On Fri, Jun 14, 2013 at 04:57:50PM +0200, Tomasz Nowicki wrote:
W dniu 14.06.2013 16:49, Al Stone pisze:
On 06/14/2013 08:15 AM, Graeme Gregory wrote:
On Fri, Jun 14, 2013 at 08:42:47PM +0800, Hanjun Guo wrote:
On 2013-6-14 18:03, Graeme Gregory wrote:
Hi Hanjun,
Excellent on tracking this down, but we need to set it conditionally as our build must not break current x86 machines.
That would be a hard choice for us. If we don't set ACPI_REDUCED_HARDWARE as TRUE on our ARM64 platform, we will meet panic issues, but will break current x86 machines if set TRUE.
ACPICA is a dependent code both for linux and windows, so I think ACPICA will not accept config option for reduced hardware.
I think we should introduce a config option in acpi driver, how about this solution?
Is this acconfig.h actually part of the ACPICA release? If so we are going to have difficulty. If it is just a config file fed to ACPICA code then I think we are free to overide it per platform.
I do not know which files are fixed and which ones we are allowed to play with.
Al, Tomasz do you know the answer to this?
The rule of thumb is that if it is in drivers/acpi/acpica or in drivers/acpi/apei, don't touch it. Those are the ACPICA upstream files. Otherwise, feel free.
Do we have to set this in the source? IIRC, it's one of the options in one of the tables (FADT, maybe?) and I think we should be setting things up based on the flag in the table.
Correct, whether we're using reduced mode or not should be decided based on HW_REDUCED_ACPI flag in FADT table. Right now it is turn on.
Well I have stuck some debugging in fbfadt.c
acpi_gbl_reduced_hardware = FALSE; if (acpi_gbl_FADT.flags & ACPI_FADT_HW_REDUCED) { pr_err("ACPI: REDUCED HARDWARE MODE\n"); acpi_gbl_reduced_hardware = TRUE; }
And my dmesg contains.
[ 0.000000] ACPI: REDUCED HARDWARE MODE
So it looks like drivers/acpi doe not support HW_REDUCED_MODE properly yet. I guess thats another card to add.
Thanks
Graeme