Bug #167
closedX201 restarts rather than waking up
0%
Description
With coreboot 4.8.1 (and with the current git master) my X201 no longer wakes up from suspend-to-RAM, i.e. instead of resuming it boots from scratch. Sometimes (but not always) when this happens, the "moon" LED starts blinking.
The laptop works fine with coreboot 4.6 but fails entirely to boot with version 4.7 thus I can't easily figure out which commit broke it.
Is there anything I can do to help debugging this?
Updated by Arthur Heymans over 6 years ago
SeaBIOS 1.11 with CONFIG_SERCONN is known to break S3.
Could you paste a coreboot log of when coreboot rebooted instead of resumed from S3?
Updated by Stefan Ott over 6 years ago
Ah nice, thank you removing the serial port console option seems to have resolved this for me. Are you still interested in the logs?
Updated by Stefan Ott over 6 years ago
Btw, feel free to close this one, I can't seem to.
Updated by Paul Menzel over 6 years ago
- Status changed from New to Closed
The underlying problem needs still be fixing in my opinion. But it’s not clear yet, if it’s coreboot’s or SeaBIOS’ fault.
Updated by Alexander Couzens over 6 years ago
@Stefan Ott: Which serial power do you enabled? In coreboot or/and in seabios? we should exclude then from all target not supporting a serial at all.
Updated by Stefan Ott almost 6 years ago
@Alexander Couzens: Sorry for the delay, I completely forgot about this. The option that causes the issue (still does in 4.9) is CONFIG_CONSOLE_SERIAL. When I enable that, the resulting ROM will crash instead of waking up; when I disable it, everything runs just fine.
Let me know if I can do anything to analyze this further.
Updated by Paul Menzel almost 6 years ago
Thank you for your reply. Arthur debugged this, and it’s not a bug in SeaBIOS, but coreboot. You need to use latest coreboot master branch, or cherry-pick commit 804adaa1 (arch/x86/ebda: Don't trash the EBDA on the resume path) 1.