Instead of hrtimer_is_queued(), use hrtimer_active() since
it might be running concurrently, and then it's not queued
at that point in time, as suggested by Thomas Gleixner in
https://lore.kernel.org/87plqn5psu.ffs@tglx, I just never
got to this for ages.
I think the concurrency is otherwise fine since we'll get
to cancel if we're actually removing things, and otherwise
we just send a beacon at slightly the wrong time or so.
Reported-by: syzbot+41e4341f493f1155aa3d@syzkaller.appspotmail.com
Link: https://patch.msgid.link/20241011145230.5a4d38d4ff9b.Iac0ec316a0c9a7b2619abe52ddc8e04c25d8c7e1@changeid
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
if (!data->started || !link_data->beacon_int) {
hrtimer_cancel(&link_data->beacon_timer);
- } else if (!hrtimer_is_queued(&link_data->beacon_timer)) {
+ } else if (!hrtimer_active(&link_data->beacon_timer)) {
u64 tsf = mac80211_hwsim_get_tsf(hw, NULL);
u32 bcn_int = link_data->beacon_int;
u64 until_tbtt = bcn_int - do_div(tsf, bcn_int);
info->enable_beacon, info->beacon_int);
vp->bcn_en = info->enable_beacon;
if (data->started &&
- !hrtimer_is_queued(&link_data->beacon_timer) &&
+ !hrtimer_active(&link_data->beacon_timer) &&
info->enable_beacon) {
u64 tsf, until_tbtt;
u32 bcn_int;