Re: [PATCH] Remove broken by design and by implementation devtmpfs maintenance disaster

[ebiederm@xmission.com (Eric W. Biederman)]

Arjan van de Ven <arjan@infradead.org> writes:

> On Fri, 18 Sep 2009 06:54:39 -0700
> ebiederm@xmission.com (Eric W. Biederman) wrote:
>
>> 
>> > I don't understand. Udev applies the final policy including
>> > permissions/ownership, just as before. There is no differrence. It's
>> > just that you can bring up a box without complex userspace to
>> > bootstrap /dev. And that's a big win on its own.
>> 
>> udev is too complex to use?  That sounds like a userspace bug.
>> 
>> This I guess is where I am baffled.   The argument for devtmpfs
>> always seem to boil down to: udev sucks let's write some kernel
>> code instead.
>> 
>> I have been trying to ask for a long time why we can't just fix
>> udev to not suck. 
>> 
>> >  And things like
>> > "modprobe loop; losetup /dev/loop0" will just work, which it doesn't
>> > with todays async udev. Again, please make yourself familiar how
>> > things work, and what the problems are.
>> 
>> I guess I don't understand why 
>> modprobe loop; losetup /dev/loop0 is an interesting case.
>> When you can just as easily do:
>> modprobe loop; udevadm settle; losetup /dev/loop0.
>
> frankly, modprobe should call the settle.
> And not just this one, but we can use this to settle other things as
> well... and then it can get an --async command line option for the
> cases where you know you don't want to synchronize.

I think this would be a bit of a pain when I modprobe a network
driver and the udev scripts trigger a blocking dhcp on the device.

If this is a major pain point in initscripts I can see how it would
make sense.

Eric