Re: Bug#403136: udev bug ... (d-i fails to create /dev/sd* devices)

[Sven Luther <sven.luther@wanadoo.fr>]

On Sun, Dec 17, 2006 at 02:48:24PM +0100, Marco d'Itri wrote:
> On Dec 16, Sven Luther <sven.luther@wanadoo.fr> wrote:
> 
> > Well, i never dealt with udev, i have not even the faintest idea where to
> > start, and the initramfs-tools or d-i environment i have to play with is
> > really not all that user friendly, so please help me or give me guidance on
> > how to further investigate ? 
> OTOH looks like you were really quick to blame udev for this.
> 
> > Can you tell me how i can make udev log all received events, or something
> > such. Once i get the hand, it is usually already too late, and everything
> > happened.
> Uncomment the last line of /etc/udev/hotplug.rules .

Well, in the d-i context, or probably even the initramfs-tools context, this
is probably not going to do us much good.

I did build a modified udev-udeb, in which i added a hotplug.rules which
includes only the :

  # Log every event to /dev/hotplug.log (for debugging).
  RUN+="logger.agent"

lines, and i added /lib/udev/logger.agent to it too.

I am unsure if this will be enough though, because lot of stuff are missing
from the udev-udeb, and possibly also whatever is used to run the
hotplug.rules.

Would a udev-dbg-udev make sense ? It would make debugging this kind of stuff
much easier.

Now, i wonder if it is udev related, since the udev-udeb is much more
lightweight than the full udev, is d-i using something else to create the
devices ? 

Anyway, thanks for your help.

Friendly,

Sven Luther