svn commit: r328257 - in head/sys: arm/broadcom/bcm2835 dts/arm modules

[Poul-Henning Kamp]

--------
In message <20180122094629.f292f9f239174f0865b046ea at bidouilliste.com>, Emmanuel
 Vadot writes:

>> + * Copyright (C) 2013-2015 Daisuke Aoyama <aoyama at peach.ne.jp>
>
> Was is based on some file ?
> Also someone will probably complain about SDPX tag here :)

Yes, I grabbed the blurp from a nearby file and forgot to edit it,
it's fixed now.

>> +	if (unit != BCM_PWM_CLKSRC) {
>> +		device_printf(sc->sc_dev,
>> +		    "Unsupported unit 0x%x", unit);
>> +		return (0);
>> +	}
>
> Uhg, this is ... ugly ...

Yes, very much so.

> The proper way will be to introduce a real clock manager exposing a
>clock domain and clocks using the extres/clk framework.

Yes, Somebody[tm] should really do that[1], just like Somebody[tm] should
write drivers for all sorts of other stuff on all sorts of platforms.

Eventually Somebody[tm] may even get around to that, but in the
meantime people who play with FreeBSD on Rpi's can at least get
the PWM output to work.

Poul-Henning

[1] Are you tired of Nice & Accurate[tm] chip documentation which
leaves no room for creativity and exploration?

Do you love to extrapolate and second-guess chip-designers mental
process from random snippets of incompetently written code ?

If so, you may find welcome relief trying to write a device driver
for the "BCM283x - Documentation - Scmockumentation!" family of
devices!

-- 
Poul-Henning Kamp       | UNIX since Zilog Zeus 3.20
phk at FreeBSD.ORG         | TCP/IP since RFC 956
FreeBSD committer       | BSD since 4.3-tahoe    
Never attribute to malice what can adequately be explained by incompetence.