Hmmm, that guide could use a section how to update coreboot from then on. But other than that it seems absolutely fine. Simon, all your problems beyond that (not knowing how to get flashrom to write the update, not being able to flash...Nico Huber
Looks like the PCIe link is completely down. I wouldn't say it's likely with an Intel card, but it's possible that this happens when coreboot enables the link power management (assuming the link worked without pm). A coreboot log would t...Nico Huber
Brian L wrote in #note-16: > > If you want to get libgfxinit working again, a log would really be the next best step. > ... Thanks. Not what I expected. Gave me some ideas, though, even if it feels a bit desparate: There's a timeout wa...Nico Huber
Brian L wrote in #note-13: > This is not correct, if you've read the above, SeaBios is the only way that the option rom does work. I am not sure what the confusion is here. I've restated multiple times the environments I've tested and w...Nico Huber
One very important thing: Do you remember the exact coreboot version / config that you used when it broke? Many of my ealier thoughts were based on the assumption that it was stock HEADS and that libgfxinit worked before with your pan...Nico Huber
Brian L wrote in #note-10: > Yup, so I'm like 99% sure now. Coreboot does not initialize my internal display whether using libgfxinit or vga blob. However, if SeaBios runs the vgablob, then it does work > ... There are plenty of things...Nico Huber
Patrick Rudolph wrote in #note-4: > UEFI firmware patches the VBT, while coreboot does not. Maybe a wrong version was checked in causing this issue? This could indeed be part of the problem: ``` $ intel_vbt_decode --file=src/mainbo...Nico Huber
Why explicitly specify `libcurl4` anyway? If it's a dependency, `apt-get` can solve it. If we want to build anything with it, we need one of the `-dev` variants instead.Nico Huber
> /home/zhongli/coreboot/util/crossgcc/xgcc/bin/x86_64-elf-ld.bfd: build/romstage/console/vtxprintf.o: in function `number': Actually, this looks like it's using the reference toolchain, but the 64-bit one. `xcompile` picks up 64-bit ...Nico Huber
Valerii Huhnin wrote in #note-14: > * What are well-formed regions? > ... Definitely yes. > * The last address of the region +1 should be representable as a `size_t` value? This won't be necessary and would only complicate th...Nico Huber