Heyla,
I've been getting "free inode /usr/38950 had 144 blocks" type errors
from kernels built over the last week or so.  It's causing major errors
with real files so I get big blocks of 0's in files (maybe the real
files become sparse until I examine them?) destroying data.  This has
occured on different machines and now I've had this happen on /tmp which,
for me, is an mfs.  The only kernel config change I've made in the last
week is enabling INET6 and co.  It's a bit hard to see what's changed
recently, a side effect of checking when kernel RCSID's are being
added everywhere :-(

Anyone else seen this?

Regards,
-- 
Geoff Wing   gcw
main email :  " @pobox.com
rxvt stuff :  " @rxvt.org
zsh stuff  :  " @zsh.org