maemo.org - Talk

maemo.org - Talk (https://talk.maemo.org/index.php)
-   Maemo 5 / Fremantle (https://talk.maemo.org/forumdisplay.php?f=40)
-   -   [Announce] kernel-power stable v49 in Extras-Testing (https://talk.maemo.org/showthread.php?t=71879)

thedarkus 2012-02-03 09:23

Re: [Announce] kernel-power stable v49 in Extras-Testing
 
2 Attachment(s)
Quote:

Originally Posted by freemangordon (Post 1159850)
Well, that is useless. Please, do it again with all applications closed, i.e. no imageviewer or Calendar running. Just follow the steps bellow:

1. reboot
2. wait 10 minutes device to settle down.
3. open xterm
4. do sudo gainroot
5. start powertop and immediately lock the device
6. wait about 90 secs.
7. post powertop result here.

Attached both lsmod.log (vi_ asked me for it) and powertop.log, saved after a reboot (i also closed calendar and mediaplayer widget before rebooting)

d!

vi_ 2012-02-03 09:25

Re: [Announce] kernel-power stable v49 in Extras-Testing
 
Actually you want to put the phone in offline mode, shutdown, pull the battery, power up, stay in offline mode.

Wait 4 minutes.

do:

powertop > output; powertop >> output; powertop >> output; powertop >> output

immediately lock teh phone, wait 3.5 minutes then paste contents of file 'output'

freemangordon 2012-02-03 09:32

Re: [Announce] kernel-power stable v49 in Extras-Testing
 
@vi_ - any idea what those sleeps are? Never seen powertop output like that, i.e. with 14 sleep processes running, seems like another "optimization" script :(

thedarkus 2012-02-03 09:39

Re: [Announce] kernel-power stable v49 in Extras-Testing
 
Ok, i'm doing what vi_ asked for.

freemangordon, i can assure you that i DON'T have installed any kind of optimization script. Nothing at all.

I have simply reflashed, installed pk49, installed cssu updates, installed rootsh

Then i opened the xterminal and wrote:

sudo gainroot
kernel-config load ulv
kernel-config limits 250 805
kernel-config save 805ulv
kernel-config default 805ulv

This is all

After that i started to install app from repository, nothing special, something like kmplayer, advanced power monitor, advanced interface switcher, dropn900, pyradio, koffice, the usual stuff that i daily use since i bought n900 almost 2 yrs ago


d!

thedarkus 2012-02-03 09:40

Re: [Announce] kernel-power stable v49 in Extras-Testing
 
double post, sorry


d!

freemangordon 2012-02-03 09:51

Re: [Announce] kernel-power stable v49 in Extras-Testing
 
Quote:

Originally Posted by thedarkus (Post 1159867)
Ok, i'm doing what vi_ asked for.

freemangordon, i can assure you that i DON'T have installed any kind of optimization script. Nothing at all.

I have simply reflashed, installed pk49, installed cssu updates, installed rootsh

Then i opened the xterminal and wrote:

sudo gainroot
kernel-config load ulv
kernel-config limits 250 805
kernel-config save 805ulv
kernel-config default 805ulv

This is all

After that i started to install app from repository, nothing special, something like kmplayer, advanced power monitor, advanced interface switcher, dropn900, pyradio, koffice, the usual stuff that i daily use since i bought n900 almost 2 yrs ago


d!

Well, your initial statement was that you have a freshly flashed device with only KP49 installed. Now it turns that you actually have a lot of stuff on it besides KP49. And because you are the first one to report battery life got worse by installing KP49, it is obvious that some software installed on your device is incompatible with KP49. Noone is saying that it is your fault, so take it easy. And there were several reports that some of those CPU/memory monitoring tools are causing troubles with KP49.

thedarkus 2012-02-03 10:00

Re: [Announce] kernel-power stable v49 in Extras-Testing
 
1 Attachment(s)
Quote:

Originally Posted by freemangordon (Post 1159873)
Well, your initial statement was that you have a freshly flashed device with only KP49 installed. Now it turns that you actually have a lot of stuff on it besides KP49. And because you are the first one to report battery life got worse by installing KP49, it is obvious that some software installed on your device is incompatible with KP49. Noone is saying that it is your fault, so take it easy. And there were several reports that some of those CPU/memory monitoring tools are causing troubles with KP49.

