Skip site navigation (1)Skip section navigation (2)
Date:      Thu, 22 Jul 1999 12:55:50 -0600
From:      Nate Williams <nate@mt.sri.com>
To:        Terry Lambert <tlambert@primenet.com>
Cc:        nate@mt.sri.com (Nate Williams), scheper@beast.toad.net, freebsd-smp@FreeBSD.ORG
Subject:   Re: SMP + XDM = keyboard lockup
Message-ID:  <199907221855.MAA03429@mt.sri.com>
In-Reply-To: <199907221850.LAA10852@usr05.primenet.com>
References:  <199907221845.MAA03196@mt.sri.com> <199907221850.LAA10852@usr05.primenet.com>

next in thread | previous in thread | raw e-mail | index | archive | help
> > > I believe the issue is the system is still starting up, and thus
> > > has the kernel poking at things.
> > 
> > It shouldn't be.  The kernel is completely done 'starting up', else all
> > the /etc/rc.* files wouldn't have been run and init wouldn't have
> > started spawning off gettys.
> 
> I'm referring to things that result in a lot of I/O activity
> which are started in the rc files (NFS, DNS, sendmail, etc.).

Sure, but none of those are going to compete with X for resources that
'conflict' or would cause lockups.

> > > This jibes with his statement that manually starting the xdm from
> > > a root shell prompt on a relatively quiescent system "works".  It
> > > is likely that it "works" only because there is not simultaneous
> > > activity.
> > 
> > Or there is a a race in the video/keyboard driver that is causing
> > problems.  X calls into the driver, so it's possible a missing spl() or
> > somesuch is causing problem.
> > 
> > I doubt very much the problem is related to inb/outb, and is instead a
> > just a 'bug'. :)
> 
> Maybe.  I think we should hold off investing too much of ourselves
> into our respective theories until after he has done some of the
> tests I suggested.

Your tests won't prove anything one way or the other.

> You should feel free to suggest tests, too, of course, since I could
> be wrong, no matter how "right" it feels to me.

I don't think we can 'test' this one to test any particular theory.  Any
test complex enough to point to a particular problem would be almost as
complex as fixing the bug.

Instead, we should find someone with the knowledge (Kazu and/or Soren)
and hardware to recreate and fix the problem.



Nate


To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-smp" in the body of the message




Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199907221855.MAA03429>