Matt DeVillier has posted comments on this change. ( https://review.coreboot.org/c/coreboot/+/39863 )
Change subject: mb/google/glados: disable serial console by default ......................................................................
Patch Set 2:
(1 comment)
https://review.coreboot.org/c/coreboot/+/39863/1/src/mainboard/google/glados... File src/mainboard/google/glados/Kconfig:
https://review.coreboot.org/c/coreboot/+/39863/1/src/mainboard/google/glados... PS1, Line 98: config CONSOLE_SERIAL : bool : default n
Tianocore experiences significant boot delays due to trying to send console output to a non-existent port
This is what is confusing me. I'm well aware of the fact that booting will be awkwardly slow if you tell Tianocore to output debug messages. But I don't see the mentioned causality. It's not slower because the port is not connected, is it?
I have a google/chell boards both with and without a servo header. On the board with the header, there's minimal impact to boot time (ie, 1-2s) if I leave CONSOLE_SERIAL enabled, with or without the servo connected. On the one without, if I leave it enabled then the boot time is measured in minutes.
If not, it seems like a made up argument to hide the actual intentions. Pretending that there is something special about this board while it's actually just that somebody wants to change a default for convenience and wants to avoid further discussion by not changing the default globally. For instance why not just let CONSOLE_SERIAL default to n for Tianocore?
Sorry if I'm off the track here, I don't want to imply any bad intentions.
it's not a convenience issue, on one release I accidentally left it enabled and many users thought their device was bricked due to the ~5 min boot time. defaulting CONSOLE_SERIAL to n for Tianocore would also seem like a reasonable solution, but didn't want to make a large change like that when the specific issue of ~5 min boot times appeared limited to glados boards.