I'm taking it easy, freemangordon :)

I think this is useful to find the cause of my problem. I can give you the exact list of the application that i have installed. A lot of stuff, but nothing related to tweaking or modding of the system, just only what that came from this community (like cssu) or general purpose application, like media player, opera, cuteTube-qml and stuff like that


Btw, here is the powertop output

d!

freemangordon 2012-02-03 10:48

Re: [Announce] kernel-power stable v49 in Extras-Testing
 
Quote:

Originally Posted by thedarkus (Post 1159874)
I'm taking it easy, freemangordon :)

Good to hear, keep it that way :)

Quote:


I think this is useful to find the cause of my problem. I can give you the exact list of the application that i have installed. A lot of stuff, but nothing related to tweaking or modding of the system, just only what that came from this community (like cssu) or general purpose application, like media player, opera, cuteTube-qml and stuff like that


Btw, here is the powertop output

d!
Well, someone is running those sleeps like mad, I would recommend to remove advanced power, qcpufreq and all other monitoring tools (including interface swither if installed) and run powertop (as vi_ recommended) again.

In the meantime please post here the output of:

maemo-list-user-packages

Estel 2012-02-03 11:27

Re: [Announce] kernel-power stable v49 in Extras-Testing
 
thedarkus, You said you've installed some things, then, uninstalled it. Depending, on how good (or bad) postremoval scripts are, it's possible that some crap remains are still on Your device, even though You "formally" uninstaled it.

My suggestion, would be to do full reflash (even eMMC - not related to case, but will give You a "like new" layout), apply things like repartitioning - if You use it, then, install KP, CSSU, rootsh, and nothing more.

Then, and *only* after such step - no other things installed - re-try powertop test as suggested by freemangordon and vi_.

/Estel

ibrakalifa 2012-02-03 11:31

Re: [Announce] kernel-power stable v49 in Extras-Testing
 
how bout random reboot while watch some video??, is that only happened to me??

thedarkus 2012-02-03 11:41

Re: [Announce] kernel-power stable v49 in Extras-Testing
 
Ok guys, i will try all the solution you suggested to me. Btw, now my battery is empty and i'm recharging the device, after that, i wll try first removing interface switcher and advanced power (the only app related to system), posting here the output requested.

If the problem is still unsolved, i will reflash the device.

I really would solve the problem in the first way, 'cos i want to know why this **** happened, i don't wanna encounter the same problem again.


For the moment, a big THANK YOU to all of you guys


d!

vi_ 2012-02-03 11:47

Re: [Announce] kernel-power stable v49 in Extras-Testing
 
Code:

Powertop 1.13.3
Sleeping for 11 seconds before sampling
Collecting data for 30 seconds
Sample interval was 00m 30s 32896us

C#      | Ratio  | Avg/dura | Frequency | Ratio
--------+--------+----------+-----------+--------+
    C0 |  0.1% |          |  1150 MHz |  nan% |
    C1 |  0.0% |    0.2ms |
    C2 |  0.2% |    2.0ms |
    C3 |  0.0% |          |
    C4 |  99.7% | 3741.6ms |

IRQ#    | Activity  | Type          | Name
--------+------------+----------------+---------------------------
    56 |        90 |          INTC | i2c_omap
    37 |        24 |          INTC | gp
    57 |        22 |          INTC | i2c_omap
    11 |        18 |          INTC | prcm

PID#    | Activity  | Name          | Function Entry (Expire)
--------+------------+----------------+---------------------------
      0 |        13 |  <kernel core> | tick_nohz_restart_sched_tick (tick_sched_timer)
    39 |          6D|            awk | cpufreq_governor_dbs (delayed_work_timer_fn)
    760 |          4 |      bme_RX-51 | sys_timer_settime (posix_timer_fn)
      1 |          3 |  <kernel core> | queue_delayed_work (delayed_work_timer_fn)
      0 |          2 |  <kernel core> | hrtimer_start (tick_sched_timer)
    721 |          2 |          dsme | do_nanosleep (hrtimer_wakeup)
    721 |          2 |          dsme | __enqueue_rt_entity (sched_rt_period_timer)
    760 |          2 |      bme_RX-51 | sys_timer_settime (posix_timer_fn)
    66 |          2 |  cifsdnotifyd | schedule_timeout (process_timeout)
    742 |          2D|<kernel module> | queue_delayed_work (delayed_work_timer_fn)
    760 |          2 |      bme_RX-51 | do_nanosleep (hrtimer_wakeup)
    760 |          2 |      bme_RX-51 | schedule_timeout (process_timeout)
    618 |          1 |          mmcqd | queue_delayed_work (delayed_work_timer_fn)
  1124 |          1 |          iphbd | schedule_hrtimeout_range (hrtimer_wakeup)
  3054 |          1 |      powertop | do_nanosleep (hrtimer_wakeup)

