Re: invalid requirement from ethtool?

[Ben Hutchings <bhutchings@solarflare.com>]

On Thu, 2011-07-28 at 00:37 -0700, David Miller wrote:
> From: Eli Cohen <eli@dev.mellanox.co.il>
> Date: Thu, 28 Jul 2011 10:23:26 +0300
> 
> > On Wed, Jul 27, 2011 at 10:43:09PM -0700, David Miller wrote:
> >> From: Eli Cohen <eli@dev.mellanox.co.il>
> >> Date: Tue, 26 Jul 2011 15:42:22 +0300
> >> 
> >> > I think both params zero should be allowed and mean coalescing is not
> >> > operational, thus we can remove these comments from ethtool.h
> >> 
> >> The existing precendence has existed for more than 10 years.  You can't
> >> just change it like this.
> >> 
> >> You'll need to find a new way to encode "disabled" coalescing.
> > 
> > I can't see the text explicitly specifies how to "disable" coalescing.
> > If I ignore the specific comment that disallows both params 0, I could
> > interpret the text such that when they're both zero, the feature is
> > disabled.
> 
> The documentation does not determine what the rules are, the cpu
> does not execute the documentation it executes the code, and that's
> what determines the rules.

The ethtool core doesn't check the values in struct ethtool_coalesce, so
the rules are really driver-specific.

Ben.

-- 
Ben Hutchings, Senior Software Engineer, Solarflare Communications
Not speaking for my employer; that's the marketing department's job.
They asked us to note that Solarflare product names are trademarked.