Re: [PATCH] xen: Modify domain_crash() to take a print string

[Julien Grall <julien@xxxxxxx>]

Hi,

On 03/02/2022 13:38, Andrew Cooper wrote:
> diff --git a/xen/include/xen/sched.h b/xen/include/xen/sched.h
> index 37f78cc4c4c9..38b390d20371 100644
> --- a/xen/include/xen/sched.h
> +++ b/xen/include/xen/sched.h
> @@ -736,10 +736,15 @@ void vcpu_end_shutdown_deferral(struct vcpu *v);
>    * from any processor.
>    */
>   void __domain_crash(struct domain *d);
> -#define domain_crash(d) do {                                              \
> -    printk("domain_crash called from %s:%d\n", __FILE__, __LINE__);       \
> -    __domain_crash(d);                                                    \
> -} while (0)
> +#define domain_crash(d, ...)                            \
> +    do {                                                \
> +        if ( count_args(__VA_ARGS__) == 0 )             \
> +            printk("domain_crash called from %s:%d\n",  \
> +                   __FILE__, __LINE__);                 \
I find a bit odd that here you are using a normal printk but...


> +        else                                            \
> +            printk(XENLOG_G_ERR __VA_ARGS__);           \
here it is XENLOG_G_ERR. In fact, isn't it ratelimited? If so, wouldn't 
it be better to only use XENLOG_ERR so they can always be seen? (A 
domain shouldn't be able to abuse it).
Cheers,

--
Julien Grall