Finally i've made some experiments with unloading modules. Results are
insane. Unloading nearly all does not changed amount of strange wakeups...
Lsmod - modules that cannot be removed (os booted from encrypted lvm on
sata disk):
Module Size Used by
cpufreq_stats 7104 0
binfmt_misc 12808 1
ipv6 267780 10
acpi_cpufreq 10796 2
freq_table 5536 2 cpufreq_stats,acpi_cpufreq
ext3 136712 0
jbd 48404 1 ext3
mbcache 9600 1 ext3
reiserfs 239616 1
aes_i586 33536 2
cbc 5504 1
blkcipher 8324 1 cbc
sd_mod 30720 2
ahci 28420 1
libata 159344 1 ahci
scsi_mod 151436 2 sd_mod,libata
dm_crypt 15364 1
dm_mod 62660 8 dm_crypt
processor 36872 3 acpi_cpufreq
fuse 50580 1
powertop:
PowerTOP 1.9 (C) 2007 Intel Corporation
Collecting data for 15 seconds
Cn Avg residency
C0 (cpu running) (48.3%)
C1 0.0ms ( 0.0%)
C2 0.0ms (51.7%)
C3 0.0ms ( 0.0%)
P-states (frequencies)
2.41 Ghz 0.0%
2.40 Ghz 0.0%
2.00 Ghz 0.0%
800 Mhz 100.0%
Wakeups-from-idle per second : 36761.6 interval: 15.0s
Power usage (ACPI estimate): 26.2W (1.4 hours)
Top causes for wakeups:
31.8% ( 0.5) <kernel module> : neigh_table_init_no_netlink
(neigh_periodic_timer)
27.3% ( 0.4) <kernel core> : neigh_table_init_no_netlink
(neigh_periodic_timer)
9.1% ( 0.1) <interrupt> : ahci
9.1% ( 0.1) <kernel IPI> : Rescheduling interrupts
9.1% ( 0.1) kcryptd : blk_plug_device (blk_unplug_timeout)
4.5% ( 0.1) <interrupt> : PS/2 keyboard/mouse/touchpad
4.5% ( 0.1) <kernel core> : page_writeback_init (wb_timer_fn)
4.5% ( 0.1) sm-notify : do_journal_end (delayed_work_timer_fn)
26 W on idle system without Xorg... wired is not enough...
Of course it's the same kernel - 2.6.24-12-generic
i've noticed strange thing yesterday. After resume from sleep and
unloading some modules mentioned here I was able to get into C3. Fan
turned off some times (which has never happened before on linux), amount
of wakeups was about 200~300 during normal work what is accpetable for me.
After reboot everything changed into old 40k wakeups per second.
barteq
_______________________________________________
Power mailing list
[email protected]
http://www.bughost.org/mailman/listinfo/power
