Subject: Re: CVS commit: syssrc/sys/net
To: None <itojun@iijlab.net>
From: Atsushi Onoe <onoe@sm.sony.co.jp>
List: tech-net
Date: 09/26/2002 22:35:35
> m->m_type
> m->m_nextpkt
> are left uninitialized.
Yes, I know. Unlike Jason and Darren, I don't intend to make a complete
mbuf for bpf. It is still _hack_ anyway. Even if these members are also
initialized, the mbuf is still illegal, since it cannot be free..
I just think counting the whole length in bpf_mtap() is not necessary
most cases and make it easier by putting m_pkthdr.len.
But after scanning source tree for dummy mbuf, I would agree that the
change in bpf.c is too expensive than the effectiveness.
I won't modify other files I listed.
Atsushi Onoe