Bug 14045 - Sporadic hrtimer / HPET initialization problem: "interrupt too slow", system becomes unresponsive/sluggish
Summary: Sporadic hrtimer / HPET initialization problem: "interrupt too slow", system ...
Status: CLOSED OBSOLETE
Alias: None
Product: Timers
Classification: Unclassified
Component: Realtime Clock (show other bugs)
Hardware: All Linux
: P1 normal
Assignee: timers_realtime-clock
URL:
Keywords:
Depends on:
Blocks:
 
Reported: 2009-08-23 20:16 UTC by Ulrich
Modified: 2012-06-13 15:33 UTC (History)
1 user (show)

See Also:
Kernel Version: 2.6.31-rc7
Subsystem:
Regression: No
Bisected commit-id:


Attachments
Output of "dmesg" when error occurs (39.84 KB, text/plain)
2009-08-23 20:16 UTC, Ulrich
Details
Output of "dmesg" when system boots normally (39.82 KB, text/plain)
2009-08-23 20:17 UTC, Ulrich
Details
My .config for 2.6.31-rc7 (83.25 KB, text/plain)
2009-08-23 20:19 UTC, Ulrich
Details
Output of "lspci -vv" (29.90 KB, text/plain)
2009-08-23 20:20 UTC, Ulrich
Details
Output of cat /proc/cpuinfo (1.48 KB, text/plain)
2009-08-23 20:21 UTC, Ulrich
Details

Description Ulrich 2009-08-23 20:16:09 UTC
Created attachment 22819 [details]
Output of "dmesg" when error occurs

Hi,


with recent mainline kernels, including 2.6.31-rc7, the system sporadicly (approx. once every five or so system starts) boots into a state with limited system responsiveness (e.g. second-long latency to console keyboard input, gaps in video playback under X).                  


Everytime (and only) when this occurs, dmesg displays the following line:

"hrtimer: interrupt too slow, forcing clock min delta to [...] ns"


Additional detail: when the system has booted with the error, system responsiveness is noticeably affected by mouse movements, even on the console without X11 running. E.g. during system shutdown (which takes more than twice the time than normal), the mere act of moving the mouse considerably speeds up the system and the further shutdown process. This is always reproducible.

(I'm guessing this has to do with the associated hardware interrupts)


I have tickless system enabled in .config and HPET enabled in system's BIOS.


My system:

Linux 2.6.31-rc7 compiled for 64 bit (.config attached)

AMD Athlon 64 X2 4400+ (/proc/cpuinfo attached)

Mainboard with NVidia nForce 630a (lspci -vv attached)

dmesg attached, one time with and one time without the error

Please feel free to ask if you need further information.
Comment 1 Ulrich 2009-08-23 20:17:07 UTC
Created attachment 22820 [details]
Output of "dmesg" when system boots normally
Comment 2 Ulrich 2009-08-23 20:19:41 UTC
Created attachment 22821 [details]
My .config for 2.6.31-rc7
Comment 3 Ulrich 2009-08-23 20:20:45 UTC
Created attachment 22822 [details]
Output of "lspci -vv"
Comment 4 Ulrich 2009-08-23 20:21:21 UTC
Created attachment 22823 [details]
Output of cat /proc/cpuinfo
Comment 5 Alan 2012-06-13 15:33:01 UTC
If this is seen with modern kernels please re-open the bug

Note You need to log in before you can comment on or make changes to this bug.