Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

t440p/w541 marked UNTESTED (boot in 50s) : CONFIG_CBFS_SIZE of 8mb (changed to CONFIG_CBFS_SIZE=0xBE4FFF by need) #1825

Open
2 tasks
tlaurion opened this issue Oct 30, 2024 · 6 comments · Fixed by #1826

Comments

@tlaurion
Copy link
Collaborator

tlaurion commented Oct 30, 2024

Why master was not enforcing CONFIG_CBFS_SIZE=0xBE4FFF as all other 12mb opaque flash chips from ivy bridge? (was 8mb). Merged as emergency fix: untested. Unfortunately reason was that anything CBFS bigger then 8mb will make Haswell boot in 50s, thanks to improper MRC blob we depend on.

Solution would be to switch to NRI #1711, which is not merged upstream.
Board owners: time to participate testing/reviewing/commenting upstream your need/appreciation of memory initialization being free under #1711.


OLD
#1826 was merged moving those boars to untested to prevent users from potential bricking without external programmers.

Testing needed from boards testers (BOARDS_TESTERS.md):

TESTED: Master roms will boot, but after 50s see #1825 (comment)

Therefore, t440p/w541 boards now build UNTESTED tagged board roms. Be warned.

@tlaurion tlaurion reopened this Oct 30, 2024
@tlaurion tlaurion changed the title t440p/w541 : CONFIG_CBFS_SIZE is 8mb and prevents stitching roms on master t440p/w541 : CONFIG_CBFS_SIZE was 8mb and prevented stitching roms on master. moved to untested. TESTING NEEDED! Oct 30, 2024
@tlaurion tlaurion pinned this issue Oct 30, 2024
@tlaurion
Copy link
Collaborator Author

Pinned issue

@tlaurion tlaurion changed the title t440p/w541 : CONFIG_CBFS_SIZE was 8mb and prevented stitching roms on master. moved to untested. TESTING NEEDED! UNTESTED: t440p/w541 : CONFIG_CBFS_SIZE was 8mb and prevented stitching roms on master. moved to untested. TESTING NEEDED! Oct 30, 2024
@ResendeGHF
Copy link

Why master was not enforcing CONFIG_CBFS_SIZE=0xBE4FFF as all other 12mb opaque flash chips from ivy bridge? (was 8mb).

CONFIG_CBFS_SIZE=0xBE4FFF leads to longer RAM initialization times. Relevant information about this issue can be found in #1282. As icon noted, this is likely due to the following reasons:

I assume with the bigger size, not everything in CBFS can be cached. There’s CAR (Cache as RAM) at 0xff7c0000 to 0xff7effff, which is used to store data before the DRAM is initialized. The BIOS region is mapped directly below 4G (at 0x100000000). Therefore, making the CBFS too large could conflict with CAR. I don't believe this can be changed without also modifying the MRC blob. The best option is to set CBFS_SIZE to 8MiB. One can still make the rest of the BIOS region available using a custom FMAP, though it would not be part of CBFS.

Additionally, I've just tested on the W541 board from UNTESTED_w541-maximized (The path under UNTESTED_w541-hotp-maximized.config is inconsistent due to renaming to UNTESTED). As expected, boot time is over 50 seconds.

@tlaurion
Copy link
Collaborator Author

@ResendeGHF Maybe its time to switch to native ram init #1711 ? (With its documented downsides)
Point is: past value for CBFS (8mb) is not big enough anymore as can be seen from CircleCI failed builds :(

@tlaurion tlaurion changed the title UNTESTED: t440p/w541 : CONFIG_CBFS_SIZE was 8mb and prevented stitching roms on master. moved to untested. TESTING NEEDED! t440p/w541 : CONFIG_CBFS_SIZE of 8mb (changed to CONFIG_CBFS_SIZE=0xBE4FFF by need) now boots in 50 seconds Nov 5, 2024
@tlaurion
Copy link
Collaborator Author

tlaurion commented Nov 5, 2024

@ResendeGHF op updated.

@tlaurion tlaurion changed the title t440p/w541 : CONFIG_CBFS_SIZE of 8mb (changed to CONFIG_CBFS_SIZE=0xBE4FFF by need) now boots in 50 seconds t440p/w541 marked UNTESTED (boot in 50s) : CONFIG_CBFS_SIZE of 8mb (changed to CONFIG_CBFS_SIZE=0xBE4FFF by need) Nov 5, 2024
@tlaurion
Copy link
Collaborator Author

tlaurion commented Nov 6, 2024

Also note that participating in upstream issue would make this go forward https://ticket.coreboot.org/issues/457

@tlaurion
Copy link
Collaborator Author

tlaurion commented Nov 7, 2024

Proposed collaboration in s3 resume from suspend patchset at https://review.coreboot.org/c/coreboot/+/81897/comments/d4d3ed3a_d6c46137

If you see replies faster than I do, please ping me here.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Projects
None yet
2 participants