Distribution: Debian Sarge Hardware Environment: Acer TM382 (notebook i386, PM) Software Environment: Linux Problem Description: I get syslkog filled up w/ msg like below; they are triggered by cpufreqd checking /proc/acpi/battery/BAT0/state; may happen also by hand: normally it would read: root# cat /proc/acpi/battery/BAT0/state present: yes capacity state: ok charging state: charged present rate: unknown remaining capacity: 4400 mAh present voltage: 11100 mV but may happen (1-2 times/min on a 1s loop) also: root# cat /proc/acpi/battery/BAT0/state present: yes ERROR: Unable to read battery status when this happens, syslog record msgs here at bottom. Apparently cpufreqd does its job, ie freq tracks load. I see an annoying and unusual ~1s cyclic cpu peak activity donnow why - no problem except playing mmedia, and hiccup on fast scrolling. W/ and w/out cpufreqd. I've started playing on this machine w/ kernel 2.6.10-2 Debian, then compiled myself current kernel, which is patched with latest acpi-20050408-2.6.11.diff.bz2. Patches for swsusp2 are in place, and S4 works. /proc/acpi/fan/ is void root# cat /proc/acpi/info version: 20050408 root# modinfo battery filename: /lib/modules/2.6.11.7-ck1-bs-ss-pm.6/kernel/drivers/acpi/battery.ko author: Paul Diefenbaugh description: ACPI Battery Driver license: GPL vermagic: 2.6.11.7-ck1-bs-ss-pm.6 preempt PENTIUMM gcc-3.3 depends: I'll post any other info upon request. Steps to reproduce: run cpufreqd; root# tail -500 /var/log/syslog|grep ACPI-|sort -k7|uniq -f7 Apr 25 16:06:19 ngb kernel: ACPI-0304: *** Error: Looking up [ECCT] in namespace, AE_ALREADY_EXISTS Apr 25 16:06:19 ngb kernel: ACPI-0667: *** Error: Looking up [ECCT] in namespace, AE_NOT_FOUND Apr 25 16:06:19 ngb kernel: ACPI-1172: *** Error: Method execution failed [\_SB_.AC__._PSR] (Node deec3740), AE_NOT_FOUND Apr 25 17:03:28 ngb kernel: ACPI-1172: *** Error: Method execution failed [\_SB_.BAT0.UPBS] (Node deec3860), AE_NOT_FOUND Apr 25 16:06:16 ngb kernel: ACPI-1172: *** Error: Method execution failed [\_SB_.BAT0._BIF] (Node deec2900), AE_NOT_FOUND Apr 25 17:03:28 ngb kernel: ACPI-1172: *** Error: Method execution failed [\_SB_.BAT0._BST] (Node deec28e0), AE_NOT_FOUND Apr 25 16:06:19 ngb kernel: ACPI-1172: *** Error: Method execution failed [\_SB_.PCI0.LPCB.EC__.COMD] (Node deec5bc0), AE_ALREADY_EXISTS Apr 25 16:06:16 ngb kernel: ACPI-1172: *** Error: Method execution failed [\_SB_.PCI0.LPCB.EC__.COMD] (Node deec5bc0), AE_NOT_FOUND
Could you please try latest kernel? It includes some ACPI interpreter fixes.
yay! 2.6.12-rc3 solved the ACPI-* syslog msg problem indeed, thanks. Had to hack a bit to get swsusp2 2.1.8.5 patches in, but now acpi + cpufreqd seems to work ok, and hibernate too (except for a bunch of incompatible modules). Still don't know why I get cpu usage spikes at ~1s freq - that breaks mmedia play. Don't know if it's acpi-related, but this systems w/ these kernels (CPU=P4 of course) don't show such problem in an Toshiba w/ P4 notebook. Ho can I best track down this problem?
closing this bug as it seems to be solved -- the usage spikes should be put into a different bug if it is indeed a kernel bug.