Dan Price wrote:
> On Fri 13 Nov 2009 at 02:24PM, Garrett D'Amore - sun microsystems wrote:
>>   63: S_IFREG mode:0600 dev:400,1 ino:254311 uid:23805 gid:10 size:24000
>>       O_RDWR|O_APPEND|O_CREAT|O_LARGEFILE FD_CLOEXEC
>>       /home/dmick/.sh_history offset:24000 
> 
> +1 from me.  I have wanted this many times-- I'm only sorry I didn't
> implement it years ago!  My only question would be to Roger-- whether he
> intended for the filename to live alone on a line, or could offer
> guidance about where offset: should live.
> 
> But I don't really know where else to stick this info.  The first line
> seems to be properties of the file.  The next, properties of the way the
> fd was opened (or fcntl'd).  So, it's not obvious that either of these
> lines would be right either.
> 
> It would help to see an example of the output for some more exotic types
> of things, such as sockets and doors-- I assume that for some of these,
> this output is elided?  How about a pfiles of, say, inetd?
> 
>         -dp
> 

Here are all the offsets from my inetd:

       /devices/pseudo/mm at 0:null offset:0
       /var/svc/log/network-inetd:default.log offset:0
       /var/svc/log/network-inetd:default.log offset:0
       /devices/pseudo/sysevent at 0:sysevent offset:0
       /devices/pseudo/sysevent at 0:sysevent offset:0
       /devices/pseudo/tl at 0:ticotsord offset:0
       /devices/pseudo/tl at 0:ticots offset:0
       /devices/pseudo/tl at 0:ticotsord offset:0
       /devices/pseudo/tcp at 0:tcp offset:0
       /devices/pseudo/tcp6 at 0:tcp6 offset:0
       /devices/pseudo/tl at 0:ticlts offset:0
       /devices/pseudo/tl at 0:ticotsord offset:0
       /devices/pseudo/tl at 0:ticotsord offset:0
       /devices/pseudo/tl at 0:ticlts offset:0
       /devices/pseudo/udp at 0:udp offset:0
       /devices/pseudo/udp6 at 0:udp6 offset:0
       /proc/1019/psinfo offset:0

so basically none of them are 'useful'.  It's possible that the information 
that 
none of the devices increment a seek offset is itself useful, though, which is 
why I decided to output them rather than try to invent a heuristic for which 
dev 
types mattered and which didn't.

Reply via email to