short white screen after kernel boots

Mike (mwester) mwester at dls.net
Fri Oct 31 23:51:24 CET 2008


Michael 'Mickey' Lauer wrote:
> Am Friday 31 October 2008 15:28:41 schrieb Andy Green:
>> But the problems with white screen on boot are going to be solved a
>> better way: Qi does not turn on the backlight and when the
>> balaji-tracking stuff is used the backlight device is a child of the
>> framebuffer, meaning the backlight will not be initialized until the
>> framebuffer has been initialized.
> 
> Can we find out whether we have been jumpstarted by Qi or U-Boot? We might 
> want to "polish" the turn-on-experience for the users still using U-Boot.

Nothing specifically identifying the bootloader or its config is passed
into the kernel.  However, something like this could be done in the
command line passed to the kernel.  Qi might add "bootloader=Qi",
signaling it's presence to user-space (and kernel).  Another possibility
is to attempt to divine this information by checking the entire ATAG
list passed in to the kernel; something as simple as the ordering of the
tags might be sufficient to use as a "signature" to differentiate one
from the other.

(But I would prefer the command-line parameter, I think.)

Mike (mwester)



More information about the openmoko-kernel mailing list