Date: Wed, 8 Mar 2006 10:12:10 +0300 From: Yar Tikhiy <yar@comp.chem.msu.su> To: Max Laier <mlaier@FreeBSD.org> Cc: freebsd-pf@FreeBSD.org Subject: Re: cvs commit: src/etc/rc.d pflog src/sys/contrib/pf/net if_pflog.c if_pflog.h pf_ioctl.c src/sys/modules Makefile src/sys/modules/pf Makefile src/sys/modules/pflog Makefile Message-ID: <20060308071210.GA99290@comp.chem.msu.su> In-Reply-To: <200603061610.k26GAJKR071335@repoman.freebsd.org> References: <200603061610.k26GAJKR071335@repoman.freebsd.org>
next in thread | previous in thread | raw e-mail | index | archive | help
[moving this thread to freebsd-pf]
On Mon, Mar 06, 2006 at 04:10:19PM +0000, Max Laier wrote:
> mlaier 2006-03-06 16:10:19 UTC
>
> FreeBSD src repository
>
> Modified files: (Branch: RELENG_6)
> etc/rc.d pflog
> sys/contrib/pf/net if_pflog.c if_pflog.h pf_ioctl.c
> sys/modules Makefile
> sys/modules/pf Makefile
> Added files: (Branch: RELENG_6)
> sys/modules/pflog Makefile
> Log:
> MFC:
> Make pflog a seperate module. As a result pflog_packet() becomes a
> function pointer that is declared in pf_ioctl.c
>
> Requested by: yar (as part of the module build reorg)
>
> Approved by: re (scottl)
>
> Revision Changes Path
> 1.5.2.2 +3 -3 src/etc/rc.d/pflog
> 1.13.2.2 +8 -1 src/sys/contrib/pf/net/if_pflog.c
> 1.6.2.1 +14 -0 src/sys/contrib/pf/net/if_pflog.h
> 1.20.2.2 +5 -0 src/sys/contrib/pf/net/pf_ioctl.c
> 1.450.2.12 +2 -0 src/sys/modules/Makefile
> 1.7.2.1 +0 -3 src/sys/modules/pf/Makefile
> 1.4.2.1 +27 -0 src/sys/modules/pflog/Makefile (new)
Thanks a lot!
BTW, are the DEV_PF and DEV_PFLOG #defines really called for in the
Makefiles? DEV_PF is not used explicitly in the whole /sys while
the explicit use of DEV_PFLOG has been eliminated from our kernel
by this change. When the kernel is built w/o static pf or pflog,
they will be undefined. I'd rather remove the #defines from the
Makefiles when doing my part of the job, integrating KERNBUILDDIR
stuff from HEAD to RELENG_6 in pf/Makefile. Just tested whether
the status of DEV_PF or DEV_PFLOG would affect the pf.ko or pflog.ko
binaries -- the result was negative.
I also fail to recall if I proposed the attached patch to get rid
of the inclusion of <net/pfvar.h> from if_pflog.h. Both pf.ko and
pflog.ko build well with this patch applied, and it's a well-known
approach in general.
--
Yar
Index: if_pflog.h
===================================================================
RCS file: /home/ncvs/src/sys/contrib/pf/net/if_pflog.h,v
retrieving revision 1.7
diff -u -p -r1.7 if_pflog.h
--- if_pflog.h 5 Feb 2006 17:17:32 -0000 1.7
+++ if_pflog.h 8 Mar 2006 07:05:11 -0000
@@ -71,8 +71,9 @@ struct old_pfloghdr {
#ifdef _KERNEL
#ifdef __FreeBSD__
-/* XXX */
-#include <net/pfvar.h>
+struct pf_rule;
+struct pf_ruleset;
+struct pfi_kif;
typedef int pflog_packet_t(struct pfi_kif *, struct mbuf *, sa_family_t,
u_int8_t, u_int8_t, struct pf_rule *, struct pf_rule *,
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20060308071210.GA99290>