[Bug 943845] New: X pegs CPU at 100%
http://bugzilla.opensuse.org/show_bug.cgi?id=943845 Bug ID: 943845 Summary: X pegs CPU at 100% Classification: openSUSE Product: openSUSE Factory Version: 201505* Hardware: Other OS: Other Status: NEW Severity: Normal Priority: P5 - None Component: X.Org Assignee: xorg-maintainer-bugs@forge.provo.novell.com Reporter: rjschwei@suse.com QA Contact: xorg-maintainer-bugs@forge.provo.novell.com Found By: --- Blocker: --- Created attachment 645652 --> http://bugzilla.opensuse.org/attachment.cgi?id=645652&action=edit Xorg log file from run where condition was observed.
rpm -qa | grep -i xorg xorg-x11-Xvnc-1.5.0-2.1.x86_64 xorg-x11-fonts-7.6-31.4.noarch xorg-x11-essentials-7.6_1-16.1.noarch xorg-x11-server-7.6_1.16.2-1.3.x86_64 xorg-scripts-1.0.1-10.1.noarch xorg-x11-util-devel-7.6_1-10.1.noarch xorg-cf-files-1.0.5-5.1.noarch xorg-x11-server-extra-7.6_1.17.2-3.1.x86_64 xorg-x11-driver-video-7.6_1-15.1.x86_64 xorg-x11-driver-input-7.6_1-13.1.noarch xorg-sgml-doctools-1.11-9.1.noarch xorg-x11-libX11-ccache-7.6-20.2.noarch xorg-x11-fonts-core-7.6-31.4.noarch xorg-x11-devel-7.6-47.1.noarch
On startup X loads the CPU 100% and makes the machine unresponsive. The condition has been observed to last 15 minutes after which I turned the machine off, forced off via power switch, shutdown did not work. The condition does luckily not exist 100% of the time, it appears that it occurs on start up, then after reboot by triggering a power cycle the 100% CPU condition only occurs for a short period of time, approximately 1 to 2 minutes. # hwinfo --gfxcard 09: PCI 02.0: 0300 VGA compatible controller (VGA) [Created at pci.366] Unique ID: _Znp.W2XlUwjUxL4 SysFS ID: /devices/pci0000:00/0000:00:02.0 SysFS BusID: 0000:00:02.0 Hardware Class: graphics card Model: "Intel 3rd Gen Core processor Graphics Controller" Vendor: pci 0x8086 "Intel Corporation" Device: pci 0x0166 "3rd Gen Core processor Graphics Controller" SubVendor: pci 0x17aa "Lenovo" SubDevice: pci 0x21f9 Revision: 0x09 Driver: "i915" Driver Modules: "drm" Memory Range: 0xf0000000-0xf03fffff (rw,non-prefetchable) Memory Range: 0xe0000000-0xefffffff (ro,non-prefetchable) I/O Ports: 0x4000-0x403f (rw) IRQ: 28 (385 events) Module Alias: "pci:v00008086d00000166sv000017AAsd000021F9bc03sc00i00" Driver Info #0: Driver Status: i915 is active Driver Activation Cmd: "modprobe i915" Config Status: cfg=new, avail=yes, need=no, active=unknown Primary display adapter: #9 I am using XFCE as my desktop -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.opensuse.org/show_bug.cgi?id=943845 Stefan Dirsch <sndirsch@suse.com> changed: What |Removed |Added ---------------------------------------------------------------------------- Priority|P5 - None |P3 - Medium Status|NEW |CONFIRMED -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.opensuse.org/show_bug.cgi?id=943845 http://bugzilla.opensuse.org/show_bug.cgi?id=943845#c6 --- Comment #6 from Egbert Eich <eich@suse.com> --- Robert? Any news? -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.opensuse.org/show_bug.cgi?id=943845 http://bugzilla.opensuse.org/show_bug.cgi?id=943845#c7 --- Comment #7 from Robert Schweikert <rjschwei@suse.com> --- So the command/application that is most likely involved at this point is Emacs. As soon as I switch to the desktop where Emacs is autostarted, as it was running there when I shut down the machine, the CPU load from X goes down to "regular" levels. I will now attempt to confirm this with the instructions provided in comment #5. -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.opensuse.org/show_bug.cgi?id=943845 http://bugzilla.opensuse.org/show_bug.cgi?id=943845#c8 Robert Schweikert <rjschwei@suse.com> changed: What |Removed |Added ---------------------------------------------------------------------------- Flags|needinfo?(rjschwei@suse.com | |) | --- Comment #8 from Robert Schweikert <rjschwei@suse.com> --- Using gdb did not work. When I tried to attach to the X server process the whole system locked up. I waited 10 or so minutes to make certain I was not being impatient but the system did not recover and I ended up doing a hard reset. I did confirm my suspicion that it is emacs related. After starting back up I closed all the emacs sessions and then rebooted. The condition was not reproducible, i.e. with emacs not starting automatically when I log in the X server no longer pegs the CPU. -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.opensuse.org/show_bug.cgi?id=943845 http://bugzilla.opensuse.org/show_bug.cgi?id=943845#c9 --- Comment #9 from Stefan Dirsch <sndirsch@suse.com> --- Robert, seems Egbert forgot to mention that the Xserver needs to be debugged via ssh. -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.opensuse.org/show_bug.cgi?id=943845 http://bugzilla.opensuse.org/show_bug.cgi?id=943845#c10 --- Comment #10 from Egbert Eich <eich@suse.com> --- (In reply to Stefan Dirsch from comment #9)
Robert, seems Egbert forgot to mention that the Xserver needs to be debugged via ssh.
