RE: [PATCH 1/3] ACPICA: Add acpi_gbl_force_rsdt variable

[Thomas Renninger <trenn@suse.de>]

On Thu, 2008-05-08 at 09:32 -0700, Moore, Robert wrote:
> Has there any investigation into how windows handles this?
It's too late now, anyway.
This is about 3 years old machines, we do not want to change the
behavior XSDT vs RSDT in general and produce some regressions as a
side-effect, but just fix up some known broken machines.

I expect that XP did violate a bit the spec and tended to prefer the
RSDT. I don't know how to check all this, but I expect that Vista is
preferring the XSDT again.
Or there always were bugs in acpica setting up the FADT via XSDT causing
this.

Providing OS a knob at this place shouldn't be a big deal?

   Thomas

> >-----Original Message-----
> >From: Thomas Renninger [mailto:trenn@suse.de]
> >Sent: Thursday, May 08, 2008 8:10 AM
> >To: linux-acpi; Len Brown
> >Cc: Zhao, Yakui; me@markdoughty.co.uk; Moore, Robert; Starikovskiy,
> Alexey
> >Y
> >Subject: [PATCH 1/3] ACPICA: Add acpi_gbl_force_rsdt variable
> >
> >I expect separate patches should now be patched against
> >acpica when possible?
> >
> >Does this also make sense with patches which have dependencies
> >to other patches for the Linux kernel? In this case there is not a
> >"does not compile" dependency, but at least they belong together.
> >
> >I wonder what works out for acpica and Linux kernel maintainers best,
> >please advise.
> >
> >
> >----
> >
> >ACPICA: Add acpi_gbl_force_rsdt variable
> >
> >Signed-off-by: Thomas Renninger <trenn@suse.de>
> >Tested-by: Mark Doughty <me@markdoughty.co.uk>
> >
> >
> >---
> > drivers/acpi/tables/tbutils.c     |    3 ++-
> > drivers/acpi/utilities/utglobal.c |    1 +
> > include/acpi/acglobal.h           |    1 +
> > 3 files changed, 4 insertions(+), 1 deletion(-)
> >
> >Index: linux-acpi-
> >2.6_video_native_vs_vendor/drivers/acpi/utilities/utglobal.c
> >===================================================================
> >--- linux-acpi-
> >2.6_video_native_vs_vendor.orig/drivers/acpi/utilities/utglobal.c
> >+++
> linux-acpi-2.6_video_native_vs_vendor/drivers/acpi/utilities/utglobal.c
> >@@ -76,6 +76,7 @@ u8 acpi_gbl_method_executing = FALSE;
> > /* System flags */
> >
> > u32 acpi_gbl_startup_flags = 0;
> >+int acpi_gbl_force_rsdt = 0;
> >
> > /* System starts uninitialized */
> >
> >Index: linux-acpi-2.6_video_native_vs_vendor/include/acpi/acglobal.h
> >===================================================================
> >--- linux-acpi-2.6_video_native_vs_vendor.orig/include/acpi/acglobal.h
> >+++ linux-acpi-2.6_video_native_vs_vendor/include/acpi/acglobal.h
> >@@ -246,6 +246,7 @@ ACPI_EXTERN u8 acpi_gbl_system_awake_and
> >
> > extern u8 acpi_gbl_shutdown;
> > extern u32 acpi_gbl_startup_flags;
> >+extern int acpi_gbl_force_rsdt;
> > extern const char *acpi_gbl_sleep_state_names[ACPI_S_STATE_COUNT];
> > extern const char *acpi_gbl_highest_dstate_names[4];
> > extern const struct acpi_opcode_info
> >acpi_gbl_aml_op_info[AML_NUM_OPCODES];
> >Index:
> linux-acpi-2.6_video_native_vs_vendor/drivers/acpi/tables/tbutils.c
> >===================================================================
> >--- linux-acpi-
> >2.6_video_native_vs_vendor.orig/drivers/acpi/tables/tbutils.c
> >+++ linux-acpi-2.6_video_native_vs_vendor/drivers/acpi/tables/tbutils.c
> >@@ -421,7 +421,8 @@ acpi_tb_parse_root_table(acpi_physical_a
> >
> > 	/* Differentiate between RSDT and XSDT root tables */
> >
> >-	if (rsdp->revision > 1 && rsdp->xsdt_physical_address) {
> >+	if (rsdp->revision > 1 && rsdp->xsdt_physical_address
> >+	    && !acpi_gbl_force_rsdt) {
> > 		/*
> > 		 * Root table is an XSDT (64-bit physical addresses). We
> must
> >use the
> > 		 * XSDT if the revision is > 1 and the XSDT pointer is
> present,
> >as per
> >
>