NetBSD-Bugs archive

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index][Old Index]

Re: kern/57155: OpenVPN (tap and tun) doesn't run as expected on 10.0_BETA



The following reply was made to PR kern/57155; it has been noted by GNATS.

From: Ryota Ozaki <ozaki-r%netbsd.org@localhost>
To: =?UTF-8?Q?BERTRAND_Jo=C3=ABl?= <joel.bertrand%systella.fr@localhost>
Cc: gnats-bugs%netbsd.org@localhost, kern-bug-people%netbsd.org@localhost, gnats-admin%netbsd.org@localhost, 
	netbsd-bugs%netbsd.org@localhost
Subject: Re: kern/57155: OpenVPN (tap and tun) doesn't run as expected on 10.0_BETA
Date: Thu, 5 Jan 2023 19:45:48 +0900

 On Thu, Jan 5, 2023 at 3:50 PM BERTRAND Jo=C3=ABl <joel.bertrand@systella.f=
 r> wrote:
 >
 > Ryota Ozaki a =C3=A9crit :
 > >  Oh, sorry. You also need to grep with "discard" to see all packet drop=
 s.
 >
 > Main server:
 >
 > legendre# netstat -s | grep discard
 >                 0 discarded for bad checksums
 >                 0 discarded for bad header offset fields
 >                 0 discarded because packet too short
 >         0 output packets discarded due to no route
 >         0 output packets discarded due to reject route
 >                 0 packets discarded for bad interface
 >                 0 packets discarded for wrong TTL
 >                 0 packets discarded for bad checksum
 >                 0 packets discarded with a bad version
 >                 0 discarded because packet was too short
 >                 0 packets discarded for bad authentication
 >                 0 packets discarded for bad vhid
 >                 0 packets discarded because of a bad address list
 >         276 output packets discarded due to no route
 >         0 packets discarded due to too many headers
 >         0 output packets discarded due to reject route
 >                 0 discarded for bad checksums
 >                 0 discarded for bad header offset fields
 >                 0 discarded because packet too short
 >
 > VM:
 >
 > netbsd-test1# netstat -s | grep discard
 >                 0 discarded for bad checksums
 >                 0 discarded for bad header offset fields
 >                 0 discarded because packet too short
 >         0 output packets discarded due to no route
 >         0 output packets discarded due to reject route
 >                 0 packets discarded for bad interface
 >                 0 packets discarded for wrong TTL
 >                 0 packets discarded for bad checksum
 >                 0 packets discarded with a bad version
 >                 0 discarded because packet was too short
 >                 0 packets discarded for bad authentication
 >                 0 packets discarded for bad vhid
 >                 0 packets discarded because of a bad address list
 >         0 output packets discarded due to no route
 >         0 packets discarded due to too many headers
 >         0 output packets discarded due to reject route
 >                 0 discarded for bad checksums
 >                 0 discarded for bad header offset fields
 >                 0 discarded because packet too short
 
 Thanks.
 
 "packets discarded due to no route" may be related to the issue,
 but I'm not sure for now.
 
 >
 >
 > >  Anyway thank you for the outputs and the explanation on the other repl=
 y.
 > >  I've got the situation.
 > >
 > >  I have one question. Does this issue happen with a vanilla kernel?
 > >  Or only with your custom kernel?
 >
 >         legendre runs with a customized kernel (GENERIC + altq + iscsi).
 > netbsd-test1 runs with a GENERIC kernel (from iso install image).
 
 I've tried a kernel with the same configuration. It seems to me that
 ALTQ and iscsi are not related to the issue.
 
 Anyway I've finally found a setup to reproduce the issue (or something).
 At least it's a regression.
 
 It's a tap version of your simple setup:
 [server] openvpn --dev tap --ifconfig 10.4.0.1 255.255.255.0 --verb 1
 --local 10.6.0.9 --proto tcp-server
 [client] openvpn --remote 10.6.0.9 --dev tap --ifconfig 10.4.0.2
 255.255.255.0 --proto tcp-client --verb 1 --float --ping 10
 
 It works with NetBSD 9 & Linux, however, doesn't work with NetBSD 10 & Linu=
 x
 and NetBSD 10 & NetBSD 9.
 
 I haven't reproduced the issue with tun but anyway I'm investigating the is=
 sue
 with the setup first.
 
 Thanks,
   ozaki-r
 


Home | Main Index | Thread Index | Old Index