Thinkpad T22 - X.org/GDM freeze GPU when docked (Flight 6 +)

Bug #39333 reported by floid on 2006-04-12
4
Affects Status Importance Assigned to Milestone
Ubuntu
Medium
Unassigned

Bug Description

It'd be tempting to file this under xserver-xorg-driver-savage, but I get the impression this is more likely to be a kernel issue. Software concerned is Flight 6 with all updates available 11-Apr-06, including kernel 2.6.15-20, root partition mounted by label to work around the IDE enumeration headaches.

Undocked, the T22 boots fine, framebuffer works, and gdm generally loads, initializes the display, and life is good.

Attached to a Dock II, IBM type 2631, with or without hardware in the dock's PCI slot, the framebuffer works, but the display wedges with a frozen (text console) cursor when gdm runs.

Sometimes gdmplay can be heard working, consoles can be switched blind, and a clean shutdown -r can be performed. In this case, the GPU is left in such a state that the BIOS can't reinit it (leaving nothing but the unblinking cursor image), complains with a dah-dit-dit sequence of error beeps, and otherwise continues to boot with the frozen display.

Other times, the machine appears to wedge hard.

This occurs equally with Flight 6 straight and the updated kernel. 5.10 had no problems.

---

To make matters worse, I've used JFS for the root slice, and various issues therewith are blocking instrumentation of the issue. What diagnostic info I've got will be forthcoming as soon as I find my power brick to run the machine without the dock.

Matthew Lange (matthewlange) wrote :

Thank you for your report.
Does this still happen with Dapper RC?

> Does this still happen with Dapper RC?

Just installed from CD this evening, and *undocked* I'm experiencing a
50% chance of booting into GDM and a 50% chance of a hang at a frozen
(textmode cursor) screen.

Have yet to try it docked or investigate more heavily. Did just
confirm that trying to use suspend-on-lid-close successfully suspends
the machine, but still provides no way to wake it up.

Pardon my laziness (or distraction), but what would those chasing this
issue most like to see? lspci output, ACPI dump(s), anything else?

Michael Vogt (mvo) wrote :

I recently tested dapper final release on a t22 and had not problems there (with or without dock). So is this still a issue for you in dapper/final?

--- Michael Vogt <email address hidden> wrote:

> I recently tested dapper final release on a t22 and had not problems
> there (with or without dock). So is this still a issue for you in
> dapper/final?

I'm pretty sure I gave up and went to the VESA driver for the time
being; it turns out this is less dock-specific than some sort of severe
"pickiness" on the Savage driver's part. (Also, it seems resetting the
ECSD can turn a never-working configuration into a sometimes-working
one, and the only way to do that requires pulling the lithium backup
battery.)

Docking is, of course, one way to add hardware that'll rearrange the
configuration space and possibly show up the problem. I've also been
using various Cardbus 802.11 cards and a combination Firewire/USB 2
card in the dock's PCI slot, all of which work but push on whatever
resource allocation goes on during boot.

This should probably be closed out in favor of
https://launchpad.net/bugs/33617 or similar.

To post a comment you must log in.
This report contains Public information  Edit
Everyone can see this information.

Other bug subscribers