Page 27 of 74 FirstFirst ... 17222324252627282930313237 ... LastLast
Results 521 to 540 of 1472
Like Tree4Likes
  1. #521  
    #~~~~~~~~~~~~~~~~~~~~
    #
    # EDIT:
    #
    # The below ONLY happens when running on battery.
    # It does NOT happen when plugged in.
    #
    #~~~~~~~~~~~~~~~~~~~~


    Psycho,

    Please see the attached log. This is what F104 is doing to my phone. You have any ideas?

    It's in a CONSTANT state (as you'll see in the log) of waking up and adjusting the CPU. I have Starfighter running your stick settings right now, but, I've tried 500/900 (closest to Thunderbird without using Thunderbird)...

    Psycho, I've Doctored my phone 5 times since last Friday, Tuesday afternoon being the most recent. After I OTA'd to 1.4.5, everything was fine. When I Doctored for 1.4.5, it has (literally) being doing this since. 24/7. Literally.

    Thunderbird does the same thing, I already tried that, but not since the last doctoring. I'll put it on and update this post. I've used Thunderbird since early summer with ZERO issues prior to the first doctoring last Friday so I decided to try Starfighter and get the same thing.


    Sprint Pre (not launch)
    webOS 1.4.5
    Startfighter-110






    messages.txt.zip


    Anything you need to know, let me know. I've trusted your judgment in this realm since late Jan / early Feb.


    M.
    Last edited by Xanadu73; 09/10/2010 at 10:11 PM.
  2.    #522  
    Quote Originally Posted by Xanadu73 View Post
    Psycho,

    Please see the attached log. This is what F104 is doing to my phone. You have any ideas?

    It's in a CONSTANT state (as you'll see in the log) of waking up and adjusting the CPU. I have Starfighter running your stick settings right now, but, I've tried 500/900 (closest to Thunderbird without using Thunderbird)...

    Psycho, I've Doctored my phone 5 times since last Friday, Tuesday afternoon being the most recent. After I OTA'd to 1.4.5, everything was fine. When I Doctored for 1.4.5, it has (literally) being doing this since. 24/7. Literally.

    Thunderbird does the same thing, I already tried that, but not since the last doctoring. I'll put it on and update this post. I've used Thunderbird since early summer with ZERO issues prior to the first doctoring last Friday so I decided to try Starfighter and get the same thing.


    Sprint Pre (not launch)
    webOS 1.4.5
    Startfighter-110

    Anything you need to know, let me know. I've trusted your judgment in this realm since late Jan / early Feb.


    M.
    It could be either battery scaling or temp scaling, but it is suppose to say if it is doing one of those things. It's just bouncing.

    But after looking at it the prcm subsystem is calling __set_cpu_policy() after waking up. From your logs and my phone as well the policy doesnt change. I enabled more debugging with the latest releases to dump policy changes during battery, voltage, and temp scaling to see what it's doing.

    For example: you will see it calls the policy change always after:
    Code:
    [54662.100000] 00091454.438: enter suspend
    [54662.100000] 00091546.791: leave suspend
    [54662.100000] WAKEUP STATE/REASON
    [54662.100000] PM_WKST_WKUP      00000108
    ....then immediately followed by policy change.

    Old versions of the kernels I didnt have it dumping all that and all you saw was:
    Code:
    [54662.100000] screenstate: policy change
    So this is normal when the phone goes into deep sleep and gets woken up from doing voltage scaling. This is an old version the kernel omap driver and newer kernels do it somewhat more gracefully, but it correctly should call the cpufreq functions to notify the governors on "suspected" frequency changes since it indeed was doing voltage scaling behind its back.

    It is a waste of maybe 4 cpu clock cycles to do all that tho.
    Live free or DIE!
  3. #523  
    extra logging is a Good Thing during testing, but my system load is *constantly* through the roof. The 20's as shown above is kinda low, it's been in the 30's and 40's most of the time. This makes my 1350mAh battery last 5 or 6 hours (I don't know because I've had to plug it in every time I'm back at my desk). *This* can't be "normal" / your expected extra load, is it?

