Re: iwm(4) broken on latest -current kernel

[Kevin Bloom <ktnb%posteo.net@localhost>]

(Michael van Elst) <mlelstv%serpens.de@localhost> wrote:

> ktnb%posteo.net@localhost (Kevin Bloom) writes:
> 
> >(Michael van Elst) <mlelstv%serpens.de@localhost> wrote:
> >> Can you please provide the type of your iwm device. E.g. output
> >> from dmesg like:
> 
> >Sure thing:
> 
> >[     1.040837] iwm0 at pci2 dev 0 function 0: Intel Dual Band Wireless AC 8260 (rev. 0x3a)
> >[     1.040837] iwm0: interrupting at msi2 vec 0
> >[     6.265043] iwm0: hw rev 0x200, fw ver 22.361476.0, address xx:xx:xx:xx:xx:xx
> 
> 
> This gets "interesting". A similar device was already tested successfully,
> but there are at least two versions of AC 8260 that might behave differently.
> 
> One test would be to use a more recent firmware for it. You should already
> have (in /libdata/firmware/if_iwm/):
> 
> SHA1 (iwlwifi-8000C-22.ucode) = 7ff722a5d1f968c20a6592c13f5a64d35e7da178
> SHA1 (iwlwifi-8000C-36.ucode) = a6a01d48822dac433f8fcddf0eaa43fb9c45314f
> 
> and the driver selects the older version as the newer failed for us, probably
> because it requires more features in the driver.
> 
> So one test would be to build a kernel that selects the newer firmware.
> 
> The driver source contains:
> 
> #if 0
> /* 8000C-36 */
>                 sc->sc_fwname = "iwlwifi-8000C-36.ucode";
> #else 
> /* 8000C-22 */
>                 sc->sc_fwname = "iwlwifi-8000C-22.ucode";
> #endif
> 
> and by changing '#if 0' to 'if 1' it will select the new firmware.
> 
> 
> Another difference of course could be your environment. Do you use
> a simple WPA2 net with pre-shared key or something more complex ?
> 

It appears to happen about 50% of the time when I boot my machine up.
I'll change that line of code and get back to you as soon as I can!