[Bug 222174] New: XEN Kernel 2.6.16.21-0.25 x86_64+Intel 945M+X11 hangs with part of the screen distorted.
https://bugzilla.novell.com/show_bug.cgi?id=222174 Summary: XEN Kernel 2.6.16.21-0.25 x86_64+Intel 945M+X11 hangs with part of the screen distorted. Product: SUSE Linux 10.1 Version: Final Platform: x86-64 OS/Version: SuSE Linux 10.1 Status: NEW Severity: Critical Priority: P5 - None Component: Xen AssignedTo: cgriffin@novell.com ReportedBy: Wojciech.Szenajch@bull.com.pl QAContact: qa@suse.de On Dell laptop D820 Intel T7200 Core2 Duo 2GHz 4MB Cache, 2GB RAM, Video Intel 945M, 100GB HDD SATA. Suse 10.1 x86_64 XEN installs and starts successfully. After some activity i.e. executing sum command on 0.5 GB file, down of the screen is blurred (probably by disk buffers allocated in RAM used also by video card) and quite often system hangs some time after. With standard 2.6.16.21-0.25-smp x86_64 kernel this system ont the same laptop works without any problems including executing sum command on the same file as previously and on 4GB sized file also. Additional information: SuSE10.1 installed from Open SuSE 10.1 Remastered DVD x86_64 and all relevant patches for day 17.11.2006 were installed also. Tested with kde and gnome. Video card: 32: udi = '/org/freedesktop/Hal/devices/pci_8086_27a0' info.bus = 'pci' info.linux.driver = 'agpgart-intel' info.parent = '/org/freedesktop/Hal/devices/computer' info.product = 'Mobile 945GM/PM/GMS/940GML and 945GT Express Memory Controller Hub' info.udi = '/org/freedesktop/Hal/devices/pci_8086_27a0' info.vendor = 'Intel Corporation' linux.hotplug_type = 1 (0x1) linux.subsystem = 'pci' linux.sysfs_path = '/sys/devices/pci0000:00/0000:00:00.0' linux.sysfs_path_device = '/sys/devices/pci0000:00/0000:00:00.0' pci.device_class = 6 (0x6) pci.device_protocol = 0 (0x0) pci.device_subclass = 0 (0x0) pci.linux.sysfs_path = '/sys/devices/pci0000:00/0000:00:00.0' pci.product = 'Mobile 945GM/PM/GMS/940GML and 945GT Express Memory Controller Hub' pci.product_id = 10144 (0x27a0) pci.subsys_product = 'Unknown (0x01cc)' pci.subsys_product_id = 460 (0x1cc) pci.subsys_vendor = 'Dell' pci.subsys_vendor_id = 4136 (0x1028) pci.vendor = 'Intel Corporation' pci.vendor_id = 32902 (0x8086) 33: udi = '/org/freedesktop/Hal/devices/pci_8086_27a2' info.bus = 'pci' info.parent = '/org/freedesktop/Hal/devices/computer' info.product = 'Mobile 945GM/GMS/940GML Express Integrated Graphics Controller' info.udi = '/org/freedesktop/Hal/devices/pci_8086_27a2' info.vendor = 'Intel Corporation' linux.hotplug_type = 1 (0x1) linux.subsystem = 'pci' linux.sysfs_path = '/sys/devices/pci0000:00/0000:00:02.0' linux.sysfs_path_device = '/sys/devices/pci0000:00/0000:00:02.0' pci.device_class = 3 (0x3) pci.device_protocol = 0 (0x0) pci.device_subclass = 0 (0x0) pci.linux.sysfs_path = '/sys/devices/pci0000:00/0000:00:02.0' pci.product = 'Mobile 945GM/GMS/940GML Express Integrated Graphics Controller' pci.product_id = 10146 (0x27a2) pci.subsys_product = 'Unknown (0x01cc)' pci.subsys_product_id = 460 (0x1cc) pci.subsys_vendor = 'Dell' pci.subsys_vendor_id = 4136 (0x1028) pci.vendor = 'Intel Corporation' pci.vendor_id = 32902 (0x8086) 34: udi = '/org/freedesktop/Hal/devices/pci_8086_27a6' info.bus = 'pci' info.parent = '/org/freedesktop/Hal/devices/computer' info.product = 'Mobile 