x86: log non-responding CPUs in fatal_trap()

Commit Message

Jan Beulich July 13, 2022, 6:51 a.m. UTC

This eases recognizing that something odd is going on.

Signed-off-by: Jan Beulich <[e-mail]>

Comments

Andrew Cooper July 18, 2022, 3:55 p.m. UTC | #1

On 13/07/2022 07:51, Jan Beulich wrote:
> This eases recognizing that something odd is going on.
>
> Signed-off-by: Jan Beulich <[e-mail]>
>
> --- a/xen/arch/x86/traps.c
> +++ b/xen/arch/x86/traps.c
> @@ -845,6 +845,9 @@ void fatal_trap(const struct cpu_user_re
>                      msecs = 10;
>                  }
>              }
> +            if ( pending )
> +                printk("Non-responding CPUs: %*pbl\n",
> +                       CPUMASK_PR(&show_state_mask));

Prevailing style elsewhere is {%*pbl}.

Preferably with that adjusted, Acked-by: Andrew Cooper
<[e-mail]>

Patch

--- a/xen/arch/x86/traps.c
+++ b/xen/arch/x86/traps.c
@@ -845,6 +845,9 @@  void fatal_trap(const struct cpu_user_re
                     msecs = 10;
                 }
             }
+            if ( pending )
+                printk("Non-responding CPUs: %*pbl\n",
+                       CPUMASK_PR(&show_state_mask));
         }
     }