Rare project developers

[mc36 <>]

hi,

On 6/14/22 11:27, Alexander Gall wrote:
> On Tue, 14 Jun 2022 11:09:24 +0200, mc36 <> said:
>
> > hi,
> > On 6/14/22 10:20, Alexander Gall wrote:
> > > That doesn't happen. When the port is configured, the oper state is
> > > never reported back. All I see is (apart from some other "rx" messages
> > > unrelated to the port state)
> > so it starts getting more and more interesting:
>
>
> > sid(cfg)#show interfaces sdn1 | first 1
> > % no such interface
> > sid(cfg)#interface sdn1
> > sid(cfg-if)#show interfaces sdn1 | first 1
> > sdn1 is down
>
> > sid(cfg-if)#^Z
> > sid(cfg)#server p4lang p4
> > sid(cfg-server)#show interfaces sdn1 | first 1
> > sdn1 is down
>
> > sid(cfg-server)#export-port sdn1 1
> > sid(cfg-server)#show interfaces sdn1 | first 1
> > sdn1 is down
>
> > sid(cfg-server)#show interfaces sdn1 | first 1
> > sdn1 is down
>
> > sid(cfg-server)#info prtTcp.connectionRcvd:prtTcp.java:756 got future 
> > acknowledge number p4lang loopback0 9080 -> 10.8.255.1 37192
> > debug servP4langConn.doNegot:servP4langConn.java:184 fwd0 rx: dynrange 666 
> > 6666
> > debug servP4langCfg.sendLine:servP4langCfg.java:1000 fwd0 tx: ports_add 1 0 0 
> > 0 0
> > warning servP4langConn.doNegot:servP4langConn.java:239 neighbor 10.8.255.1 up
> > debug servP4langCfg.sendLine:servP4langCfg.java:1000 fwd0 tx: state 1 1 0 0 0 0
> > debug servP4langCfg.sendLine:servP4langCfg.java:1000 fwd0 tx: mtu 1 1500
> > debug servP4langCfg.sendLine:servP4langCfg.java:1000 fwd0 tx: mylabel4_add 
> > 1025147 0
> > sid(cfg-server)#
> > sid(cfg-server)#show interfaces sdn1 | first 1
> > sdn1 is down
>
> > sid(cfg-server)#debug servP4langConn.doReports:servP4langConn.java:256 fwd0 
> > rx: state 1 1
> > sid(cfg-server)#
> > sid(cfg-server)#show interfaces sdn1 | first 1
> > sdn1 is up
>
> > sid(cfg-server)#
>
> This is on an ASIC, not the model, right? The port actually reports
> going UP, which doesn't happen on the model (at least for me). If this
> *is* on the model, I'd like to understand why it behaves differently
> from mine.
so this test was done by hand ( spot the dynrange 666 6666 :))) )
just to validate freerouter behaviour... now i spin up my model vm
on a test case: java -Xmx512m -jar rtr.jar test tester p4lang-rout001 other 
p4lang301.ini wait
and as you can see i got 260 state messages almost immediately after the 
forwarder started:


r1#show config-differences
server p4lang p4
 api-stat
 exit
r1#show p4lang p4 api-rx
message       count  last                 ago
capabilities  1      2022-06-14 09:41:33  00:02:21
counter       2200   2022-06-14 09:43:53  00:00:00
dynrange      1      2022-06-14 09:41:33  00:02:21
platform      1      2022-06-14 09:41:33  00:02:21
portname      128    2022-06-14 09:41:33  00:02:21
state         260    2022-06-14 09:41:34  00:02:20

r1#



> > that is, freerouter, up until it gets the state 1 1 reports the port
> > down, that is,
>
> 1 is "up", no?
yess, that's it....

so i still keep my opinion, if you got any ping messages, i got the "state 
<portid> 1" message,
because if a port stays down, then bfforwarder will send nothing then 
freerouter will leave the
sdnX in down state....

so, could you please try the debug serv p4 rx to see if it reports or not to 
freerouter?

thanks,
cs


> > unless that message, it's impossible that you pinged anything successfully????
>
> I can actually ping as long as freertr thinks the port is up. That's
> because the model doesn't care about the port oper status and packets
> can flow in both directions. They only get dropped by freertr once it
> believes the port is down and re-programs the tables.