Power domain activity breakdown
Domain  | % of time spent in states
--------+---------+---------+---------+---------+----------
usbhost |OFF: 100%|RET:  0%|INA:  0%| ON:  0%| now:(OFF)
    sgx |OFF: 100%|RET:  0%|INA:  0%| ON:  0%| now:(OFF)
    per |OFF:  99%|RET:  0%|INA:  0%| ON:  0%| now:(ON)
    dss |OFF: 100%|RET:  0%|INA:  0%| ON:  0%| now:(OFF)
    cam |OFF: 100%|RET:  0%|INA:  0%| ON:  0%| now:(OFF)
  core |OFF:  99%|RET:  0%|INA:  0%| ON:  0%| now:(ON)
  neon |OFF:  99%|RET:  0%|INA:  0%| ON:  0%| now:(ON)
    mpu |OFF:  99%|RET:  0%|INA:  0%| ON:  0%| now:(ON)
  iva2 |OFF: 100%|RET:  0%|INA:  0%| ON:  0%| now:(OFF)

Clock activity breakdown at end of period
Domain  | Active clocks
--------+---------------+---------------+------------------
  core |          SDRC | HSOTGUSB_IDLE |      OMAPCTRL
        |    MAILBOXES |
  wkup |          GPT1 |      32KSYNC |        GPIO1
        |          WDT1 |          SR1 |          SR2
        |
  ckgen |          CORE |          PERI |          96M
        |          48M |          12M |          54M
        |      EMU_CORE |
    per |        GPIO2 |        GPIO3 |        GPIO4
        |        GPIO5 |        GPIO6 |

Total wakeups  199,  6.6/s | IRQ  154,  5.1/s | Timers  45,  1.5/s
HW wakeups      18,  0.6/s |    Real gp_timers expired  24,  0.8/s

This is my device in offline mode. You will notice the lack of all those 'sleep' things. You have installed some a55 somewhere along the line.

also, output of 'dpkg --get-selections' now plox.

thedarkus 2012-02-03 11:48

Re: [Announce] kernel-power stable v49 in Extras-Testing
 
Quote:

Originally Posted by Estel (Post 1159904)
thedarkus, You said you've installed some things, then, uninstalled it. Depending, on how good (or bad) postremoval scripts are, it's possible that some crap remains are still on Your device, even though You "formally" uninstaled it.


/Estel

The only thing installed and removed until now is qCpuFreq. I want to ask you if enabling the smartreflex function could have caused the problem i encountered



d!

thedarkus 2012-02-03 11:50

Re: [Announce] kernel-power stable v49 in Extras-Testing
 
Quote:

Originally Posted by vi_ (Post 1159913)
Code:

Powertop 1.13.3
Sleeping for 11 seconds before sampling
Collecting data for 30 seconds
Sample interval was 00m 30s 32896us

C#      | Ratio  | Avg/dura | Frequency | Ratio
--------+--------+----------+-----------+--------+
    C0 |  0.1% |          |  1150 MHz |  nan% |
    C1 |  0.0% |    0.2ms |
    C2 |  0.2% |    2.0ms |
    C3 |  0.0% |          |
    C4 |  99.7% | 3741.6ms |

IRQ#    | Activity  | Type          | Name
--------+------------+----------------+---------------------------
    56 |        90 |          INTC | i2c_omap
    37 |        24 |          INTC | gp
    57 |        22 |          INTC | i2c_omap
    11 |        18 |          INTC | prcm

PID#    | Activity  | Name          | Function Entry (Expire)
--------+------------+----------------+---------------------------
      0 |        13 |  <kernel core> | tick_nohz_restart_sched_tick (tick_sched_timer)
    39 |          6D|            awk | cpufreq_governor_dbs (delayed_work_timer_fn)
    760 |          4 |      bme_RX-51 | sys_timer_settime (posix_timer_fn)
      1 |          3 |  <kernel core> | queue_delayed_work (delayed_work_timer_fn)
      0 |          2 |  <kernel core> | hrtimer_start (tick_sched_timer)
    721 |          2 |          dsme | do_nanosleep (hrtimer_wakeup)
    721 |          2 |          dsme | __enqueue_rt_entity (sched_rt_period_timer)
    760 |          2 |      bme_RX-51 | sys_timer_settime (posix_timer_fn)
    66 |          2 |  cifsdnotifyd | schedule_timeout (process_timeout)
    742 |          2D|<kernel module> | queue_delayed_work (delayed_work_timer_fn)
    760 |          2 |      bme_RX-51 | do_nanosleep (hrtimer_wakeup)
    760 |          2 |      bme_RX-51 | schedule_timeout (process_timeout)
    618 |          1 |          mmcqd | queue_delayed_work (delayed_work_timer_fn)
  1124 |          1 |          iphbd | schedule_hrtimeout_range (hrtimer_wakeup)
  3054 |          1 |      powertop | do_nanosleep (hrtimer_wakeup)

