Bug#2351: xdm doesn't detect gpm mouse lock
On Sun, 18 Feb 1996, Kevin M Bealer wrote:
> When I run gpm (from init.d in my case) then run xdm, xdm either:
> a) general protection faults listed as 0000
This is a kernel bug.
> b) repeatedly runs and lets die the X server (at least it looks like
> this is what is happening. I needed to reboot to stop it,
> then it gives a gp fault.
This is a bit of bad design in xdm. I don't know how easy it would be to
fix it; I'll have a look at the source when I have some time.
> I recommend that xdm detect fatal conditions such as unavailable resources,
> or that the X servers be given a way to communicate to xdm that they will
> not recover from whatever error by reloading. Perhaps there are some
> considerations of a user creating such a condition in order to terminate xdm
> and thus breach security, esp. if xdm is started manually by the root user.
I think the ideal situation would be that an X server that couldn't start
properly would exit with a particular exit code, which xdm would notice
and wouldn't attempt to start the server again, possibly putting a message
in the syslog or its own log file.
I'm sure that xdm does this already in some cases, but I haven't had time
to look at the source so I'll shut up for now and post more when I know
what I'm talking about.
> When xdm is cycling X, hitting the cntrl-alt-del allows me to view a general
> protection fault before reboot:
> general protection: 0000
> EIP: 0010:0014a115
> EFLAGS: 00010206
> eax: 6e6f632d ebx: 00215910 ecx: 007c1dd0 edx: 002c5000
> esi: 00330f54 edi: 00330f54 ebp: 007c1e8c esp: 007c1e00
> ds: 0018 es: 0018 fs: 002b gs: 002b ss: 0018
> Process xdm (pid: 621, process nr 6, stackpage=007c10000
> or is it 007c|0000?
Definitely a kernel problem of some sort.
> I hit scroll-lock during reboot and the system just sat there until I was
> done... should it listen for scroll lock while it's rebooting?
This is a question for somebody else. If you're worried about it, I
suggest you post a separate message. Presumably the shutdown script is
blocking on output to the console.