Subject: Re: gcore! Where'd it go? ...and docs for procfs?
To: None <>
From: der Mouse <mouse@Rodents.Montreal.QC.CA>
List: current-users
Date: 02/14/2000 01:23:04
> I haven't had call to use gcore for over twelve years now.
> Imagine my surprise when I found that NetBSD does not have such a
> program in its core tree!

I noticed that a while ago, but I want gcore so seldom it didn't really
bother me.

> I have a locked tape drive now [...]  I'm trying to convince the
> process that it can go away now.  Since it's in 'DL' state where's gcore come into it?  It doesn't sound as though getting a
userland core would be of any use.

> I'm trying to force something in this process to trigger so that it
> will blow up and go away.

This is an issue you'll have to negotiate with the relevant driver for.
Somewhere in the kernel (probably in a driver) that process is sleeping
without PCATCH.  There's an option to ps to print the wchan, I think,
which *ought* to allow you to identify the sleep call....

Whether you can do anything about it short of dropping into ddb (or a
ROM debugger-alike) and poking at kernel memory, that's another
question.  If the process isn't paying any attention to signals you're
not likely to be able to do very much to it.

ISTR you said that the device it was sleeping on was an Exabyte tape
drive.  You might be able to do something by power-cycling the drive.
In general I hesitate to advise doing something like that "hot" (you
could suffer unfortunate effects, from panicking the kernel by
disrupting a transfer in progress to silently corrupting a transfer in
progress to some other device to puffing a terminator fuse to frying
SCSI bus drivers, roughly in my estimate of most to least probable),
but nothing else comes to mind as having any real hope of waking it up.

					der Mouse

		     7D C8 61 52 5D E7 2D 39  4E F1 31 3E E8 B3 27 4B