Hi Jan,

> On 4 Nov 2024, at 08:46, Jan Beulich <[email protected]> wrote:
> 
> On 04.11.2024 08:19, Bertrand Marquis wrote:
>> Hi Andrew,
>> 
>>> On 1 Nov 2024, at 12:05, Andrew Cooper <[email protected]> wrote:
>>> 
>>> On 16/10/2024 10:21 am, Bertrand Marquis wrote:
>>>> diff --git a/xen/arch/arm/tee/ffa_partinfo.c 
>>>> b/xen/arch/arm/tee/ffa_partinfo.c
>>>> index fde187dba4e5..d699a267cc76 100644
>>>> --- a/xen/arch/arm/tee/ffa_partinfo.c
>>>> +++ b/xen/arch/arm/tee/ffa_partinfo.c
>>>> @@ -77,7 +77,21 @@ void ffa_handle_partition_info_get(struct cpu_user_regs 
>>>> *regs)
>>>>    };
>>>>    uint32_t src_size, dst_size;
>>>>    void *dst_buf;
>>>> -    uint32_t ffa_sp_count = 0;
>>>> +    uint32_t ffa_vm_count = 0, ffa_sp_count = 0;
>>>> +#ifdef CONFIG_FFA_VM_TO_VM
>>>> +    struct domain *dom;
>>>> +
>>>> +    /* Count the number of VM with FF-A support */
>>>> +    rcu_read_lock(&domlist_read_lock);
>>>> +    for_each_domain( dom )
>>>> +    {
>>>> +        struct ffa_ctx *vm = dom->arch.tee;
>>>> +
>>>> +        if (dom != d && vm != NULL && vm->guest_vers != 0)
>>>> +            ffa_vm_count++;
>>>> +    }
>>>> +    rcu_read_unlock(&domlist_read_lock);
>>>> +#endif
>>> 
>>> ...
>>> struct domain *dom;
>>> 
>>> if ( IS_ENABLED(CONFIG_FFA_VM_TO_VM) )
>>> {
>>>    /* Count the number of VM with FF-A support */
>>>    rcu_read_lock(&domlist_read_lock);
>>>    ...
>>>    rcu_read_unlock(&domlist_read_lock);
>>> }
>>> 
>>> drops the explicit ifdef.  Hiding function-level variable declarations
>>> behind an ifdef like that works exactly once, and it doesn't make
>>> pleasant code.
>> 
>> Ack I will fix that.
> 
> While at that, please also name struct domain * type variables "d", not
> "dom". For consistency with (almost) all other code we have.

Sure, I will fix that.

Cheers
Bertrand

> 
> Jan



Reply via email to