Re: git: 637c0bace261 - main - sysutils/pftop: Fix build on 14.0-CURRENT
- In reply to: Michael Gmelin : "Re: git: 637c0bace261 - main - sysutils/pftop: Fix build on 14.0-CURRENT"
- Go to: [ bottom of page ] [ top of archives ] [ this month ]
Date: Sun, 11 Jun 2023 12:40:53 UTC
On Sun, 11 Jun 2023 12:12:44 +0200
Michael Gmelin <grembo@freebsd.org> wrote:
> > On 11. Jun 2023, at 12:00, Herbert J. Skuhra <herbert@gojira.at>
> > wrote: On Sun, 11 Jun 2023 11:25:45 +0200, Michael Gmelin wrote:
> >>
> >>
> >>
> >>> On 11. Jun 2023, at 10:27, Herbert J. Skuhra <herbert@gojira.at>
> >>> wrote: On Sat, 10 Jun 2023 12:06:09 +0200,
> >>> Michael Gmelin <grembo@FreeBSD.org> wrote:
> >>>> The branch main has been updated by grembo:
> >>>> URL:
> >>>> https://cgit.FreeBSD.org/ports/commit/?id=637c0bace26138529a36232e948549ad59342ba9
> >>>> commit 637c0bace26138529a36232e948549ad59342ba9 Author:
> >>>> Michael Gmelin <grembo@FreeBSD.org> AuthorDate: 2023-06-10
> >>>> 10:03:39 +0000 Commit: Michael Gmelin <grembo@FreeBSD.org>
> >>>> CommitDate: 2023-06-10 10:03:39 +0000
> >>>> sysutils/pftop: Fix build on 14.0-CURRENT
> >>>> ---
> >>>> sysutils/pftop/Makefile | 10 ++++++++--
> >>>> sysutils/pftop/files/extra-patch-config.h | 6 +++++-
> >>>> 2 files changed, 13 insertions(+), 3 deletions(-)
> >>>> diff --git a/sysutils/pftop/Makefile b/sysutils/pftop/Makefile
> >>>> index f3c6d879f637..cba2ecd65aeb 100644
> >>>> --- a/sysutils/pftop/Makefile
> >>>> +++ b/sysutils/pftop/Makefile
> >>>> @@ -1,6 +1,6 @@
> >>>> PORTNAME= pftop
> >>>> PORTVERSION= 0.8
> >>>> -PORTREVISION= 2
> >>>> +PORTREVISION= 3
> >>>> CATEGORIES= sysutils net
> >>>> MAINTAINER= grembo@FreeBSD.org
> >>>> @@ -22,7 +22,13 @@ EXTRA_PATCHES+=
> >>>> ${FILESDIR}/extra-patch-bpf_dump.c \
> >>>> ${FILESDIR}/extra-patch-sf-gencode.h MAKE_ARGS=
> >>>> LOCALBASE="${PREFIX}" \
> >>>> - OSLEVEL=45
> >>>> +
> >>>> +.if ${OPSYS} == FreeBSD && ${OSVERSION} >= 1400090
> >>>> +MAKE_ARGS+= OSLEVEL=46
> >>>> +.else
> >>>> +MAKE_ARGS+= OSLEVEL=45
> >>>> +.endif
> >>>> +
> >>>> CFLAGS+= -DHAVE_SNPRINTF=1 -DHAVE_VSNPRINTF=1 \
> >>>> -DHAVE_FINE_GRAINED_LOCKING=1
> >>>> diff --git a/sysutils/pftop/files/extra-patch-config.h
> >>>> b/sysutils/pftop/files/extra-patch-config.h index
> >>>> 6d2873c42ab1..d24f88179718 100644 ---
> >>>> a/sysutils/pftop/files/extra-patch-config.h +++
> >>>> b/sysutils/pftop/files/extra-patch-config.h @@ -1,7 +1,7 @@
> >>>> $OpenBSD: patch-config_h,v 1.4 2008/12/20 04:36:11 canacar Exp $
> >>>> --- config.h.orig Tue Nov 6 22:34:18 2007
> >>>> +++ config.h Fri Dec 19 20:28:01 2008
> >>>> -@@ -74,11 +74,20 @@
> >>>> +@@ -74,11 +74,24 @@
> >>>> #define HAVE_PFSYNC_STATE
> >>>> #endif
> >>>> @@ -11,7 +11,11 @@ $OpenBSD: patch-config_h,v 1.4 2008/12/20
> >>>> 04:36:11 canacar Exp $ +#endif
> >>>> +
> >>>> #ifdef HAVE_PFSYNC_STATE
> >>>> ++#if OS_LEVEL > 45
> >>>> ++typedef struct pfsync_state_1400 pf_state_t;
> >>> Are you sure that this is correct?
> >>> If I replace pfsync_state_1400 with pfsync_state_1301 the port
> >>> builds and the output looks sane.
> >>
> >> Hi, thanks for reporting, could you please add some details (like,
> >> how the output differs)?
> >
> > With your change:
> >
> > sctp Out (null)[13715] (null)[0]
> > 0:255 0 * * * 237 In (null)[0]
> > (null)[0] NO_TRAFFIC:NO_TRAFFIC 0 0
> > * * ip In (null)[0] (null)[0]
> > NO_TRAFFIC:NO_TRAFFIC 0 9324h * * ip In
> > (null)[7185] (null)[512] NO_TRAFFIC:NO_TRAFFIC
> > 0 53 * 2048G ip In (null)[4097]
> > (null)[49408] NO_TRAFFIC:NO_TRAFFIC 0 0 0
> > 1 ip In (null)[9732] (null)[4992]
> > 9:0 0 0 42 73728M cpnx In (null)[0]
> > (null)[0] 255:0 0 *
> > * * ipenca In (null)[0] (null)[0]
> > NO_TRAFFIC:NO_TRAFFIC 0 0 * * ip In
> > (null)[0] (null)[0] 0:9
> > * * * *
> >
> > Garbage?
> >
> > With attached patch I see ipv[46] addresses and port numbers again.
> > :-)
> >
> > I am running main-n263493-4e8d558c9d1c.
> >
>
> The question is if pftop shouldn’t use pfsync_state_1400, or if there
> is a kernel problem with that struct. I’ll look into it.
>
> Thanks
Hi Herbert,
Apparently what happened is that I tested using 1301, which worked,
then changed to 1400, which also seemed to work and committed the
change. Apparently I tested the wrong binary, as rebuilding and testing
I see the same results as you do.
Therefore I'll switch to using 1301, thanks again for reporting.
Best
Michael
--
Michael Gmelin