wireless status report

From: Bjoern A. Zeeb <bz_at_FreeBSD.org>
Date: Mon, 08 Apr 2024 22:26:45 UTC
Hi,

(in random order)

- I got a private report for a net80211 crash (which smells similar to
   other once we tracked down in the last months) but is due to PS locks.
   It had started with suspend/resume "fixing" and the workaround from
   PR  263632 seems to have helped there too).  I just got more data
   I need to work through tomorrow.
   Might put it all into a PR if the solution does not turn out to be obvious.

- I've held back on MFCs to stable/13 as I need to figure out the state
   of EPOCH there.  Some of the changes which did not go into 13.3-R
   are still pending because of that.

- Opened a drm-kmod pull request to build with LINT kernels (I do a
   universe build for LinuxKPI changes with drm-kmod to make sure we do
   not break [the build] for that)

- In addition to an updated iwlwifi (incl. firmware -- to be figured out
   how to install in the future, plain /boot/firmware or as .ko still),
   I have rtw88, rtw89 (scans once but no packets yet), and mt76 updated
   locally.  I opened some LinuxKPI reviews for those and will keep
   working on getting all pre-reqs out.  mt76 will need more work in that
   regard.  Once the review queue is smaller, I'll update the github
   tree again.  I do not anticipate to merge that into main before 14.1
   is on its way as it would be too disturbing at this point.

- Having received my USB debug cable I made the latest iwlwifi d3.c
   compile as well.  I have not yet been able to test the setup.
   I also currently assume that the LinuxKPI suspend/resume bits are
   not all hooked up.  If these need changing a lot of more testing will
   be needed also for drm-kmod.  Hopefully all repreq for USB debugging
   can be sorted soon one way or another as that would be a nice "fix"
   to have for 14.1-R still. We'll see how much needs doing and if we
   can meet the timeline.

- On my testbed (I have an arm64 machine with a alot of WiFi cards)
   2 out of 7 iwlwifi cards worked after I fixed a request_module()
   hang on kldload.  The others fail to download firmware.  I have a
   similar experience with two iwn(4) cards where one works and the other
   does not (there we have to UP the interface to notice).  I am trying
   to sort out if it is my local setup or something else.  If anyone is
   testing wireless on arm64 with iwlwifi/rtw88 please let me know.
   I have the chnage queued to add iwn(4) to arm64 builds given at least
   one card seems to work as fine as on amd64.
   My rtw88/rtw89 cards seem fine by the way and so is ath(4) or are
   USB dongles so I am unsure if it's "Intel"-specific or not.

- The iwn(4) crash made me look through older iwn(4) PRs and categorize
   them a bit or start cleanup/feedback.

- I also, still locally, made LinuxKPI 802.11 debugging dependent on the
   net80211 compile time flag so we either have it all or nothing.
   Also noticed that arm64 kernels do not enable that so have a change
   for that too.

Lots of health,
Bjoern

-- 
Bjoern A. Zeeb                                                     r15:7