Re: [PATCH] ARM: dts: meson8: fix the size of the PMU registers

[Kevin Hilman]

Martin Blumenstingl <martin.blumenstingl@xxxxxxxxxxxxxx> writes:

> The PMU registers are at least 0x18 bytes wide. Meson8b already uses a
> size of 0x18. The structure of the PMU registers on Meson8 and Meson8b
> is similar but not identical.
>
> Meson8 and Meson8b have the following registers in common (starting at
> AOBUS + 0xe0):
>   #define AO_RTI_PWR_A9_CNTL0 0xe0 (0x38 << 2)
>   #define AO_RTI_PWR_A9_CNTL1 0xe4 (0x39 << 2)
>   #define AO_RTI_GEN_PWR_SLEEP0 0xe8 (0x3a << 2)
>   #define AO_RTI_GEN_PWR_ISO0 0x4c (0x3b << 2)
>
> Meson8b additionally has these three registers:
>   #define AO_RTI_GEN_PWR_ACK0 0xf0 (0x3c << 2)
>   #define AO_RTI_PWR_A9_MEM_PD0 0xf4 (0x3d << 2)
>   #define AO_RTI_PWR_A9_MEM_PD1 0xf8 (0x3e << 2)
>
> Thus we can assume that the register size of the PMU IP blocks is
> identical on both SoCs (and Meson8 just contains some reserved registers
> in that area) because the CEC registers start right after the PMU
> (AO_RTI_*) registers at AOBUS + 0x100 (0x40 << 2).
>
> The upcoming power domain driver will need to read and write the
> AO_RTI_GEN_PWR_SLEEP0 and AO_RTI_GEN_PWR_ISO0 registers, so the updated
> size is needed for that driver to work.
>
> Fixes: 4a5a27116b447d ("ARM: dts: meson8: add support for booting the secondary CPU cores")
> Signed-off-by: Martin Blumenstingl <martin.blumenstingl@xxxxxxxxxxxxxx>

Queued as a fix for v5.5-rc,

Kevin