[Bug 967938] New: Kernel panics / reboot / logout / frozen processes, when watching some Youtube videos
http://bugzilla.opensuse.org/show_bug.cgi?id=967938 Bug ID: 967938 Summary: Kernel panics / reboot / logout / frozen processes, when watching some Youtube videos Classification: openSUSE Product: openSUSE Tumbleweed Version: 2015* Hardware: Other OS: Other Status: NEW Severity: Normal Priority: P5 - None Component: Kernel Assignee: kernel-maintainers@forge.provo.novell.com Reporter: sonichedgehog_hyperblast00@yahoo.com QA Contact: qa-bugs@suse.de Found By: --- Blocker: --- After approximately a few weeks ago, I've been experiencing one of the strangest issues since using openSUSE: While watching certain videos on Youtube under Firefox, there is a probability that the operating system will crash. This only seems to happen while watching certain videos... I can play some clips as much as I want and nothing happens, but if I play others the OS may crash typically somewhere between 2 and 10 minutes. What's stranger is that there's a completely different effect every time, and there is no single crash... the list of possible outcomes I've experienced so far includes: - The computer simply freezes: I can no longer move the mouse cursor, use any keys (including toggling the NumLock / CapsLock leds), etc. Strangely enough, the audio of the video continues to play, even if the image is frozen. - The machine reboots by itself. The screen suddenly goes black for a few seconds, then I seamlessly find myself at the BIOS startup screen. The power LED stays up and I don't hear the typical "click" sound which normally occurs when you power off a computer... it acts as if software triggered. - I am logged out and immediately find myself in the login manager. It's exactly as if I clicked the "log out" button, except I don't press any keys nor have any applications that could automatically issue this command. - The Firefox process freezes, but with a bizarre side effect: Any process that tries to kill or create a new process from there on will itself freeze immediately! For example: If I open Konsole and type in "dolphin" to open the file manager, the cursor goes one line down and stays there... same thing if I try to kill a process with the kill command. If I try to start up a program from the kickoff launcher, plasmashell will freeze. If I try to launch something from krunner, krunner will freeze. I even logged off my normal user account, pressed control + alt + f1 to start a new runlevel, logged in as root, and tried to kill or start processes there... same thing. - In the best case scenario, Firefox simply crashes and I can start it back up again. Note: I do not believe this to be caused by faulty hardware. I use many programs on this machine, including hardware intensive games... I do not experience such crashes with anything else. Also I've recently changed video cards, and moved from the r600 driver to radeonsi... I remember that the issue used to happen on the old one as well, although I think it stopped at some point, possible after a certain package update. I don't use the proprietary video driver, and am running openSUSE Tumbleweed. -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.opensuse.org/show_bug.cgi?id=967938 Mircea Kitsune <sonichedgehog_hyperblast00@yahoo.com> changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |sonichedgehog_hyperblast00@ | |yahoo.com -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.opensuse.org/show_bug.cgi?id=967938 http://bugzilla.opensuse.org/show_bug.cgi?id=967938#c1 Takashi Iwai <tiwai@suse.com> changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |tiwai@suse.com Flags| |needinfo?(sonichedgehog_hyp | |erblast00@yahoo.com) --- Comment #1 from Takashi Iwai <tiwai@suse.com> --- Since the audio continued playing, it's likely a GPU hangup that resulted later in a kernel panic. Check the kernel log of the previous boot. You might have the kernel stack trace there. In anyway, please give more hardware details (e.g. hwinfo --gfxcard output) and the kernel messages including the Oops if available. -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.opensuse.org/show_bug.cgi?id=967938 http://bugzilla.opensuse.org/show_bug.cgi?id=967938#c2 --- Comment #2 from Mircea Kitsune <sonichedgehog_hyperblast00@yahoo.com> --- Created attachment 666773 --> http://bugzilla.opensuse.org/attachment.cgi?id=666773&action=edit The ~/.xsession-errors-:0 file -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.opensuse.org/show_bug.cgi?id=967938 http://bugzilla.opensuse.org/show_bug.cgi?id=967938#c3 --- Comment #3 from Mircea Kitsune <sonichedgehog_hyperblast00@yahoo.com> --- Created attachment 666774 --> http://bugzilla.opensuse.org/attachment.cgi?id=666774&action=edit Output of hwinfo --gfxcard (In reply to Takashi Iwai from comment #1) Which Kernel log please? There are a lot of system logs about various things, and they tend to get confusing. I attached the output of the command you mentioned for now. Also the xsession-errors file, since I remember people asked for it over similar reports. -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.opensuse.org/show_bug.cgi?id=967938 http://bugzilla.opensuse.org/show_bug.cgi?id=967938#c4 --- Comment #4 from Takashi Iwai <tiwai@suse.com> --- (In reply to Mircea Kitsune from comment #3)
Created attachment 666774 [details] Output of hwinfo --gfxcard
(In reply to Takashi Iwai from comment #1)
Which Kernel log please? There are a lot of system logs about various things, and they tend to get confusing.
