[broken counters was.. Re: [nsp] Re: CiSCO IOS 12.* source code stolen]

[Tony Li]

Ummm...  a different scheduler will not fix this problem.  The only way 
to ensure accurate
counters is accurate hardware implementation.

Tony



On May 17, 2004, at 6:34 PM, Andrew Fort wrote:

> On 16/05/2004 7:08 AM, jlewis at lewis.org wrote:
>
>> Hmm...maybe the counter bugs will finally get fixed...but we'll have 
>> to
>> run bootleg IOS to get the fixes.
>>
>
> heh.  I asked Phil Harris about this at a networkers recently (too 
> good an opportunity and all).  His answer was essentially "well, the 
> software interrupt called after packet send/receive (i.e., the one 
> that does counters, etc) can get masked, so if it does, your counters 
> don't get updated..."; this doesn't seem to add up entirely to some of 
> the types of counter bugs I'm sure we've all seen, but it's one 
> explanation.  I guess it's just one of the most obvious signs of the 
> aging IOS 'scheduler'.  Bring on that nonstop kernel, I say :).
>
> -afort
> _______________________________________________
> cisco-nsp mailing list  cisco-nsp at puck.nether.net
> https://puck.nether.net/mailman/listinfo/cisco-nsp
> archive at http://puck.nether.net/pipermail/cisco-nsp/
>