Hi,
Le jeu., oct. 13 2022 at 08:46:39 +0200, Arnd Bergmann <arnd@xxxxxxxx>
a écrit :
On Thu, Oct 13, 2022, at 8:37 AM, Siarhei Volkau wrote:
пн, 10 окт. 2022 г. в 01:29, kernel test robot
<lkp@xxxxxxxxx>:
config: ia64-allyesconfig
config: arm64-randconfig-r035-20221010
> 142 #define CGU_REG_CPCCR ((void *)CKSEG1ADDR(0x10000000))
0-DAY CI Kernel Test Service
I know CKSEG1ADDR is MIPS specific, might be it needed to disable
COMPILE_TEST
on the driver?
Since early syscon isn't mainlined yet I don't see any other way at
the moment.
Any suggestions on that, folks?
This looks like some setup that belongs into the bootloader. If you
are
handing over the console from bootloader to kernel, the hardware
should
already be in a working state, with no need to touch it during early
boot.
If you are dealing with broken bootloaders that are not under your
control,
having this code in the architecture specific early boot as a fixup
would be better than putting it into the driver.
Agreed. I am not fond of having a driver poking into an unrelated
subsystem's memory area.
Just disable the divider in ingenic_fixup_fdt() in
arch/mips/generic/board-ingenic.c.
Cheers,
-Paul
