Page 3 of 4 FirstFirst 1234 LastLast
Results 41 to 60 of 80
  1. #41  
    Did you try powertop:
    From webos internals wiki:
    Code:
    cd /opt/bin
    ./ipkg-opt update
    ./ipkg-opt install powertop
    /opt/sbin/powertop
    EDIT: some explanation for powertop.
  2.    #42  
    Just ran powertop for 30 seconds on two different runs.

    Run 1) 30 seconds while I kept the phone awake. 200 wakeups-from-idle per second. 21% from i2c_omap, 18% from DMA, 16% from touchscreen, 15% from gp timer, 12% from OMAP HDQ, and 8% from prcm. The rest were single digit stuff.

    Run 2) 30 seconds after I hit the power button. 235 wakeups-from-idle pre second. 50% from i2c_omap, 15% from gp timer, 14% from prcm, 8% from OMAP HDQ, 3% from hsmmc-omap, 3% from ohci_hcd:usb1, and the rest are lower then that.

    What do you get? I ran the command "powertop -d -t 30" from terminal. One keeping the system awake and one putting it to sleep (and counting to 32) right after starting the program.
  3. #43  
    Switched on:
    Code:
    Wakeups-from-idle per second : 60.8	interval: 30.0s
    Top causes for wakeups:
      45.4% ( 78.4)       <interrupt> : i2c_omap 
      14.9% ( 25.8)       <interrupt> : prcm 
      13.8% ( 23.9)       <interrupt> : gp timer 
       9.7% ( 16.7)       <interrupt> : hsmmc-omap 
       4.8% (  8.3)       <interrupt> : DMA 
       3.0% (  5.1)       <interrupt> : WIFI_GPIO_INT 
       2.3% (  4.0)         modem_rmw : usb_hcd_poll_rh_status (rh_timer_func) 
       1.4% (  2.4)     <kernel core> : omap_mmc_request_done (omap_mmc_timer_cb) 
       1.2% (  2.1)            powerd : schedule_timeout (process_timeout) 
       1.2% (  2.0)       mediaserver : schedule_timeout (process_timeout) 
       0.6% (  1.0)       <interrupt> : twl4030-madc
    It will not run when switched off. I tried ssh and terminal. It just waits until I switch it on again.
  4. #44  
    Is it your battery? The "Dallas Semiconductor 1-Wire protocols" is used to communicate with the battery.

    hdq: documentation for OMAP HDQ

    Add a brief document about omap2430/3430 HDQ/1-wire driver.

    The HDQ/1-Wire module of TI OMAP2430/3430 platforms implement the hardware
    +protocol of the master functions of the Benchmark HDQ and the Dallas
    +Semiconductor 1-Wire protocols. These protocols use a single wire for
    +communication between the master (HDQ/1-Wire controller) and the slave
    +(HDQ/1-Wire external compliant device).
    +
    +A typical application of the HDQ/1-Wire module is the communication with battery
    +monitor (gas gauge) integrated circuits.
    You have 12% wake ups from that. I have 0.2% (not in the listing above)
  5.    #45  
    The number of interrupts for i2c_omap is the same between our phones. Everything else looks to be +3 times greater for my phone. Yea! My phone is interrupt happy.

    As far as running while "asleep". You can't run powertop because your phone actually goes to sleep. Which is what I thought should happen. Mine keeps running because it can't get to sleep. The interesting interrupt numbers for my runs on moving from "on" to "sleep" ...

    i2c_omap goes from 87.9 to 326.2
    gp timer goes from 65 to 101.8
    prcm goes from 39.4 to 89.0
    OMAP OHCI (usb device usb1) goes from 0.5 to 6.3


    So ... OMAP is the the system board. i2c (Inter-Integrated Circuit) is used to talk to sensors and such. So what does the Pre use i2c for? Is this a stuck accelerometer?