945GM/GMS/940GML Express Integrated Graphics Controller' info.udi = '/org/freedesktop/Hal/devices/pci_8086_27a6' info.vendor = 'Intel Corporation' linux.hotplug_type = 1 (0x1) linux.subsystem = 'pci' linux.sysfs_path = '/sys/devices/pci0000:00/0000:00:02.1' linux.sysfs_path_device = '/sys/devices/pci0000:00/0000:00:02.1' pci.device_class = 3 (0x3) pci.device_protocol = 0 (0x0) pci.device_subclass = 128 (0x80) pci.linux.sysfs_path = '/sys/devices/pci0000:00/0000:00:02.1' pci.product = 'Mobile 945GM/GMS/940GML Express Integrated Graphics Controller' pci.product_id = 10150 (0x27a6) pci.subsys_product = 'Unknown (0x01cc)' pci.subsys_product_id = 460 (0x1cc) pci.subsys_vendor = 'Dell' pci.subsys_vendor_id = 4136 (0x1028) pci.vendor = 'Intel Corporation' pci.vendor_id = 32902 (0x8086) I will be able to answer any additional information requests after 26.11.2006. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is.
https://bugzilla.novell.com/show_bug.cgi?id=222174 jdouglas@novell.com changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |NEEDINFO Info Provider| |Wojciech.Szenajch@bull.com.pl ------- Comment #1 from jdouglas@novell.com 2006-11-20 13:52 MST ------- Thanks for reporting this issue. We think this is fixed in openSUSE 10.2. Could you please retest with openSUSE 10.2 Release Candidate 1 (available 11/24/06)? Thanks. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is.
https://bugzilla.novell.com/show_bug.cgi?id=222174 Wojciech.Szenajch@bull.com.pl changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEEDINFO |ASSIGNED Info Provider|Wojciech.Szenajch@bull.com.p| |l | ------- Comment #2 from Wojciech.Szenajch@bull.com.pl 2006-11-22 01:15 MST ------- Is it fixed in SLES10? I am evaluating OpenSuSE 10.1 as preliminary tests for SLES10 for company usage. I want to make xen presentations outside of the office. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is.
https://bugzilla.novell.com/show_bug.cgi?id=222174 lbendixs@novell.com changed: What |Removed |Added ---------------------------------------------------------------------------- AssignedTo|cgriffin@novell.com |jdouglas@novell.com Status|ASSIGNED |NEW QAContact|qa@suse.de |jdouglas@novell.com -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is.
https://bugzilla.novell.com/show_bug.cgi?id=222174 lbendixs@novell.com changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |ASSIGNED -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is.
https://bugzilla.novell.com/show_bug.cgi?id=222174 ------- Comment #3 from Wojciech.Szenajch@bull.com.pl 2006-11-28 08:06 MST ------- The same bug is in OpenSuSE 10.2 RC1. Tested with kernels XEN: 2.6.18.2-23 and 2.6.18.2-33. With default kernels 2.6.18.2-23-default and 2.6.18.2-33-default it works correctly. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is.
https://bugzilla.novell.com/show_bug.cgi?id=222174 lbendixs@novell.com changed: What |Removed |Added ---------------------------------------------------------------------------- AssignedTo|jdouglas@novell.com |jbeulich@novell.com Status|ASSIGNED |NEW -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is.
https://bugzilla.novell.com/show_bug.cgi?id=222174 lbendixs@novell.com changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |ASSIGNED -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is.
