occasionally corrupted network stats in /proc/net/dev

Hi,

I've found that occasionally (not that often) there is bogus data
reported from /proc/net/dev. Look at the following samples taken at 1
second intervals:

Thu Dec 18 08:06:02 EET 2008
eth0:60874488128 375488575 0 0 0 0 0 9671228
157337361468 370012081 0 0 0 0 0 0
eth1:2734320074 30326964 0 0 0 0 0
1 426 6 0 0 0 0 0 0
Thu Dec 18 08:06:03 EET 2008
eth0: 0 372085566 0 0 0 0 0 360340043
698134474 373414640 0 0 0 0 0 0
eth1:2734320325 30326967 0 0 0 0 0
1 426 6 0 0 0 0 0 0
Thu Dec 18 08:06:04 EET 2008
eth0:60874560653 375488975 0 0 0 0 0 9671230
157337474626 370012483 0 0 0 0 0 0
eth1:2734320453 30326969 0 0 0 0 0
1 426 6 0 0 0 0 0 0
Thu Dec 18 08:06:05 EET 2008
eth0:60874561921 375488990 0 0 0 0 0 9671231
157337476380 370012497 0 0 0 0 0 0

This problem is also reported in one linux kernel mailing group given
as below:

https://kerneltrap.org/mailarchive/l...08/1/14/566936

I just want to know is there some other mechanism other than reading /
proc/net/dev to get the network statistics.

Br
VAndy