Kalle Valo <kvalo@xxxxxxxxxx> writes: > Baochen Qiang <quic_bqiang@xxxxxxxxxxx> writes: > >> On 9/19/2024 2:10 AM, Kalle Valo wrote: >>> @@ -4310,7 +4301,7 @@ static void ath12k_sta_rc_update_wk(struct work_struct *wk) >>> >>> spin_unlock_bh(&ar->data_lock); >>> >>> - mutex_lock(&ar->conf_mutex); >>> + wiphy_lock(ath12k_ar_to_hw(ar)->wiphy); >> >> note in patch 4/4 ath12k_sta::update_wk is converted to use >> wiphy_work. While a wiphy work item is running wiphy lock is held >> already. So here try to acquire wiphy lock once again will lead to a >> deadlock. > > Ouch again, thanks for catching this! This time I actually tested > changing bitrates and it shouldn't deadlock in v3. But I did notice > sleeping while atomic warnings (even without this patchset) and decided > to fix those in the same patchset as well. > > Oh, and WCN6855 firmware was also crashing whenever I tried to change > the bitrates. But let's handle that separately. Bah, I of course mean WCN7850. Too many chipsets... -- https://patchwork.kernel.org/project/linux-wireless/list/ https://wireless.wiki.kernel.org/en/developers/documentation/submittingpatches
