Mon, Aug 12, 2019 at 06:02:18AM CEST, da...@davemloft.net wrote:
>From: David Ahern <dsah...@kernel.org>
>Date: Tue, 6 Aug 2019 12:15:17 -0700
>
>> From: David Ahern <dsah...@gmail.com>
>>
>> Prior to the commit in the fixes tag, the resource controller in netdevsim
>> tracked fib entries and rules per network namespace. Restore that behavior.
>>
>> Fixes: 5fc494225c1e ("netdevsim: create devlink instance per netdevsim
>> instance")
>> Signed-off-by: David Ahern <dsah...@gmail.com>
>
>Applied, thanks for bringing this to our attention and fixing it David.
>
>Jiri, I disagree you on every single possible level.
>
>If you didn't like how netdevsim worked in this area the opportunity to do
>something about it was way back when it went in.
Yeah, I expressed my feelings back then. But that didn't help :(
>
>No matter how completely busted or disagreeable an interface is, once we have
>committed it to a release (and in particular people are knowingly using and
>depending upon it) you cannot break it.
I understand it with real devices, but dummy testing device, who's
purpose is just to test API. Why?
>
>It doesn't matter how much you disagree with something, you cannot break it
>when it's out there and actively in use.
>
>Do you have any idea how much stuff I'd like to break because I think the
>design turned out to be completely wrong? But I can't.
Sure, me too :) But that is for real devices. That is a different story
as I see it. Apparently, I'm wrong...
