Shouldn't that be ${_iwx} instead, with $_iwx being defined in the optional section near line 569 like $_iwlwifi is?

missing static?

(although at this point I'd be putting these routines in an if_iwx_debug.c file.)

Let's get rid of this?

Hm leave this one in, we'll need to go see if addba stop works right

And this too now?

We really should migrate this to an opt_iwx.h file and include it in sys/conf/options . I don't mind seeing it fixed after we land though, this driver is going to need a bunch of work anyway!

Can we can -O0 before we land it?

There is no need for an opt_ for an individual driver's debug flags. Normally if single file we just define them in the .c otherwise if built as module only I also added them to the Makefile. No one but developers (or users debugging but by that they are working on code) will turn them on so no need to spam kernel configs with that.

unlikely; iwlwifi is restricted by wrking LinuxKPI bts

I think given it's a native driver and not restricted by working LinuxKPI currently you ca probably build it unconditionally just as iwx; I am never sure why iwn and iwm were never added as such -- do we expect big/little endian problem with them?

No longer needed.

@thj do you just want to put a # in front of DEBUG_FLAGS?

I've split debug out to its own file

I've moved it to be with iwm. iwn and iwm are only build for systems that support sourceless drivers - i.e. things that need firmware.

I don't really know if this is needed so I default to being consistent

If anyone claims to have done 700Mbit/s with this, then the driver internally doesn't even adhere to the channel config:
ifconfig -v wlan0 list chan
shows max VHT/20.

Try this patch as a starter... but we really want to add RADAR/DFS bits for whatever they mean in the defines here.

https://people.freebsd.org/~bz/tmp/20250327-01-iwx.diff

I locally moved that somewhere to at least allow it on arm64 as well but I think for now it sits good where it is.

Given this is still on I wonder why I do not get register dumps...

The total copyright and provenance of this driver is a little weak (though it's all there in the OpenBSD history to dig out).

All the OpenBSD stuff is ISC, as is the FreeBSD port.

It's unclear the extent to which reference driver material was included in the original iwn (or earlier?) driver and how much of that remains after it was copied to iwm and then iwx. The most favorable reading is that they just read the code and included the reference because they read it quite intensely and may have therefore had a resulting driver that was too close. It's unclear of code was copied, or just know-how and if code was copied if it was a fair use under various copyright doctrines that ensure that facts aren't covered by copyright, but creative expression is. It would take a rather extensive analysis to judge this, and the result might be ambiguous and unsatisfying.

My advice would be to add a SPDX-License-Identifier: BSD AND ISC since there's an intent to declare that the Damien and/or Stefan was opting to take whatever IP from iwlwifi using the BSD licensing only and creating the derivative work that way. If you want to be more cautious, (BSD OR GPL) AND ISC wouldn't be terrible. I'd add whatever you decide as a separate commit AFTER we push the driver in since it's advisory and we can explain the nuance of our decision in the commit message (which seems out of place for the initial port)..

I'd also consider a 2025 copyright date since this is still in 'draft' form.

So this is wrong. You need it to describe the entire file, which is either BSD-3-Clause AND ISC (since it declares it was derived from the BSD-licensed file, opting for only BSD for the original dual licensed file, which is an option). Or it's (GPL-2.0-only OR BSD-3-Clause) AND ISC.

Note: Since there's so many licenses here, this can't be used as a shortcut for these changes.

This needs a license. And since we have so many other licenses in the file, the project policy is you can't use simple SPDX here.
I'd get the Foundations' permission to just use ISC copyright, since that's what all the others are (in reality).

* Permission to use, copy, modify, and distribute this software for any
 * purpose with or without fee is hereby granted, provided that the above
 * copyright notice and this permission notice appear in all copies.
 *
 * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES
 * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
 * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR
 * ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
 * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
 * ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
 * OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.

See comments on if_iwx.c and do the same here.

See comments on if_iwx.c and do the same here.

See comments on if_iwx.c and do the same here.

See comments on if_iwx.c and do the same here.

I don't know where history of all this started, as I always said tracing any of this back to the original is virtually impossible given I am not aware of any git hashes from Linux in the OpenBSD history this was originally based on.

For what it's worth, the non-GPL-only code in iwlwifi got "GPL-2.0 OR BSD-3-Clause", no "-only". The files which are GPL-only are "SPDX-License-Identifier: GPL-2.0-only" of which there are three or so which would still be relevant for this work I assume but if started from iwn or earlier might have been almost exclusively that?

We could delete the extra blank line here