Page 35 of 145 FirstFirst ... 2530313233343536373839404585135 ... LastLast
Results 681 to 700 of 2889
Like Tree17Likes
  1. #681  
    I would guess they went and got the ipk on their own from another feed (see the URL I posted) and used WebOS Quick Install to put it in. If they are running fine, then I cab't imagine it'd be much longer until it gets put in the proper feed for general consumption.

    I am in no way "in" with any of the (talented!) folks who make these decisions, so I can't sit hit and give *any* kind of ETA.


    M.
  2.    #682  
    Quote Originally Posted by cooliosam View Post
    thanks for the quick reply xanadu! i just saw some comments that ppl have been using the 800MHz already on the AT&T... i guess they probably downgraded to 1.4.1 to do it??
    If you're talking about previous overclocking solutions, then they are running a webOS 1.4.0 kernel on a webOS 1.4.2 device.

    We know that presents a problem for Verizon PrePlus keyboards, we haven't tested that combination and don't have the source code for the 1.4.2 kernel from Palm yet (see opensource.palm.com and send them an email requesting the 1.4.2 kernel source if you want to expedite the process).

    -- Rod
  3. #683  
    Well...

    While the cat is away, the mice will play... 'Psycho released a new rev Friday afternoon based on the 1.4.1.x Palm Patch rev. I specifically asked him if Friday's kernel included the updated kernel module, and he says it does.

    But that aside, Rod is right on the money. Everything besides Uber and whatever Psycho is calling his these days is not a kernel that really should be used (other than us Sprint folks) because of the different hardware and the different kernel driver.


    M.

    P.S.
    I don't know what bromwich bases his kernels on, so I can't comment there.
  4.    #684  
    Quote Originally Posted by Xanadu73 View Post
    Well...

    While the cat is away, the mice will play... 'Psycho released a new rev Friday afternoon based on the 1.4.1.x Palm Patch rev. I specifically asked him if Friday's kernel included the updated kernel module, and he says it does.

    But that aside, Rod is right on the money. Everything besides Uber and whatever Psycho is calling his these days is not a kernel that really should be used (other than us Sprint folks) because of the different hardware and the different kernel driver.


    M.

    P.S.
    I don't know what bromwich bases his kernels on, so I can't comment there.
    As you can see from git.webos-internals.org Git - preware/build.git/blob - testing-kernels/sbromwich-kernel-pre/Makefile and git.webos-internals.org Git - preware/build.git/blob - nonworking/psycho-kernel-pre/Makefile all the kernel developers are playing in the same animal-house

    Steve has his kernel fully integrated into the autobuild/release infrastructure, Marco is waiting on me (I was away for the weekend) to get the final touches on his kernel integration folded in.

    We all base the kernels on the latest source code available from Palm, which at this point in time is 1.4.1(.1) - you can send an email to the address at opensource.palm.com to politely request that they expedite the release of the source code for the 1.4.2 kernel so we can build the three kernel streams for it.

    -- Rod
    WebOS Internals and Preware Founder and Developer
    You may wish to donate by Paypal to donations @ webos-internals.org if you find our work useful.
    All donations go back into development.
    www.webos-internals.org twitter.com/webosinternals facebook.com/webosinternals
  5. #685  
    Yea, I read you were out, so I put the first line in as humor (humour).

    That's good to know all three are "on the same page". I only know of the existence of Steve's kernel. His upstart script on the wiki gave me a few things to think about. I haven't tried his at all yet.

    As far as Palm goes, how responsive are they to community prodding? I mean, they *clearly* are very supportive of the (F)OSS community and don't need the GLP thrown at them (not that I'd go there anyway - I'm not Linus (or RMS...)).


    M.
  6.    #686  
    Quote Originally Posted by Xanadu73 View Post
    Yea, I read you were out, so I put the first line in as humor (humour).

    That's good to know all three are "on the same page". I only know of the existence of Steve's kernel. His upstart script on the wiki gave me a few things to think about. I haven't tried his at all yet.

    As far as Palm goes, how responsive are they to community prodding? I mean, they *clearly* are very supportive of the (F)OSS community and don't need the GLP thrown at them (not that I'd go there anyway - I'm not Linus (or RMS...)).


