Udev in b6_0: to be or not to be

Nathan Coulson conathan at conet.dyndns.org
Tue Jun 1 21:39:13 PDT 2004


> Ian Molton wrote:
>
>> although, lets be realisticv here - theres no reason single-user boot
>> and a quick fixup cant be done, and the patch (if deemed the 'right
>> way') will be folded into the kernel in time.
>
> If by "quick fixup" you mean rebuild the kernel with the patch included
> then I would agree. If by that you mean get rid of the file that's
> causing the bootscript to skip running udevstart, that's harder, because
> it will just come back, and somewhat randomly.
>
> To be honest, when I posted the bug I was really expecting the response
> to be "let's just remove the short-circuit check in the bootscript and
> always run udevstart". I would be OK with that, I was the one who
> originally asked for the short-circuit check and my own personal need
> for it has gone anyway :-) However, it is technically better to not
> re-run udevstart if it's not needed, and the underlying problem with the
> kernel expecting userspace to be ready for hotplug events the instant
> the root filesystem is mounted is a bigger issue.
>
> For now, I'd rather see the bootscript get modified than the
> incorporation of the kernel patch; for the vast (99.9% or more) majority
> of systems that will take care of the problem and noone will have any
> clue that their system handled some too-early hotplug events.

I can yank the check for now, but still if you have a readwrite root
filesystem at bootup, you'll get a few garbage devices underneath the
ramfs you mounted on /dev.  I dont think that is acceptable in the
longterm.

[Just throwing this out before I make this change]



More information about the lfs-dev mailing list