Power domain activity breakdown
Domain  | % of time spent in states
--------+---------+---------+---------+---------+----------
usbhost |OFF: 100%|RET:  0%|INA:  0%| ON:  0%| now:(OFF)
    sgx |OFF: 100%|RET:  0%|INA:  0%| ON:  0%| now:(OFF)
    per |OFF:  99%|RET:  0%|INA:  0%| ON:  0%| now:(ON)
    dss |OFF: 100%|RET:  0%|INA:  0%| ON:  0%| now:(OFF)
    cam |OFF: 100%|RET:  0%|INA:  0%| ON:  0%| now:(OFF)
  core |OFF:  99%|RET:  0%|INA:  0%| ON:  0%| now:(ON)
  neon |OFF:  99%|RET:  0%|INA:  0%| ON:  0%| now:(ON)
    mpu |OFF:  99%|RET:  0%|INA:  0%| ON:  0%| now:(ON)
  iva2 |OFF: 100%|RET:  0%|INA:  0%| ON:  0%| now:(OFF)

Clock activity breakdown at end of period
Domain  | Active clocks
--------+---------------+---------------+------------------
  core |          SDRC | HSOTGUSB_IDLE |      OMAPCTRL
        |    MAILBOXES |
  wkup |          GPT1 |      32KSYNC |        GPIO1
        |          WDT1 |          SR1 |          SR2
        |
  ckgen |          CORE |          PERI |          96M
        |          48M |          12M |          54M
        |      EMU_CORE |
    per |        GPIO2 |        GPIO3 |        GPIO4
        |        GPIO5 |        GPIO6 |

Total wakeups  199,  6.6/s | IRQ  154,  5.1/s | Timers  45,  1.5/s
HW wakeups      18,  0.6/s |    Real gp_timers expired  24,  0.8/s

This is my device in offline mode. You will notice the lack of all those 'sleep' things. You have installed some a55 somewhere along the line.

also, output of 'dpkg --get-selections' now plox.


How can I format the output text in that way? Mine are almost unreadable


d!

vi_ 2012-02-03 11:53

Re: [Announce] kernel-power stable v49 in Extras-Testing
 
Quote:

Originally Posted by thedarkus (Post 1159914)
The only thing installed and removed until now is qCpuFreq. I want to ask you if enabling the smartreflex function could have caused the problem i encountered



d!

No, smart reflex is the single greatest battery saving thing for the n900 ever. It is cpu voltage 'autoscalling' with SR enabled you no longer should be using custom voltage profiles.

Wait you have interface switcher and advanced power installed? I thought you said this was a fresh device! Well herp derp no wonder there is a bunch of hardware polling and r3tarded scripts running in the background.


//edit in the formatting options there is one for 'code'. Actually the best way to do this is to install 'pastebinit' then froim command line type 'pastebinit output', if you are connected to the internet it will automatically upload the contents of the file for you and you just have to paste a url here. No copy/pasting stuff to/from your n900 to the net.

thedarkus 2012-02-03 12:05

Re: [Announce] kernel-power stable v49 in Extras-Testing
 
Quote:

Originally Posted by vi_ (Post 1159917)
No, smart reflex is the single greatest battery saving thing for the n900 ever. It is cpu voltage 'autoscalling' with SR enabled you no longer should be using custom voltage profiles.

So, are you telling me that ith smatreflex enabled i don't need to load a custom voltage profile like LV or ULV or starving? How can i enable smartreflex without using qCpuFreq app?

Quote:

Originally Posted by vi_ (Post 1159917)
Wait you have interface switcher and advanced power installed? I thought you said this was a fresh device! Well herp derp no wonder there is a bunch of hardware polling and r3tarded scripts running in the background.

Yes, i have installed these app. Now they are gone, removed as suggested some post ago. When i said that mine is a fresh and clean device, I said this without knowing that these apps can cause problem like mine ^^


Quote:

Originally Posted by vi_ (Post 1159917)
//edit in the formatting options there is one for 'code'. Actually the best way to do this is to install 'pastebinit' then froim command line type 'pastebinit output', if you are connected to the internet it will automatically upload the contents of the file for you and you just have to paste a url here. No copy/pasting stuff to/from your n900 to the net.

Intersting, i will later, don't want to install any other app until the cause of the problem is discovered :)

d!

vi_ 2012-02-03 12:22

Re: [Announce] kernel-power stable v49 in Extras-Testing
 
pastebinit is a safe program. It is stuff that autoruns and deals with hardware/settings you have to watch.

LITTLE KNOWN FACT!


