sockstat: fix column length for PROTO

[tuexen]

Summary

The computation of the length was not taking into account that IPv6 endpoints, which are not IPv6 only, have a suffix of 46.

[asomers]

Can you give an example of the incorrect output? Does it matter whether "-q" is in use?

[tuexen]

In D53212#1215533, @asomers wrote:

Can you give an example of the incorrect output? Does it matter whether "-q" is in use?

Here is the output without the fix:

tuexen@ampere32:~ % sockstat -Psctp
USER   COMMAND     PID FD PROTO LOCAL ADDRESS         FOREIGN ADDRESS      
nobody thttpd     3572  4 sctp46 10.0.1.86:80          *:*                  
??     ??           ?? ?? ??    192.168.1.86:80       ??                   
??     ??           ?? ?? ??    2a02:c6a0:4015:11::86 ??                   
??     ??           ?? ?? ??    fe80::6a05:caff:fe92: ??                   
??     ??           ?? ?? ??    212.201.121.86:80     ??                   
??     ??           ?? ?? ??    127.0.0.1:80          ??                   
??     ??           ?? ?? ??    fe80::1%lo0:80        ??                   
??     ??           ?? ?? ??    ::1:80                ??

Here is the output without the fix and using -q:

tuexen@ampere32:~ % sockstat -qPsctp
nobody thttpd     3572  4 sctp46 10.0.1.86:80          *:*                  
??     ??           ?? ?? ??    192.168.1.86:80       ??                   
??     ??           ?? ?? ??    2a02:c6a0:4015:11::86 ??                   
??     ??           ?? ?? ??    fe80::6a05:caff:fe92: ??                   
??     ??           ?? ?? ??    212.201.121.86:80     ??                   
??     ??           ?? ?? ??    127.0.0.1:80          ??                   
??     ??           ?? ?? ??    fe80::1%lo0:80        ??                   
??     ??           ?? ?? ??    ::1:80                ??

In both cases the problem is that sctp46 is too long and you have the misalignment in the first row. In case sctp4 and sctp6 it would look OK.
With the fix the output is:

tuexen@ampere32:~ % sockstat -Psctp
USER   COMMAND     PID FD PROTO  LOCAL ADDRESS         FOREIGN ADDRESS      
nobody thttpd     3572  4 sctp46 10.0.1.86:80          *:*                  
??     ??           ?? ?? ??     192.168.1.86:80       ??                   
??     ??           ?? ?? ??     2a02:c6a0:4015:11::86 ??                   
??     ??           ?? ?? ??     fe80::6a05:caff:fe92: ??                   
??     ??           ?? ?? ??     212.201.121.86:80     ??                   
??     ??           ?? ?? ??     127.0.0.1:80          ??                   
??     ??           ?? ?? ??     fe80::1%lo0:80        ??                   
??     ??           ?? ?? ??     ::1:80                ??

The problem is that the PROTO column is computed with:

snprintf(buf, bufsize, "%s%s%s",
        s->protoname,
        s->vflag & INP_IPV4 ? "4" : "",
        s->vflag & INP_IPV6 ? "6" : "");

whereas the length is computed by:

len = strlen(s->protoname);
if (s->vflag & (INP_IPV4 | INP_IPV6))
        len += 1;

So in the 46 case, you are off by one character.

[asomers]

That explains it. I was never testing with SCTP, and I doubt that Damin was, either.

[tuexen]

In D53212#1215579, @asomers wrote:

That explains it. I was never testing with SCTP, and I doubt that Damin was, either.

the same issue will come up with udplite (udplite4, udplite6, and udplite46), since the protocol name can be longer than PROTO.
I plan to add this soon.