Re: iwlwifi(4) on riscv64 (intel ax200)

On Wed, Jan 14, 2026 at 06:43:46AM +0000, Bjoern A. Zeeb wrote:
Hi,

yeah, this looks good. Different way to make it compile, so I guess the 
#ifdefs I added are not needed, but keeping the linux headers around was 
the key.

Do you need any help w/ testing?

> On Mon, 5 Jan 2026, Tassilo Philipp wrote:
>
> Hi,
>
>> I thought I simply share this here, maybe it's useful to someone. I 
>> got an intel wifi-6 ax200 m.2 slot card working on a HiFive 
>> Unmatched board w/ fbsd 15.0. It seems to work fine for me, so far, 
>> but it's in no way thoroughly tested...
>>
>> In order to build the if_iwlwifi kernel module I needed to disable 
>> ACPI in the module's Makefile, and patch a source file to add some 
>> missing ACPI related #ifdefs.
>>
>> Also, some firmware files need to be built and installed via ports' 
>> net/wifi-firmware-iwlwifi-kmod. Those firmware files can be built on 
>> any platform - when building them on riscv64 themselves, a small 
>> patch needs to be applied to net/wifi-firmware-kmod so riscv64 isn't 
>> excluded, though.
>>
>> Anywho, to have a seemingly working iwlwifi(4) on my riscv64 setup, 
>> the following steps were needed:
>>
>> 1) make sure src is installed
>> 2) apply attached patch (which does the tweaks described above)
>> 3) build/install kernel module from source:
>>   cd /usr/src/sys/modules/iwlwifi
>>   make
>>   make install clean 
>> 4) install firmware port:
>>   cd /usr/ports/net/wifi-firmware-iwlwifi-kmod
>>   make install clean 
>> 5) kldload wlan_ccmp wlan_amrr if_iwlwifi
>>
>> As said, thought that I share this here, happy that it's working, 
>> but no idea or guarantee about stability and correctness.
>
>
> There is a stack of 5 changes open for review as of last night:
> 	https://reviews.freebsd.org/D54694
>
> One is for the firmware in ports.
>
> The others are (a) ACPI related for RISC-V so we get the headers on __riscv 
> as well so we do not have to change other driver code, and (b) module Makefile 
> related to enable ACPI based on DEV_ACPI and factor out the LinuxKPI wireless 
> drivers in their own block also compiling them on RISC-V.  Turns out rtw89 is 
> broken requiring ACPI unconditionally (also on Linux it seems) so it's currently 
> disabled until I get to the point of fixing that in an upstreamable way.
>
> Hope they can land soon.
>
> Thanks a lot again for letting us know!
>
> /bz
>
> --
> Bjoern A. Zeeb                                                     r15:7