http://bugzilla.kernel.org/show_bug.cgi?id=10196
------- Comment #31 from [EMAIL PROTECTED] 2008-04-25 06:49 ------- I did some more bisection. Unfortunately I didn't pin down the commit responsible. [It ended with a run of 8 good revisions, which is somewhat improbable. It blamed v2.6.23-2600-g117494a, but checking out that commit and rebuilding it generated a "good" kernel]. Now I still had the "bad" kernel, as "2.6.23-g117494a.old". There was one significant difference between "good" and "bad"; at some point during the bisection CONFIG_HIGH_RES_TIMERS had become unset (likely a mistake on my part). I rebuilt again to be sure (to exclude a couple of other config changes which had crept in). v2.6.23-2600-g117494a with a certain config and CONFIG_HIGH_RES_TIMERS=y is fine. But v2.6.23-2600-g117494a with the exact same config except CONFIG_HIGH_RES_TIMERS *unset* is broken. And the same is true of my 2.6.25; it was only broken because I'd built it with a particular configuration. I think the problem is that I left CONFIG_HIGH_RES_TIMERS unset but selected CONFIG_NO_HZ=y. If I deselect CONFIG_NO_HZ then there's no problem. CONFIG_NO_HZ was added to x86_64 for 2.6.24, disabled by default, so this isn't a basic regression; it looks more like an unusual new configuration that didn't get hibernation tested. I don't have an alternative x86_64 machine, so it would be great if someone could reproduce on another machine using the "broken" 2.6.25 config I attached earlier. BTW the broken kernel hung after hibernation. It didn't show any calltraces. Though I did notice processes apparently soft-locked in softirqs. (I may not have waited long enough for the lockup detection to trigger). When I resumed, "kicker" (kde panel) was at "9999% CPU" according to top, and 99.9% of CPU time was in state "si" - which I now assume means soft-interrupt. I killed it, and after maybe 30 seconds it went away and I could restart it without problems. I also noticed something new. I started "konqueror" to try and report this. After another half-minute or so, konqueror locked up in exactly the same way as kicker did. So the lockup can affect processes started after resuming from hibernation. -- Configure bugmail: http://bugzilla.kernel.org/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. ------------------------------------------------------------------------- This SF.net email is sponsored by the 2008 JavaOne(SM) Conference Don't miss this year's exciting event. There's still time to save $100. Use priority code J8TL2D2. http://ad.doubleclick.net/clk;198757673;13503038;p?http://java.sun.com/javaone _______________________________________________ acpi-bugzilla mailing list acpi-bugzilla@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/acpi-bugzilla
