F2A85-V PRO as good target?

[Rudolf Marek]

Hi all,

I think it is time to finally get new FM2 board! I have in mind some version of 
the F2A85-V PRO. Best would be to coordinate bit the efforts and choose some 
board which is long term supported. Also, everything costs quite some money and 
it would be nice if maybe someone already owns it to get some lspci/flash dumps 
etc in advance.

So far I have learned:

1) it has socketed BIOS
2) there seems to be some hw way how to unbrick that, but question is if it will 
accept our images

USB BIOS Flashback - Easy, worry-free USB BIOS

I suspect it is done by some 8051 controller, possibly some of those chips:
http://detail.zol.com.cn/picture_index_975/index9747187_0_p336189.shtml

Looks like independent of any other chips, it could be cool feature for coreboot 
users.

3) it has serial port
4) it has NCT6779D which has no public datasheet
5) I was able to decompress the EFI firmware (skip first 2KB) with the 
bios_extract with some enhancements for compressed volumes. I think I got this 
from Patrick? I found there at least one 8051 firmware which looks like it is 
providing the EPU function. I suspect this runs inside southbridge, because the 
usual AMD signature is wiped with "xxxxxxxx" at 0x2000 ;) I think this can be a 
PITA but I dont know if it will work without this. Or if we would have to 
provide a way how to include that in the resulting image. I did not find a 
decompressed version of this fw. I suspect it lives iniside the flash after 
first flashing or so...

5) some of firmware files are quite funny and include cryptographic keys for 
microsoft and canonical also, it seems the "secure part" of the image updater is 
packed with UPX and this is then used by flashers (maybe extra other SMI layer 
is involved I dont know) (look for ASEX_update_bios_firmware does the trick)

All in all, I'm nearly convinced I will try to do it, but of course it is hard 
not to imagine some showstoper and all the money invested to new coreboot hw as 
wasted.

Thanks
Rudolf

[Peter Stuge]

Rudolf Marek wrote:
> there seems to be some hw way how to unbrick that, but question
>
> is if it will accept our images
>
> USB BIOS Flashback - Easy, worry-free USB BIOS
>
> I suspect it is done by some 8051 controller, possibly some of those chips:
> http://detail.zol.com.cn/picture_index_975/index9747187_0_p336189.shtml
>
> Looks like independent of any other chips, it could be cool feature
> for coreboot users.

I was curious about what this really was and looked into the manual.

--8<-- Page 2-19, "Onboard switches"
5. BIOS-FLBK switch

The BIOS Flashback buttton offers the most convenient way to update
the BIOS. It allows overclockers to update new BIOS versions without
having to enter their existing BIOS or operating system. Just plug in
a USB storage and press the BIOS Flashback button for 3 seconds, and
the BIOS is automatically flashed using standby power. Hassle-free
updating for the ultimate convenience!

[mainboard drawing showing BIOS_FLBK button location]
F2A85-V PRO BIOS Flashback switch

1. Download the USB BIOS Flashback wizard from the ASUS service
   website (www.asus.com). Save it to a USB portable disk, and place
   it in the root directory.
2. Plug the USB flash drive into the bottom port of ESATA6G_USB3_12
   connector in back I/O.
3. Press the USB BIOS Flashback button for 3 seconds, and the light
   will begin to flash (on the third second).
4. The Flashback function is enabled once the light starts to flash.
5. The flashing rate of the light signal accelerates along with the
   updating speed.
6. BIOS update is complete once the light signal goes out.

[image showing the lower USB 3.0 port marked in red]

* Do not unplug the portable disk, power off the system or press the
  CLR_CMOS button while BIOS update is ongoing, otherwise update will
  be interrupted. In case of interruption, please follow the
  recommended procedures again to complete BIOS update.

* If the light flashes for five seconds and turns into a stable
  light, this means that the USB BIOS Flashback is not operating
  properly. This may be caused by:
  1. Improper installation of the portable disk.
  2. Error in file name or incompatibility in file format. If this is
  the case, please restart the system to turn off the light.

* Updating BIOS may have risks. If the BIOS program is damaged during
  the process causing the system unable to reboot, please contact the
  local service station for help.
-->8--

The fact that one of the USB 3.0 ports are used for this, and the
fact that they write "If the BIOS program is damaged .. please
contact the local service station" makes me believe that this is
still the same old story, but instead of flashing code being in
the bootblock and running on the CPU the flashing code is maybe in
the IMC firmware instead, running before the CPU.

//Peter

[Rudolf Marek]

Hi all,

I got the superio datasheet from Nuvoton. It came the other day I wrote them. I 
think this was fastest ever request fulfilled I made to any east asia company ;)

I'm still considering buying this mainboard with A10 CPU and some RAM. Problem 
is that it costs quite a lot of money (280E) which are at risk because of following:

1) lack of documentation of NB

The BKDG does not contain any docu for PCI devices < 0x18
Anone knows if this changes? I'm OK with some free software friendly NDA

2) lack of SB documenation

the hudson D4 has completely different PCI ids and I could not find that D4 
is supported by current AGESA.
Anone knows if this changes? I'm OK with some free software friendly NDA

3) there is still the firmware stuff I'm afraid of.

