Description of problem:

netstat outputs the following error, once per client association address:

SCTP error in line: <address line num>

rather than the actual output, which can still be found in /proc/net/sctp/assocs

See below for a specific example.

NOTE: this was working properly in FC16




Version-Release number of selected component (if applicable):

FC17 3.3.4-5 x86_64


How reproducible:

100%



Steps to Reproduce:
1. Create a server listening on an SCTP (IPv4) socket
2. Connect from a client to the server on an SCTP (IPv4) socket
3. run "netstat --sctp" and notice that the association info (which can be found in /proc/net/sctp/assocs) results in errors from netstat
  


Actual results:

netstat generates "SCTP error in line: <address line num>" errors - one per line of association info (i.e. per address)



Expected results:

A human-readable interpretation of the data in /proc


Additional info:

This issue appears to have been introduced with FC17 as it was working 
properly with FC16.

Basically, when listing SCTP associations via netstat (i.e. "netstat 
--sctp"), an error is displayed (see below).  Servers appear to be listed 
without issue - although that could just be due to the specific parameters 
of my application.  In any case, the servers and associations appear to be 
listed properly in /proc/net/sctp/eps and /proc/net/sctp/assocs, so 
presumably (?) the issue is just with netstat's interpretation of the data 
in /proc.

Example:

After starting an SCTP server:

the netstat output is correct:

[peter@peter ~]$ netstat -nap --sctp
(Not all processes could be identified, non-owned process info
 will not be shown, you would have to be root to see it all.) 
Active Internet connections (servers and established) 
Proto Recv-Q Send-Q Local Address           Foreign Address         State  
    PID/Program name 
sctp       0      0 10.91.20.100:4105       0.0.0.0:*               LISTEN 
     1602/server


and it matches that in /proc:

[peter@peter ~]$ cat /proc/net/sctp/eps 
 ENDPT     SOCK   STY SST HBKT LPORT   UID INODE LADDRS 
ffff880222fdac00 ffff880221d504c0 2   10  9    4105   1000 27667 
10.91.20.100 10.91.10.100 
 
[peter@peter ~]$ cat /proc/net/sctp/assocs 
 ASSOC     SOCK   STY SST ST HBKT ASSOC-ID TX_QUEUE RX_QUEUE UID INODE 
LPORT RPORT LADDRS <-> RADDRS HBINT INS OUTS MAXRT T1X T2X RTXC



Whereas after an association is established by a client:

the netstat output reports an error where the two client lines would 
normally be output (2 lines due to 2 addresses on client side of 
association):

[peter@peter ~]$ netstat -nap --sctp
(Not all processes could be identified, non-owned process info
 will not be shown, you would have to be root to see it all.)
Active Internet connections (servers and established)
Proto Recv-Q Send-Q Local Address           Foreign Address         State  
    PID/Program name
sctp       0      0 10.91.20.100:4105       0.0.0.0:*               LISTEN 
     1602/server
SCTP error in line: 1
SCTP error in line: 2


the server entry matches that in /proc:

[peter@peter ~]$ cat /proc/net/sctp/eps
 ENDPT     SOCK   STY SST HBKT LPORT   UID INODE LADDRS
ffff880222fdac00 ffff880221d504c0 2   10  9    4105   1000 27667 
10.91.20.100 10.91.10.100


and the client entries are also reported in /proc (even though they yielded the aforementioned errors when using netstat):

[peter@peter ~]$ cat /proc/net/sctp/assocs
 ASSOC     SOCK   STY SST ST HBKT ASSOC-ID TX_QUEUE RX_QUEUE UID INODE 
LPORT RPORT LADDRS <-> RADDRS HBINT INS OUTS MAXRT T1X T2X RTXC
ffff880223e51000 ffff880222fd0000 2   1   3  6426    8        0        0  
1000 10777 4105   4106  10.91.20.100 10.91.10.100 <-> *10.91.0.100   30000 
   10    10    1    0    0        0
ffff880223e56800 ffff880221d50980 2   1   3  6681    7        0        0  
1000 27692 4106   4105  10.91.0.100 <-> *10.91.20.100 10.91.10.100   30000 
   10    10 59885    0    0        0