The output of "journalctl -k". You can get the exact boot log for one session with -b option. For example, the previous boot is -b-1, two times before is -b-2, etc.
I attached the output of the command you mentioned for now. Also the xsession-errors file, since I remember people asked for it over similar reports.
These look OK through a quick glance. Also, try the kernel in OBS Kernel:stable repo. It contains the latest kernel that will be submitted to FACTORY/TW later. There are known issues about drm vblank handling that hits radeon, too. This will be fixed soon, maybe tomorrow or so, on Kernel:stable repo. -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.opensuse.org/show_bug.cgi?id=967938 http://bugzilla.opensuse.org/show_bug.cgi?id=967938#c5 --- Comment #5 from Mircea Kitsune <sonichedgehog_hyperblast00@yahoo.com> --- Created attachment 666779 --> http://bugzilla.opensuse.org/attachment.cgi?id=666779&action=edit Output of journalctl -k (In reply to Takashi Iwai from comment #4) Submitted the output of "journalctl -k". I don't think messing with the Kernel repositories is something I wish to risk doing now, especially if a future Kernel will fix the issue and I only need to wait a few days. I'm no longer installing package updates while watching Youtube, so there should be no danger of permanent damage to the system. -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.opensuse.org/show_bug.cgi?id=967938 http://bugzilla.opensuse.org/show_bug.cgi?id=967938#c6 --- Comment #6 from Mircea Kitsune <sonichedgehog_hyperblast00@yahoo.com> --- Created attachment 667233 --> http://bugzilla.opensuse.org/attachment.cgi?id=667233&action=edit top screenshot Attached a photo of top running in a console, taken after a new system freeze triggered by Youtube. After approximately a minute of being frozen like this, the machine rebooted by itself. -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.opensuse.org/show_bug.cgi?id=967938 http://bugzilla.opensuse.org/show_bug.cgi?id=967938#c7 --- Comment #7 from Takashi Iwai <tiwai@suse.com> --- Could you try radeon-quickfix-kmp in OBS home:tiwai:bnc955096 repo with 4.4.x kernel? It contains the backports from 4.5-rc regarding radeon driver. It doesn't contain backports for drm core, so something might be still missing, though. There are for openSUSE_Factory and Kernel-stable: the former is for the latest openSUSE TW kernel while the latter is for the latest OBS Kernel:stable. -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.opensuse.org/show_bug.cgi?id=967938 Markus Kolb <novell@tower-net.de> changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |novell@tower-net.de -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.opensuse.org/show_bug.cgi?id=967938 http://bugzilla.opensuse.org/show_bug.cgi?id=967938#c9 --- Comment #9 from Mircea Kitsune <sonichedgehog_hyperblast00@yahoo.com> --- This took a while, but it seems I've finally found the bizarre cause: The problem apparently happens only when watching 1080p and / or 60FPS videos. It stopped occurring after I disabled webm support in my browser, and now watch videos through mp4 instead. In Firefox, you must go in about:config and disable media.mediasource.webm.enabled. Does anyone know why this might be? Is there a vulnerability in the kernel or the free video drivers, which allows webm to cause kernel panics? -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.opensuse.org/show_bug.cgi?id=967938 http://bugzilla.opensuse.org/show_bug.cgi?id=967938#c10 Mircea Kitsune <sonichedgehog_hyperblast00@yahoo.com> changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution|--- |INVALID --- Comment #10 from Mircea Kitsune <sonichedgehog_hyperblast00@yahoo.com> --- Marking this as Resolved; It appears the problem was not a bug in the Kernel or video driver. It had to do with my overclock settings! The memories were clocked slightly too high, which introduced a small probability that the system would crash precisely while watching 1080p / 60FPS videos. I lowered the RAM clocks in BIOS, and there has been no crash after 10 days of uptime now. It's beyond me why a frequency setting would crash a machine only while watching Youtube, even after I had +4 days of uptime without problems... in any case, this wasn't what I thought it was. -- You are receiving this mail because: You are on the CC list for the bug.
participants (1)
-
bugzilla_noreply＠novell.com