There is a file called:

Code:

/etc/pmconfig
This contains almost ALL the settings you can possibly want to set WRT CPU speeds 'n sh1t.

(warning if you set some crazy nonsense in pmconfig you could end up with a non booting device (derp), in the event of a no booting device due to pmconfig buggery, use flasher from a PC to restore the nokia kernel, boot, fix it, reinstall power kernel.)

Just edit that file, set max/min CPU speed and set smart reflex to '1'.

This removes the need for kernel-power-settings/qcpufreq/101 other badly written python appz that muck around with your CPU speed/govenor altogether (almost).

I will say this again from kernel power v49 freemangordon has done the impossible and smashed TMO dogma. Now Smart reflex works UP TO 900MHz. Smart reflex saves battery life, the most battery life. If you want to undervolt and save battery power (which you do) enable smart reflex and NEVER load a voltage underclock profile again. XLV, ULV, STARVING, whatever is now dead/depreciated/deceased/removed/cease to be/extinct. DO NOT USE THEM!

thedarkus 2012-02-03 14:09

Re: [Announce] kernel-power stable v49 in Extras-Testing
 
Ok, I'm giving up guys, removed almost every applcation, rebooted after every single removal, did powertop each time without no effect.

I'm waiting for the battery goes fully charged, i will backup some data and i will reflash again -.-'

Meanwhile, guys, pls, can you post me here a step by step guide to edit /etc/pmconfig for having a smartreflex working without loading any custom profile?

d!

thedarkus 2012-02-03 20:23

Re: [Announce] kernel-power stable v49 in Extras-Testing
 
O, refleshed some minutes ago, now i'm installing power kernel, after that i will install CSSU updates. Suggestion to enable smartreflex without using qcpufreq?



d!

vi_ 2012-02-03 21:09

Re: [Announce] kernel-power stable v49 in Extras-Testing
 
Quote:

Originally Posted by thedarkus (Post 1160103)
O, refleshed some minutes ago, now i'm installing power kernel, after that i will install CSSU updates. Suggestion to enable smartreflex without using qcpufreq?



d!

Install rootsh, midnight commander (mc), sudser.

Start a root shell with the command:

Code:

root
start midnight commander with the commmand:

Code:

mc
You will then be in a file manager type interface. Navigate to directory /etc/

Highlight the file 'pmconfig' and press the edit button at the bottom of the screen.

Change smart reflex to '1'. Set your frequencys. DO NOT exceed 900MHz!!

Save the file and quit MC.

reboot.

job done, now post some tits.

thedarkus 2012-02-03 21:30

Re: [Announce] kernel-power stable v49 in Extras-Testing
 
This is what there is inside

enable_off_mode 1
sleep_while_idle 1
sr_vdd1_autocomp 0
sr_vdd2_autocomp 0
clocks_off_while_idle 1
voltage_off_while_idle 1
scaling_governor ondemand
scaling_max_freq 600000
scaling_min_freq 250000
slee_ind 1



Are the fileds that i have to edit the fileds related to vdd1 and vdd2?

To set a max freq of 850 MHz have I to type 850000, right?




Sorry for the stupid questions, guys, i'm a newbie. N900 is my first approach to the linux world and i'm still learning everything



d!

vi_ 2012-02-03 21:53

Re: [Announce] kernel-power stable v49 in Extras-Testing
 
Quote:

Originally Posted by thedarkus (Post 1160140)
This is what there is inside

enable_off_mode 1
sleep_while_idle 1
sr_vdd1_autocomp 0
sr_vdd2_autocomp 0
clocks_off_while_idle 1
voltage_off_while_idle 1
scaling_governor ondemand
scaling_max_freq 600000
scaling_min_freq 250000
slee_ind 1



Are the fileds that i have to edit the fileds related to vdd1 and vdd2?

To set a max freq of 850 MHz have I to type 850000, right?




Sorry for the stupid questions, guys, i'm a newbie. N900 is my first approach to the linux world and i'm still learning everything



d!

For 250-850 with smart reflex do this:


Code:

enable_off_mode 1
sleep_while_idle 1
sr_vdd1_autocomp 1
sr_vdd2_autocomp 1
clocks_off_while_idle 1
voltage_off_while_idle 1
scaling_governor ondemand
scaling_max_freq 850000
scaling_min_freq 250000
slee_ind 1


thedarkus 2012-02-03 22:08

Re: [Announce] kernel-power stable v49 in Extras-Testing
 
Ok, man, it seems all is workin' fine!



d!

corduroysack 2012-02-04 11:33

Re: [Announce] kernel-power stable v49 in Extras-Testing
 
yesterday after reading this thread i followed vi_ advice
http://talk.maemo.org/showpost.php?p...postcount=2736
removed qcputweak and kernel-power-settings (gave up on wonder patches a few flashes ago) and i'm getting some great battery life including usage and standby with power v49. cheers to everyone involved with kernel power and to vi_ for this information :D

thedarkus 2012-02-04 11:54

Re: [Announce] kernel-power stable v49 in Extras-Testing
 
Should I remove kernel power SETTINGS too? I've installed it..

