Bug#465278: Appletouch failure upon resume after suspend2ram
Severity: grave I am increasing the gravity of this bug, because it makes the apple iBook G4 laptop pretty much unusable without an external mouse, and because people seems to just ignore this PowerPC specific bug. A very easy way to systematicaly reproduce this bug is to close the lid (or make it to sleep somehow), and then touch continously the mousepad during the wakeup. If you don't touch the trackpad, it might wakeup properly (maybe 1 chance on 3 to still have a working mouse). I also tried the same with the 2.6.18 kernel, and got similar results but with a significant difference: with the 2.6.18 kernel, I can make the mouse working again by simply doing rmmod appletouch and then modprobe -a appletouch. If I use the 2.6.26 kernel instead, I get the already reported following errors and the only way I know to make the mouse working again is rebooting. [45692.025239] appletouch: Could not do mode read request from device (Geyser Raw mode) [45692.025289] appletouch: probe of 1-2:1.0 failed with error -5 [45692.025347] usbcore: registered new interface driver appletouch Please note that xserver-xorg-input-synaptics is not even installed, and that the bug is reproducible in runlevel 1 with just enough things running to be able to put the laptop in sleep mode. My guess is that there is 2 bugs involved: a new bug in the appletouch driver in the kernel = 2.6.24 (maybe before) that prevent removing and adding again the driver properly in some conditions, and another possibly old bug that, for some reasons, showup systematically in Lenny while it was seldom appearing in Etch. In Etch, you could sometime lost the mouse after a wakeup, but it happened maybe just once a month, and since you could just remove and add again the driver, it was mostly just a nuisance. This bug is really annoying, reducing enough the usability of the iBook to make me seriously considering downgrading to Etch or looking at another distribution if it fix my problem. Simon Valiquette -- To UNSUBSCRIBE, email to debian-kernel-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Bug#465278: Appletouch failure upon resume after suspend2ram
Severity 465278 grave By searching a little on mailing lists, I saw that someone reported that the problems might be related with this patch, or at least related with this part of the code, but nobody seems to have seriously looked at it: http://www.linuxhq.com/kernel/v2.6/24/drivers/input/mouse/appletouch.c http://www.mail-archive.com/debian-powe...@lists.debian.org/msg59480.html From the Debian PowerPC mailing list, I deduce that the problem appeared in Debian Sid with the kernel 2.6.24. Simon Valiquette -- To UNSUBSCRIBE, email to debian-kernel-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Bug#465278: Appletouch failure upon resume after suspend2ram
Package: linux-image-2.6-vserver-powerpc Version: 2.6.26-13 Followup-For: Bug #465278 After upgrading from Etch to Lenny, I started to get exactly the same problem than what Cyril Brulebois experienced. The laptop I am using is the latest iBook G4. The error message I get when I try to add again the appletouch module is almost identical to this: | Feb 11 17:04:52 evy kernel: drivers/input/mouse/appletouch.c: Could not do mode read request from device (Geyser Raw mode) | Feb 11 17:04:52 evy kernel: appletouch: probe of 1-2:1.0 failed with error -12 | Feb 11 17:04:52 evy kernel: usbcore: registered new interface driver appletouch except that in my case, I get the error -5 instead of -12: [45692.025239] appletouch: Could not do mode read request from device (Geyser Raw mode) [45692.025289] appletouch: probe of 1-2:1.0 failed with error -5 [45692.025347] usbcore: registered new interface driver appletouch Just to make things very clear, I don't think that this bug is the same than the one where removing and adding again the appletouch module is enough to make the mouse working again. The later bug appears quite infrequently, and usually the mouse is still able to move (even if in an unusable fashion). Also, from memory the error messages were very different. In the current bug, the mouse don't react anymore even in the text console and seems to happens systematically once the laptop goes to suspend mode. Also, it seems that the only way to make it working again is by rebooting the laptop. Some people have mentioned that the problem could be in the package xserver-xorg-input-synaptics and not in the kernel. Unless he was talking about the other mouse bug, I am sceptical about it and will try to reproduce the bug without running the X server at all during the boot process and report the result once I have time. Simon Valiquette http://gulus.USherbrooke.ca -- To UNSUBSCRIBE, email to debian-kernel-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Bug#465278: Appletouch failure upon resume after suspend2ram
Simon Valiquette un jour écrivit: Some people have mentioned that the problem could be in the package xserver-xorg-input-synaptics and not in the kernel. Unless he was talking about the other mouse bug, I am sceptical about it and will try to reproduce the bug without running the X server at all during the boot process and report the result once I have time. I can now confirm that this bug have nothing to do with xserver-xorg-input-synaptics, or xorg for that matter. I booted in single user mode and basically have done the following: /bin/mount -a /etc/init/udev start /etc/init/bootmisc.sh start /etc/init/pmud restart /etc/init/hibernate restart /etc/init/gpm start # to have a working mouse in the console /usr/sbin/pm-suspend And that was more than enough to reliably reproduce this old bug with very few running programs. Simon Valiquette -- To UNSUBSCRIBE, email to debian-kernel-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org