I have a fairly large but mostly simple patch which changes the stats collected in uvmexp for faults, intrs, softs, syscalls, and traps from 32 bit to 64 bits and puts them in cpu_data (in cpu_info). This makes more accurate and a little cheaper to update on 64bit systems. I hvaen't looked at your patch, but I noticed that in netbsd-5 that vmstat -M is broken for per-cpu stats. It would be good to verify that your patch doesn't cause any similar regressions.
Attachment:
pgpq_YJiKVBTC.pgp
Description: PGP signature