So, get this, guys. When I unplug the power cable, the screen goes crazy. The only thing semi-consistent is the mouse pointer, everything else looks almost like line noise is being piped into the screen buffer. More appropriately, I suppose, it looks like the current windows are being put into a blender--I don't see a spectrum of colors, I see a mutated version of my current windows.

Now, here's the best part. Plug the power cable back in? And I'm golden. The moment the blue light indicating power comes up, I'm all good, windows and screens go back to normal. The problem is infinitely replicable and works no matter how many times I do it, no matter how long I leave the machine on battery before I bring it back.

The severity of the problem seems to depend on a variety of factors. Right now, with X freshly restarted and no compositing window manager, I just see some white lines dancing around the screen in a few places, all horizontal. I can't give specific details about what makes the problem worse or better because I need time to test them. If they'll be helpful, let me know.

My kernel is fed the 'noapic' parameter, because the HP dv6000 has a flawed APIC and without this the OS will not boot. I'm using Ubuntu 7.04 Beta, with kernel build 2.6.20-15-generic, in 64 bit. I am using the nvidia-glx driver package from the repositories. Here's some data . . .

My CPU Model Name: AMD Turion(tm) 64 X2 Mobile Technology TL-50

I sincerely apologize for any of this information not being relevant, but I honestly don't understand this problem in any capacity, so I don't know what 'relevance' means at this point. Any ideas, anyone? I'm listening for any suggestions . . .

~Dakota

I suppose from this that you have a laptop with a battery? Most desktops with which I am familiar don't have the display go crazy when you unplug the power cord; they go dead.

Most likely you have a battery problem; voltage is low. If not that, then you have some voltage regulation problem such that the voltage delivered to the system is below the regulation point.

I'm certain that the voltage levels are perfectly fine, because this laptop is a few days old and the battery was at a full charge when I tested this. Incidentally, I'm on it with only battery power right now, because I resumed from a hibernate without the power cable in and it's working fine. Just now, I plugged in the power cable, and nothing happened, then unplugged it, and the screen garbling began again.

It really seems like I'm the first person to ever have this problem, which is very odd . . .

You may be the first, but not the only one to have this problem. I have a
dv6000 and could not get hibernation to work at all until I loaded kubuntu 7.04
two days ago. I have tried various sequences of power to battery and battery
to power transitions under various sequences of complete shutdown vs
hibernation. My summary, as well as I can put it, is that after a complete
shutdown I can boot on battery or adapter and swap between them with no problem
until I hibernate. Coming out of hibernation on battery or adapter makes no
difference: the screen always looks fine, but the problem is set up and waiting
to happen when the adapter is unplugged. Only the transition from adapter to
battery while the machine is running is the last straw, and on the adapter the
screen is always clear. Once a hibernation has been done, the only way to get
back to what it should do is a complete shutdown. I can hibernate, unplug the
adapter, and awaken from hibernate just fine. I can then plug in the adapter
just fine, but unplugging while on shows the problem is still there. I have
even tried starting hibernation while on battery and an unreadable screen by
CTRL-ALT-DEL then ALT-H. It hibernates, it comes out of hibernation fine, but
the problem still happens on the unplug. I also set the power configuration to
hibernate when the screen closes. It's the same hibernation either way. As
you have said, is quite deterministic. I don't know what causes it, but
obviously it was not anticipated by those who designed the hardware and wrote
the drivers. All I can say is that if I want to go from the adapter to the
battery I have to pay a time penalty while the machine goes through a full
hibernate-awake cycle. But when I do transition to battery it is usually going
to be when I want to move the notebook to a different location, and I would
have to admit I should not be doing that with the machine running anyway.
Still, it would be nice to have things working as they should.