New Beta Bootscript Package 20040601

Nathan Coulson conathan at conet.dyndns.org
Wed Jun 2 06:25:47 PDT 2004


> Alexander E. Patrakov wrote:
>
>>> The udev script will either pass udev or hotplug to
>>> /proc/sys/kernel/hotplug (both are currently untested as of yet, but
>>> 99%
>>> sure they work).
>>
>> The solution is wrong, because there is still a small time window for
>> unexpected hotplug events.
>
> Just trying this now. Got the patch in, and a modified udev script.
> Also pcimodules and which.
>
> A few complaints from the keyboard/mouse stuff that there is
> "No hotplug agent" defined; but everything works.
>
> Looking forward to the next installment.
>
> Chris

well, so much for that idea );  back to the ol drawing board.

So hotplug calls scripts which do everything but create the device, and
then call udev which creates the device...

I wonder if we can disable the udev part until after we mount a ramfs on
/dev...  (Another crazy unresearched idea from yours truly... that's 2 in
one week).



More information about the lfs-dev mailing list