From: Peter Hessler Subject: Re: sys/qwz: initialize on Snapdragon but fails on scan To: Mark Kettenis Cc: "Kirill A. Korinsky" , tech@openbsd.org Date: Sun, 29 Mar 2026 22:25:59 +0200 On 2026 Mar 29 (Sun) at 22:20:51 +0200 (+0200), Mark Kettenis wrote: :> Date: Sun, 29 Mar 2026 21:23:36 +0200 :> From: Kirill A. Korinsky :> :> tech@, :> :> I'd like to share some work which I pocking around with attempt to move :> forward with qwz driver. :> :> I've attached 3 diffs where: :> :> 1. fixed a copy and typo in DTS for my Honor machine; :> 2. updated qwz firmware to the last version; :> 3. a bit of work in qwz. :> :> Right now, after boot qwz reads like this: :> :> qwz0: flags=8802 mtu 1500 :> lladdr 3c:0a:f3:d3:55:c5 :> index 1 priority 4 llprio 3 :> groups: wlan :> media: IEEE802.11 autoselect :> status: no network :> ieee80211: nwid "" :> :> and attempt to up an interface leads to firmware crash. :> :> Scope of qwz work: :> 1. Fixed a crash on my machine which was suggested as: :> https://marc.info/?l=openbsd-tech&m=177410294328488&w=2 :> 2. Implemented world reg domain like ath12k :> 3. Fully filled flags and follow ath12k logic with peer-metadata :> 4. Increased RDDM size and synced timeouts with ath12k :> :> This work was like read how ath12k works, and figure out what we do the :> wrong way. Right now qwz works more or less the same way, and I run out of :> ideas why it crashing. Any suggestion and ideas are welcome. :> :> Also, I'd like to commit all three diffs, because it safe and qwz is :> disabled anyway. Anyone willing to OK it? : :The two ports diffs are ok kettenis@ (but try to get an ok from ports :person as well for the ports mechanics). : ports bits, OK :I'll try to have a look at the other diff soon. : :> [2:text/plain Show Save:qwz.wrk.diff (13kB)] :> :> :> [3:text/plain Show Save:qwz-firmware.diff (1kB)] :> :> :> [4:text/plain Show Save:arm64-qcom-dtb.diff (2kB)] :> : -- If God had not given us sticky tape, it would have been necessary to invent it.