Status updates

From: Bjoern A. Zeeb <>
Date: Sun, 17 Apr 2022 22:18:03 UTC
Hi everyone,

I wanted to give you a summary about multiple things given I'll likely
not be very responsive the next two weeks.  I'll still try to check
every other day or so if I can find an hour.


- Thanks for all the "success" reports from some of you still coming
   in (privately).  Thanks should go to the FreeBSD Foundation,
   bug reports you can keep sending to me :)

- Thanks to all the people who keep testing, applying patches,
   doing early cherry-picks to stable/13 and report back and are
   patiently waiting for the next bits to come.

Thanks to all of you!


- most pressing issues currently:

   * cards crashing firmware on start (mostly 9xxx/8xxx/7xxx) and
     almost all iwm(4) supported cards.
     Thanks to the people who send me traces and reports.  I have
     nothing conclusive yet but may send out a patch to try still
     tomorrow or as soon as possible.

   * Need to update documentation before release gets out the door.

   * The UP / DOWN (beacon|connectivity)loss reports / UP on startup.
     There has also a lot of reports on Linux for that over the last
     two years.  I haven't found a cure but I'll let you know in case
     I hear anything.  Could be FW related which we cannot fix.

- while started on rate control I found that the driver reports a
   HT MCS 0 instead of the legacy rates which is unfortunate.  I'll
   need to trace back to see if this comes out of FW or the driver.
   This needs finishing and I hope that the media feedback from
   ifconfig will be more realistic once done; then we can start
   talking "speed".


- current state as I understand it from reports:

   * people with more than 4GB of physical memory need a patch still.
     This still needs a more general solution but the latest patch at
     least allows people to run with more than 4GB of physical memory
     usable for the remainder of the system.

   * I have one out-of-memory report which I need to reply to.

   * 8822CE and newer firmware I keep jotting away to implement
     missing bits.  The firmware downgrade seems to have helped for

   * ``Works as "well" as it does on Linux''

   * Is in "main" (CURRENT) only and not yet in stable/13 but changes
     very self-contained.  See list of hased on wiki page.


- it compiles and loads but needs some LinuxKPI support not there yet.
   I hope this will likely appear in May.

"one more":

- I have one more driver compiling already but that seems to require
   more work for legacy STA to work and I consider it "queued" once
   the above list is shorter.

LinuxKPI & net80211:

- Lots of work to do for LinuxKPI and I keep jotting away filling gaps
   as I spot them and see them needed while also add stuff needed for
   other driver or more functionality.

- There's a wake_tx problem and a more underlying general TX problem
   for which I have a WIP in a tree;  I need to do some more
   investigations both towards the driver but also with other LinuxKPI
   components which I have patches locally in a way not suitable for
   upstreaming yet.

- WME and HW_KEY support should be (re-)added and finished in
   LinuxKPI.  There's #ifdefs from months ago but were left out as
   we were tracking too many issues.

- big on the todo lists are also power saving and suspend/resume.

- net80211 gets small changes still only at this point and I am
   collecting TODOs but will keep larger churn away until we really
   need it as it'll also require way more testing on in-tree drivers.

- MFCs outstanding for the latest changes (hopefully to come

Anything done in this category will benefit all the above drivers
in theory. H aving more drivers does seem to help overall "stability"
currently as different drivers require different details which do
not show up on the radar (somehow masked) for others but still
improve things.

Realtek USB drivers:

I've had multiple requests for various Realtek USB chipsets not
curently supported.  I do own one or two myself.  At this point I
do not want to go near these drivers for multiple reasons:
(a) GPL-only to my best knowledge (b) I consider the various sources
and duplicate versions of drivers a "mess" and (c) I haven't checked
if OtherBSDs have improved what is rtwn_usb for us for at least some
of them.  In other words: I'd have to do my homework first.

I think I'll leave it at this for tonight.

Have blessed times,

Bjoern A. Zeeb                                                     r15:7