cvs commit: src/sbin/ipfw ipfw.8 ipfw2.c src/sys/netinet ip_divert.c ip_fw.h ip_fw2.c

[Brian Fundakowski Feldman]

On Fri, May 13, 2005 at 01:58:25AM +0400, Gleb Smirnoff wrote:
>   Brian,
> 
> On Thu, May 12, 2005 at 11:17:49AM -0400, Brian Fundakowski Feldman wrote:
> B> On Thu, May 12, 2005 at 03:11:30PM +0000, Brian Feldman wrote:
> B> > green       2005-05-12 15:11:30 UTC
> B> > 
> B> >   FreeBSD src repository
> B> > 
> B> >   Modified files:        (Branch: RELENG_5)
> B> >     sbin/ipfw            ipfw.8 ipfw2.c 
> B> >     sys/netinet          ip_divert.c ip_fw.h ip_fw2.c 
> B> >   Log:
> B> >   MFC: IPFW ALTQ(4) classification support, diverted traffic match rules,
> B> >        and the TCP packet data length match rule.
> B> >   
> B> >   Revision   Changes    Path
> B> >   1.150.2.7  +55 -2     src/sbin/ipfw/ipfw.8
> B> >   1.54.2.5   +259 -33   src/sbin/ipfw/ipfw2.c
> B> >   1.98.2.4   +16 -12    src/sys/netinet/ip_divert.c
> B> >   1.89.2.4   +17 -3     src/sys/netinet/ip_fw.h
> B> >   1.70.2.11  +92 -1     src/sys/netinet/ip_fw2.c
> B> 
> B> Gleb, would you like to merge the subsequent ip_divert.c changes?
> 
> Definitely, I would like!
> 
> You have broken ng_ksocket operation in RELENG_5 the same way you did it in
> HEAD. However, HEAD is for untested changes, so I saw no problem here. After
> some silence from you, I have fixed it in rev. 1.109.
> 
> However, RELENG_5 is for _tested_ changes, not for stepping on the same
> errors again. Since you asked me for merge, this means that you knew about
> breakage you are committing.
> 
> Why didn't you incorporated 1.109 into MFC?
> 
> If this moment is not clear to you, why didn't you send me
> email _before_ committing? I usually respond quickly.
> 
> What was the reason to rush in and break things in RELENG_5?
> 
> P.S. I can't quickly merge it now, because it needs testing and I have no time
> and testbox now at home at 2 PM. So this will be left unfixed until tomorrow. If
> you want, you can test and commit it yourself. The testcase is the same I've sent
> you long time ago:
> 
> http://lists.freebsd.org/pipermail/freebsd-net/2004-November/005562.html

I'm trying not to step on anyone's toes by merging "their" code
unless it's a dire emergency.  I noticed the issue in question after
doing the MFC and someone else saying something orthogonal to it,
then happening upon it in the CVS history.  It can be difficult to
figure out why someone did piecemeal merges, so after realizing what
happened, and considering that you were the only one affected, I
concluded that you would be the best person to sort it out.

So, in any case, sorry for missing a part of the MFC that is essential
to you.  I would prefer it if you would do what you propose and fix it
at your leisure, because it would take me a whole day just to figure
out how to work with netgraph again enough to know what it is that
I am verifying.  It looks like you've already MFCed the rest of the
changes that you had made just after those to ip_divert.c?

-- 
Brian Fundakowski Feldman                           \'[ FreeBSD ]''''''''''\
  <> green at FreeBSD.org                               \  The Power to Serve! \
 Opinions expressed are my own.                       \,,,,,,,,,,,,,,,,,,,,,,\