On Nov 2,  6:13pm, bouyer@antioche.lip6.fr (Manuel Bouyer) wrote:
-- Subject: kern/1515 [Re: PR-7148 (causes kernel panic)]

| Hi,
| here are diffs for the 64bit interface counter. I basically followed
| Christos's suggestion about versionning the sysctl and routing socket message.
| The thing I don't know how to deal with is rt_ifmsg(). If it sends both
| messages (RTM_OIFINFO and RTM_IFINFO) it will complain about out-of-sync
| ifaddr, but I think it only does at startup (not checked yet).
| I didn't check what gated does yet, I will tomorow.
| 
| I found a mail from Darren Reed on the topic (tech-net, Mon, 7 Jun 1999
| 17:04:11 +1000 (EST), <199906070704.RAA00417@avalon.reed.wattle.id.au>,
| "Re: kern/1515 - Changing netstat variables to 64 bit") whith a patch
| included where is increases the per-protocols stats to 64bit as well.
| As this is used only by kmem growvelers no compat problem here :),
| so I did it as well (and for ip6 too).
| 

Looks fine to me, two points:
1. why 'struct if_odata' as opposed to 'struct if_data14'?
2. what is the decision between %qd and %lld? Isn't the second more
   portable?

Regards,

christos