[coreboot] coreboot-2 on a DFI NF570 motherboard

Corey Osgood corey.osgood at gmail.com
Wed Oct 29 20:27:15 CET 2008


On Wed, Oct 29, 2008 at 2:30 PM, Chris Lingard <chris at stockwith.co.uk>
wrote:
>
> Motherboards DFI NF-570 range, (NF570SLI-M2/G, NF570-M2/G and NF550-M2/G).
>
>
> This is a suggestion to port coreboot to the NF570 series. I have already
posted a patch, but this just duplicates the entire M57SLI stuff just for a
two line change.  If the build script could set a flag, shown in the example
as dfi_nf570, then coreboot would work on the above motherboards.
>
> After discussion on IRC this would be a better implementation
>
>
> From mptable.c
>
>        PCI_INT(0,sbdn+6,1, 23); /* HD Audio */
>        PCI_INT(0,sbdn+5,0, 20); /* SATA */
>        PCI_INT(0,sbdn+5,2, 22); /* SATA */
>        PCI_INT(0,sbdn+5,1, 23); /* SATA */
>
> #ifdef   dfi_nf570
>
>        PCI_INT(0,sbdn+5,2, 22); /* SATA */
>
>        PCI_INT(0,sbdn+8,0, 21); /* GBit Ether */
>
> #else
>
>        PCI_INT(0,sbdn+5,2, 21); /* SATA */
>
>        PCI_INT(0,sbdn+8,0, 22); /* GBit Ether */
>
> #endif
>
>        /* The PCIe slots, each on its own bus */
>
> Taken and modified from the Gigabyte M57SLI tree
> signed-of-by: Chris Lingard   chris at stockwith.co.uk
> Tested on a DSI NF570-M2/G motherboard
>
> The use of this motherboard eases the learning path into coreboot, because
it has a socketed BIOS, and needs no modification.  I had never hot plugged
a chip nor flashed a BIOS until this week

I don't like this for one simple reason, I don't like the idea of trying to
explain to users that they have to compile target x with config variable y
set to make it work on board z, because once we start down that road it's
going to get messy quick.

If we really want to avoid duplicated code, my proposal would be a mainboard
Config.lb like the following. I'm stuck in windows atm, so no way to test if
it actually would build correctly, but I think it should. There are reasons
we've avoided doing stuff like this in the past though, mainly because of
the possibility of a change to one target breaking multiple other targets
using the same code, and I'm much more comfortable with a bit of duplicated
code.

-Corey

1##
2## This file is part of the coreboot project.
3##
4## Copyright (C) 2007 AMD
5## Written by Yinghai Lu <yinghailu at amd.com> for AMD.
6##

<snip>

72##
73## Build the objects we have code for in this directory.
74##
75

# These files duplicated, reflect different mainboard name and irq tables

76driver mainboard.o
77#needed by irq_tables and mptable and acpi_tables
78object get_bus_conf.o
79
80if HAVE_MP_TABLE object mptable.o end
81if HAVE_PIRQ_TABLE object irq_tables.o end
82#object reset.o
83if USE_DCACHE_RAM
84
85        if CONFIG_USE_INIT

# These don't have to be

86                makerule ../../gigabyte/m57sli/cache_as_ram_auto.o
87                        depends "../../gigabyte/m57sli/cache_as_ram_auto.c
option_table.h"
88                        action "$(CC) $(DISTRO_CFLAGS) -I$(TOP)/src -I.
$(CPPFLAGS) ../../gigabyte/m57sli/cache_as_ram_auto.c -Os -nostdinc
-nostdlib -fno-builtin -Wall -c -o $@"
89                end
90        else
91                makerule ../../gigabyte/m57slicache_as_ram_auto.inc
92                        depends "../../gigabyte/m57sli/cache_as_ram_auto.c
option_table.h"
93                        action "$(CC) $(DISTRO_CFLAGS) -I$(TOP)/src -I.
$(CPPFLAGS) ../../gigabyte/m57sli/cache_as_ram_auto.c -Os -nostdinc
-nostdlib -fno-builtin -Wall -c -S -o $@"
94                        action "perl -e 's/.rodata/.rom.data/g' -pi $@"
95                        action "perl -e 's/.text/.section .rom.text/g' -pi
$@"
96                end
97        end
98
99end
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.coreboot.org/pipermail/coreboot/attachments/20081029/4b7eb467/attachment.html>


More information about the coreboot mailing list