coreboot VS thinkpad-firmware-patches

Why BIOS (did you mean UEFI?) when it runs the best boot loader, which is Coreboot¹. Many users would love to re-flash their bios/uefi for it, if it’s supported.

1: https://www.coreboot.org/

Coreboot is 0.6% C++.

NovaCustom ; some models come with Dasharo a coreboot distribution.

There may be other protections to restrict SPI flash access for security reasons (so you might not be able to flash your custom firmware in the OS), but worst case you can use a HW flasher (or maybe USB flashback). Still, this doesn't address the elephant in the room - platform initialization code might be open-sourced, but that isn't everything. You'll still need to figure out the board-specific stuff (the Super I/O chip, chipset GPIOs, other peripherals, etc.). Using coreboot as an example, Intel provides the Firmware Support Package blob to handle platform initialization. I think AGESA is somewhat similar to this, though Intel publicly releases the binaries for use in coreboot/etc. Thanks to the FSP, coreboot has support for recent Intel chipsets. However, there is only support for two recent consumer boards: the MSI PRO Z690-A WiFi DDR4 and DDR5.

coreboot/devicetree.cb at bcc9879151c0ff61dfa3f2206b538d5c8be619ba · coreboot/coreboot (github.com)

You usually can't take existing firmware and run it on a new CPU, because the new CPU requires different bringup code anyway. Take a look at https://github.com/coreboot/coreboot/tree/master/src/soc/int... to get some idea of how many different implementations there are for modern Intel alone (there's a bunch more for the pre-SoC style Intels). If you already have to port your firmware to a new CPU, you can deal with the CPU starting in a different mode - it is entirely an implementation detail that can be handled in the firmware.

NEM/CAR is used to avoid hacks like romcc. you need memory for a stack/etc if you want to run C code; romcc worked around this by compiling C code to asm that only used registers (no stack usage at all). this isn't needed if you can use NEM/CAR. either way I wouldn't say this massively reduces latency, since code is still being backed by SPI flash MMIO

Novacustom. Some models come with Dasharo an open source firmware based on coreboot.

It seems that Intel is currently 404ing the I remember in the 82579LM datasheet regarding firmware made a vague reference to the Intel ME being used the PXE boot. If the ME parition has neutered or the generated GBE partition is being used, this shouldn't be a concern.

Coreboot/Libreboot (recently merged with OSboot, so presumably that means the policy problems have been fixed) as far as I know do not benefit from any such standardization and often have to resort to reverse-engineering.

"skyra1n" ThinkPad Firmware Patches from digmorepaka's GitHub a list of UEFI patches for ThinkPads, these are the special sauce for adding features to our BIOS. If you have a ThinkPad 30 series device, you do not need to follow this guide for flashing skyra1n, you can flash your BIOS with the internal flashing software.

https://github.com/digmorepaka/thinkpad-firmware-patches *

I'm considering to buy an used T480 which is getting pretty inexpensive these days. I'm planning to buy one and flash its BIOS with the image where Intel ME is neutralized with me_cleaner and wlan whitelist is removed with this patch.

Note that I had no such issues with it on my ThinkPad X131e, though I didn't have the TPM enabled at the time. Later, when I did have it enabled, the laptop would beep quite a bit upon boot (but it would still reach the OS) - though this may have been due to the patched BIOS I had installed, which is known to break the TPM.