Re: ppp routing bug ?

2017-10-30 Thread Eugene Grosbein
On 30.10.2017 20:57, Mike Tancsa wrote: > On 10/26/2017 6:16 PM, Eugene Grosbein wrote: >> >> That makes sense: ppp send bogus request to the routing socket and >> the request has not RTF_HOST flag nor RTA_NETMASK address. >> It seems, earlier kernel code masked this bug somehow but it does not now

Re: ppp routing bug ?

2017-10-30 Thread Mike Tancsa
On 10/26/2017 6:16 PM, Eugene Grosbein wrote: > > That makes sense: ppp send bogus request to the routing socket and > the request has not RTF_HOST flag nor RTA_NETMASK address. > It seems, earlier kernel code masked this bug somehow but it does not now. > > Anyway, we have two bugs here: ppp sen

Re: ppp routing bug ?

2017-10-27 Thread Mike Tancsa
On 10/26/2017 6:16 PM, Eugene Grosbein wrote: > 27.10.2017 1:04, Mike Tancsa пишет: >> >> the borked route monitor looks like >> >> got message of size 124 on Thu Oct 26 12:21:12 2017 >> RTM_ADD: Add Route: len 124, pid: 29597, seq 2, errno 0, >> flags: >> locks: inits: >> sockaddrs: >> 192.168.

Re: ppp routing bug ?

2017-10-26 Thread Eugene Grosbein
27.10.2017 1:04, Mike Tancsa пишет: > On 10/26/2017 12:01 PM, Eugene Grosbein wrote: >> >> I would re-run ppp under ktrace to make sure while having "route monitor" >> running around. >> Then compare pids with kdump output. > > I wonder if I copied and pasted from 2 different test sessions. Anywa

Re: ppp routing bug ?

2017-10-26 Thread Mike Tancsa
On 10/26/2017 12:01 PM, Eugene Grosbein wrote: > > I would re-run ppp under ktrace to make sure while having "route monitor" > running around. > Then compare pids with kdump output. I wonder if I copied and pasted from 2 different test sessions. Anyways, same PID when I re-ran the test and more

Re: ppp routing bug ?

2017-10-26 Thread Eugene Grosbein
26.10.2017 21:25, Mike Tancsa wrote: > On 10/26/2017 9:59 AM, Eugene Grosbein wrote: >> On 26.10.2017 20:45, Mike Tancsa wrote: >> >>> vs >>> >>> got message of size 124 on Thu Oct 26 09:37:40 2017 >>> RTM_ADD: Add Route: len 124, pid: 24236, seq 1, errno 0, >>> flags: >>> locks: inits: >>> socka

Re: ppp routing bug ?

2017-10-26 Thread Mike Tancsa
On 10/26/2017 9:59 AM, Eugene Grosbein wrote: > On 26.10.2017 20:45, Mike Tancsa wrote: > >> vs >> >> got message of size 124 on Thu Oct 26 09:37:40 2017 >> RTM_ADD: Add Route: len 124, pid: 24236, seq 1, errno 0, >> flags: >> locks: inits: >> sockaddrs: >> 192.168.136.1 64.7.128.7 >> >> got me

Re: ppp routing bug ?

2017-10-26 Thread Eugene Grosbein
On 26.10.2017 20:45, Mike Tancsa wrote: > vs > > got message of size 124 on Thu Oct 26 09:37:40 2017 > RTM_ADD: Add Route: len 124, pid: 24236, seq 1, errno 0, > flags: > locks: inits: > sockaddrs: > 192.168.136.1 64.7.128.7 > > got message of size 196 on Thu Oct 26 09:37:40 2017 > RTM_CHANGE

Re: ppp routing bug ?

2017-10-26 Thread Mike Tancsa
On 10/26/2017 6:17 AM, Eugene Grosbein wrote: >> >> >> Any ideas ? > > Are you sure that: Hi, Yes I have rebuilt all a number of times. Also, this is i386. I think I will try and build and AMD64 image to see if it follows there too. > > Try using "route -n monitor" in the background t

Re: ppp routing bug ?

2017-10-26 Thread Eugene Grosbein
On 26.10.2017 04:10, Mike Tancsa wrote: > On 10/25/2017 4:54 PM, Mike Tancsa wrote: >> While testing out our nanobsd RELENG11 images, I noticed a strange >> routing issue. Using a standard pppoe config that has been working fine >> on RELENG8,9,10 I am getting bogus routing entries. >> having the

Re: ppp routing bug ?

2017-10-25 Thread Mike Tancsa
On 10/25/2017 4:54 PM, Mike Tancsa wrote: > While testing out our nanobsd RELENG11 images, I noticed a strange > routing issue. Using a standard pppoe config that has been working fine > on RELENG8,9,10 I am getting bogus routing entries. > having the commands in ppp.linkup > > pppoe: > add 192.