Re: [PATCH net-next v2 00/10] genetlink: support per-command policy dump

On Fri, 02 Oct 2020 17:13:28 +0200 Johannes Berg wrote:

On Fri, 2020-10-02 at 08:09 -0700, Jakub Kicinski wrote:

quoted

On Fri, 02 Oct 2020 17:04:11 +0200 Johannes Berg wrote:  

quoted

quoted

quoted

Yeah, that'd work. I'd probably wonder if we shouldn't do

[OP_POLICY]
  [OP] -> (u32, u32)

in a struct with two u32's, since that's quite a bit more compact.    

What do we do if the op doesn't have a dump or do callback?
0 is a valid policy ID, sadly :(    

Hm, good point. We could do -1 since that can't ever be reached though.

But compactness isn't really that necessary here anyway, so ...  

Cool, sounds like a plan.

This series should be good to merge, then.  

I suppose, I thought you wanted to change it to have separate dump/do
policies? Whatever you like there, I don't really care much :)

I just want to make the uAPI future-proof for now.

At a quick look ethtool doesn't really accept any attributes but
headers for GET requests. DO and DUMP are the same there so it's 
not a priority for me.

But I can also change my patches later to separately advertise dump/do
policies, and simply always use the same one for now.

Right that was what I was thinking. Basically:

	if ((op.doit && nla_put_u32(skb, CTRL_whatever_DO, idx)) ||
	    (op.dumpit && nla_put_u32(skb, CTRL_whatever_DUMP, idx)))
		goto nla_put_failure;

But this series does conflict with the little bugfix I also sent, could
you please take a look?

https://lore.kernel.org/netdev/20201002094604.480c760e3c47.I7811da1539351a26cd0e5a10b98a8842cfbc1b55@changeid/ (local)

I'm not really sure how to handle.

Yeah, just noticed that one now :S

Dave, are you planning a PR to Linus soon by any chance? The conflict
between this series and Johannes's fix would be logically simple to
resolve but not trivial :(