Anyway, is going better for me now!


d!

Estel 2012-02-04 15:21

Re: [Announce] kernel-power stable v49 in Extras-Testing
 
Quote:

Originally Posted by vi_ (Post 1159926)
This contains almost ALL the settings you can possibly want to set WRT CPU speeds 'n sh1t.

Not exactly true - afaik, if You want to edit correlation between CPU speed and max DSP speed, You need kernel-power-settings (see 720p video recording). If I'm wrong, feel free to correct me.

/Estel

woody14619 2012-02-05 03:10

Re: [Announce] kernel-power stable v49 in Extras-Testing
 
Quote:

Originally Posted by Estel (Post 1160447)
Not exactly true - afaik, if You want to edit correlation between CPU speed and max DSP speed, You need kernel-power-settings (see 720p video recording). If I'm wrong, feel free to correct me.

/Estel

AFAIK: Kernel-power-settings is mainly scripts and a command line-tool used to ease maintenance of multiple kernel configurations. It has about 8 built-in configs, and the kernel-config tool that lets you load configs or change them on the fly. You can do everything this package does by editing /etc/pmconfig and rebooting, or by touching /dev or /sys files. The tool just makes it a little easier for us humans.

Do you need it? No. Does it affect anything if it's there (or not)? Not really. It exists only to make user configuration easier.

ibrakalifa 2012-02-05 03:35

Re: [Announce] kernel-power stable v49 in Extras-Testing
 
Quote:

Originally Posted by woody14619 (Post 1160662)
AFAIK: Kernel-power-settings is mainly scripts and a command line-tool used to ease maintenance of multiple kernel configurations. It has about 8 built-in configs, and the kernel-config tool that lets you load configs or change them on the fly. You can do everything this package does by editing /etc/pmconfig and rebooting, or by touching /dev or /sys files. The tool just makes it a little easier for us humans.

Do you need it? No. Does it affect anything if it's there (or not)? Not really. It exists only to make user configuration easier.

random reboot when set up over than 850Mhz, :confused:


example
kernel-config limits 500 850(working fine)

kernel-config limits 500 1000(disaster begin):(

woody14619 2012-02-05 04:33

Re: [Announce] kernel-power stable v49 in Extras-Testing
 
Quote:

Originally Posted by ibrakalifa (Post 1160667)
random reboot when set up over than 850Mhz, :confused:

example
kernel-config limits 500 850(working fine)

kernel-config limits 500 1000(disaster begin):(

Uh? What's that got to do with any of this?

The default N900 is like an oven without a dial, just a big "auto-cook" button. The K## packages provide lots of thin metal cylinders you can turn to change the settings, but most can't grip them well enough to use them directly. The settings package is "the big plastic knob" on the front that makes it easier to use. But if you turn the oven up to twice it's rated value and it catches on fire, it's not the plastic knobs fault.

Some devices simply can't operate at faster speeds. It has nothing to do with the tools that configure the system. It's a hardware limitation. Nokia/TI tested these to work up to 600Mhz. Some reboot at 605Mhz, some reboot at 1200Mhz. If your device explodes randomly going over 850Mhz, then don't set it above that. Done.

praveenchand 2012-02-05 08:34

Re: [Announce] kernel-power stable v49 in Extras-Testing
 
Quote:

Originally Posted by vi_ (Post 1160131)
Install rootsh, midnight commander (mc), sudser.

Start a root shell with the command:

Code:

root
start midnight commander with the commmand:

Code:

mc
You will then be in a file manager type interface. Navigate to directory /etc/

Highlight the file 'pmconfig' and press the edit button at the bottom of the screen.

Change smart reflex to '1'. Set your frequencys. DO NOT exceed 900MHz!!

Save the file and quit MC.

reboot.

job done, now post some tits.

i followed your steps. By doing this will the default profile will change to this, i mean after reboot max freq. will change to 850 MHz? It didn't happen with me.

sulu 2012-02-05 17:40

Re: [Announce] kernel-power stable v49 in Extras-Testing
 
Hi,

I recently upgraded from v42 to v49. My profile, looks like this (which is basically the old "ideal" profile and worked fine under v42) :
Code:

# kernel configuration file generated by /usr/sbin/kernel-config
MINFREQ=500
MAXFREQ=850
FREQS="0:30,90 250:38,180 500:30,360 550:33,400 600:38,430 720:45,430 805:48,430 850:48,500 900:54,500 950:54,500 1000:48,500 1100:72,520 1150:72,520 "
SMARTREFLEX_VDD1=0
SMARTREFLEX_VDD2=0
GOVERNOR=ondemand
IGNORE_NICE_LOAD=1
UP_THRESHOLD=75
SAMPLING_RATE=150000
POWERSAVE_BIAS=0

Please note that I've already removed the 700MHz entry and changed the one for 750MHz to 720MHz.

