[coreboot] [dbm690t] The new acpi table doesn't seem to be correct.

Bao, Zheng Zheng.Bao at amd.com
Tue May 12 13:51:23 CEST 2009

Please check the coreboot log.
Based on the dmesg, The kernel can find the RSD PTR in 0xf0400. But who
knows what happens next?


Devices initialized
High Tables Base is 3fff0000.
Writing IRQ routing tables to 0xf0000...write_pirq_routing_table done.
Writing IRQ routing tables to 0x3fff0000...write_pirq_routing_table
ACPI: Writing ACPI tables at 3fff0400...
ACPI: added table 1/9 Length now 40
ACPI: added table 2/9 Length now 44
processor_brand=AMD Athlon(tm) 64 X2 Dual Core Processor 3400+
Pstates Algorithm ...
Pstate_freq[0] = 1800MHz        Pstate_vid[0] = 18      Pstate_volt[0] =
1100mv Pstate_power[0]
 = 35000mw
Pstate_freq[1] = 1000MHz        Pstate_vid[1] = 22      Pstate_volt[1] =
1000mv Pstate_power[1]
 = 16069mw
ACPI: added table 3/9 Length now 48
ACPI:    * DSDT @ 3fff07ee Length 27c8
pm_base: 0x0800
ACPI: added table 4/9 Length now 52
ACPI: done.
rom_table_end=f0400, rsdt_location=3fff0424
rom_table_end=f0430, rsdt_location=3fff0424
Wrote the mp table end at: 00000020 - 00000134
Wrote the mp table end at: 3fff3410 - 3fff3524
Moving GDT to 0x3fff3800...ok
Multiboot Information structure has been written.
Writing high table forward entry at 0x00000500
Wrote coreboot table at: 00000500 - 00000518  checksum 83df
New low_table_end: 0x00000500
Now going to write high coreboot table at 0x3fff3c00
rom_table_end = 0x3fff3c00
Adjust low_table_end from 0x00000500 to 0x00001000
Adjust rom_table_end from 0x3fff3c00 to 0x40000000
Adding high table area
uma_memory_start=0x38000000, uma_memory_size=0x0
Wrote coreboot table at: 3fff3c00 - 3fff3df4  checksum 1bc3

elfboot: Attempting to load payload.
rom_stream: 0xfffc0000 - 0xfffdffff
Found ELF candidate at offset 0
header_offset is 0


-----Original Message-----
From: Patrick Georgi [mailto:patrick at georgi-clan.de] 
Sent: Tuesday, May 12, 2009 7:23 PM
To: Bao, Zheng
Cc: coreboot at coreboot.org
Subject: Re: [coreboot] [dbm690t] The new acpi table doesn't seem to be

Am 12.05.2009 10:38, schrieb Bao, Zheng:
> Hi, all,
> The new auto-generated acpi table doesn't seem to be correct on
> Dose anybody have any idea?

> [    0.000000] ACPI: RSDP signature @ 0xFFFF8100000F0400 checksum 0
> [    0.000000] ACPI: RSDP 000F0400, 0024 (r2 CORE  )
With both HAVE_HIGH_TABLES and HAVE_LOW_TABLES, the code writes the ACPI

tables to high memory (64kb somewhere in top memory), and merely writes 
the RSDP to low memory.
With only HAVE_HIGH_TABLES, the low memory RSDP is not written.
With only HAVE_LOW_TABLES, the entire ACPI block is written to low

At least that's the plan.

A coreboot log with high log level (debug or spew, so 8 or 9) should 
give some hints what coreboot does.


More information about the coreboot mailing list