Re: [PATCH 6/6] brcmfmac: fallback mechanism to determine monitor mode features

[Arend van Spriel <arend.vanspriel@broadcom.com>]

On 6/24/2018 4:08 PM, Rafał Miłecki wrote:
> On Fri, 22 Jun 2018 at 20:45, Arend van Spriel
> <arend.vanspriel@broadcom.com> wrote:
>> Firmwares may not provide all monitor mode features in the "cap" iovar.
>> For those this fallback mechanism uses "sta_monitor" iovar. If firmware
>> is compiled with stamon, this iovar will fail with BCME_NOTUP; Otherwise
>> it fails with BCME_UNSUPPORTED.
>
> It's probably not the first time ever, but it appears your research
> (theory) doesn't match my experience (practice) ;) I'm afraid you
> missed some important check when analyzing firmware code.

It was not all theory ;-) but apparently I did not cover all bases. I 
only checked with 4366c0 (actually with 43664 aka 4366E) on the release 
branch I am working on.

> I've just tested all firmwares I got (for 43602a1, 4366b1 and 4366c0)
> and all of them return -4 (BCME_NOTUP) for "sta_monitor" when
> firmware/interface is down. It appears this test requires bringing
> firmware/interface up to make it reliable. Apparently even firmwares
> *without* sta_monitor return -4 (BCME_NOTUP) when firmware/interface
> is down.

That is crap. So back to the drawing board. Thanks for keeping the taps 
on this.

Regards,
Arend