On 2013-9-14 4:16, Moore, Robert wrote:
-----Original Message----- From: Moore, Robert Sent: Friday, September 13, 2013 6:08 AM To: Hanjun Guo; Rafael J. Wysocki; Len Brown Cc: Zheng, Lv; linux-acpi@vger.kernel.org; patches@linaro.org; linaro- kernel@lists.linaro.org; linaro-acpi@lists.linaro.org Subject: RE: [PATCH] ACPICA / hwreg: Use acpi_gbl_reduced_hardware to prevent accessing PM registers
NOT_IMPLEMENTED refers to the software only; therefore AE_SUPPORT should be returned.
Otherwise, seems like this may be a good idea.
On the other hand, could one not argue that the host OS should darn well know that it is executing on a hardware-reduced platform and not call these functions in the first place?
Well, there is a macro ACPI_REDUCED_HARDWARE and used to switch off the compiling of these functions on hardware-reduced platform, if ACPI_REDUCED_HARDWARE is TRUE, those functions will never be called.
But is that reasonable? When somebody want to use the ACPI code on hardware-reduced platform, they should introduce a patch just like:
- #define ACPI_REDUCED_HARDWARE FALSE + #define ACPI_REDUCED_HARDWARE TRUE in include/acpi/acconfig.h
and this patch does not exist in upstream, and it will never be accepted by upstream too, it is weird. so we should proposal another way.
I thought about introducing a kernel config such as CONFIG_ACPI_REDUCED_HARDWARE to enable ACPI_REDUCED_HARDWARE, but it is just the same as a macro, and I drop it.
So, this is a problem should be solved for hardware-reduced platform, use falg acpi_gbl_reduced_hardware seems a good solution.
Thanks Hanjun
Bob