Re: Quartz64 (RK3566) clock driver and patches.

On 24 Jul 2021, at 13.56, Emmanuel Vadot <manu@bidouilliste.com> wrote:
> 
> On Sat, 24 Jul 2021 13:26:07 +0200
> Søren Schmidt <soren.schmidt@gmail.com> wrote:
> 
>> Hi 
>> 
>> As promised, the clk driver I did for the RK3566 used on the Pine Quartz64 board:
>> 
>> https://people.freebsd.org/~sos/Q64/ <https://people.freebsd.org/~sos/Q64/>
>> 
>> Included the buld script, kernel config and u-boot I have used to get there.
> 
> That will be much easier to review on phab, or even github PR or git
> send-email, basically anything else that this raw diff.
> I've only glance at it a bit and swa this :
> Index: sys/arm64/rockchip/clk/rk_clk_fract.c
> ===================================================================
> --- sys/arm64/rockchip/clk/rk_clk_fract.c	(revision 570)
> +++ sys/arm64/rockchip/clk/rk_clk_fract.c	(working copy)
> @@ -191,8 +191,12 @@
> 	if (sc->denominator == 0) {
> 		printf("%s: %s denominator is zero!\n",
> clknode_get_name(clk), __func__);
> +#if 0
> 		*freq = 0;
> 		return(EINVAL);
> +#else
> +		return (0);
> +#endif
> 
> which I don't understand why it's needed.

The denominator is defaulting to zero in the chip, and u-boot in its current form doesn’t set a value -> panic..

> I won't look/comment at all the clocks definition in this form.

I’m just sharing here in the good old fashioned way, I dont force you to do anything :)

>> I still use the ?all inclusive? DTS and not the shiny new ?official? one (last wont boot :) ) thats for later?
> 
> What does that mean ?

Good question, the dts derived from the DTB included with the rockchip u-boot works just fine, but the latest greatest from the linux crowd does not even produce the ---<<BOOT>>—- line, I havn’t had time to look into it..

--
Søren Schmidt
sos@deepcore.dk / sos@freebsd.org
"So much code to hack, so little time"