    It would be hilarious if the fix is "shake your Pre really really hard and then run really fast around the parking lot in a circle".

    while_on.txt

    while_off.txt
  6. #46  
    Quote Originally Posted by arrasmith View Post
    The interesting interrupt numbers for my runs on moving from "on" to "sleep" ...

    i2c_omap goes from 87.9 to 326.2
    gp timer goes from 65 to 101.8
    prcm goes from 39.4 to 89.0
    OMAP OHCI (usb device usb1) goes from 0.5 to 6.3


    So ... OMAP is the the system board. i2c (Inter-Integrated Circuit) is used to talk to sensors and such. So what does the Pre use i2c for? Is this a stuck accelerometer?

    It would be hilarious if the fix is "shake your Pre really really hard and then run really fast around the parking lot in a circle".
    Yea looks terrible. I assume you already tried to disable everything you can (GPS,DATA ...). Do you have a second battery to test?
  7.    #47  
    I've swapped out a Centro battery as well as the one that came with it a few days ago (I wanted two full batteries for when I can't be sitting next to power). Problem stayed.

    Are you sure that the Pre uses HDQ as a monitory for a battery monitor? Where is the light sensor at? I just tried to trigger that and the screen stays the same brightness.

    edit = added

    Also I repeat tests in airplane mode with GPS off and I've still keep getting the same results. I think you are on the right track with HDQ and I2C. Some sensor has (maybe) flipped out. What would account for the problem staying past a Doctor and why others see this issue intermittently as well.
    Last edited by arrasmith; 09/30/2010 at 11:28 AM. Reason: typo OHCI versus HDQ
  8. #48  
    I said OMAP HDQ. Is this the same as OHCI? I'm 100% sure it's using the "Dallas Semiconductor 1-Wire". You can find some info on it in my thread about Dr.Battery or http://forums.precentral.net/palm-pr...r-battery.html
    EDIT: But since you tried with two batteries it will not be the reason.
    What I've read (and can remember) the sensor is nearby the earpiece.
    I just tried it and yes near the earpiece.
  9.    #49  
    If I'm getting interrupts on HDQ does that mean that the micro-controller talking to the battery is sending them to the board? Or is it information from the battery? You think I can block one of the three wires and test it.

    Looks like the light sensor is working. If I hold it up to a light and then move it away it will lower brightness after a bit. Holding my hand over the ear piece area does nothing at all though. How did it work for you?
  10. #50  
    Quote Originally Posted by arrasmith View Post
    If I'm getting interrupts on HDQ does that mean that the micro-controller talking to the battery is sending them to the board? Or is it information from the battery? You think I can block one of the three wires and test it.

    Looks like the light sensor is working. If I hold it up to a light and then move it away it will lower brightness after a bit. Holding my hand over the ear piece area does nothing at all though. How did it work for you?
    You need all tree wires +,-,Communication.
    If I bring the earpiece into light the light on the pre goes up. If I cover it with my hand it goes down.
    Sorry but I think I can't help you anymore. Since you already doctored it looks like a hardware problem?
    Did you already tried the battery removal and hold the powerbutton > 15s?
  11.    #51  
    Thanks for all the help.

    As far as hardware goes, others have seen these sleep/wakeup events. And they have had it come and go with different kernels. Right now I feel like a dog with a chew toy. I just don't want to let it go. :/

    I have a Sprint family plan with 4 Pre's right now. And from the list it seems that this is a common problem. I know enough to write a script that would check if someone's Pre is stuck in a sleep/wake loop. But, without an answer as to WHY it will probably drive me nuts.
  12.    #52  
    @somline ...

    What do you get when you type "hwclock"?
  13. #53  
    Quote Originally Posted by arrasmith View Post
    @somline ...

    What do you get when you type "hwclock"?
    Code:
    sommac:~ somline$ ssh root@palm "hwclock"
    hwclock: can't open '/dev/misc/rtc': No such file or directory
  14.    #54  
    Bleh. I was hoping for an obvious rtc issue. I've noticed some negative times in my logs and the fastpath log time is rather different from the RTC time fastpath lists (RTC is 15 seconds behind the listed system clock).

