Re: [PATCH] x86/p2m: also clear defer_nested_flush on error

["Egger, Christoph" <chegger@amazon.de>]

On 15.04.14 16:10, Jan Beulich wrote:
> While presumably benign (domain creation ought to fail when P2M setup
> fails) let's be on the safe side and clear the flag as intended. And
> actually, the code can be streamlined quite a bit by recognizing that
> the only difference between the success and error cases is the message
> printed in each case. With that, a stray spin_unlock() also goes away.
> 
> Signed-off-by: Jan Beulich <jbeulich@suse.com>

Nice cleanup!

Reviewed-by: Christoph Egger <chegger@amazon.de>

> 
> --- a/xen/arch/x86/mm/p2m.c
> +++ b/xen/arch/x86/mm/p2m.c
> @@ -435,18 +435,12 @@ int p2m_alloc_table(struct p2m_domain *p
>      p2m->defer_nested_flush = 1;
>      rc = p2m_set_entry(p2m, 0, _mfn(INVALID_MFN), PAGE_ORDER_4K,
>                         p2m_invalid, p2m->default_access);
> -    if ( rc )
> -        goto error;
>      p2m->defer_nested_flush = 0;
> -
> -    P2M_PRINTK("p2m table initialised for slot zero\n");
> -    p2m_unlock(p2m);
> -    return 0;
> -
> -    spin_unlock(&p2m->domain->page_alloc_lock);
> - error:
> -    P2M_PRINTK("failed to initialise p2m table for slot zero. rc:%d\n", rc);
>      p2m_unlock(p2m);
> +    if ( !rc )
> +        P2M_PRINTK("p2m table initialised for slot zero\n");
> +    else
> +        P2M_PRINTK("failed to initialise p2m table for slot zero (%d)\n", rc);
>      return rc;
>  }