Messages in this thread | | | Date | Thu, 29 May 2008 12:15:16 -0500 | From | James Cammarata <> | Subject | Re: [PATCH] net: add ability to clear stats via ethtool - e1000/pcnet32 |
| |
> For one quite a few of the network cards keep the stats in hardware and > don't neccessarily have a way to reset them. In other cases there is a > mix of OS accumulated stats bulk updated by overflow events on the device > itself.
Yes, and there's no good way to get around that, so the best thing for that hardware would probably to not add this functionality, in which case ethtool would simply report that it's not available.
> Its a lot of complexity, and changes all over the places for the sake of > a trivial userspace change. I would suggest you instead write a quick bit > of perl or python that fetches the stats and then updates every second > with the changes.
I would disagree with you on the complexity issue. In general, it is one new function (trivial in length) and an addition to the ethtool ops. SNMP and sar cover the reporting needs, but there have been multiple times when troubleshooting under the gun that this would have been extremely useful to help diagnose the issue (in one instance, we were getting floods of jumbo frames and other errors, and we did not know how quickly they were accumulating, because the switch was not seeing the same problem). My day job is in the financial industry, so we are often under very short SLA's to diagnose issues, and having this built into the kernel/existing tools would be a boon. I'd be more than happy to change this to a kernel config option, experimental and default off for those who would not want it.
| |