    M.
    I think an email showing your appreciation for the speed at which previous kernel source code has been published, and a polite request for expediting the 1.4.2 source code release, is entirely appropriate.

    Remember that Palm has already completely fulfilled their GPL obligations with the PDF file on your USB drive area, so there is absolutely no GPL issue here at all. This is a polite request from Palm's users to assist the homebrew community in keeping those users happy by expediting the publishing of the kernel source code on the opensource.palm.com site.

    -- Rod
    WebOS Internals and Preware Founder and Developer
    You may wish to donate by Paypal to donations @ webos-internals.org if you find our work useful.
    All donations go back into development.
    www.webos-internals.org twitter.com/webosinternals facebook.com/webosinternals
  7. #687  
    I think I could've word that better since I think you took me the wrong way. I'm full aware they're making good on their GPL obligations. That's what I was trying to say. That there's no need to go there (not that I would anyway). I was more trying to say that since they *are* good for it, how responsive are they to people bothering them about it? Ya know, like *your* stuff is free and you're pretty parient with people bugging you for updates. Are Palm as understaing (for a business, I mean...)?


    M.
  8.    #688  
    Quote Originally Posted by Xanadu73 View Post
    I think I could've word that better since I think you took me the wrong way. I'm full aware they're making good on their GPL obligations. That's what I was trying to say. That there's no need to go there (not that I would anyway). I was more trying to say that since they *are* good for it, how responsive are they to people bothering them about it? Ya know, like *your* stuff is free and you're pretty parient with people bugging you for updates. Are Palm as understaing (for a business, I mean...)?


    M.
    Oh, I took your post the right way - I was mostly restating that for other readers. Sorry I didn't make that clear.

    I think any business would be happy to receive emails expressing how passionate people are about the platform, and how the business can easily assist them to continue to get an even better experience from it.

    -- Rod
    WebOS Internals and Preware Founder and Developer
    You may wish to donate by Paypal to donations @ webos-internals.org if you find our work useful.
    All donations go back into development.
    www.webos-internals.org twitter.com/webosinternals facebook.com/webosinternals
  9. #689  
    Quote Originally Posted by rwhitby View Post
    I think an email showing your appreciation for the speed at which previous kernel source code has been published, and a polite request for expediting the 1.4.2 source code release, is entirely appropriate.

    Remember that Palm has already completely fulfilled their GPL obligations with the PDF file on your USB drive area, so there is absolutely no GPL issue here at all. This is a polite request from Palm's users to assist the homebrew community in keeping those users happy by expediting the publishing of the kernel source code on the opensource.palm.com site.

    -- Rod
    Thanks rwhitby! I'm new to the Pre scene so I'm very excited =) I will join the efforts and ask opensource.palm.com to expedite the kernel source code!
  10. #690  
    Im Loving this new feature
  11. #691  
    I place my Pre on the touchstone before I got to bed, sometimes when i get back to it in the morning the screen is "on" but just a blank/black screen. I have to remove the battery to get it to restart.

    I am not sure when this happens, but I always get the sweet spot, and it says it is charging before i leave it. The Pre is not hot, works fine sometimes.

    Pre does not charge when this happens... Any ideas? I have had the same touchstone/back since release on sprint. My Pre is a refurb I got last month, but seems to be working fine.

    Uber-Kernel, Govnah, and CPUscaler Ultimate. Set to on-demand 250-800. I never had the issue before uber-kernel, was using 800MHz Ipk.

    Also using preset reset @ 1am every morning, could be a contributing factor. The weird thing is it only happens sometimes, not every night.

    I posted this here because I did not have the issue before uber-kernel, is this a know issue?