The problem is that when I load this profile I get the following error messages:
Code:

Nokia-N900-02-8:~# kernel-config load myprofile
loading /home/user/.kernel/myprofile1
sh: write error: Invalid argument
sh: write error: Invalid argument
successfully loaded.

I tracked it down to these two lines in kernel-config:
Code:

    echo $vsel > $pwr/vdd1_opps_vsel                         
    echo $rate > $pwr/dsp_opps_rate

The reason seems to be that $vsel and $rate still contain entries for the 700MHz setting since it's one value more than the files they are written to contain:
Code:

Nokia-N900-02-8:~# cat /sys/power/vdd1_opps_vsel
30 38 48 54 60 60 60 60 60 60 60 72 72
Nokia-N900-02-8:~# cat /sys/power/dsp_opps_rate
90 180 360 400 430 520 520 520 520 520 520 520 520

outputs of the variables in kernel-config just before they are written to the files:
Code:

vsel 30 38 38 30 33 38 45 48 48 54 54 48 72 72
rate 90 180 180 360 400 430 430 430 500 500 500 500 520 520

So it seems like my profile is not the only file I have to edit to remove the 700MHz setting. Can somebody please tell me where else I have to look for it?

majaczek 2012-02-05 18:11

Re: [Announce] kernel-power stable v49 in Extras-Testing
 
i'm not sure sulu, but i believe it's the only file you need to edit.
notice when you remove 700 and 750 freq from the list and add the 720 one, you have exactly one less in frequencies count, so guess what it means and remove correct one from vsel and rate lists...

am i smart enough ? :)

and sorry all (including pali, freemangordon, estel, vi_) about some stupid things i said once in history since i am an computer pro but rather linux noob so there is sometimes things got crazy i said as a result - well i can't guess all linux technicalies so while posing pro i may *sometimes* say very nooob things. And all people which I said sorry to, I hope you don't have me at blackilist/ignore_list, or whatever the forum delivers, yet.

EDIT: my concerns about changing KP versions and multiboot was really dumb, and most of it i could checked myself (there was just old entries invalidated by update). BTW (for pure curiosity though) is it possible install both newer and older version of KP obtainable through multiboot ? (I'm curious if there exist some package maintaining trick or if it is just impossible with same named packages and different versions of them). Of course all my problems regarding my actual multiboot config was easily fixed (it was just to remove entries for older KP and older KBFS from the list since upgrade of packages removed things needed by old entries - what a my derp... - i have written about 4 posts about it some time ago while it was pretty obvious)

sulu 2012-02-05 19:17

Re: [Announce] kernel-power stable v49 in Extras-Testing
 
Quote:

Originally Posted by majaczek (Post 1160936)
notice when you remove 700 and 750 freq from the list and add the 720 one, you have exactly one less in frequencies count, so guess what it means and remove correct one from vsel and rate lists...

My thoughts exactly. ;)
But how do I do that (without adding evil hacks to the kernel-config script)?

vi_ 2012-02-05 19:37

Re: [Announce] kernel-power stable v49 in Extras-Testing
 
Quote:

Originally Posted by sulu (Post 1160955)
My thoughts exactly. ;)
But how do I do that (without adding evil hacks to the kernel-config script)?

Then you need to make a patch for a new version of KP!

FWIW!

I do not use KP settings, I gave up on it a long time ago. There is some kinky bug that appears some times that makes the whole thing spazz out leading to massive battery drain leaving you no option but to power off and pull the battery.

