On 2/28/2025 6:54 PM, Johannes Berg wrote:
+++ b/include/net/cfg80211.h
@@ -2158,6 +2158,9 @@ struct link_station_info {
* @local_pm: local mesh STA power save mode
* @peer_pm: peer mesh STA power save mode
* @nonpeer_pm: non-peer mesh STA power save mode
+ * @is_per_link_stats_support: 0- for non-ML STA and for ML STA,if driver
+ * offload link decisions and do not provide per-link statistics.
+ * 1- if driver provides per-link statistics.
Seems like that should be false/true since it's bool, but I also don't
really see why you need to list the values at all.
I intended for true-1, false-0, will add actual true/false as it's
misleading.
Or even need this extra value at all since you have
WIPHY_FLAG_SUPPORTS_MLO_STA_PER_LINK_STATS?? And perhaps that should
rather be an nl80211 flag so userspace knows what to expect? Not sure it
would care though.
Yes, Flag directly can be used everywhere but at some use case we are
not having wiphy structure, there either we have to pass an extra
argument, or need to fetch back wiphy structure and check.
So instead, added a boolean in sinfo structure.
For eg: in cfg80211_sinfo_release_content() only sinfo is passed, so
instead of passing an additional argument everywhere, added a flag in
sinfo itself.
+++ b/net/mac80211/sta_info.c
Don't mix that in where not needed for API changes.
Sure, will split this patch.
johannes