https://bugzilla.novell.com/show_bug.cgi?id=222174 ------- Comment #4 from Wojciech.Szenajch@bull.com.pl 2006-11-29 02:53 MST ------- XEN kernel from 32 bit SuSE 10.1 works correctly on the same hardware on which x86_64 SuSE10.1/10.2 XEN versions failed as reported above. Tested with kernel 2.6.16.13-4-xen i686 and sum command on about 3.7 GB file. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is.
https://bugzilla.novell.com/show_bug.cgi?id=222174 jbeulich@novell.com changed: What |Removed |Added ---------------------------------------------------------------------------- Status|ASSIGNED |NEEDINFO Info Provider| |Wojciech.Szenajch@bull.com.pl ------- Comment #5 from jbeulich@novell.com 2006-11-29 03:33 MST ------- Are you seeing the screen distortion also on 10.2? In either case, we will need to see any potential Xen messages, so we need to ask you to attach a second machine via serial cable to collect Xen output. If that is impossible (due to the lack of a serial connector on the laptop), try running the sum command from a text console (if you do this on 10.2, you would also want to enlarge screen space by specifying vga=text-80x50,keep or vga=text-80x60,keep). Also, did you check that the kernel hangs (rather than crashes)? -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is.
https://bugzilla.novell.com/show_bug.cgi?id=222174 ------- Comment #6 from jbeulich@novell.com 2006-11-29 03:34 MST ------- *** Bug 224170 has been marked as a duplicate of this bug. *** -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is.
https://bugzilla.novell.com/show_bug.cgi?id=222174 ------- Comment #7 from jbeulich@novell.com 2006-11-29 05:05 MST ------- Also, please provide - boot.msg for both a native and a Xen kernel boot - a list of modules active at the time of the hang -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is.
https://bugzilla.novell.com/show_bug.cgi?id=222174 ------- Comment #8 from Wojciech.Szenajch@bull.com.pl 2006-11-29 07:27 MST ------- Created an attachment (id=107396) --> (https://bugzilla.novell.com/attachment.cgi?id=107396&action=view) Requested lsmod and boot.msg files -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is.
https://bugzilla.novell.com/show_bug.cgi?id=222174 Wojciech.Szenajch@bull.com.pl changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEEDINFO |ASSIGNED Info Provider|Wojciech.Szenajch@bull.com.p| |l | ------- Comment #9 from Wojciech.Szenajch@bull.com.pl 2006-11-29 07:32 MST ------- Answers: "Are you seeing the screen distortion also on 10.2?" Yes, always. Down bottom is distorted up to about default panel height, but it is panel independent (I moved it at the top of the screen to test this). After some time usually shell-console title bar is distorted and cursor area. SuSE 10.2 behaves a little differently than SuSE 10.1 but on SuSE 10.2 I am not able to set native screen resolution and I have to use lower one. (I will report this as separate bug.) "... try running the sum command from a text console (if you do this on 10.2, you would also want to enlarge screen space by specifying vga=text-80x50,keep or vga=text-80x60,keep)." I did this on 10.2 from text console as described. It hanged (no any message) after NFS copying 3.7 GB file to local disk and running sum on it next. It is easier to catch the problem running sum from kde shell-console. "Also, did you check that the kernel hangs (rather than crashes)?" Computer hanged completely without any messages. The only key which worked it was power key kept for about 7 seconds to switch computer off :(. Requested files in attachement above: lsmod.xen was made without starting any X session, lsmod.xen.x was made after reboot from X kde session and lsmod.default was made from X kde session. boot.msg and boot.omsg for default and xen also included. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is.
https://bugzilla.novell.com/show_bug.cgi?id=222174 jbeulich@novell.com changed: What |Removed |Added ---------------------------------------------------------------------------- Status|ASSIGNED |NEEDINFO Info Provider| |Wojciech.Szenajch@bull.com.pl ------- Comment #10 from jbeulich@novell.com 2006-11-29 08:23 MST ------- Not having seen any messages when running the command from a text console doesn't mean anything, yet: (a) If the kernel crashes, you'd have to watch the syslog console to observe anything. (b) If Xen spits out messages or crashes, you have to force it to keep generating output to the VGA console. Finally, if indeed neither shows anything, then you'd need to enable the debug hypervisor (which is installed alongside the non-debug one, so you just need to edit /boot/grub/menu.lst) to see if that one complains about anything. And again, we will want to see the complete (beginning at boot) output from (normal or debug) Xen collected via a serial cable. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is.
