svn commit: r282280 - in head/sys/dev: e1000 ixgbe ixl

[John Baldwin]

On Friday, May 01, 2015 02:27:37 AM Gleb Smirnoff wrote:
>   John,
> 
> On Thu, Apr 30, 2015 at 06:23:39PM +0000, John Baldwin wrote:
> J> Author: jhb
> J> Date: Thu Apr 30 18:23:38 2015
> J> New Revision: 282280
> J> URL: https://svnweb.freebsd.org/changeset/base/282280
> J> 
> J> Log:
> J>   Various fixes to the stats in igb(4), ixgbe(4), and ixl(4).
> J>   - Use hardware counters for ifnet stats in igb(4) when possible.  This
> J>     ensures these stats include packets that bypass the regular stack via
> J>     netmap.
> J>   - Don't derefence values off the end of the igb(4) VF stats structure.
> J>     Instead, add a dedicated if_get_counter method for igb(4) VF interfaces.
> J>   - Report missed packets on igb(4) as input queue drops rather than an
> J>     input error.
> J>   - Report bug_ring drop counts as output queue drops for igb(4) and ixgbe(4).
> J>   - Export the buf_ring drop stats for individual rings via sysctl on
> J>     ixgbe(4).^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
>        ^^^^^^^^^
> 
> While you are here, let me remind you about this plan:
> 
> https://lists.freebsd.org/pipermail/svn-src-head/2014-October/063575.html
> 
> We can prototype the API to userland now, write down a utility that uses
> it, or add the functionality to an existing utility. And start with Intel
> drivers, that seem to be most interested in extra stats.

So the importaing thing here is that if_get_counter() is still doing
per-ifnet stats.  The stat you underlined above is per-queue instead.
We well need more explicitly knowledge of queues outside of drivers
and in the stack itself to support a generic framework for per-queue
stats.

-- 
John Baldwin