Timing issue (race) with kdm and hal during boot?

[Michael Biebl]

Petter Reinholdtsen wrote:
> When I enabled readahead-fedora on a freshly installed unstable KDE
> desktop with dependency based boot sequencing, a strange thing
> happened.  KDM started X, but the keyboard did not work.  I tried two
> things based on the idea that kdm depend on something started later,
> and suspecting it was hal.  First I tried to make kdm depend on
> network-manager.  This got the keyboard working.  network-manager
> depend on hal, so the time separation between kdm and hal was
> increased.  Next, I dropped that dependency and made hal interactive,
> pushing it further ahead in the boot sequence.  This worked too, and
> likewise made hal and kdm start further apart.
> 
> I do not know what the cause of this problem is, but suspect hal is
> not yet operational when its init.d script is started, and X need an
> operational hal to find its keyboard settings.  Not sure how we should
> solve this.  Race conditions like this will make it impossible to
> speed up the boot.  I suspect we will find more of them when the boot
> become faster, and thus exposing latent problems currently hidden
> because of the slowness of the current boot.

kdm should at least depend on dbus and hal, I'd say. It shouldn't matter though
when hal comes up exactly, as Xorg will retry as soon as hal is available.
This means, it might happen, that you don't have a not working mouse or keyboard
for a few seconds, but they should become functional as soon as hal is there.

Btw, you might check Xorg.0.log, regarding any hal/dbus related problems.

Cheers,
Michael

-- 
Why is it that all of the instruments seeking intelligent life in the
universe are pointed away from Earth?

-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 261 bytes
Desc: OpenPGP digital signature
URL: <http://lists.alioth.debian.org/pipermail/initscripts-ng-devel/attachments/20090812/66944f0c/attachment.pgp>