    And hwclock was the only way I knew to get the RTC from command line.
  15.    #55  
    Don't know if anyone cares to know about this, but I got tired to messing with finding the issue (hardware? software?) so I went to the Sprint Store. Pre has 1 year warranty and according to Palm and Sprint I voided it by putting on non-App Catalog applications (Preware stuff).

    So I guess I'm back to trying to fix this issue. And I suppose when I have to bring a phone to the store I should Doctor it and lie.
  16. #56  
    Quote Originally Posted by arrasmith View Post
    Don't know if anyone cares to know about this, but I got tired to messing with finding the issue (hardware? software?) so I went to the Sprint Store. Pre has 1 year warranty and according to Palm and Sprint I voided it by putting on non-App Catalog applications (Preware stuff).

    So I guess I'm back to trying to fix this issue. And I suppose when I have to bring a phone to the store I should Doctor it and lie.
    According to Palm? I don't believe that. And this is not what palm said about homebrew. I think the Sprint store didn't know better.
    I would contact palm about it.
  17.    #57  
    Quote Originally Posted by somline View Post
    According to Palm? I don't believe that. And this is not what palm said about homebrew. I think the Sprint store didn't know better.
    I would contact palm about it.
    The specific response was that Palm let Sprint know that if you "root" your phone and install applications outside of the catalog path it voided the warranty. I'll fire off Palm an email to check. But, I wouldn't be surprised if there are two lines. One for the Sprint Stores and a different line for people who push a little harder. Saves Palm money that way.

    OR, maybe more likely, Sprint wants to sell insurance and collect the extra $100 on top of that you have to pay to get a replacement.


    P.S. The powertop advise helped give me a different route to consider on educating myself on the possible problem. I might not know who to contact at Palm about powerd, but powertop is opensource and has an active mail list for me to bug.
    Last edited by arrasmith; 10/01/2010 at 01:49 PM. Reason: I'm an idiot.
  18. #58  
    I'm really sorry about your story. What I've read (was a lot) palm will cover your problems.
    I think you know: there is no rooting of a pre/pixi. This is not an stupid iPhone.
    Palm maintains homebrew. THE reason I bought again a palm (pre) (coming from the palm pilot and many palm devices between).
    If you have problems with palm please post in the forum. You will get help from this unique community.
    EDIT: you may contact member heardbeatz (don't know how he is written). He is connected to palm.
    EDIT2: Sorry but this is to important. I started a thread here.
    Last edited by somline; 10/01/2010 at 02:15 PM.
  19.    #59  
    @somline ...

    Starting a new thread is fine.

    Two new ideas on the sleep issue ( and it may effect all the people complaining about alarms ).

    One: The powerd script in /etc/events.d uses a call hwclock to set the rtc to system time. hwclock isn't working as you know. So if your system's rtc drifts then you should start to see problems.

    Two: My phone has fewer powertop interrupts on the charger then not, because it does to a different power saver state ( not extreme saver ). I wonder if I can change powerd.conf to allow the same state on battery. It may make my batter drain be a bit better.
  20. #60  
    Quote Originally Posted by arrasmith View Post
    @somline ...

    Starting a new thread is fine.

    Two new ideas on the sleep issue ( and it may effect all the people complaining about alarms ).

    One: The powerd script in /etc/events.d uses a call hwclock to set the rtc to system time. hwclock isn't working as you know. So if your system's rtc drifts then you should start to see problems.

    Two: My phone has fewer powertop interrupts on the charger then not, because it does to a different power saver state ( not extreme saver ). I wonder if I can change powerd.conf to allow the same state on battery. It may make my batter drain be a bit better.
    I "analysed" powerd once for my battery investigations. I found one setting not in the default powerd.conf by analyzing the strings in the executable.
    But yes, there are some timing setting in it. I have no clue what they are about.
Page 3 of 4 FirstFirst 1234 LastLast

Posting Permissions