[U-Boot] [PATCH 00/10] sunxi: PSCI implementation rewrite in C

[Hans de Goede]

Hi,

On 23-05-16 14:41, Chen-Yu Tsai wrote:
> Hi everyone,
>
> This series rewrites the Allwinner/sunxi PSCI implementation in C, to make
> it easier to maintain and extend for the currently unsupported multi-cluster
> SoCs. The SMP code in the BSP kernels are in C. Having the PSCI code in C
> as well will make it easier to work on.
>
> To be able to convert the platform bits to C, some common PSCI functions
> have to be fixed up according to the ARM calling conventions. Function
> declarations are also needed.
>
> This series is based on sunxi/next. Parts of it will likely conflict with
> the effort to support PSCI 1.0 on the Freescale LS102xA.
>
> Patch 1 fixes up psci_get_cpu_stack_top.
>
> Patch 2 fixes up the PSCI version of v7_flush_dcache_all.
>
> Patch 3 adds function declarations for some of the common PSCI functions.
>
> Patch 4 fixes issues with reserving memory for the secure section.
>
> Patch 5 unifies the CPUCFG_BASE macro names for various sunxi platforms.
>
> Patch 6 groups cpu core related controls together into one struct per core.
> This makes it straightforward to access the controls by the cpu index.
>
> Patch 7 adds a missing header to cpucfg.h
>
> Patch 8 adds some missing fields to cpucfg, which were used in the assembly
> code.
>
> Patch 9 adds the base address for the GIC.
>
> Patch 10 is the new PSCI implementation in C. Almost all of the code is
> converted, with the exception of initial setup of the stack.

Thanks for your work on this, from a sunxi pov it looks good
(once Marc's remarks are fixed).

Also many thanks to Marc for the thorough review. I've been treating the
PSCI stuff as a black-box, only doing mostly style / sanity reviews, so
the thorough review is appreciated a lot.

Lets do a v2 and try to land this soon-ish ?

Regards,

Hans