[Bug 12788] ~40,000 wakeups/sec after suspend/resume with Tickless kernel; idle=nomwait or max_cstate=1 helps

https://bugzilla.kernel.org/show_bug.cgi?id=12788 Alan changed: What|Removed |Added Status|RESOLVED|CLOSED --- Comment #69 from Alan 2012-05-30

[Bug 12788] ~40,000 wakeups/sec after suspend/resume with Tickless kernel; idle=nomwait or max_cstate=1 helps

https://bugzilla.kernel.org/show_bug.cgi?id=12788 Alan changed: What|Removed |Added Status|NEEDINFO|RESOLVED CC|

[Bug 12788] ~40,000 wakeups/sec after suspend/resume with Tickless kernel; idle=nomwait or max_cstate=1 helps

https://bugzilla.kernel.org/show_bug.cgi?id=12788 --- Comment #68 from James Ettle 2012-01-19 12:09:16 --- (In reply to comment #67) > can you please verify if the problem still exists in the latest upstream > kernel? Using kernel-3.2.1-4.jhe.fc16.x86_64 (built from Fedora sources), with t

[Bug 12788] ~40,000 wakeups/sec after suspend/resume with Tickless kernel; idle=nomwait or max_cstate=1 helps

https://bugzilla.kernel.org/show_bug.cgi?id=12788 --- Comment #67 from Zhang Rui 2012-01-18 01:41:36 --- It's great that kernel bugzilla is back. James, can you please verify if the problem still exists in the latest upstream kernel? -- Configure bugmail: https://bugzilla.kernel.org/userp

[Bug 12788] ~40,000 wakeups/sec after suspend/resume with Tickless kernel; idle=nomwait or max_cstate=1 helps

https://bugzilla.kernel.org/show_bug.cgi?id=12788 --- Comment #66 from ykzhao 2010-12-06 03:30:14 --- Hi, James Sorry for the late response about the issue. It is very strange that the number of wakeup is increased very significantly after suspend/resume when the mwait is used to enter t

[Bug 12788] ~40,000 wakeups/sec after suspend/resume with Tickless kernel; idle=nomwait or max_cstate=1 helps

https://bugzilla.kernel.org/show_bug.cgi?id=12788 --- Comment #65 from James Ettle 2010-11-10 11:22:59 --- (In reply to comment #63) > One difference is the default clocksource gets set to hpet > rather than tsc, and tsc is no longer available. dmesg attached. OK, I now understand this is

[Bug 12788] ~40,000 wakeups/sec after suspend/resume with Tickless kernel; idle=nomwait or max_cstate=1 helps

https://bugzilla.kernel.org/show_bug.cgi?id=12788 --- Comment #64 from James Ettle 2010-11-10 09:01:56 --- Strange... in the 40 minutes or so since I last posted, wakeups seem to have calmed to around 2500-4000 per second. No idea what's changed in that time... -- Configure bugmail: https

[Bug 12788] ~40,000 wakeups/sec after suspend/resume with Tickless kernel; idle=nomwait or max_cstate=1 helps

https://bugzilla.kernel.org/show_bug.cgi?id=12788 --- Comment #63 from James Ettle 2010-11-10 08:19:03 --- Created an attachment (id=37022) --> (https://bugzilla.kernel.org/attachment.cgi?id=37022) dmesg when booted with acpi_skip_timer_override Some more info: If I boot with acpi_skip_ti

[Bug 12788] ~40,000 wakeups/sec after suspend/resume with Tickless kernel; idle=nomwait or max_cstate=1 helps

https://bugzilla.kernel.org/show_bug.cgi?id=12788 --- Comment #62 from Vitus Jensen 2010-11-07 11:54:02 --- I do have a similar problem with tickless kernel but in a totally different machine: single Pentium M, acpi_idle. See https://bugzilla.kernel.org/show_bug.cgi?id=22332 if interested.

[Bug 12788] ~40,000 wakeups/sec after suspend/resume with Tickless kernel; idle=nomwait or max_cstate=1 helps

https://bugzilla.kernel.org/show_bug.cgi?id=12788 --- Comment #61 from James Ettle 2010-11-05 18:04:10 --- Created an attachment (id=36252) --> (https://bugzilla.kernel.org/attachment.cgi?id=36252) lspci -vxxx on 2.6.36 after resume -- Configure bugmail: https://bugzilla.kernel.org/userp

[Bug 12788] ~40,000 wakeups/sec after suspend/resume with Tickless kernel; idle=nomwait or max_cstate=1 helps

https://bugzilla.kernel.org/show_bug.cgi?id=12788 --- Comment #60 from James Ettle 2010-11-05 18:03:36 --- Created an attachment (id=36242) --> (https://bugzilla.kernel.org/attachment.cgi?id=36242) lspci -vxxx on 2.6.36 before suspend -- Configure bugmail: https://bugzilla.kernel.org/use