    Sprint Pre- Meta-Doctor 2.1.0 w/Flash

    Everything is Amazing & Nobody is Happy, "People with their mobile phones, "uh... oh... it won't..."... GIVE IT A SECOND... IT'S GOING TO SPACE!" Louis C.K.
  12. #692  
    Quote Originally Posted by Xanadu73 View Post
    I don't know what bromwich bases his kernels on, so I can't comment there.
    Just to expand on what Rod said... All the kernel developers upload their patches to the repository. The Makefile (that creates the ipkg you download and install for the new kernel) includes a list of which patches to apply; the Makefile is (at the source level) the only thing that exists to build the kernel for each stream. When the autobuilder runs it just extracts the kernel, applies the Palm kernel patch, then applies whichever set of patches the kernel developer wants to run and then compiles the kernel accordingly. If I wanted to include compcache in my kernels, for example, it would be a matter of editing the Makefile for my kernel to include it, and the autobuilder would build it in next time a build was run.

    Cheers, Steve
  13. Dolio's Avatar
    Posts
    48 Posts
    Global Posts
    54 Global Posts
    #693  
    Greetings, just wanted to add my observations. First off thanks so much to everyone involved in bringing the Pre community these awesome kernel hacks and all the other homebrew apps! I only wish that I could keep up with a fraction of what's happing on these forums.

    I started a little late with the "Super PreKernel" app and service along with the "CPUScalerUltimate" app. Both worked very well with no major issues. Later I reverted to the stock kernel (leaving both the SPK and Scaler apps installed) and then installed the UberKernel and Govnah, both of which also seem to be working well.

    In the graph below (more about the graph itself here http://forums.precentral.net/2459224-post1.html) the first half of the week was on the SPK and Scaler (OnDemand 500-800MHz), the second half using the UberKernel. The dark patch at the center of the week was running Govnah with OnDemand 125-800MHz, there after I've been using primarily Screen State.

    I noticed that the Govnah and Scaler's OnDemand settings are not the same and the Govnah's version of OnDemand produced an observable "problem". That being that while playing video and even with video paused and "minimized" in a non active card that there were video output static/glitches or random and rapidly flashing vertical lines. I tried various OnDemand settings and frequency ranges but this video glitch was always present in OnDemand, even when attempting to mimic the Scaler OnDemand settings. To be clear, the Scaler OnDemand settings do not produce this video glitch problem even with the same kernel, and interestingly I am able to save the Scaler OnDemand settings as a profile in Govnah, but can not create that profile from Govnah itself (shrugs).

    Another thing I noticed is that even though the Govnah Screen State setting says (125-800MHz) the kernel never appears to drop to frequencies below 500MHz. The Screen State setting does not produce the video glitch described above. On the up side it does seem to be running at 800MHz when it needs to, that is it's snappy when I'm actively doing things.

    I have not played around with many of the other Govnah settings. the conservative and powersave to a greater extent just seems to make things unacceptably laggy, as expected since they are weighted towards slower frequencies I don't have any data as to if they actually save power, yet...

    While I really like the graphs that Govnah draws I with they were slightly more static in scale and had Y axes labels and lines.

  14.    #694  
    Quote Originally Posted by Dolio View Post
    I noticed that the Govnah and Scaler's OnDemand settings are not the same and the Govnah's version of OnDemand produced an observable "problem". That being that while playing video and even with video paused and "minimized" in a non active card that there were video output static/glitches or random and rapidly flashing vertical lines. I tried various OnDemand settings and frequency ranges but this video glitch was always present in OnDemand, even when attempting to mimic the Scaler OnDemand settings. To be clear, the Scaler OnDemand settings do not produce this video glitch problem even with the same kernel, and interestingly I am able to save the Scaler OnDemand settings as a profile in Govnah, but can not create that profile from Govnah itself (shrugs).
    Did you perform this test with Govnah running, or with Govnah closed? Govnah does not need to be running - you should save the parameters and then close it. If you leave it running, then it will definitely affect performance since it will be waking up the processor and context switching to the Govnah app and redrawing a graph every second. So just don't do that if you want to get quality measurements.

    Please extract the kernel settings for each from the Linux command line and post them here for comparison.

    Another thing I noticed is that even though the Govnah Screen State setting says (125-800MHz) the kernel never appears to drop to frequencies below 500MHz. The Screen State setting does not produce the video glitch described above. On the up side it does seem to be running at 800MHz when it needs to, that is it's snappy when I'm actively doing things.
    Yes, that is clearly stated in many many places in the documentation and in these forum threads - the screenstate governor has a 500MHz lower limit at the moment.


