Re: [PATCH net-next v5 1/3] net: devmem: allow rx-buf-size > PAGE_SIZE per dmabuf binding

[Jakub Kicinski <kuba@kernel.org>]

On Fri, 24 Jul 2026 09:39:54 -0700 Bobby Eshleman wrote:
> > BTW did you add both min and max checks? Cause the only risk with using
> > a dummy value would be that the policy will be rendered inline, and
> > inline policy is u16 so 64k wouldn't fit. But your sample above has a
> > max of u32_max which forces the out-of-line policy, which is what we
> > want.  
> 
> Yep, u32_max:
> 
> +	 name: rx-page-size
> ...
> +        checks:
> +          min: page-size
> +          max: u32-max
> 
> Sorry, probably should have just sent the whole patch instead of
> replying hunk-by-hunk.

Ack, LG, just double checking.

> BTW, how expressive do we want these policies? For example, would
> absorbing the power_of_2 check into a policy be valid in the future? or
> is that too bespoke?

Power-of-2 could be useful (it's implicitly one bit set, which is also
potentially useful for validating one-hot flags). The trickiness is
combining power-of-2 and the min check :S We have one validation per
field. I was wondering if we would be better off defining the field
as a shift instead, then we only have to check min. But I thought
that it'd be a little unusual for uAPI and possibly maybe one day
we will want the non-power of 2? So I figured checking min using
the existing facilities and open coding power of two check is good
enough for now.