[Mactel-linux-devel] [Fwd: Re: C4 state - Status report]

[Thomas M. <tho...@we...>]

Sven Anders schrieb:

> Nicolas Boichat schrieb:
>   
>> Hi,
>>
>> Wow, impressive report, that's really unfortunate and frustrating it
>> doesn't really helps us to improve power management on Linux...
>>
>>     
>>>> The following hopes I have:
>>>>
>>>>  1) The auto USB suspend framework will solve the too-many interrupts problems.
>>>>         
>> Have you tried this kernel option?
>> CONFIG_USB_SUSPEND: USB selective suspend/resume and wakeup (EXPERIMENTAL)
>>     
>
> What I meant was the new "autosuspend" parameter of the "usbcore":
>
> anders@applecar:~$ modinfo usbcore
> filename:       /lib/modules/2.6.21-rc6-mm1-mactel/kernel/drivers/usb/core/usbcore.ko
> license:        GPL
> vermagic:       2.6.21-rc6-mm1-mactel SMP preempt mod_unload CORE2
> parm:           usbfs_snoop:true to log all usbfs traffic (bool)
> parm:           use_both_schemes:try the other device initialization scheme if the first one fails (bool)
> parm:           old_scheme_first:start with the old device initialization scheme (bool)
> parm:           blinkenlights:true to cycle leds on hubs (bool)
> parm:           autosuspend:default autosuspend delay (int)
>
> The last one here - I'm currently testing is...
>
>   
>> Looks interesting... You might be able to disable some devices and see
>> if it improves power consumption...
>>
>>     
>>>>  2) The dyn-ticks patches will allow the processor to sleep longer.
>>>>         
>> At least in the current version, it doesn't help at all unfortunately.
>>     
Why not?

Additionally i encountered this, look at the number of interrupts:

procs -----------memory---------- ---swap-- -----io---- -system--
----cpu----
 r  b   swpd   free   buff  cache   si   so    bi    bo   in   cs us sy
id wa

  0  0      0 315128  66764 360508    0    0     0     0   49  109  4  0
97  0
 0  0      0 315252  66764 360508    0    0     0     0   59 2095  4  1
96  0
 0  0      0 315252  66764 360508    0    0     0     0   44   99  4  0
97  0
 0  0      0 315252  66764 360508    0    0     0     0   64  147  3  0
97  0
 0  0      0 315252  66764 360508    0    0     0     0   42   85  4  0
96  0
 0  0      0 315252  66764 360508    0    0     0     0   69  175  3  0
97  0
 0  0      0 315252  66764 360508    0    0     0     0   52 2082  4  0
96  0
 0  0      0 315252  66764 360508    0    0     0     0   89  217  4  0
97  0
 0  0      0 315252  66772 360500    0    0     0    25   52  116  4  0
84 13
 0  0      0 315252  66772 360508    0    0     0     0   64  151  3  0
97  0
 0  0      0 315252  66772 360508    0    0     0     0   38   84  4  0
96  0
 0  0      0 315252  66772 360508    0    0     0     0   63 2089  3  0
96  0
 0  0      0 315252  66780 360500    0    0     0    36   54  115  4  0
96  1
 0  0      0 315252  66780 360500    0    0     0    27   69  149  4  0
97  0
 0  0      0 315252  66780 360500    0    0     0     0   46  109  4  0
97  0
 0  0      0 315252  66780 360500    0    0     0     0   68  148  4  0
97  0
 0  0      0 315252  66780 360500    0    0     0     0   39 2033  4  1
96  0
 0  0      0 315252  66788 360508    0    0     0    36  180  411  4  0
96  1
 0  0      0 315252  66788 360508    0    0     0     0   43  109  3  0
97  0
 0  0      0 315252  66788 360508    0    0     0     0   63  148  4  0
96  0
 0  0      0 315252  66788 360508    0    0     0     0   44  113  4  0
97  0
 0  0      0 315252  66788 360508    0    0     0     0   65 2097  4  1
96  0
 0  0      0 315252  66788 360508    0    0     0     0   49  103  3  0
97  0
 0  0      0 315252  66788 360508    0    0     0     0   61  143  4  0
96  0
 0  0      0 315252  66788 360508    0    0     0     0   46  111  4  0
97  0
 0  0      0 315252  66788 360508    0    0     0     0   63  144  4  0
97  0
 0  0      0 315252  66788 360508    0    0     0     0   41 2019  4  0
97  0
 0  0      0 315252  66788 360508    0    0     0     0   66  169  3  0
96  0
 0  0      0 315252  66788 360508    0    0     0     0   77  245 11  0
89  0

Here i first used the touchpad:
 0  0      0 315236  66788 360508    0    0     0     0  195  202  5  0
95  0
 0  0      0 315236  66788 360508    0    0     0     0  170  150  5  0
95  0
procs -----------memory---------- ---swap-- -----io---- -system--
----cpu----
 r  b   swpd   free   buff  cache   si   so    bi    bo   in   cs us sy
id wa
 1  0      0 315236  66788 360508    0    0     0     0  191 2601  5  0
95  0
 0  0      0 315236  66788 360508    0    0     0     0  175  139  5  0
95  0
 0  0      0 315268  66788 360508    0    0     0     0  200  200  5  1
95  0
 0  0      0 315268  66788 360508    0    0     0     0  170  128  5  0
95  0
 0  0      0 315268  66788 360508    0    0     0     0  195  191  5  0
95  0
 0  0      0 315268  66788 360508    0    0     0     0  171 2535  5  0
95  0
 0  0      0 315284  66800 360496    0    0     4    16  309  458  5  0
82 13
 0  0      0 315284  66800 360496    0    0     0     0  173  128  5  0
95  0
 0  0      0 315284  66800 360508    0    0     0     0  197  196  5  0
95  0
 0  0      0 315284  66800 360508    0    0     0     0  174  128  5  1
95  0
 0  0      0 315292  66800 360508    0    0     0     0  198 2646  5  0
95  0
 0  0      0 315292  66800 360508    0    0     0     0  175  142  5  0
95  0
 0  0      0 315292  66808 360500    0    0     0    36  208  247  4  0
92  3
 0  0      0 315292  66808 360500    0    0     0     0  168  122  5  0
95  0
 0  0      0 315292  66808 360508    0    0     0     0  193  216  5  0
95  0

And from here on the number of interrupts stays this high. only after
doing rmmod uhci_hcd && modprobe uhci_hcd the number of interrupts goes
back to below 100 per second. maybe the touchpad driver is missing a
"deactive function".

with kind regards
thomas