Re: [PATCH 3/3] sunxi: psci: implement PSCI on R528

[Andre Przywara <andre.przywara@arm.com>]

On Tue, 15 Aug 2023 13:17:33 -0600
Sam Edwards <cfsworks@gmail.com> wrote:

Hi Sam,

> On 8/14/23 08:16, Andre Przywara wrote:
> > Hi Sam,
> >   
> >> This patch adds the necessary code to make nonsec booting and PSCI
> >> secondary core management functional on the R528/T113.  
> > 
> > Unfortunately this patch breaks the build on older 32-bit SoCs, as
> > SUNXI_CPUX_BASE is not defined there. That's a typical problem of the
> > "C-if" approach, but typically there is a clean, albeit not trivial,
> > solution:
> > 
> > It seems like SUNXI_CPUX_BASE and SUNXI_CPUCFG_BASE are mutually
> > exclusive, and I actually carry a "#define SUNXI_CPUCFG_BASE 0" hack in my
> > patches already.
> > So I wonder if we should unify those two into SUNXI_CPUCFG_BASE, with the
> > following rework:  
> 
> The SUNXI_CPUX_BASE -> SUNXI_CPUCFG_BASE rename worked excellently. 
> We're having the same problem with SUNXI_R_CPUCFG_BASE as well, though. 
> How do you want to handle that?

So that's a bit more nasty indeed. I don't even know if R_CPUCFG really
makes sense here, as the _R_ term typically refers to the management
processor, which the D1/R528 don't have. Or at least the always-on power
domain, but then again this hardly relates to the secondary entry
point. I think the name was just used because the address matches the
one used in the H6. Anyway, I got the impression that Allwinner just
uses registers wherever they find them, and that they don't care too
much about logical grouping or compatibility.

So taking a step back, I wonder if we should actually just define a
CONFIG_SUNXI_CPU_SOFT_ENTRY (or so) *Kconfig* symbol, which holds that
address, and let the per-SoC definition be solved in Kconfig instead.
Because SUNXI_R_CPUCFG_BASE and also SUNXI_RTC_BASE seem to be just
used as the base address for that purpose, with some magic offset
added, across all of U-Boot (ARMv8 FEL and v7 PSCI).

So can you try to work on that base? I will take care of
armv8/fel_utils.S, which uses some post-increment assembly trick to
keep the code small, which wouldn't work anymore. But I have an idea
how to solve this.

Cheers,
Andre