Date: Thu, 14 Dec 1995 15:49:25 -0800 From: deborah@microunity.com (Deborah Gronke Bennett) To: questions@freebsd.org, xfree86@xfree86.org Subject: Summary: Re: can I run XFree86 on FreeBSD 2.0.5 from a mounted filesystem? Message-ID: <9512141549.ZM11738@gallifrey.microunity.com>
next in thread | raw e-mail | index | archive | help
Earlier today, I submitted a question to both these lists like this:
>
> I have a network of 5 identical lab machines which are all
> Gateway 90MHz pentium machines with ATI Mach64 graphics cards.
> For administration reasons, I would prefer to have them mount
> /usr/X11R6 from a network location. I am trying to run XFree86 this
> way, and having no luck.
>
> I have created an XF86Config file that works
> fine on one of the machines which happens to have a full XFree86
> installation. I have copied that XF86Config file to /etc/XF86Config
> on one of the other machines. I have also made the symbolic link
> from usr/X11R6/bin/X to usr/X11R6/bin/XF86_Mach64 on the server.
> I have mounted usr/X11R6 from the server on /usr/X11R6 on the client.
> When I run startx, I get this error message:
> ld.so: xinit: libXmu.so.0: Undefined error: 0
Boyd Faulkner (faulkner@mpd.tandem.com) wins the prize for the first
and correct answer. Yes, you can XFree86 from NFS. The error
message above is caused by the fact that ldconfig is not adding
/usr/X11R6. The root cause was that I was trying to access /usr/X11R6
through a symlink, which referred to an nfs-mounted filesystem I get
to with amd (an automounter). When I switched to mounting /usr/X11R6
in /etc/fstab, ldconfig found it OK and I got startx to run.
For anyone else attempting this kind of config, here's what I did
to the mounted usr/X11R6 tree:
- make a link on the server from usr/X11R6/bin/X to usr/X11R6/bin/XF86_Mach64
(all of my machines are identical, and have ATI Mach64 graphics)
- I ran xf86config on the one machine which has XFree86 installed, then
put a copy of the XF86Config file it created into
usr/X11R6/lib/X11/XF86Config on the server.
(Again, I can do this since all my machines are identical).
Unfortunately, I am now having kernel reliability problems. Sometimes,
with some kernel configurations, this nfs-mounted X configuration
works just fine. But sometimes when I run startx,
I am apparently falling over some kind of bug in the vnode or vm system,
where I get this stack traceback:
panic: vm_object_deallocate: object deallocated too many times
Debugger("panic")
Stopped at _Debugger+0x2b: movb $0,_in_Debugger.110
db> trace
_Debugger(f0116f6b,f0116f60,f018800b,efbffe8c,f06f2080) at _Debugger+0x2b
_panic(f018800b,f06f7680,0,f06ec000,efbffec0) at _panic+0x3b
_vm_object_deallocate(f06f2080,f06ec000,f06f4900,f01b918c,f06f7680) at
_vm_object_deallocate+0x21
_vn_close(f06f7680,1,f06e5300,f06ec000,efbfff24) at _vn_close+0x70
_vn_closefile(f06f4900,f06ec000,f06d7e80,4,f06ec000) at _vn_closefile+0x19
_closef(f06f4900,f06ec000,efbfff94,f06ec000,0) at _closef+0x12c
_close(f06ec000,efbfff94,efbfff8c,817a060,8184fc4) at _close+0x83
_syscall(27,27,0,8184fc4,efbfd2f8) at _syscall+0x161
I am attempting to isolate this problem now, and I will communicate
with the FreeBSD bugs mailing list.
Thanks again to all the helpful people who took time to respond so
quickly to my question.
-deborah bennett
----------
Deborah Gronke Bennett (WD5HJH) kernel and device drivers engineer
deborah@microunity.com (408)-734-8100
MicroUnity Systems Eng., 255 Caspian Drive, Sunnyvale, CA 94089-1015 USA
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?9512141549.ZM11738>