[coreboot] Kontron 986LCD-M support

Petr Cvek petrcvekcz at gmail.com
Fri Dec 7 05:50:49 CET 2018

sorry for delay it seems I didn't receive your answer. 

My card is SAPPHIRE NITRO Radeon RX 460 4G D5 OC. I've read the sapphire cards usually lacks dual vbios, so using the coreboot is probably the only way.


> > I was trying to upgrade my Kontron 986LCD-M setup to AMD Radeon RX460
> > and I've found the vendor BIOS lack support for the video BIOS (RX460
> > requires UEFI). This incompatibility leads to x16 PCIe slot disable and
> > no graphics in linux at all.
> > I was able to re-enable the PCIe slot until now when I tried to upgrade
> > to 2+2 GiB DDR2 modules (little more MB and dual channel). The RAM space
> > is shared with PCI space and limited by 32bit so the top part of RAM is
> > eaten by the regions of PCI devices. Problem is when vendor BIOS
> > disables the card's slot it doesn't leave a hole for my GPU and it sets
> > the TOLUD register (top of used RAM). This causes kernel to crash in GPU
> > initialization.
> > There is probably no way to fix TOLUD (and ACPI tables, SMI and other
> > stuff) or force vendor's BIOS to reserve region for GPU, so I'm
> > considering to switch to coreboot.
> > My question is: does anybody know advanced information about TPM and SPI
> > connectors? In the Kontron 986LCD-M datasheet (sections 4.16 and 4.17)
> > there are only pin names and caption "unsupported". I understand these
> > signals are likely buses from chipset, but it would be nice to know more
> > details. Namely if the default LPC flash chip needs to be disconnected
> > when using these busses and which pins of the chipset are connected to
> > SPI pins BOOT0 and BOOT1 (I suppose they are LPC/SPI/PCI priority?).

> I run a RX580 without UEFI - note that only some vendors are silly and
> fail to implement a dual VBIOS what is your cards OEM/model?

More information about the coreboot mailing list