Re: m_get_rcvif and symmetry

[Ryota Ozaki <ozaki-r%netbsd.org@localhost>]

On Sat, Feb 4, 2017 at 2:04 AM, J. Lewis Muir <jlmuir%imca-cat.org@localhost> wrote:
> On 02/03, Ryota Ozaki wrote:
>> During fixing the usages I remember what I wanted to achieve by the API
>> and found what I should fix is m_get_rcvif, not m_put_rcvif.
>>
>> The change is below. By the change, a caller must call m_put_rcvif only if
>> a returned rcvif is non-NULL. If rcvif is NULL, a caller don't need to call
>> it while a caller can call it safely. So codes of callers sometimes can be
>> simplified. And the behavior is the same as other APIs such as
>> m_get_rcvif_psref/m_put_rcvif_psref and if_get/if_put.
>
> Just a few English grammar suggestions for the comment:
>
>> --- a/sys/sys/mbuf.h
>> +++ b/sys/sys/mbuf.h
>> @@ -1006,16 +1006,23 @@ void m_print(const struct mbuf *, const char
>> *, void (*)(const char *, ...)
>>  /*
>>   * Get rcvif of a mbuf.
>>   *
>> - * The caller must call m_put_rcvif after using rcvif. The caller cannot
>> - * block or sleep during using rcvif. Insofar as the constraint is satisfied,
>> - * the API ensures a got rcvif isn't be freed until m_put_rcvif is called.
>> + * The caller must call m_put_rcvif after using rcvif if the returned rcvif
>> + * isn't NULL. If the return rcvif is NULL, the caller doesn't need to call
>
> s/return/returned/
>
>> + * m_put_rcvif (calling it is safe). The caller cannot block or sleep during
>
> s/calling it is safe/although calling it is safe/
> s/during/while/
>
>> + * using rcvif. Insofar as the constraint is satisfied, the API ensures a got
>
> s/Insofar as the constraint is satisfied, t/T/
> s/got/returned/
>
>> + * rcvif isn't be freed until m_put_rcvif is called.
>
> s/be//

Thanks! I revised the comment.

  ozaki-r