On 5/15/2025 4:55 PM, Johannes Berg wrote:
On Thu, 2025-05-15 at 11:18 +0530, Sarika Sharma wrote:+ for_each_valid_link(sinfo, link_id) { + link_sinfo = sinfo->links[link_id]; + link = nla_nest_start(msg, link_id + 1); + if (!link) + goto nla_put_failure; + + if (nla_put_u8(msg, NL80211_ATTR_MLO_LINK_ID, + link_id)) + goto nla_put_failure; + + if (link_sinfo && + !is_zero_ether_addr(link_sinfo->addr) && + nla_put(msg, NL80211_ATTR_MAC, ETH_ALEN, + link_sinfo->addr)) + goto nla_put_failure; + + if (link_sinfo && nl80211_fill_link_station(msg, rdev, + link_sinfo)) + goto nla_put_failure; + + nla_nest_end(msg, link);This seems wrong, should be is_valid_ether_addr(), and WARN_ON(!link_sinfo), and that really shouldn't even be possible after iterating the bitmap. In all of those cases it also really shouldn't build the nested attribute at all. (I had made those changes.)
Sure, let me pull those patches. Thankyou!
johannes
