Red Hat Bugzilla – Bug 20255
kickstart: /dev/mouse is created with 'mouse none'
Last modified: 2005-10-31 17:00:50 EST
If you specify 'mouse none' in ks.cfg, /dev/mouse link is still created, to ttyS0.
I don't think this should be done.
I haven't had a chance to see if this is still an issue in RHL7.
Passed to QA to reproduce.
If you happen to have a serial mouse plugged in, it's
probably getting created in post-install by kudzu.
There was no mouse plugged in at all.
hmmm ... testing Red Hat Linux 7 here w/joe random machine (with mouse NOT
plugged in) ...
/dev/mouse -> /dev/psaux
is this a problem ...? what circumstances may occur that would make this
Well, I'm not sure if there's any harm in that, but ttyS0 was bad because
we used that for serial console...
What if the system doesn't have psaux port (old AT case) or it's disabled?
I wonder if the link would be created, and pointing where.
I would agree that ttyS0 would be a bad link to make ... :)
what is your hardware configuration on the machine that made the link to ttyS0
The system was P3 w/ i815-based board in ATX case.
Assigning to a developer.
This could be made to go away (at least partly) if /dev/mouse goes away, see bug
23499 (my patch to exchange /dev/mouse in favor of an entry in the X config and
Still present in internal builds.
I looked at this for quite some time today and the answer is not clear. I'm
reluctant to change code at this point, especially for something that is not
critical. This must be fixed in the future, though. Deferring to a future release.
Brent please verify against roswell.
Problem still present in Roswell.
Bill, would kudzu potentially detect the mouse when we run it in the postinstall
even if the installer has been set up with no mouse?
Almost certainly, if there is one there. I think the initial bug report is about
a link for a serial mouse being created when there wasn't one present, which is
a somewhat different problem...
Yeah, that was the original behavior and the bug has evolved over time... that
at least answers the later bit of /dev/mouse pointing to /dev/psaux and thus
it's fairly safe to say that this is fixed in the current release (at least, it
won't cause conflicts with serial consoles)