No, I assumed that Robert understands enough about software that he would not run a debugger whose output is displayed by a program that it tries to debug and stop :) -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.opensuse.org/show_bug.cgi?id=943845 http://bugzilla.opensuse.org/show_bug.cgi?id=943845#c11 --- Comment #11 from Egbert Eich <eich@suse.com> --- Robert, - does this only happen after emacs is autostarted and not when it is started fresh? - you said, the load goes down, when you switch to the desktop where emacs is running. If I read your initial description right, it doesn't come back when you switch to away from this desktop again, does it? -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.opensuse.org/show_bug.cgi?id=943845 http://bugzilla.opensuse.org/show_bug.cgi?id=943845#c12 --- Comment #12 from Robert Schweikert <rjschwei@suse.com> --- (In reply to Egbert Eich from comment #11)
Robert, - does this only happen after emacs is autostarted and not when it is started fresh?
Correct
- you said, the load goes down, when you switch to the desktop where emacs is running. If I read your initial description right, it doesn't come back when you switch to away from this desktop again, does it?
Correct, the symptom only occurs when I login and emacs autostarts. Once I switch to the desktop where emacs is running the load goes down and remains at "normal" level for the rest of the session, including switching away from the desktop. -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.opensuse.org/show_bug.cgi?id=943845 http://bugzilla.opensuse.org/show_bug.cgi?id=943845#c13 --- Comment #13 from Robert Schweikert <rjschwei@suse.com> --- (In reply to Egbert Eich from comment #10)
(In reply to Stefan Dirsch from comment #9)
Robert, seems Egbert forgot to mention that the Xserver needs to be debugged via ssh.
No, I assumed that Robert understands enough about software that he would not run a debugger whose output is displayed by a program that it tries to debug and stop :)
Yes, I guess that should have been rather obvious. However, given that I do not work in this area on a more or less regular basis I guess it could have been reasonably predictable that I would make that mistake. Anyway, here is the confirmation that it is emacs that is triggering the problem: Breakpoint 1, ProcGetInputFocus (client=0x1805900) at events.c:4870 4870 events.c: No such file or directory. (gdb) p *client->clientIds $1 = {pid = 15289, cmdname = 0x1805c80 "/usr/bin/emacs", cmdargs = 0x1805ca0 "--smid=2d8141d58-12db-49a6-8893-2addd79140b3 --no-splash --chdir=/work/rjschwei/devel/smt apache2/conf.d/smt_mod_perl.conf"} -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.opensuse.org/show_bug.cgi?id=943845 http://bugzilla.opensuse.org/show_bug.cgi?id=943845#c14 Egbert Eich <eich@suse.com> changed: What |Removed |Added ---------------------------------------------------------------------------- Status|CONFIRMED |NEW Assignee|xorg-maintainer-bugs@forge. |werner@suse.com |provo.novell.com | Summary|[Intel IVB] X pegs CPU at |Emacs auto-started on |100% |non-current desktop hammers | |Xserver with requests | |pushing CPU to 100% Severity|Normal |Minor --- Comment #14 from Egbert Eich <eich@suse.com> --- Thanks for following up, Robert! (In reply to Robert Schweikert from comment #13)
Yes, I guess that should have been rather obvious. However, given that I do not work in this area on a more or less regular basis I guess it could have been reasonably predictable that I would make that mistake.
