Patch Set 1:

> If log_target_create_gsmtap() is called with add_sink==true (which
 > it is from logging_vty.c) then the GSMTAP packets are both
 > originated and received by the program.  No additional tool should
 > be required.

What about sysmoBTS sending gsmtap to a different machine, and possibly also 
osmo-x sending gsmtap to a different local interface?

In my practical experience (all of osmo-bts-sysmo and all bsc + core net 
components configured to do gsmtap logging) that's not covered and I got 50% 
ICMP in my pcap with considerable amount of GSMTAP dropped, until I started 
this tool to listen on 'any'.

I grant that I wasn't aware of the "originating program" field in gsmtap to 
filter by, and started out sending gsmtap to a different 127.0.0.* address per 
program for filtering. So now the scenario is sysmoBTS sending to 
and the components on my laptop sending to IIUC I'll need at least 
one additional sink to cover both?

 > would argue it shouldn't just print the "." but should actually
 > print the log messages to stdout/stderr?

Granted, but I see that as an additional feature to add later. I see all of it 
in wireshark, so I really just want to save me the ICMP and packet drops so 
far... The dots are just a gimmick to see that something is happening, and not 
the default behavior.

I can drop it or shift it to noinst if you think all this is dumb :)

To view, visit
To unsubscribe, visit

Gerrit-MessageType: comment
Gerrit-Change-Id: I85499af06bc5a0d757f6071779075e57bb435930
Gerrit-PatchSet: 1
Gerrit-Project: libosmocore
Gerrit-Branch: master
Gerrit-Owner: Neels Hofmeyr <>
Gerrit-Reviewer: Harald Welte <>
Gerrit-Reviewer: Jenkins Builder
Gerrit-Reviewer: Neels Hofmeyr <>
Gerrit-HasComments: No

Reply via email to