[dm-devel] [PATCH] dm: Change return type of pg_init_all_paths() from int into void

[Bart Van Assche]

On 10/01/16 00:53, Hannes Reinecke wrote:
> On 09/30/2016 07:44 PM, Bart Van Assche wrote:
>> None of the callers of pg_init_all_paths() check its return value.
>> Hence change the return type of pg_init_all_paths() from int into
>> void.
>>
>> Signed-off-by: Bart Van Assche <bart.vanassche at sandisk.com>
>> ---
>>  drivers/md/dm-mpath.c | 7 ++-----
>>  1 file changed, 2 insertions(+), 5 deletions(-)
>>
>> diff --git a/drivers/md/dm-mpath.c b/drivers/md/dm-mpath.c
>> index dd66653..8e2341d 100644
>> --- a/drivers/md/dm-mpath.c
>> +++ b/drivers/md/dm-mpath.c
>> @@ -378,16 +378,13 @@ static int __pg_init_all_paths(struct multipath *m)
>>  	return atomic_read(&m->pg_init_in_progress);
>>  }
>>
>> -static int pg_init_all_paths(struct multipath *m)
>> +static void pg_init_all_paths(struct multipath *m)
>>  {
>> -	int r;
>>  	unsigned long flags;
>>
>>  	spin_lock_irqsave(&m->lock, flags);
>> -	r = __pg_init_all_paths(m);
>> +	__pg_init_all_paths(m);
>>  	spin_unlock_irqrestore(&m->lock, flags);
>> -
>> -	return r;
>>  }
>>
>>  static void __switch_pg(struct multipath *m, struct priority_group *pg)
>>
> Hmm. But this would invalidate your earlier patch, now would it not?
>
> And at least in my code we _do_ check the return value in
> pg_init_done(), and it's return value is used to check if we need to
> wake up for submitting waiting I/O.
>
> Which tree is the patch made against?

Hello Hannes,

This patch applies on top of kernel v4.8-rc8. In that kernel version the 
return value of __pg_init_all_paths() is used but the return value of 
pg_init_all_paths() is not used by any caller.

Bart.