Re: [PATCH v2 1/8] ARM: at91: pm: rework cpu detection

[Alexandre Belloni <alexandre.belloni@free-electrons.com>]

Hi,

On 16/01/2015 at 02:44:39 +0000, Yang, Wenyou wrote :
> > -	/* AT91RM9200 SDRAM low-power mode cannot be used with self-refresh.
> > */
> > -	if (cpu_is_at91rm9200())
> > +	at91_pm_data.memctrl = AT91_MEMCTRL_SDRAMC;
> > +
> > +	if (of_machine_is_compatible("atmel,at91rm9200")) {
> > +		/*
> > +		 * AT91RM9200 SDRAM low-power mode cannot be used with
> > +		 * self-refresh.
> > +		 */
> >  		at91_ramc_write(0, AT91RM9200_SDRAMC_LPR, 0);
> > -
> > +
> > +		at91_pm_data.uhp_udp_mask = AT91RM9200_PMC_UHP |
> > +					    AT91RM9200_PMC_UDP;
> > +		at91_pm_data.memctrl = AT91_MEMCTRL_MC;
> > +	} else if (of_machine_is_compatible("atmel,at91sam9260") ||
> > +		   of_machine_is_compatible("atmel,at91sam9g20") ||
> > +		   of_machine_is_compatible("atmel,at91sam9261") ||
> > +		   of_machine_is_compatible("atmel,at91sam9g10") ||
> > +		   of_machine_is_compatible("atmel,at91sam9263")) {
> > +		at91_pm_data.uhp_udp_mask = AT91SAM926x_PMC_UHP |
> > +					    AT91SAM926x_PMC_UDP;
> > +	} else if (of_machine_is_compatible("atmel,at91sam9g45")) {
> > +		at91_pm_data.memctrl = AT91_MEMCTRL_DDRSDR;
> > +	}
> > +
> About the memory controller type, I prefer to get it during the memory controller initialization, from the sram_ids[].data which defined in the setup.c,
> As this,
> static const struct at91_ramc_of_data at91rm9200_ramc_of_data = {
> 	.ramc_type = AT91_MEMCTRL_MC,
> };
> ...
> static struct of_device_id ramc_ids[] = {
> 	{ .compatible = "atmel,at91rm9200-sdramc", .data = &at91rm9200_ramc_of_data},
> 	... ...
> 	{ /*sentinel*/ }
> };
> 
> What about you? 

Yes, we agreed that using of_machine_is_compatible is not nice and that
is why I remove that usage in patch 4.

We still have to fill the uhp_udp_mask and that would mean adding a
match on the pmc compatible string. I would prefer not doing that. Or
maybe we can just remove the check, I don't think it it necessary
anymore.

At some point in time, I would like to be able to get rid of the
ramc_ids in mach-at91 but I'm not sure how yet.

Maybe we can do what you suggest after
http://lists.infradead.org/pipermail/linux-arm-kernel/2015-January/316771.html
Because then, the ram detection is local to pm.c


-- 
Alexandre Belloni, Free Electrons
Embedded Linux, Kernel and Android engineering
http://free-electrons.com