igb interrupt moderation

[Pyun YongHyeon]

On Sun, Jan 03, 2010 at 05:13:41PM +0100, Michael T?xen wrote:
> On Jan 3, 2010, at 3:28 PM, Mike Tancsa wrote:
> 
> > At 08:55 AM 1/3/2010, Michael T?xen wrote:
> >> Hi Barney, Hi Jack,
> >> 
> >> some comments and some more questions inside...
> > 
> > 
> > Just a separate datapoint about this driver, unless I apply
> > 
> > http://people.freebsd.org/~yongari/igb/igb.buf.patch6
> > 
> > the driver is not really usable for me in RELENG_8 on the dual port version of the card
> Could you elaborate on what you mean by "not really usable"?
> 

The patch was generated before Jack committed new code in HEAD.
At that time I saw several issues like too many interrupts under
load, unresponsible system under heavy RX load and spurious
watchdog timeouts etc.
To fix large number of interrupts I just disabled AIM feature as it
seems to cause more TX completion interrupts under heavy RX load
(64 bytes UDP frames). It seems the controller can be configured to
intelligently generate interrupts based on load but it seems we
need some more experimentation. Just disabling AIM yielded about
8-9k interrupts/sec and I was satisfied with the result at that
time.
I haven't had time to experiment new igb(4) code in HEAD but new
code might have fixed some known issues.