Re: Gkrellm locking

On Wed, 28 Nov 2007 18:39:10 -0300
"Salatiel Filho" <[email protected]> wrote:

> > Well, thanks for the effort, but I'm sorry to say the stack trace doesn't
> > help.
> > I hoped it would show a backtrace into gkrellmms functions, but it
> > doesn't.
> > Now I don't know what is going on.
> 
> anything else i can do help to find a fix , or do i just have to get use to
> this ? :/

The thing I do know is that there can be an interaction with threads
and libxmms calling setuid().  So even though the stack trace didn't
isolate your problem specifically to that, I do wonder if it really is the
same problem.

What I've done in cases where it affects me is to intercept the setuid()
calls so they don't do anything and that allows gkrellmms to run on my
boxes without a problem.   There appears to be no side effects so I don't
know why libxmms needs to do the setuid().

You can try the same experiment if you want to compile a gkrellm.  In the
just released 2.3.1, you can edit the gkrellm-2.3.1/src/deprecated.c file
and towards the bottom of the file change the line
	#if 0
to
	#if 1
to enable compiling in a couple of setuid() functions.

If you can then run gkrellmms without gkrellm hanging, then it's the same
problem but I don't know the right fix since it seems to be a low level
interaction of a Linux thread exiting during a setuid() call.
Until I know more you could just run your compiled version.

But if you still have the problem, it's something new and it would take more
digging to track down.  If this is the case, how long can you typically run
before it hangs?

Bill

______________________________________________
Gkrellm mailing list
[email protected]
Read archives-join-leave-set RSS feed:
http://archive.lists.netservicesgroup.com