Re: Call for testers: Atheros AR8121(L1E)/AR8113/AR8114(L2E) ethernet

On Tue, Nov 11, 2008 at 10:15:44AM +0300, Boris Samorodov wrote:
 > Pyun YongHyeon <pyunyh_at_gmail.com> writes:
 > > On Mon, Nov 10, 2008 at 01:58:54PM +0300, Boris Samorodov wrote:
 > 
 > >  > you some additional info:
 > >  > -----
 > >  > uname -a ftp://ftp.bsam.ru/pub/tmp/EeePC/uname.2
 > >  > netstat -w 1 ftp://ftp.bsam.ru/pub/tmp/EeePC/netstat.ale.2
 > >  > sysctl dev.ale.0.stats ftp://ftp.bsam.ru/pub/tmp/EeePC/sysctl.ale.stats.2
 > >  > iostat -w 1 ftp://ftp.bsam.ru/pub/tmp/EeePC/iostat.ale.2
 > >  > -----
 > >  > 
 > >  > The interesting one is a netstat one. I'm not sure what zeroes for
 > >  > packets mean while trafic exists.
 > >  > 
 > >
 > > Hmm, I also have no idea why netstat(1) shows such a non-sense
 > > value while transfer is in progress. I guess you can easily write a
 > > script that extracts interesting MAC statistics of ale(4).
 > > For example,
 > > sysctl dev.ale.0.stats.rx.good_frames
 > > sysctl dev.ale.0.stats.tx.good_frames
 > >
 > > Use 'sysctl -d dev.ale.0.stats' to get complete descriptoin of each
 > > node.
 > 
 > Here is the diff of "sysctl dev.ale.0.stats" before and after fetching a
 > ~300MB file:

[...]

 > Do you see anything suspicious?
 > 

No clear idea but the only wild guess I have is periodic updates of
if_ipackets/if_opackets in driver. Since ale(4) and age(4) supports
very detailed hardware MAC counters, both ale(4) and age(4) rely on
these hardware counters to update various ifnet counters intead of
manupulating ifnet counters with software method.
These hardware counters are updated for every one second in driver.
I thought 1 second interval would be enough because frequent
reading hardware counter registers would add additioanal burden to
CPU(e.g. it can generate too many PCI transactions). If netstat
polls these counters while no upates are made to
if_ipackets/if_opackets it may think there are no network
activities.
How about increasing polling interval of netstat to two seconds
from one second? If my theory is right it would show correct
value.

-- 
Regards,
Pyun YongHyeon