    BTW, would you like to work with us to get your rrdtool graphing scripts into Preware?

    -- Rod
    WebOS Internals and Preware Founder and Developer
    You may wish to donate by Paypal to donations @ webos-internals.org if you find our work useful.
    All donations go back into development.
    www.webos-internals.org twitter.com/webosinternals facebook.com/webosinternals
  15. Dolio's Avatar
    Posts
    48 Posts
    Global Posts
    54 Global Posts
    #695  
    Quote Originally Posted by rwhitby View Post
    Did you perform this test with Govnah running, or with Govnah closed? Govnah does not need to be running - you should save the parameters and then close it. If you leave it running, then it will definitely affect performance since it will be waking up the processor and context switching to the Govnah app and redrawing a graph every second. So just don't do that if you want to get quality measurements.
    I just verified that with both Govnah and the video playback running the entire time that switching between the default Govnah OnDemand Profile and the Scaler Saved OnDemand Profile. The video glitch appeared with the Govnah profile and then went away immediate upon switching to the Scaler OnDemand Profile. Again, both apps running the entire time so I don't think it's a problem introduced because the govnah app is active.

    Quote Originally Posted by rwhitby View Post
    Please extract the kernel settings for each from the Linux command line and post them here for comparison.
    Is this what you wanted?
    @palm-webos-device:~$ d=/sys/devices/system/cpu/cpu0/cpufreq ; for f in `ls -1 $d/` ; do if [ -f $d/$f ] ; then echo -n " $f " ; cat $d/$f ; fi ; done

    Scaler OnDemand Settings:
    affected_cpus 0
    cpuinfo_max_freq 800000
    cpuinfo_min_freq 125000
    scaling_available_frequencies 800000 720000 600000 550000 500000 250000 125000
    scaling_available_governors conservative ondemand powersave screenstate userspace performance
    scaling_cur_freq 720000
    scaling_driver omap
    scaling_governor ondemand
    scaling_max_freq 800000
    scaling_min_freq 125000
    & /sys/devices/system/cpu/cpu0/cpufreq/ondemand
    ignore_nice_load 0
    powersave_bias 1
    sampling_rate 1000000
    sampling_rate_max 100000000
    sampling_rate_min 100000
    up_threshold 11

    Govnah OnDemand Settings:
    affected_cpus 0
    cpuinfo_max_freq 800000
    cpuinfo_min_freq 125000
    scaling_available_frequencies 800000 720000 600000 550000 500000 250000 125000
    scaling_available_governors conservative ondemand powersave screenstate userspace performance
    scaling_cur_freq 800000
    scaling_driver omap
    scaling_governor ondemand
    scaling_max_freq 800000
    scaling_min_freq 125000
    & /sys/devices/system/cpu/cpu0/cpufreq/ondemand
    ignore_nice_load 0
    powersave_bias 0
    sampling_rate 200000
    sampling_rate_max 100000000
    sampling_rate_min 100000
    up_threshold 80

    I just noticed that If I change the powersave bias and up_threshold in Govnah that they do not actually appear to be changing... I changed the Sampling rate from .2 to 1 second, which did change, but notice that the other two did not change...
    ignore_nice_load 0
    powersave_bias 0
    sampling_rate 1000000
    sampling_rate_max 100000000
    sampling_rate_min 100000
    up_threshold 80

    Alright, from the console I modified the up_threshold by hand, the video continued to glitch at 70 and 60 and seemed to stop glitching at 50 or lower (Scaler uses 11)...

    Are others able to reproduce the video glitch?

    Quote Originally Posted by rwhitby View Post
    Yes, that is clearly stated in many many places in the documentation and in these forum threads - the screenstate governor has a 500MHz lower limit at the moment.
    oic, truly sorry, I've been wanting nothing more than to catchup on this thread, but alas we are on page 36 or something and it will take me a few days to get caught up...

    Quote Originally Posted by rwhitby View Post
    BTW, would you like to work with us to get your rrdtool graphing scripts into Preware?