    I'm back on Thunderbird now. I ran the Recovery Kernel after removing Starfighter Just In Case(tm). I'm down to 25% so I can't report results yet.

    Thank you for getting back to me!



    M.
  4.    #524  
    Quote Originally Posted by Xanadu73 View Post
    extra logging is a Good Thing during testing, but my system load is *constantly* through the roof. The 20's as shown above is kinda low, it's been in the 30's and 40's most of the time. This makes my 1350mAh battery last 5 or 6 hours (I don't know because I've had to plug it in every time I'm back at my desk). *This* can't be "normal" / your expected extra load, is it?

    I'm back on Thunderbird now. I ran the Recovery Kernel after removing Starfighter Just In Case(tm). I'm down to 25% so I can't report results yet.

    Thank you for getting back to me!



    M.
    The high load averages are normal when the phone is asleep. but in the 40's is bad. Try a battery pull, esp after changing from extremely different kernels.

    High load averages even happens in the stock kernel.
    Live free or DIE!
  5. #525  
    Quote Originally Posted by unixpsycho View Post
    The high load averages are normal when the phone is asleep. but in the 40's is bad. Try a battery pull, esp after changing from extremely different kernels.

    High load averages even happens in the stock kernel.

    I know it gets high sometimes, but this is a constant pegging of the system load. Also, it *only* happens when it's on battery. I never see this plugged in.


    M.
  6.    #526  
    Quote Originally Posted by Xanadu73 View Post
    I know it gets high sometimes, but this is a constant pegging of the system load. Also, it *only* happens when it's on battery. I never see this plugged in.


    M.
    Correct. That is normal. Just because the system load is high doesnt mean it's actually working on anything. When the phone sleeps certain tasks get left on the runqueue and only certain functions can wake the phone up... so the load average will increase even tho it's literally not doing anything. an RTC timed wakeup will flush out queued tasks, or WOL network traffic and such.

    The OS wont let voltage scaling and such happen while plugged in, there's no reason to. That's why you only see this happen on battery. Also why it sucks when using ssh to Pre on battery with screen off.
    Live free or DIE!
  7. #527  
    Quote Originally Posted by unixpsycho View Post
    Correct. That is normal. Just because the system load is high doesnt mean it's actually working on anything. When the phone sleeps certain tasks get left on the runqueue and only certain functions can wake the phone up... so the load average will increase even tho it's literally not doing anything. an RTC timed wakeup will flush out queued tasks, or WOL network traffic and such.

    I've just never seen my phone do this. With everything "you've" thrown at it the past 8 or so months, this is the first it's acted like this. It's even odder that it's only started after after a Doctor visit (which this is the first time in 2010, IIRC). My load's never been like nor has my bettery been plowed through like this. It's down to 52% right now, when I had to plug it back in (up to 100%) around 1:30pm (it's just after 6PM now).


    M.
  8. #528  
    Odd. I uninstalled F105 (which has been flawless, BTW!) prior to letting the 1.4.5 OTA hit, and now when I try to install either F104 or F105, I get "Kernel is not compatible with this webOS version, aborting".

    The phone info shows 1.4.5 installed correctly. Could this be a Preware bug?
  9. #529  
    Quote Originally Posted by pEEf View Post
    Odd. I uninstalled F105 (which has been flawless, BTW!) prior to letting the 1.4.5 OTA hit, and now when I try to install either F104 or F105, I get "Kernel is not compatible with this webOS version, aborting".

    The phone info shows 1.4.5 installed correctly. Could this be a Preware bug?

    Your feed is not setup properly. See the "Kernel Testing Feeds" at the bottom of this page:

    Testing Feeds - WebOS Internals


    M.
  10. #530  
    Ahh, forgot the OS version is hardcoded in the feed URL. Doh!

    Thanks Xanadu, I feel silly.
  11. #531  
    Quote Originally Posted by pEEf View Post
    Ahh, forgot the OS version is hardcoded in the feed URL. Doh!

    Thanks Xanadu, I feel silly.

    I did the same thing that's how I knew the answer so fast...



    :-)


    M.
  12. #532  
    Quote Originally Posted by pEEf View Post
    Ahh, forgot the OS version is hardcoded in the feed URL. Doh!

