I'm using a i915-compatible graphics card and upgraded from xorg-server-1.7.7 to x11-base/xorg-server-1.9.4. The new Xserver requires to build kernel mode setting (KMS) for the i915-module. Enabling the KMS by passing the parameter 'i915.modeset=1' in menu.lst resulted at the first boot in a kernel panic. After restarting, the system could boot without problems. The remaining problem is the incredible slow speed caused by the spawning of four kworker-threads consuming the entire cpu workload. I've shut down all deamons including X, but the problem remains. Changing the state of /sys/module/drm_kms_helper/parameters/poll brings no effect. I've tried both kernels sys-kernel/gentoo-sources-2.6.37-r4 and sys-kernel/git-sources-2.6.39_rc5-r2. I suppose, the kernel is misconfigured for working correctly with the graphics card. Possible fix: Reenable the xorg-server-1.7.*-branch. The system was before enable KMS fully operable.
Created attachment 271591 [details] kernel conifguration 2.6.36-gentoo-r8
Created attachment 271593 [details] kernel conifguration 2.6.39-rc5-git2
Created attachment 271595 [details] lspci output
Created attachment 271597 [details] output of `top`
Created attachment 271599 [details] emerge --info
Comment on attachment 271593 [details] kernel conifguration 2.6.39-rc5-git2 Please also attach dmesg from a kernel which has i915 kms enabled.
Created attachment 271603 [details] the kernel panic Each time, when switching from 'i915.modeset=0' to 'i915.modeset=1', I'm being greeted by a kernel panic message whenever I'm starting X in the current session. In each succeeding session, I can successfully start X. The panic message looks like a conflict between wpa_supplicant's ath9k- and the i915-driver.
Comment to the kernel panic: The kernel panic occures only when starting X after switching from 'i915.modeset=0' to 'i915.modeset=1'. In each succeeding session using 'i915.modeset=1', the system can start X successfully. I've deactivated for the next round wpa_supplicant, and the kernel panic disappeared. Instead, the complete system freezes showing only a black screen. This time it doesn't look like a kernel panic (in a casual kernel panic the keys "Sys Rq" and "Prt Scr" are constantly blinking - in this case nothing flashes). But I cannot even use ACPI (like pressing the power button for shutdown). Again, in each succeeding session, the system can start X successfully.
Created attachment 271605 [details] dmesg of 2.6.39-rc5-git2 KMS enabled
When that happens, can you still get a dmesg (through ssh or other means)?
Thank's for the idea with ssh. I've done the following steps: * Switched from 'i915.modeset=0' to 'i915.modeset=1' * System started without X, hald, dbus * I'm using ssh to invoke any commands ** First attempt: ** `startx` echoed: X.Org X Server 1.9.4 Release Date: 2011-02-04 X Protocol Version 11, Revision 0 Build Operating System: Linux 2.6.36-gentoo-r8 i686 Gentoo Current Operating System: Linux eevi 2.6.39-rc5-git2 #1 SMP Fri Apr 29 17:50:34 CEST 2011 i686 Kernel command line: root=/dev/sda1 Build Date: 29 April 2011 06:08:39PM Current version of pixman: 0.20.2 Before reporting problems, check http://wiki.x.org to make sure that you have the latest version. Markers: (--) probed, (**) from config file, (==) default setting, (++) from command line, (!!) notice, (II) informational, (WW) warning, (EE) error, (NI) not implemented, (??) unknown. (==) Log file: "/var/log/Xorg.0.log", Time: Sat Apr 30 11:13:21 2011 (==) Using config file: "/etc/X11/xorg.conf" (==) Using system config directory "/usr/share/X11/xorg.conf.d" FATAL: Module fbcon not found. (EE) SynPS/2 Synaptics TouchPad no synaptics event device found (EE) Query no Synaptics: 6003C8 (EE) SynPS/2 Synaptics TouchPad Unable to query/initialize Synaptics hardware. (EE) PreInit failed for input device "SynPS/2 Synaptics TouchPad" - resulting in a black screen. This time I could switch with Ctrl+Alt+F1 back to console. ** Second attempt ** I've started dbus and hald. `startx` prompted the same messages again, but this time I got a working workspace. Seems that the first attempt merely could not manage displaying anything. ** Third attempt ** I've stopped dbus and hald. This time `startx` freezes my session leaving this message behind: X.Org X Server 1.9.4 Release Date: 2011-02-04 X Protocol Version 11, Revision 0 Build Operating System: Linux 2.6.36-gentoo-r8 i686 Gentoo Current Operating System: Linux eevi 2.6.39-rc5-git2 #1 SMP Fri Apr 29 17:50:34 CEST 2011 i686 Kernel command line: root=/dev/sda1 Build Date: 29 April 2011 06:08:39PM Current version of pixman: 0.20.2 Before reporting problems, check http://wiki.x.org to make sure that you have the latest version. Markers: (--) probed, (**) from config file, (==) default setting, (++) from command line, (!!) notice, (II) informational, (WW) warning, (EE) error, (NI) not implemented, (??) unknown. (==) Log file: "/var/log/Xorg.0.log", Time: Sat Apr 30 11:19:26 2011 (==) Using config file: "/etc/X11/xorg.conf" (==) Using system config directory "/usr/share/X11/xorg.conf.d" FATAL: Module fbcon not found. Write failed: Broken pipe ** Considerations ** * Unfortunately, in the interesting case, the system freezes completely. In the trivial case (working or working with blank screen), there is no additional output in `dmesg`. * I've reemerged x11-drivers/xf86-input-synaptics-1.3.0 after upgrading the xorg-server. In spite of the error-message, the touchpad is working (but with very crappy motion) * In the kernel-config, the option 'CONFIG_FRAMEBUFFER_CONSOLE=y' is set. I'm supposing that I need to compile the framebuffer-support as module (fbcon)? * I'm getting the frozen system more frequent with startx and hald disabled. By using xdm-display manager and hald I get sometimes a blank screen where I can use Ctrl+Alt+F1 to switch back. My /etc/X11/xorg.conf is skimped: Section "InputDevice" Identifier "Synaptics Touchpad" Driver "synaptics" Option "SendCoreEvents" "true" Option "Device" "/dev/psaux" Option "Protocol" "auto-dev" Option "MaxTapMove" "10" Option "SHMConfig" "on" EndSection
Created attachment 271617 [details] kernel-panic without wpa_supplicant after invoking `startx` There are rare chances that the system freezes by displaying a kernel message instead displaying a blank screen. I've invoked around 10 times the command `startx`, and killed the xserver via Ctrl+C.
Any updates or changes since the last comment from April?
Unfortunately, the condition remains with the most recent configuration (kernel 3.0.6) the same.
Please try the latest kernel, vanilla sources 3.2.9 and then please post at https://bugzilla.kernel.org/ and then post the url here.
https://bugzilla.kernel.org/show_bug.cgi?id=38442 - Problem still exists for kernel version 3.2.9
Thanks, we'll keep an eye on the upstream bug and work to backport any patches identified.
