• Add a comment labelling the Doorbell
• The Device address field is 7 bits wide rather than 3 bits

It wasn't clear from out last conversation if this is ready for review or if there is more related work.

• Correct DC SBR bit definition
• Change doorbell field definition and remove get method

In D53904#1231939, @imp wrote:

So I'd like to see the loader patches too before this lands. This strikes me as doing this all wrong, but maybe I'm just missing something.

This looks great to me, some small nits for which I have made comments.

In D53893#1231126, @adrian wrote:

looks good to me but the lack of mutexes anywhere is suspicious. ;-)

This is fine to me, I'm not sure what adrian means with his outstanding item, but we can just keep hitting iwx with a stick until we are happy.

Please update the bits available comment below keymgmt when you land this

Looks good to me

I am happy with all three revisions and echo the comments from rgrimes and jtl on a single review and the impact being documented.

This seems ok to me, if it doesn't smoke and you can still do vht then its a good step forward

right after the start of stab week seems like a good time to see if iwx explodes with these

From a glance I am pretty sure this is all incorporated in the udbc patchset.

Unless someone shouts at me I will land this on Monday so it can make ALPHA3 on Thursday

• fix a typo in man page
• provide full year in copyright
• move devmethod_end onto its own line
• use interface define rather than device define

I am pretty sure I've addressed all outstanding comments, but I could easily have missed some. Could interested people check again?

• move tr_setup label to start of line
• remove body of ioctl function

• take hrs's text for the man page

• remove null check for sc
• Reference latest standard
• Remove XXX comment, this isn't that unclear
• Don't hide reserved and vendor print behind print verbose
• Add extended MSI and usb tunnel capabilities
• don't check first in loop
• spell tun as usb3_tun (and correctly)

• don't check first in loop

• Reference latest standard
• Remove XXX comment, this isn't that unclear
• Don't hide reserved and vendor print behind print verbose
• Add extended MSI and usb tunnel capabilities

• Don't split subdir += lines
• rearrange headers to build (I am sure this built in this tree before)
• Move DRIVER_MODULE to the bottom of the file
• Address review comments on makefile

I think starting wpa_supplicant will result in the interface being brought up, but I still think this is a good action to do explicitly.

• remove null check for sc

As far as I can tell this is nop, but good ground work for future stuff.

This looks to do what is advertised in some simple tests. ddb is pretty fragile for me to use on the best of days so I'm happy enough

I've done a little testing and this does what I would expect on my test laptop.

Setting the regdomain on iwlwifi cleared this out - sorry for the testing delay.

I've sent test output via email. With 160MHz configured on my ap and it beaconing iwlwifi doesn't pick a vht rate for the band, just 80211a

I'm not able to convince an ap to actually do 160MHz channels - despite support in the radios.