kurz <groug@kaod.org>]

On Tue, 17 Mar 2020 15:02:13 +1000
Nicholas Piggin <npiggin@gmail.com> wrote:

> Add some messages which explain problems and guest misbehaviour that
> may be difficult to diagnose in rare cases of machine checks.
> 
> Signed-off-by: Nicholas Piggin <npiggin@gmail.com>
> ---
>  hw/ppc/spapr_events.c | 4 ++++
>  hw/ppc/spapr_rtas.c   | 4 ++++
>  2 files changed, 8 insertions(+)
> 
> diff --git a/hw/ppc/spapr_events.c b/hw/ppc/spapr_events.c
> index 05337f0671..d35151eeb0 100644
> --- a/hw/ppc/spapr_events.c
> +++ b/hw/ppc/spapr_events.c
> @@ -807,6 +807,8 @@ static void spapr_mce_dispatch_elog(PowerPCCPU *cpu, bool recovered)
>      /* get rtas addr from fdt */
>      rtas_addr = spapr_get_rtas_addr();
>      if (!rtas_addr) {
> +        warn_report("FWNMI: Unable to deliver machine check to guest: "
> +                    "rtas_addr not found.");

Why a warning and not an error ?

Also maybe change the string to fit on one line ?

>          qemu_system_guest_panicked(NULL);
>          g_free(ext_elog);
>          return;
> @@ -848,6 +850,8 @@ void spapr_mce_req_event(PowerPCCPU *cpu, bool recovered)
>           * that CPU called "ibm,nmi-interlock")
>           */
>          if (spapr->fwnmi_machine_check_interlock == cpu->vcpu_id) {
> +            warn_report("FWNMI: Unable to deliver machine check to guest: "
> +                        "nested machine check.");

Ditto.

>              qemu_system_guest_panicked(NULL);
>              return;
>          }
> diff --git a/hw/ppc/spapr_rtas.c b/hw/ppc/spapr_rtas.c
> index 29abe66d01..12cd09701c 100644
> --- a/hw/ppc/spapr_rtas.c
> +++ b/hw/ppc/spapr_rtas.c
> @@ -462,6 +462,10 @@ static void rtas_ibm_nmi_interlock(PowerPCCPU *cpu,
>      }
>  
>      if (spapr->fwnmi_machine_check_addr == -1) {
> +        qemu_log_mask(LOG_GUEST_ERROR,
> +                      "FWNMI: ibm,nmi-interlock RTAS called with FWNMI not "
> +                       "registered.\n");

whitespace damage here   ^ best fixed by making the message a one liner.

> +
>          /* NMI register not called */
>          rtas_st(rets, 0, RTAS_OUT_PARAM_ERROR);
>          return;