[Bug 12788] ~40,000 wakeups/sec after suspend/resume with Tickless kernel; idle=nomwait or max_cstate=1 helps

https://bugzilla.kernel.org/show_bug.cgi?id=12788 --- Comment #59 from James Ettle 2010-11-05 18:03:04 --- Created an attachment (id=36232) --> (https://bugzilla.kernel.org/attachment.cgi?id=36232) dmesg on 2.6.36 after resume -- Configure bugmail: https://bugzilla.kernel.org/userprefs.c

[Bug 12788] ~40,000 wakeups/sec after suspend/resume with Tickless kernel; idle=nomwait or max_cstate=1 helps

https://bugzilla.kernel.org/show_bug.cgi?id=12788 --- Comment #58 from James Ettle 2010-11-05 18:02:40 --- Created an attachment (id=36222) --> (https://bugzilla.kernel.org/attachment.cgi?id=36222) dmesg on 2.6.36 before suspend -- Configure bugmail: https://bugzilla.kernel.org/userprefs

[Bug 12788] ~40,000 wakeups/sec after suspend/resume with Tickless kernel; idle=nomwait or max_cstate=1 helps

https://bugzilla.kernel.org/show_bug.cgi?id=12788 --- Comment #57 from James Ettle 2010-11-05 18:02:09 --- Yakui, have the data you requested. Tried it on a 2.6.36 kernel with no cstate limit; core issue remains, around 300 wups before, 1 after. Have attached dmesgs and lspcis below.(In

[Bug 12788] ~40,000 wakeups/sec after suspend/resume with Tickless kernel; idle=nomwait or max_cstate=1 helps

https://bugzilla.kernel.org/show_bug.cgi?id=12788 Vitus Jensen changed: What|Removed |Added CC||vjen...@gmx.de -- Configure bugmail:

[Bug 12788] ~40,000 wakeups/sec after suspend/resume with Tickless kernel; idle=nomwait or max_cstate=1 helps

https://bugzilla.kernel.org/show_bug.cgi?id=12788 --- Comment #56 from ykzhao 2010-11-02 07:33:34 --- HI, James Sorry for the late response. It seems that the issue still exists. In fact from the powertop data it seems that there is no change about interrupt number. But it is very weird

[Bug 12788] ~40,000 wakeups/sec after suspend/resume with Tickless kernel; idle=nomwait or max_cstate=1 helps

https://bugzilla.kernel.org/show_bug.cgi?id=12788 --- Comment #55 from James Ettle 2010-09-27 09:30:09 --- Sorry about the delay. I've just tried with a build of kernel 2.6.35.4 (OK, not the absolute latest...) and the problem still remains, albeit with a reduced number of stealth wakeups.

[Bug 12788] ~40,000 wakeups/sec after suspend/resume with Tickless kernel; idle=nomwait or max_cstate=1 helps

https://bugzilla.kernel.org/show_bug.cgi?id=12788 --- Comment #54 from Zhang Rui 2010-09-27 00:23:57 --- any update? :) -- Configure bugmail: https://bugzilla.kernel.org/userprefs.cgi?tab=email --- You are receiving this mail because: --- You are on the CC list for the bug. --

[Bug 12788] ~40,000 wakeups/sec after suspend/resume with Tickless kernel; idle=nomwait or max_cstate=1 helps

https://bugzilla.kernel.org/show_bug.cgi?id=12788 --- Comment #53 from James Ettle 2010-09-03 05:55:18 --- (In reply to comment #52) > ping James Don't worry, I've not forgotten, just been rather busy of late --- I'll get on to testing with the latest kernel as soon as I can. -- Configur

[Bug 12788] ~40,000 wakeups/sec after suspend/resume with Tickless kernel; idle=nomwait or max_cstate=1 helps

https://bugzilla.kernel.org/show_bug.cgi?id=12788 Zhang Rui changed: What|Removed |Added Status|ASSIGNED|NEEDINFO --- Comment #52 from Zhang Rui

[Bug 12788] ~40,000 wakeups/sec after suspend/resume with Tickless kernel; idle=nomwait or max_cstate=1 helps

https://bugzilla.kernel.org/show_bug.cgi?id=12788 --- Comment #51 from ykzhao 2010-08-31 00:40:49 --- HI, James Will you please try the latest kernel and see whether the issue still an be reproduced? Thanks. -- Configure bugmail: https://bugzilla.kernel.org/userprefs.cgi?tab=email ---

[Bug 12788] ~40,000 wakeups/sec after suspend/resume with Tickless kernel; idle=nomwait or max_cstate=1 helps

https://bugzilla.kernel.org/show_bug.cgi?id=12788 Zhang Rui changed: What|Removed |Added AssignedTo|venkatesh.pallip...@gmail.c |yakui.z...@intel.com |om

