-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1
Hi all,
I have started a port of AMD Mahogony board to Asrock 939A785GMH. It boots kernel, IDE works, ethernet too. No VGA output need to check why the VGA BIOS extracted from orig image hangs.
http://www.asrock.com/mb/overview.asp?Model=939A785GMH/128M&s=939
It has 1MB socketed SPI flash chip, serial header. There are some overclocking GPIO options hopefully now just set to safe values, check sio_init for details.
http://www.assembler.cz/asrock_boots2.txt
Patch is quite big because the ACPI stuff is just copied.
Signed-off-by: Rudolf Marek r.marek@assembler.cz
Thanks, Rudolf
On 3/16/10 11:27 PM, Rudolf Marek wrote:
Hi all,
I have started a port of AMD Mahogony board to Asrock 939A785GMH. It boots kernel, IDE works, ethernet too. No VGA output need to check why the VGA BIOS extracted from orig image hangs.
Nice... Looking forward to the next iterations of this
Patch is quite big because the ACPI stuff is just copied.
I think it would be better to have a clean ACPI implementation where the different parts are in the directory of each component, like on i945.
Stefan
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1
Hi all,
What about this? Looks like there will be some GSOC for ACPI cleanup/porting? Maybe I can check it in and make it usable?
Ideas?
Rudolf
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1
Ok I checked that in as 5357. The cleanup of console etc stuff included.
Add Asrock 939a785gmh motherboard. The ACPI needs more cleanup, could be done when cleaningthe Mahagony board. The SidePort mode does not work because AMD hardcoded memory type in rs780_gfx.c The UMA is enabled instead. The board boots, network and int VGA works, IDE too.
Rudolf
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1
Hi again,
So I think I'm wasting all my good luck for this year:
Index: rs780_gfx.c =================================================================== - --- rs780_gfx.c (revision 5227) +++ rs780_gfx.c (working copy) @@ -817,6 +817,11 @@ .vendor = PCI_VENDOR_ID_ATI, .device = PCI_DEVICE_ID_ATI_RS780E_INT_GFX, }; +static struct pci_driver pcie_driver_780e __pci_driver = { + .ops = &pcie_ops, + .vendor = PCI_VENDOR_ID_ATI, + .device = 0x9710, +};
Makes the VGA to Init in POST and VGA console works!!!!
I extracted with the help of bios_extract two VGA BIOS files: amipci_00_20.rom amipci_01_20.rom
I used that one which does not contain string DVI in the header, because I have only D-SUB VGA monitor.
Thanks, Rudolf
On 3/17/10 12:25 AM, Rudolf Marek wrote:
Hi again,
So I think I'm wasting all my good luck for this year:
Index: rs780_gfx.c
--- rs780_gfx.c (revision 5227) +++ rs780_gfx.c (working copy) @@ -817,6 +817,11 @@ .vendor = PCI_VENDOR_ID_ATI, .device = PCI_DEVICE_ID_ATI_RS780E_INT_GFX, }; +static struct pci_driver pcie_driver_780e __pci_driver = {
- .ops = &pcie_ops,
- .vendor = PCI_VENDOR_ID_ATI,
- .device = 0x9710,
+};
With "const" for the driver and Signed-off-by, this is Acked-by: Stefan Reinauer stepan@coresystems.de
Maybe we should use the PCI ID instead of PCI_DEVICE_ID_ATI_RS780E_INT_GFX as this would make such bugs easier to find, so you can spare some luck :-)
Stefan
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1
Yeah this is just quick hack.
Cool would be if someone in AMD can go check this for rv620 too
Rudolf
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1
Ok it is in as Committed revision 5356.
Now the motherboard stuff.
Rudolf
-
Rudolf Marek -
Stefan Reinauer