https://bugzilla.novell.com/show_bug.cgi?id=222174 ------- Comment #11 from jbeulich@novell.com 2006-11-29 08:43 MST ------- Oh, from looking at the Xen boot messages I see that you added vga= to the kernel command line, whereas this needs to go on the Xen one. Further - is your BIOS up to date? A checksum error in ACPI table scanning doesn't make any of the provided data look reliable. Also, if collecting the output as indicated in #10 doesn't provide any insights we'll need to ask you to strip down your configuration to at least cut down on the number of possible origins for the problem. Namely would I want to see the list of loaded modules signficiantly reduced (what I specifically would want to see gone are i915, drm, i2c_*, ipw3945, perhaps all the sound stuff, ideally as much as possible from the huge remainder). -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is.
https://bugzilla.novell.com/show_bug.cgi?id=222174 ------- Comment #12 from Wojciech.Szenajch@bull.com.pl 2006-11-30 04:20 MST ------- Answers: Comment #11: BIOS is the latest available from Dell. In case of vga= option - sorry - please check my files once again: xen kernel is called with suggested options. I removed as many things as it was possible at the time i could spend on this subject and I also changed testing method to more efficient. i2c_* and ipw3945 stayed but they are not used according to lsmod. New testing method for default and xen kernels: set id:1:initdefault: in /etc/inittab and (re)boot desired kernel collected files: boot.msg and lsmod.s1-* tests with Bfile sized about 600MB (enough for my 2GM of RAM). gzip Bfile guzip Bfile next manually: /etc/rc.d/xdm start collected files: lsmod.s1x-* this loads two additional drivers (for xen collected without running gzip in another boot session):
i915 28160 2 drm 94504 3 i915
For default kernel X11 starts and works without any problems but for xen the kdm login screen is distorted by vertical lines and PC hangs without possibility to login from kdm. I made additional tests by starting xdm without gzip tests from the console. I repeated gzip Bfile from kde shell-console. For default kernel everything was OK. For xen I got wonderful "RAM graphical monitor" with changing dots and patterns. Finally PC hanged. You may speed up this hang if you "clean" distorted area i.e. by moving terminal window on it to make it refreshed by X11. It seems that xen kernel together with or i915/drm incorrectly allocates/uses RAM used by operating system for other purposes. Because RAM is overwritten by graphical card and operating system together everything hangs randomly without any system message. During my past tests with SuSE10.2 (with resolution 1280x1024 instead of 1680x1050) I got sum command core dumped after several usages or even the following error was reported in another boot session: sum: SUSE-Linux-10.1-GM-DVD-i386.iso: Input/output error after several repetitions of the same: sum SUSE-Linux-10.1-GM-DVD-i386.iso command. System with lower resolution was more resistant because probably less RAM was allocated to video card or/and it was done differently. For requested files see the following attachment. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is.
https://bugzilla.novell.com/show_bug.cgi?id=222174 Wojciech.Szenajch@bull.com.pl changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEEDINFO |ASSIGNED Attachment #107396|0 |1 is obsolete| | ------- Comment #13 from Wojciech.Szenajch@bull.com.pl 2006-11-30 04:21 MST ------- Created an attachment (id=107582) --> (https://bugzilla.novell.com/attachment.cgi?id=107582&action=view) Second version of requested files boot.msg and lsmod reoports. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is.
https://bugzilla.novell.com/show_bug.cgi?id=222174 jbeulich@novell.com changed: What |Removed |Added ---------------------------------------------------------------------------- Status|ASSIGNED |NEEDINFO Info Provider| |Wojciech.Szenajch@bull.com.pl ------- Comment #14 from jbeulich@novell.com 2006-11-30 06:04 MST ------- I am still seeing the vga=text-80x60,keep on the kernel command line. I am still missing all Xen output. Are you saying that without drm and i915 loaded you are not able to bring the machine down anymore? -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is.
