Re: mbuf initialization macros

[Kengo NAKAHARA <k-nakahara%iij.ad.jp@localhost>]

Hi,

On 2016/04/19 13:29, Kengo NAKAHARA wrote:
> On 2016/04/19 0:51, Christos Zoulas wrote:
>> In article <571463BF.30607%iij.ad.jp@localhost>,
>> Kengo NAKAHARA  <k-nakahara%iij.ad.jp@localhost> wrote:
>>
>>> @@ -867,6 +867,10 @@ int        m_append(struct mbuf *, int, const void *);
>>> /* Inline routines. */
>>> static __inline u_int m_length(const struct mbuf *) __unused;
>>>
>>> +static __inline void m_hdr_init(struct mbuf *, short, struct mbuf *,
>>> +                               char *, int);
>>> +static __inline void m_pkthdr_init(struct mbuf *);
>>> +
>>
>> Looks good, but you should not need the above... Perhaps remove all
>> the static __inline decls. Is the __unused needed? Can it be moved in
>> the definition?
> 
> Oh, sorry, I was confused by the existing m_length() declaration.
> I remove my static __inline declarations.
> 
> Hmm, the static __inline m_length() declaration is added by mbuf.h r1.92
> commit and __unused attribute is added by mbuf.h r1.97 commit. It seems
> the declaration is not needed as long as I read the commit message.
> # of course, GENERIC kernel can be built without the declaration
> 
> Can I remove the m_length() declaration before adding mbuf initialization
> functions?

I change my policy according to kre@n.o's advice. I leave the declaration.
I leave it to someone(tm)...


Thanks,

-- 
//////////////////////////////////////////////////////////////////////
Internet Initiative Japan Inc.

Device Engineering Section,
IoT Platform Development Department,
Network Division,
Technology Unit

Kengo NAKAHARA <k-nakahara%iij.ad.jp@localhost>