Re: [Xen-devel] [PATCH v5 5/5] xen: add function name to lock profiling data

[Jan Beulich <jbeulich@xxxxxxxx>]

On 12.09.2019 15:28, Juergen Gross wrote:
> --- a/xen/common/spinlock.c
> +++ b/xen/common/spinlock.c
> @@ -348,6 +348,7 @@ static s_time_t lock_profile_start;
>  static struct lock_profile_anc *lock_profile_ancs;
>  static struct lock_profile_qhead lock_profile_glb_q;
>  static spinlock_t lock_profile_lock = SPIN_LOCK_UNLOCKED;
> +static const char *lock_profile_nofunc = __PRETTY_FUNCTION__;

const also for the variable itself?

> @@ -371,8 +372,10 @@ static void spinlock_profile_print_elem(struct 
> lock_profile *data,
>      printk("%s ", type);
>      if ( idx != LOCKPROF_IDX_NONE )
>          printk("%d ", idx);
> -    printk("%s: addr=%p, lockval=%08x, ", data->name, lock,
> -           lock->tickets.head_tail);
> +    printk("%s", data->name);
> +    if ( data->func && strcmp(data->func, lock_profile_nofunc) )
> +        printk("@%s", data->func);
> +    printk(": addr=%p, lockval=%08x, ", lock, lock->tickets.head_tail);

As long as this is used from a key handler only, I guess the
splitting of the printk()s isn't much of a problem.

> @@ -427,7 +430,14 @@ static void spinlock_profile_ucopy_elem(struct 
> lock_profile *data,
>  
>      if ( p->pc->nr_elem < p->pc->max_elem )
>      {
> -        safe_strcpy(elem.name, data->name);
> +        if ( data->func && strcmp(data->func, lock_profile_nofunc) )
> +        {
> +            snprintf(elem.name, sizeof(elem.name), "%s@%s", data->name,
> +                     data->func);
> +            elem.name[sizeof(elem.name) - 1] = 0;

This isn't needed after snprintf(), is it?

With the two suggested adjustments (provided you agree, and
which could probably be done while committing)
Reviewed-by: Jan Beulich <jbeulich@xxxxxxxx>

Jan

_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxxxxxxxxx
https://lists.xenproject.org/mailman/listinfo/xen-devel