[PATCH]Fix ACPI in HIGH-and-LOW tables configuration

[Patrick Georgi]

Hi,

I messed up and added some "harmless" code right before pushing this change.
Just that it wasn't harmless. I'm sorry.

Anyway, with this patch, ACPI works again for setups with both 
HAVE_HIGH_TABLES and HAVE_LOW_TABLES, as the RSDP contains valid data now 
(without it the RSDT is bogus: it is read out before it is written)

Signed-off-by: Patrick Georgi

Patrick

[Bao, Zheng]

It doesn't fix the HIGH_TABLES issue on dbm690t.

Zheng

-----Original Message-----
From: [e-mail]
[mailto:[e-mail]] On Behalf Of Patrick Georgi
Sent: Wednesday, May 13, 2009 2:46 PM
To: [e-mail]
Subject: [coreboot] [PATCH]Fix ACPI in HIGH-and-LOW tables configuration

Hi,

I messed up and added some "harmless" code right before pushing this
change.
Just that it wasn't harmless. I'm sorry.

Anyway, with this patch, ACPI works again for setups with both 
HAVE_HIGH_TABLES and HAVE_LOW_TABLES, as the RSDP contains valid data
now 
(without it the RSDT is bogus: it is read out before it is written)

Signed-off-by: Patrick Georgi

Patrick

[Myles Watson]

> -----Original Message-----
> From: [e-mail] [mailto:[e-mail]]
> On Behalf Of Patrick Georgi
> Sent: Wednesday, May 13, 2009 12:46 AM
> To: [e-mail]
> Subject: [coreboot] [PATCH]Fix ACPI in HIGH-and-LOW tables configuration
>
> Hi,
>
> I messed up and added some "harmless" code right before pushing this
> change.
> Just that it wasn't harmless. I'm sorry.
>
> Anyway, with this patch, ACPI works again for setups with both
> HAVE_HIGH_TABLES and HAVE_LOW_TABLES, as the RSDP contains valid data now
> (without it the RSDT is bogus: it is read out before it is written)

Index: src/arch/i386/boot/tables.c
===================================================================
--- src/arch/i386/boot/tables.c	(Revision 4279)
+++ src/arch/i386/boot/tables.c	(Arbeitskopie)
@@ -116,14 +116,16 @@
 #if HAVE_HIGH_TABLES == 1
 #if HAVE_LOW_TABLES == 1
    unsigned long high_rsdp=ALIGN(high_table_end, 16);
-	unsigned long rsdt_location=(unsigned
long*)(((acpi_rsdp_t*)high_rsdp)->rsdt_address);
-	acpi_write_rsdp(rom_table_end, rsdt_location);
-	rom_table_end = ALIGN(ALIGN(rom_table_end, 16) +
sizeof(acpi_rsdp_t), 16);
 #endif
    if (high_tables_base) {
    	high_table_end = write_acpi_tables(high_table_end);
    	high_table_end = (high_table_end+1023) & ~1023;
    }
+#if HAVE_LOW_TABLES == 1
+	unsigned long rsdt_location=(unsigned
long*)(((acpi_rsdp_t*)high_rsdp)->rsdt_address);
+	acpi_write_rsdp(rom_table_end, rsdt_location);
+	rom_table_end = ALIGN(ALIGN(rom_table_end, 16) +
sizeof(acpi_rsdp_t), 16);
+#endif
 #else
 #if HAVE_LOW_TABLES == 1
    rom_table_end = write_acpi_tables(rom_table_end);
> Signed-off-by: Patrick Georgi

Can you move the declaration of high_rsdp too?  One less #if.

I don't see how this changes the functionality.  Which variable are you
worried about?  Sorry if I'm being blind.

Thanks,
Myles

[Patrick Georgi]

Am 13.05.2009 14:20, schrieb Myles Watson:
> Can you move the declaration of high_rsdp too?  One less #if.

high_table_end changes in the meantime. I could store it somewhere else, 
but I'm not sure if that's a cleaner solution
> I don't see how this changes the functionality.  Which variable are you
> worried about?  Sorry if I'm being blind.

-	unsigned long rsdt_location=(unsigned 
long*)(((acpi_rsdp_t*)high_rsdp)->rsdt_address);

high_rsdp->rsdt_address is filled in by write_acpi_table().

Patrick

[Myles Watson]

On Wed, May 13, 2009 at 6:28 AM, Patrick Georgi  wrote:
> Am 13.05.2009 14:20, schrieb Myles Watson:
> > Can you move the declaration of high_rsdp too?  One less #if.
>
> high_table_end changes in the meantime. I could store it somewhere else, but
> I'm not sure if that's a cleaner solution

Yep.
> > I don't see how this changes the functionality.  Which variable are you
> > worried about?  Sorry if I'm being blind.
> unsigned long rsdt_location=(unsigned
>
> long*)(((acpi_rsdp_t*)high_rsdp)->rsdt_address);
>
> high_rsdp->rsdt_address is filled in by write_acpi_table().

Acked-by: Myles Watson

Thanks,
Myles