Calling all Asus P8x7x board owners, and Q about NCT5535D

[Keith Hui]

Hello,

I recently got my hands on boardviews of many Asus P8x7x series
mainboards, and some offshoots such as Z77-A and Sabertooth Z77, and
have also found a way to extract VBT data and the USB port config out
of the vendor firmware image (which isn't really a straight image, but
what looks like an update capsule), allowing me to attempt something
quite ambitious given I only have the P8Z77-M on hand: Bring coreboot
to the entire family.

Again because I only have the vendor firmware image and boardview to
go by, I will only be able to put out an untested new board patch,
hence I am calling for anyone with these boards, that aren't already
in the tree, to help me test them, and to get things I can't get out
of firmware image, such as HDA verbs and GPIO setup, by running
autoport on your board.

Additionally, if you have a P8Z77-V (Hello Bill!), please test this patch:
https://review.coreboot.org/c/coreboot/+/85413
And if you have a P8Z77-M PRO, please test your serial port, and if changing
it to use serial port A like P8Z77-M does, make the serial port work.

Now, during my survey of the board family, I found that:
1. They link their power LEDs to only one of two GPIOs on the PCH: 8
or 27. That's easily determined from the boardview.
2. Except for P8C WS which I can't get a boardview for, this family
only uses one of two Nuvoton super I/Os: NCT5535D or NCT6779D. 6779 is
well known, but I cannot find the datasheet for 5535, and the pinout
for the closest thing I can find, NCT5532D, have the pins shifted by
approximately 3 pins, and there may be more differences.

Does anyone know anything about NCT5535D or can hook me up with a datasheet?

Thanks
Keith

[R S]

Hi Keith,
Appreciate the undertaking. I have a Z77 Sabertooth which will be replaced
with a different system in the near future, so I would be willing to test
with a precompiled ROM. If the ROM would be publically available at one
point on Github or Gitlab where issues can be reported conveniently
similarly to what the Dasharo/3mdeb guys do that would be wonderful. Also
for the sake of being able to recover easily (without an RTE or CH341A)
from a no-POST situation I'd like to have the USB BIOS FlashBack feature
implemented *first* before doing any test. Michał Żygowski has
reverse-engineered this for the MSI Z690 boards (USB FlashBIOS) and it's
working great. I'm not an expert here, but I would assume that there are
possibly similarities in how this works across vendors?

[Keith Hui]

Hello Rene(?),

Thanks for reaching out and appreciate your offer to help test. I'll
prioritize porting Sabertooth Z77.

coreboot has an issue tracker at
https://ticket.coreboot.org/projects/coreboot. However since this is a
"new" board not previously in tree (but with support for he individual
components), I'm not sure under what category issues with this new port
should be reported. However, issues reported there are sent to the mailing
list as well, so I should see them. Having said that, email and the list is
the best way to reach me.

Also because this is a new board to coreboot, it will first appear as a
change on gerrit (https://review.coreboot.org). Usually you would have to
get the source, apply the patch, and compile. I can precompile a binary and
get it to you somehow to test, probably with a Google Drive or OneDrive
link. That can be worked out later.

Usually, I have to get a dump of the flash chip with vendor firmware to get
the SPI descriptor. A full dump is also a great backup in case the port
doesn't boot. For that you probably want one of those little CH341 SPI
programmers. You'll pop the chip out while your board is running, dump it
and put it back. Then get that dump file to me. This dump file will also
allow me to integrate the SPI descriptor and ME firmware so you can flash
my precompiled binary whole.

I'll also need to run autoport on your stock board for things that I can't
get otherwise like HDA and super IO config. It's a porting tool that's part
of coreboot sources that automates much of the necessary data dumping. It
needs Linux and Go though.

The wall of text above re chip dumps to say I haven't looked deep enough
about BIOS flashback on Asus boards. They do it with a very proprietary,
separate microcontroller that can control the SPI flash chip, and they take
the .CAP vendor firmware update file stored on a USB drive plugged into a
specific USB port, which is NOT a straight dump of the BIOS area of the SPI
chip.

Any questions feel free to ask. Again thank you for your help in testing.

Regards
Keith

On Mon, Dec 23, 2024, 17:44 R S  wrote:
> Hi Keith,
> Appreciate the undertaking. I have a Z77 Sabertooth which will be replaced
> with a different system in the near future, so I would be willing to test
> with a precompiled ROM. If the ROM would be publically available at one
> point on Github or Gitlab where issues can be reported conveniently
> similarly to what the Dasharo/3mdeb guys do that would be wonderful. Also
> for the sake of being able to recover easily (without an RTE or CH341A)
> from a no-POST situation I'd like to have the USB BIOS FlashBack feature
> implemented *first* before doing any test. Michał Żygowski has
> reverse-engineered this for the MSI Z690 boards (USB FlashBIOS) and it's
> working great. I'm not an expert here, but I would assume that there are
> possibly similarities in how this works across vendors?
> --
> Network Security Analyst
> Technology Department
> http://comicsanscriminal.com
> http://comicsanscriminal.com
> Buncombe County Schools
> ComicSans Awareness Campaign https://comicsanscriminal.com
>
> On Fri, Dec 20, 2024 at 1:46 PM Keith Hui  wrote:
> > Hello,
> >
> > I recently got my hands on boardviews of many Asus P8x7x series
> > mainboards, and some offshoots such as Z77-A and Sabertooth Z77, and
> > have also found a way to extract VBT data and the USB port config out
> > of the vendor firmware image (which isn't really a straight image, but
> > what looks like an update capsule), allowing me to attempt something
> > quite ambitious given I only have the P8Z77-M on hand: Bring coreboot
> > to the entire family.
> >
> > Again because I only have the vendor firmware image and boardview to
> > go by, I will only be able to put out an untested new board patch,
> > hence I am calling for anyone with these boards, that aren't already
> > in the tree, to help me test them, and to get things I can't get out
> > of firmware image, such as HDA verbs and GPIO setup, by running
> > autoport on your board.
> >
> > Additionally, if you have a P8Z77-V (Hello Bill!), please test this patch:
> > https://review.coreboot.org/c/coreboot/+/85413
> > And if you have a P8Z77-M PRO, please test your serial port, and if
> > changing
> > it to use serial port A like P8Z77-M does, make the serial port work.
> >
> > Now, during my survey of the board family, I found that:
> > They link their power LEDs to only one of two GPIOs on the PCH: 8
> >
> > or 27. That's easily determined from the boardview.
> > 2. Except for P8C WS which I can't get a boardview for, this family
> > only uses one of two Nuvoton super I/Os: NCT5535D or NCT6779D. 6779 is
> > well known, but I cannot find the datasheet for 5535, and the pinout
> > for the closest thing I can find, NCT5532D, have the pins shifted by
> > approximately 3 pins, and there may be more differences.
> >
> > Does anyone know anything about NCT5535D or can hook me up with a
> > datasheet?
> >
> > Thanks
> > Keith
> > _______________________________________________
> > coreboot mailing list -- [e-mail]
> > To unsubscribe send an email to [e-mail]