On Thu, 2024-03-28 at 17:17 +0100, Johannes Berg wrote: > On Thu, 2024-03-28 at 09:14 -0700, Jeff Johnson wrote: > > On 3/28/2024 9:09 AM, Johannes Berg wrote: > > > On Thu, 2024-03-28 at 20:40 +0530, Karthikeyan Periyasamy wrote: > > > > > > > > Can you point to any attribute constructed in this way from kernelspace > > > > for the reference to explore more ? > > > > > > I don't have anything directly, looking at the code finds e.g. > > > devlink_dpipe_entry_ctx_append() but honestly that's really quite > > > trivial, it just adds that new attribute while iterating whatever list > > > you have. > > > > Note that we are trying to maintain the same structure used by the current > > wiphy global advertisement since we actually refactor and reuse the existing code. > > Partially, yes. That's not true for the one I was discussing _here_, > notably nl80211_put_multi_hw_support(), however. > > It's partially true for patch 6, though even there > nl80211_put_per_hw_iface_combinations() doesn't need to do it that way, > since that's a whole new attribute (NL80211_IFACE_COMB_PER_HW_COMB) and > can define the content anew I should say "content and how it forms an array in the top-level message" here. johannes
