Bug #664
closed
Added by baptiste.ritter over 17 years ago.
Updated about 17 years ago.
Description
Hi all,
Here is an update of the patch, sent by Jonathan Buschmann on the kernel
mailing list (at the 14 march).
Differences between this patch and the first-one submitted on the kernel
mailing list :
- The announced IPv6 bug corrected (CARP interface which did not reply
to Neighbor Sollicitation)
- Some man pages was imported/updated
- LINT config file updated
- Code updated for "syncing" with CVS.
- Suppression of the code for "Report link states changes" witch was
in "conflict")
- The patch was generated (and tested) with the RSYNC/Devel snapshot
from May 20. (We had some difficulties with accessing the CVS
server for keeping our sources up to date. :-( )
- Code cleanup (nuke some commented-out code)
Notes :
- As Sacha Wildner suggest to Jonathan, we should add IFM_CARP to
ifmedia(4) man page. But, on reading other BSD man pages, they didn't
do that ; Should it be or not, and what add ? (According to the fact
that IFM_CARP is a "virtual" media")
- English is not our native language and the man page of PFSYNC from
DragonflyBSD is not the same than others BSD. So I added a paragraph
with CARP (imported from FreeBSD man page) ; It is just an example "in
the middle of nowhere in the man page" ; I find that a little bit
strange, may be there is a better choice...
- A lot of other things (I don't remember what...)
Regards,
Files
Never mind IFM_CARP. At the time I suggested this to Jonathan, I didn't
check NetBSD (which, as you noted, didn't put this in their ifmedia(4)).
Don't worry. Looks fine and we can sort the details out afterwards.
Let's see what Scott finds in testing the patch.
Sascha
On Tue, 29 May 2007 22:14:49 +0200
Baptiste Ritter <baptiste.ritter@ulp.u-strasbg.fr> wrote:
Hi,
yesterday I said that the patch applies cleanly. This is not true, I
get:
patch: ** malformed patch at line 931: @ -801,8 +821,10
@
in_arpinput(struct mbuf *m)
and it stops.
I just tried the patch on the two following tarballs:
- src-Release-1.10.tar.bz2 66315 KB 06.08.2007 13:11:00
- src-Devel.tar.bz2 66990 KB 08.08.2007 01:10:00
The both operation terminated with success.
May be, your patch files is damaged ; you should directly wget the file
from http://bugs.dragonflybsd.org/file261/carp-20070529BR01.patch
For information, the version of my patch software is 2.0-12u8-OpenBSD
and my md5sum of the "wgeted" file is fa0361adf7c2a34a92f7637278e44b3f
Regards,
On Wed, 08 Aug 2007 21:43:28 +0200
Baptiste Ritter <baptiste.ritter@ulp.u-strasbg.fr> wrote:
I was thinking the same and tried 3 different ways but looks like the
attachment got corrupted somehow, because even extracting from the mail
archive site's code failed, not to talk of FTP, Samba and other
protocols.
Either way, this works. Thank you, and I'll insult you next week if I
have errors. I did RTFM, so I know what I should do.
Se you soon,
Please update carp(4) SYNOPSIS to:
options CARP
pseudo-device carp
(not 'device carp').
-thomas
Done ; the new patch is attached.
Thanks Thomas for keeping a close watch to the patch.
Cheers,
It looks to me like this is ready enough to be committed to widen the
testing arena. Mr Ritter, do you think its good enough to go into the
tree? We can always apply more patches as they come in.
I'm sorry I can't test it myself, I'm overloaded with work at the moment.
-Matt
For me (and my colleagues Jonathan and Nicolas) the patch is good enough to be committed ;
Furthermore, we think that by committing the patch we should have more feedbacks.
And if there is a bug discovered, we are here for assuring the support and correct it.
Regards,
I have done some CARP testing; it worked as expected.
Test was using 3 pc's on LAN (2 CARP enabled & one observer (ping))
(I guess vkernels sould be just as good :).
Regarding carp(4) SYNOPSIS: I think this should include 'options CARP'.
This is used in VKERNEL & LINT
(my initial problem was that kernel didn't link w/o this).
It would be nice to have CARP GENERIC;
this not done in patch, but it is included in VKERNEL.
Thanks for porting this nice feature.
-thomas
On Wed, 15 Aug 2007 20:32:41 -0000
Thomas Nikolajsen <sinknull@crater.dragonflybsd.org> wrote:
That's my fault: I have patched with the older diff. The new one has
this fixed.
:For me (and my colleagues Jonathan and Nicolas) the patch is good enough to be committed ;
:Furthermore, we think that by committing the patch we should have more feedbacks.
:
:And if there is a bug discovered, we are here for assuring the support and correct it.
:
:Regards,
:
:--
:Baptiste Ritter
I've got it all ready to commit. I need to what attribution to put in
the commit message though.
Submitted-by: Baptiste Ritter <baptiste.ritter@ulp.u-strasbg.fr>,
Jonathan ?
Nicolas ?
Reviewed-by: Thomas Nikolajsen
Gergo Szakal <bastyaelvtars@gmail.com>
Obtained-from: OpenBSD, NetBSD, and FreeBSD
You can give me as little or as much information as you want on
your collegues.
-Matt
Matthew Dillon
<dillon@backplane.com>
:It would be nice to have CARP GENERIC;
:this not done in patch, but it is included in VKERNEL.
:
:Thanks for porting this nice feature.
:
: -thomas
I'm going to hold off making it the default in GENERIC and VKERNEL
until we can think that through a bit more.
Actually, you know what we REALLY need, we need a 'kernel' manual
page which describes all the really cool options kernels can be built
with in detail. LINT doesn't really do it justice.
-Matt
Matthew Dillon
<dillon@backplane.com>
On Wed, 15 Aug 2007 14:26:48 -0700 (PDT)
Matthew Dillon <dillon@apollo.backplane.com> wrote:
My role was testing. I cannot review this :-)
Patch #2 has been committed! I do not have it set as the default
for any kernels yet, though.
-Matt
Also available in: Atom
PDF