Always assumed you were a super-natural software wizard ;p
Anyway, here is the confirmation that it is emacs that is triggering the problem:
Breakpoint 1, ProcGetInputFocus (client=0x1805900) at events.c:4870 4870 events.c: No such file or directory. (gdb) p *client->clientIds $1 = {pid = 15289, cmdname = 0x1805c80 "/usr/bin/emacs", cmdargs = 0x1805ca0 "--smid=2d8141d58-12db-49a6-8893-2addd79140b3 --no-splash --chdir=/work/rjschwei/devel/smt apache2/conf.d/smt_mod_perl.conf"}
Right. Let's tweak the subject and reassign it to the maintainer of emacs. I'll reduce severity as the condition doesn't persist. Robert, maybe you should update your TW to make sure the condition still persists. -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.opensuse.org/show_bug.cgi?id=943845 http://bugzilla.opensuse.org/show_bug.cgi?id=943845#c19 Robert Schweikert <rjschwei@suse.com> changed: What |Removed |Added ---------------------------------------------------------------------------- Flags|needinfo?(rjschwei@suse.com | |) | --- Comment #19 from Robert Schweikert <rjschwei@suse.com> --- More than happy to provide additional information unfortunately I am also not in a position from a time perspective to debug this. Consider it took me almost 2 month to supply the information for comment 5, which is a sad state of affairs. Anyway the desktop being used on the system is XFCE. It runs emacs-gtk. Hardware is a Lenovo Carbon X1 first generation, all Intel hardware # hwinfo --gfxcard 08: PCI 02.0: 0300 VGA compatible controller (VGA) [Created at pci.366] Unique ID: _Znp.W2XlUwjUxL4 SysFS ID: /devices/pci0000:00/0000:00:02.0 SysFS BusID: 0000:00:02.0 Hardware Class: graphics card Model: "Intel 3rd Gen Core processor Graphics Controller" Vendor: pci 0x8086 "Intel Corporation" Device: pci 0x0166 "3rd Gen Core processor Graphics Controller" SubVendor: pci 0x17aa "Lenovo" SubDevice: pci 0x21f9 Revision: 0x09 Driver: "i915" Driver Modules: "drm" Memory Range: 0xf0000000-0xf03fffff (rw,non-prefetchable) Memory Range: 0xe0000000-0xefffffff (ro,non-prefetchable) I/O Ports: 0x4000-0x403f (rw) IRQ: 27 (188 events) Module Alias: "pci:v00008086d00000166sv000017AAsd000021F9bc03sc00i00" Driver Info #0: Driver Status: i915 is active Driver Activation Cmd: "modprobe i915" Config Status: cfg=new, avail=yes, need=no, active=unknown Primary display adapter: #8 # rpm -qa emacs emacs-24.5-4.1.x86_64 -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.opensuse.org/show_bug.cgi?id=943845 http://bugzilla.opensuse.org/show_bug.cgi?id=943845#c20 --- Comment #20 from Robert Schweikert <rjschwei@suse.com> --- Created attachment 661450 --> http://bugzilla.opensuse.org/attachment.cgi?id=661450&action=edit strace of X server with timestamps enabled As in the previous strace after it was running for a bit I switch to the desktop where emacs was autostarted at which point the symptom stops. This should be visible in the trace. -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.opensuse.org/show_bug.cgi?id=943845 http://bugzilla.opensuse.org/show_bug.cgi?id=943845#c21 --- Comment #21 from Egbert Eich <eich@suse.com> --- (In reply to Robert Schweikert from comment #19)
More than happy to provide additional information unfortunately I am also not in a position from a time perspective to debug this. Consider it took me almost 2 month to supply the information for comment 5, which is a sad state of affairs.
Anyway the desktop being used on the system is XFCE.
It runs emacs-gtk.
Hardware is a Lenovo Carbon X1 first generation, all Intel hardware
Robert, thanks for the info! The hardware here is not really relevant here. I believe, with systemd Werner's workload is not much lower than yours. My situation is similar as well. -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.opensuse.org/show_bug.cgi?id=943845 http://bugzilla.opensuse.org/show_bug.cgi?id=943845#c38 Stefan Seyfried <seife@novell.slipkontur.de> changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |seife@novell.slipkontur.de Flags|needinfo?(bnc-team-xfce@for | |ge.provo.novell.com) | --- Comment #38 from Stefan Seyfried <seife@novell.slipkontur.de> --- bnc-team-xfce@forge.provo.novell.com is a list that probably nobody reads. What is needed from XFCE maintainers here? I can start "xterm &" without killing my xserver performance ;-) -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.opensuse.org/show_bug.cgi?id=943845 Stefan Seyfried <seife@novell.slipkontur.de> changed: What |Removed |Added ---------------------------------------------------------------------------- CC|bnc-team-xfce@forge.provo.n | |ovell.com | -- You are receiving this mail because: You are on the CC list for the bug.
participants (1)
-
bugzilla_noreply@novell.com