    Thanks Xanadu, I feel silly.
    If the feed is named correctly, Preware updates the version automatically.

    -- Rod
  13. #533  
    Has anyone who could not previously run SR71 tried the 'newer' lower voltage SR71 and succeeded? Enquiring minds want to know!
  14. #534  
    Quote Originally Posted by MudShark22 View Post
    Has anyone who could not previously run SR71 tried the 'newer' lower voltage SR71 and succeeded? Enquiring minds want to know!
    By run do you mean not boot at all or work but not be stable?
  15. #535  
    F105 (pretty much all versions) worked great on my PrePlus. Then I tried Warthog as I wanted to try 1050MHz and <500 screen off. Before that, with the old F104 I could run 125, 250, etc., no problem. However F104A-110 does some strange stuff on my PrePlus. It'll boot fine at 1100 and screen state off to 250 no problem. It'll even wake up from 250 fine, and load Preware just fine. Strange thing with -110 is that if I try to set a Govnah profile where the "max" speed is 500 or less, or use a conservative profile, it freezes completely & requires a battery pull. Every time I try to create a profile with a max speed <700 in Govnah I get an instant freeze. And the voltage settings look higher than what I was using with F105 or Warthog. I may try to dig deeper another time, but right now I'm a bit stumped.
  16. #536  
    Quote Originally Posted by PaulB007 View Post
    By run do you mean not boot at all or work but not be stable?
    In my experience with SR71; it either works/boots or does not; there is no "using but occasional crashes".

    The intent (before I try it myself) is to see if anyone who could not keep the SR71 kernel on their phone for WHATEVER reason now have success and are using it as their regular kernel instead of F104 et al.
  17. #537  
    I've been running F105 for a wile now, but recently (as in the past 4-5 days) I have noticed that Govnah thinks I'm running at 900Mhz instead of 1Ghz. I'm not sure if this is a Govnah bug or if my phone is somehow not able to run at the 1Ghz speed it used to, anymore.
  18. #538  
    Quote Originally Posted by MudShark22 View Post
    In my experience with SR71; it either works/boots or does not; there is no "using but occasional crashes".

    The intent (before I try it myself) is to see if anyone who could not keep the SR71 kernel on their phone for WHATEVER reason now have success and are using it as their regular kernel instead of F104 et al.

    I am running the latest build of SR71 now. Previously, I constantly had issues with reboots. Luna or hard resets happened randomy and without a pattern. I have been running it again since Monday and have only had a few random reboots. This happened when I played Nesem. It also happens when I run the phone up to about 50C. When I am heavily using the phone I tend to just scale down to 1GHz and when I am moderately using it I keep it at 1.2GHz and rarely have reboots. Oh games still don't work and either reboot the phone or the card just closes after a minute or so. Hope this gives you some idea.
  19. #539  
    Quote Originally Posted by PaulB007 View Post
    I am running the latest build of SR71 now. Previously, I constantly had issues with reboots. Luna or hard resets happened randomy and without a pattern. I have been running it again since Monday and have only had a few random reboots. This happened when I played Nesem. It also happens when I run the phone up to about 50C. When I am heavily using the phone I tend to just scale down to 1GHz and when I am moderately using it I keep it at 1.2GHz and rarely have reboots. Oh games still don't work and either reboot the phone or the card just closes after a minute or so. Hope this gives you some idea.
    The 1st time I used the "older'" SR71s; it worked for 10 minutes, overheated and then went to reboothell. Considering I can run 1.1 without any trouble, I was hoping UNIX's voltage tweaks might help out.

    Can't try it @ the office...will have to do it @ home.
    Thanks for the feedback!!
  20. #540  
    Quote Originally Posted by MudShark22 View Post
    The 1st time I used the "older'" SR71s; it worked for 10 minutes, overheated and then went to reboothell. Considering I can run 1.1 without any trouble, I was hoping UNIX's voltage tweaks might help out.

    Can't try it @ the office...will have to do it @ home.
    Thanks for the feedback!!
    I hope it works, but I have a feeling that no matter what there will never be 100% stability on some of our phones running SR71 even though the latest build works better for me.

Tags for this Thread

Posting Permissions