https://bugzilla.novell.com/show_bug.cgi?id=222174 ------- Comment #15 from Wojciech.Szenajch@bull.com.pl 2006-11-30 07:14 MST ------- Created an attachment (id=107619) --> (https://bugzilla.novell.com/attachment.cgi?id=107619&action=view) Second version of b oot.msg xen. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is.
https://bugzilla.novell.com/show_bug.cgi?id=222174 Wojciech.Szenajch@bull.com.pl changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEEDINFO |ASSIGNED Info Provider|Wojciech.Szenajch@bull.com.p| |l | ------- Comment #16 from Wojciech.Szenajch@bull.com.pl 2006-11-30 07:27 MST ------- Please see attachment above. "Are you saying that without drm and i915 loaded you are not able to bring the machine down anymore?" NO. I am able to do it. Machine stops to respond (hangs completely) if you do described gzip (or similar large file) operations AND start xdm. If you do not start xdm (drm, i915 not loaded) system seems to work normally, although I did not test it for a long time without graphics. In my opinion problem is in i915/drm x86_64 xen versions. i945M uses system RAM memory and probably requests for more of it when X11 are started. From screen distortions described in #12 it seems that video obtains memory already used by operating system and there is the conflict. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is.
https://bugzilla.novell.com/show_bug.cgi?id=222174 Wojciech.Szenajch@bull.com.pl changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEEDINFO |ASSIGNED Info Provider|Wojciech.Szenajch@bull.com.p| |l | ------- Comment #18 from Wojciech.Szenajch@bull.com.pl 2006-11-30 07:37 MST ------- Possible misunderstanding: "bring machine down" for me meant to made it shutdown. You thought probably about crashing/hanging it. Anyway I confirm that I do not see any problems as long as i915/drm aren't loaded. As for #15 I am not xen kernel parameters expert. Please give me more configuration details, if more data is required. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is.
https://bugzilla.novell.com/show_bug.cgi?id=222174 ------- Comment #20 from Wojciech.Szenajch@bull.com.pl 2006-12-04 01:35 MST ------- I confirm that patch 2.6.18.4-31_xtp-xen fixed the problem. Unfortunately I had no time to test xen virtual domains yet, but all gzip/gunzip tests with file up to 1.8GB passed successfully. Thanks. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is.
https://bugzilla.novell.com/show_bug.cgi?id=222174 jbeulich@novell.com changed: What |Removed |Added ---------------------------------------------------------------------------- Status|ASSIGNED |RESOLVED Resolution| |FIXED ------- Comment #21 from jbeulich@novell.com 2006-12-04 02:09 MST ------- patch committed to 10.2, SLE10 GA, SLE10 SP1, and head. Should be available with respective next security updates for 10.2 and SLE10. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is.
https://bugzilla.novell.com/show_bug.cgi?id=222174 hmuelle@novell.com changed: What |Removed |Added ---------------------------------------------------------------------------- Status Whiteboard| |kernel:sles10 ------- Comment #22 from hmuelle@novell.com 2006-12-11 00:00 MST ------- Marked for next possible update. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is.
https://bugzilla.novell.com/show_bug.cgi?id=222174 kgw@novell.com changed: What |Removed |Added ---------------------------------------------------------------------------- Status Whiteboard|kernel:sles10 |fixreleased:kernel:sles10 ------- Comment #23 from kgw@novell.com 2007-02-09 08:25 MST ------- Patch: patches.xen/xen-x86_64-agp published in SLE10 kernelupdate 2.6.16.27-0.6, dated Dec 13, 2006 & released Dec 21, 2006. Setting Whiteboard Status -> fixreleased -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is.
participants (1)
-
bugzilla_noreply@novell.com