On 2014-6-30 18:46, Catalin Marinas wrote:
On Fri, Jun 27, 2014 at 04:49:36AM +0100, Hanjun Guo wrote:
diff --git a/drivers/acpi/Kconfig b/drivers/acpi/Kconfig index 7de5e3f..33d6dbb 100644 --- a/drivers/acpi/Kconfig +++ b/drivers/acpi/Kconfig @@ -8,10 +8,10 @@ config ACPI_SCAN_BIOS_NOT_EFI menuconfig ACPI bool "ACPI (Advanced Configuration and Power Interface) Support" depends on !IA64_HP_SIM
- depends on IA64 || X86
- depends on IA64 || X86 || ARM64
I still don't understand what the point of enabling ACPI for arm64 during this series is. Do you get any working arm64 functionality (on hardware or model) without subsequent patches? If it's just for compilation reasons, the best we could do is depending on (ARM64 && COMPILE_TEST) but even though I would not merge this patch until we have most of the arm64 required features in place (some of which are introduced by the upcoming ACPI version).
it is ok to me to merge all the patches together, but if Rafael is happy with the clean up patches (patch 1~3) for ACPI core, they can be merged first.
depends on PCI select PNP
- default y
- default y if !ARM64
For the benefit of single Image, I think you can default to y here.
It ok to me. if we default to y here, devicetree will not be unflattened in default, is it ok to you? you can refer to "[PATCH 12/13] ARM64 / ACPI: if we chose to boot from acpi then disable FDT".
Thanks Hanjun