I use 2 scripts:

/etc/X11/Xsession.post/80avoidfreqs:

Code:

#!/bin/sh
    sudo /opt/scripts/avoid_freqs.sh

and

/opt/scripts/avoid_freqs.sh:

Code:

    #!/bin/sh
    ##set some frequencies
    #avoid frequencies
    echo "125000 550000 600000 805000 950000 1000000 1100000 1150000" > /sys/devices/system/cpu/cpu0/cpufreq/ondemand/avoid_frequencies
    #up threshold
    echo 75 > /sys/devices/system/cpu/cpu0/cpufreq/ondemand/up_threshold
    #nice load
    echo "0" > /sys/devices/system/cpu/cpu0/cpufreq/ondemand/ignore_nice_load
   
    ##Set DSP freqs
    echo "90 90 180 360 400 430 520 600 600 600 600 600 600" > /sys/power/dsp_opps_rate
   
    ##FS mods
    #enable camera button
    #echo "cam"
    chmod 666 /sys/devices/platform/gpio-switch/cam_focus/disable
    echo 0 > /sys/devices/platform/gpio-switch/cam_focus/disable
    chmod 444 /sys/devices/platform/gpio-switch/cam_focus/disable
   
    #FM boost/region
    #echo "fmtx"
    chmod 666 /sys/class/i2c-adapter/i2c-2/2-0063/region
    echo 4 > /sys/class/i2c-adapter/i2c-2/2-0063/region
    chmod 644 /sys/class/i2c-adapter/i2c-2/2-0063/region

    #load injection drivers
    /opt/wifi/load.sh
         
    #setup wifi for all channels
    iw reg set JP

    exit

These fill in gaps that are left by pmconfig, it also does some other nice stuff like open wifi channel 13, maximise FMTX power, enable cam focus button through lockscreen to allow skipping tracks while locked with buttin and soon, when kp50 rolls around, disable red LED.

One day I will get around to bodging this into pmconfig, however it seems to run fine right now so uh f**kit.

Estel 2012-02-05 19:49

Re: [Announce] kernel-power stable v49 in Extras-Testing
 
It's quite funny, as I've never ever encountered any problem with kp settings, not to mention battery drain.

/Estel

sulu 2012-02-05 20:28

Re: [Announce] kernel-power stable v49 in Extras-Testing
 
KPv49 works fine for me circa one week now and v42 has been working fine since I bought my N900. I never encountered any battery drain.

It seems like there are spare values for what might have been a 125MHz setting (the 2nd value in $vsel andf $rate doesn't belong there). I tried fixing it by adding a 125MHz setting to the FREQS line in my profile but the error didn't change.

My evil hack for now is to add these two lines to kernel-config just before $vsel and $rate are written to the files to delete the 2nd value of each line:
Code:

vsel=`echo $vsel | sed s/\ [0-9]\\\+//`
rate=`echo $rate | sed s/\ [0-9]\\\+//`

It works for me but it's really just an ugly evil hack, so it doesn't qualify as a patch at all.

Edit:
Can somebody who runs v49 and uses a profile file please post his file so that I can try to find any critical differences?

xes 2012-02-05 22:41

Re: [Announce] kernel-power stable v49 in Extras-Testing
 
@freemangordon
do you have any news about the voltage boost and regulation at 805/850 Mhz?

freemangordon 2012-02-06 07:03

Re: [Announce] kernel-power stable v49 in Extras-Testing
 
Quote:

Originally Posted by xes (Post 1161019)
@freemangordon
do you have any news about the voltage boost and regulation at 805/850 Mhz?

What news do you expect? Something like that? Or I am missing something.

ibrakalifa 2012-02-06 07:35

Re: [Announce] kernel-power stable v49 in Extras-Testing
 
Quote:

Originally Posted by woody14619 (Post 1160679)
Uh? What's that got to do with any of this?

The default N900 is like an oven without a dial, just a big "auto-cook" button. The K## packages provide lots of thin metal cylinders you can turn to change the settings, but most can't grip them well enough to use them directly. The settings package is "the big plastic knob" on the front that makes it easier to use. But if you turn the oven up to twice it's rated value and it catches on fire, it's not the plastic knobs fault.

Some devices simply can't operate at faster speeds. It has nothing to do with the tools that configure the system. It's a hardware limitation. Nokia/TI tested these to work up to 600Mhz. Some reboot at 605Mhz, some reboot at 1200Mhz. If your device explodes randomly going over 850Mhz, then don't set it above that. Done.


when use kp48 i never had trouble like THIS!!! done? what ur mean then?

fw190 2012-02-06 07:41

Re: [Announce] kernel-power stable v49 in Extras-Testing
 
I did some testing. Made a clean reflash and can confirm what was said. If someone wants a fast, and saving battery N900 just launch DSP profile and that is it. Few months ago I've tested battery patch and comparing it with DPS from Freemangordon it just doesn't work.Batterygraph shows that at night the phone nearly doesn't use cpu. With BP it used a lot more. One other important thing. It is good to have a clean install with just the stuff you use. Unused crap made my phone act strange. Further more all the apps for battery stats showing, cpu saving, overclocking etc. didn't help. I can not explain it. Without them and just with DSP loaded the usage at night is close to zero.


All times are GMT. The time now is 15:34.

vBulletin® Version 3.8.8