On the other hand coreboot really deserves some modern motherboard as we all 
know. I would like to invite other folks discussion ;)

You may ask why I'm so pesky about a documentation, but the experience showed no 
code is perfect ;) I think I could buy cheapest CPU and save about 70E, but I 
had in plan to eventually use it as my new workstation.

I'm looking forward for any input,

Thanks
Rudolf

[Rudolf Marek]

Hi all

Paul Menzel found a cheaper version of the motherboard. It is ASUS F2A85-M LE it 
has a socketed BIOS and serial line. The ITE superIO is currently unknown.

Thanks
Rudolf

[Rudolf Marek]

On 22.10.2012 23:57, Rudolf Marek wrote:
> Hi all
>
> Paul Menzel found a cheaper version of the motherboard. It is ASUS F2A85-M LE it
> has a socketed BIOS and serial line. The ITE superIO is currently unknown.

According to: http://www.easycom.com.ua/mboard/asus_f2a85-m_le/?lang=ru

The superIO chip is  ITE IT8623E, and I have no clue about that.

Thanks
Rudolf
> Thanks
> Rudolf

F2A85-V PRO as good target? - VGA ROMs?

[Rudolf Marek]

Hi

Just a small update. I think all BIOSes I have seen have UEFI only VGA BIOS. Now 
question is what we will have to do? I found in the EFI file the 0xaa 0x55 style 
BIOS (20KB big) but the init code has only retf and most likely it contains only 
an atom bios. The int10 may be handled  through int10smm.efi module.

I think we either need VGA BIOS blob or implement the atomBIOS interpreter into 
Seabios...

Thanks
Rudolf

F2A85-V PRO as good target? - VGA ROMs?

[Denis 'GNUtoo' Carikli]

On Thu, 25 Oct 2012 08:59:04 +0200
Rudolf Marek  wrote:
> I think we either need VGA BIOS blob or implement the atomBIOS
> interpreter into Seabios...

I've done a little experiment with SeaBIOS and an ATI card:
The following "patch" resulted in the display beeing black until the
radeon driver was loaded by Xorg:
diff --git a/src/optionroms.c b/src/optionroms.c
index 00697b2..dec7a37 100644
--- a/src/optionroms.c
+++ b/src/optionroms.c
@@ -25,6 +25,10 @@ static void
 __callrom(struct rom_header *rom, u16 offset, u16 bdf)
 {
     u16 seg = FLATPTR_TO_SEG(rom);
+#if 1
+    dprintf(1, "Skipping option rom at %04x:%04x\n", seg, offset);
+    return;
+#else
     dprintf(1, "Running option rom at %04x:%04x\n", seg, offset);

struct bregs br;
@@ -40,6 +44,7 @@ __callrom(struct rom_header *rom, u16 offset, u16 bdf)
     farcall16big(&br);
     finish_preempt();

+#endif
     debug_serial_setup();
 }

Denis.

F2A85-V PRO as good target? - VGA ROMs?

[Rudolf Marek]

On 25.10.2012 08:59, Rudolf Marek wrote:
> Hi
>
> Just a small update. I think all BIOSes I have seen have UEFI only VGA BIOS. Now
> question is what we will have to do? I found in the EFI file the 0xaa 0x55 style
> BIOS (20KB big) but the init code has only retf and most likely it contains only
> an atom bios. The int10 may be handled  through int10smm.efi module.
>
> I think we either need VGA BIOS blob or implement the atomBIOS interpreter into
> Seabios...

Still no reply from AMD folks :/ are there some on this list?

In the meanwhile I tried to dump the BIOS style rom I found yesterday with a 
atombios dumper and it does indeed show atombios tables.

So, If we don't find a legacy VGA BIOS for trinity, we could just install this 
as fake VGA bios and then let linux to post the GPU. Or if someone likes really 
adventures we could possibly build into coreboot an AtomBIOS interpreter and do 
the legacy VGA modeset ourself.

Thanks
Rudolf

F2A85-V PRO as good target? - VGA ROMs?

[Rudolf Marek]

Hi all,

Just an update. I think I go for F2A85-M (no sufixes), because the LE board will 
not be available some time here in Czech Republic. I keep you posted what I have 
bought at the end. This board is similar to LE board but with 4 DIMM slots and 
has LPC (TPM) connector.

I checked the SEC phase of BIOS and it does indeed something with TPM. Moreover 
there is some device at io 0x804 which has to be 0xCxx in order not to panic 
very early.

I don't know what superio it is, I checked some photos and it looks like it has 
the same one.

I compared the bootblocks and the "LE" bootblock has also this "0xCxx" thingie. 
Anyone has contacts at ITE? It would be handy to get the datasheet once I know 
what exact ITE chip is at the non LE board.

Thanks
Rudolf

F2A85-V PRO as good target? - update

[Rudolf Marek]

Hi,

I bought F2A85-M (no sufixes). So far status is no serial, no port 80 in PCI. 
ITE superio is IT8603. Nothing works. I event dont know if CPU is executing 
something. I guess I will need another thread ;)

Thanks
Rudolf