    -- Rod
    Absolutely! Catch me via the thread on my hack, or PM. I have not spent any time with the optware rrd packages that I saw. And I only spent about a half an hour with the optware mini snmp package which I could not get to work, and wouldn't be the right route anyway... I have not written any real Pre apps nor services at this point but would very much like to...
    Last edited by Dolio; 05/18/2010 at 05:37 AM.
  16.    #696  
    Quote Originally Posted by Dolio View Post
    I just verified that with both Govnah and the video playback running the entire time that switching between the default Govnah OnDemand Profile and the Scaler Saved OnDemand Profile. The video glitch appeared with the Govnah profile and then went away immediate upon switching to the Scaler OnDemand Profile. Again, both apps running the entire time so I don't think it's a problem introduced because the govnah app is active.
    Please re-run the test with the Govnah application CLOSED.

    Keeping the Govnah application open causes periodic wake-ups, context switches, service calls (including dbus activity), graph redraws, etc. Of course it is going to cause video glitches at low scaling frequencies.

    The CPU Scaler app does nothing while it is open. The two apps are completely different in their operation.

    You CANNOT take correct measurements with Govnah open. Simply save the settings, then close the app, THEN start measuring. The settings remain until they are changed or until the next reboot.

    -- Rod
    WebOS Internals and Preware Founder and Developer
    You may wish to donate by Paypal to donations @ webos-internals.org if you find our work useful.
    All donations go back into development.
    www.webos-internals.org twitter.com/webosinternals facebook.com/webosinternals
  17.    #697  
    Quote Originally Posted by Dolio View Post
    oic, truly sorry, I've been wanting nothing more than to catchup on this thread, but alas we are on page 36 or something and it will take me a few days to get caught up...
    That's the trouble with forums, and the reason why WebOS Internals uses wiki pages for documentation instead of expecting people to read through hundreds of pages of forum threads.

    The wiki pages are always linked as the Homepage link in the Preware description page for the package. Why do people never read them first *before* installing an application?

    Application:Govnah - WebOS Internals

    Application:UberKernel - WebOS Internals

    -- Rod
    WebOS Internals and Preware Founder and Developer
    You may wish to donate by Paypal to donations @ webos-internals.org if you find our work useful.
    All donations go back into development.
    www.webos-internals.org twitter.com/webosinternals facebook.com/webosinternals
  18.    #698  
    Quote Originally Posted by Dolio View Post
    Scaler OnDemand Settings:
    powersave_bias 1
    sampling_rate 1000000
    up_threshold 11

    Govnah OnDemand Settings:
    powersave_bias 0
    sampling_rate 200000
    up_threshold 80
    If you want to compare two kernels, then you need to set the parameters the same. Use Govnah to set exactly the same parameters as CPU Scaler sets, and you should expect exactly the same results. Set different parameters (as you have here), and you should expect different results.

    -- Rod
    WebOS Internals and Preware Founder and Developer
    You may wish to donate by Paypal to donations @ webos-internals.org if you find our work useful.
    All donations go back into development.
    www.webos-internals.org twitter.com/webosinternals facebook.com/webosinternals
  19. #699  
    I had the UberKernel installed but was having some issues. I removed it to run with the stock kernel for a while to see if the issues resolved themselves. It seems the kernel wasn't the issue. I tried to reinstall the kernel but am getting an IPKG error. I tried installing the stock kernel from the feed, rebooting, then reinstalling UberKernel but still get the IPKG error. When I removed UberKernel I did it by Alt-tapping the icon instead of through PreWare. Did I remove it wrong? Is there a manual cleanup I can do and try the install again?
  20.    #700  
    Quote Originally Posted by cdbillups View Post
    I had the UberKernel installed but was having some issues. I removed it to run with the stock kernel for a while to see if the issues resolved themselves. It seems the kernel wasn't the issue. I tried to reinstall the kernel but am getting an IPKG error. I tried installing the stock kernel from the feed, rebooting, then reinstalling UberKernel but still get the IPKG error. When I removed UberKernel I did it by Alt-tapping the icon instead of through PreWare. Did I remove it wrong? Is there a manual cleanup I can do and try the install again?
    Install the Palm Kernel, not Stock Kernel.

    -- Rod

Tags for this Thread

Posting Permissions