(WIP) update coreboot #568
Conversation
|
coreboot patches 0001/0002/0003 need all instances of 'IS_ENABLED(CONFIG_XXX)' replaced with 'CONFIG(XXX)'. I count 6 lines / 8 instances that need to be changed across the 3 patches |
done that and updated. thanks |
merge
force-pushed
the
coreboot_next
branch
5 times, most recently
from
298ff66
to
b4d5bad
Compare
|
I've been working on this from my end as well, and still seeing issues with coreboot 4.10 + Xen/Qubes. On an x230, Qubes will fail to start sys-net and sys-firmware at boot, same symptoms as issue #536. On the Librems, Xen freezes on the kexec. (there's also a patch needed for all SKL/KBL devices or the system is completely broken) I'm wondering if we should just jump ahead to 4.11 at this point given it was just tagged |
sure we should; and stick with that as long as it's convenient here. |
merge
force-pushed
the
coreboot_next
branch
2 times, most recently
from
248d77c
to
bf76111
Compare
|
@MrChromebox any info on what happens at kexec? |
negative, the screen just locks after printing the kernel command line info and kexec'ing |
|
@sylv-io what approach are you taking for measured boot support? Using vboot w/4.11? backporting the patch that separates out measured boot from vboot? adapting the existing Heads measured boot patch? |
Using measured boot implementation from coreboot 4.11. So we can create the missing target boards. Are there any missing patch in 4.11 for heads? |
|
@sylv-io so using vboot with A/B booting. We can't use that across the board as some (many) devices lack the necessary room in flash to store multiple copies of the kernel. Also, Qubes doesn't boot with coreboot 4.11 (4.10+ actually) + Heads currently, at least on the devices I've tested. |
|
@MrChromebox : have you played with https://review.coreboot.org/c/coreboot/+/35077 ? |
Yes, this is a big problem. Therefore I'm using vboot only with a ro partition and reduce the ME region to the minimal size. In this way it does fit (tested with a x220 with 8MiB).
too bad.. do you know the reason why ? |
|
@tlaurion I have, that patch has gone thru a lot of changes, and can't be applied to coreboot 4.11 as it depends on a lot of commits since the tagged release. I tried an older patchset at the suggestion of the author but it too would require a lot of massaging to apply against 4.11 @sylv-io I'm working on the Qubes issue |
|
@PatrickRudolph #568 (comment) ? |
The maximum CBFS size with VBOOT+measured boot is 0x7a0000 |
|
At the moment vboot is the only useful solution since 35077 never gets merged and the previous patch of trammel is incomplete. The only solution IMHO is to strip down HEADS as I said before. Wasting more and more space in the SPI flash makes it unusable anyway, even without vboot. I would recommend starting to identify the space consumption of all HEADS components Update: Getting rid of libgcrypt and replacement for gpg would be a good way to safe 2MB in total. |
Will look, thanks. |
DO NOT MERGE! This is for testing only! Update coreboot and configs from 4.8.1 to 4.11 and port the needed patches. For an overview of the changes in coreboot, see https://doc.coreboot.org/releases/coreboot-4.9-relnotes.html https://doc.coreboot.org/releases/coreboot-4.10-relnotes.html https://doc.coreboot.org/releases/coreboot-4.11-relnotes.html tested on an X230 running Debian. TODO: * coreboot defconfig update for all boards! (configure FSP_USE_REPO?) Needed for skylake/kabylake * test
Update librem coreboot defconfigs after rebasing on coreboot 4.10-pre Test: build/boot Librem 13v2 and verify Heads functionality working Signed-off-by: Matt DeVillier <[email protected]>
|
Closing per agreement |
(WIP) update coreboot to 4.11