[Bug 12788] ~40,000 wakeups/sec after suspend/resume with Tickless kernel; idle=nomwait or max_cstate=1 helps

http://bugzilla.kernel.org/show_bug.cgi?id=12788 --- Comment #50 from Thomas Renninger 2009-12-11 10:16:03 --- Thanks for the test. I don't know enough to go on further, having a lapic suspend function for one CPU only looked bogus and I hoped the machine behaved better without any or with e

[Bug 12788] ~40,000 wakeups/sec after suspend/resume with Tickless kernel; idle=nomwait or max_cstate=1 helps

http://bugzilla.kernel.org/show_bug.cgi?id=12788 James Ettle changed: What|Removed |Added Summary|~40,000 wakeups/sec after |~40,000 wakeups/sec after

[Bug 12788] ~40,000 wakeups/sec after suspend/resume with Tickless kernel; idle=nomwait helps

http://bugzilla.kernel.org/show_bug.cgi?id=12788 --- Comment #48 from James Ettle 2009-12-09 21:36:21 --- OK, I tried the patch with apic=verbose. Without disable_apic_suspend_hooks, it didn't suspend properly (hard disc turned off, then it froze); with disable_apic_suspend_hooks, it suspen

[Bug 12788] ~40,000 wakeups/sec after suspend/resume with Tickless kernel; idle=nomwait helps

http://bugzilla.kernel.org/show_bug.cgi?id=12788 James Ettle changed: What|Removed |Added Summary|~40,000 wakeups/sec after |~40,000 wakeups/sec after

[Bug 12788] ~40,000 wakeups/sec after suspend/resume with Tickless kernel

http://bugzilla.kernel.org/show_bug.cgi?id=12788 --- Comment #46 from Thomas Renninger 2009-12-07 12:04:31 --- And please attach /proc/interrupts if it still does not work (with some thousand irqs logged in there). -- Configure bugmail: http://bugzilla.kernel.org/userprefs.cgi?tab=email --

[Bug 12788] ~40,000 wakeups/sec after suspend/resume with Tickless kernel

http://bugzilla.kernel.org/show_bug.cgi?id=12788 --- Comment #45 from Thomas Renninger 2009-12-07 12:00:42 --- Created an attachment (id=24074) --> (http://bugzilla.kernel.org/attachment.cgi?id=24074) Debug patch: Cleanup ups and enhancements for apic suspend routines -- Configure bugmail

[Bug 12788] ~40,000 wakeups/sec after suspend/resume with Tickless kernel

http://bugzilla.kernel.org/show_bug.cgi?id=12788 --- Comment #44 from Thomas Renninger 2009-12-07 11:58:55 --- Quick summary: If you use idle=nomwait the irq activity drops from 4 to 800 which you say is high, but this looks pretty normal if you still have some things running. So the roo

[Bug 12788] ~40,000 wakeups/sec after suspend/resume with Tickless kernel

http://bugzilla.kernel.org/show_bug.cgi?id=12788 --- Comment #43 from James Ettle 2009-12-04 19:18:55 --- No improvement with 2.6.32 (kernel-2.6.32-1.fc13.x86_64). -- Configure bugmail: http://bugzilla.kernel.org/userprefs.cgi?tab=email --- You are receiving this mail because: ---

[Bug 12788] ~40,000 wakeups/sec after suspend/resume with Tickless kernel

http://bugzilla.kernel.org/show_bug.cgi?id=12788 --- Comment #42 from ykzhao 2009-11-18 07:20:39 --- Hi, James Will you please try the latest kernel(2.6.32-rc7) and see whether the issue still exists? Thanks. -- Configure bugmail: http://bugzilla.kernel.org/userprefs.cgi?tab=email --

[Bug 12788] ~40,000 wakeups/sec after suspend/resume with Tickless kernel

http://bugzilla.kernel.org/show_bug.cgi?id=12788 --- Comment #41 from James Ettle 2009-08-13 21:38:41 --- No improvement with kernel 2.6.30.4, still ~40,000 wups after resume. Still none the wiser as to where they are coming from... -- Configure bugmail: http://bugzilla.kernel.org/userpre

[Bug 12788] ~40,000 wakeups/sec after suspend/resume with Tickless kernel

http://bugzilla.kernel.org/show_bug.cgi?id=12788 Zhang Rui changed: What|Removed |Added Status|NEEDINFO|ASSIGNED AssignedTo|cpuf...@vger.kern

[Bug 12788] ~40,000 wakeups/sec after suspend/resume with Tickless kernel

http://bugzilla.kernel.org/show_bug.cgi?id=12788 --- Comment #40 from James Ettle 2009-05-26 20:00:31 --- No improvement with kernel-2.6.30-0.91.rc7.git1.fc12.x86_64. -- Configure bugmail: http://bugzilla.kernel.org/userprefs.cgi?tab=email --- You are receiving this mail because:

[Bug 12788] ~40,000 wakeups/sec after suspend/resume with Tickless kernel

http://bugzilla.kernel.org/show_bug.cgi?id=12788 James Ettle changed: What|Removed |Added Summary|After suspend/resume with |~40,000 wakeups/sec after