Page 1 of 2 12 LastLast
Results 1 to 20 of 28
  1.    #1  
    DO NOT RUN THIS KERNEL UNLESS YOU KNOW WHAT YOU ARE DOING. BEWARE OF PINK FLUFFY UNICORNS. RED LORRY, YELLOW LORRY, RED LORRY, YELLOW LORRY. IF YOU SERIOUSLY WANT TO RUN A KERNEL COMPILED BY SOMEONE WHO WRITES WARNINGS LIKE THIS YOU DESERVE EVERYTHING YOU GET.

    Features in this kernel:

    * Probable fix for resume from suspend hang. Note that the price to be paid for this is sometimes it can take a little extra time for the phone to wake up fully if there are a lot of apps running (including background tasks such as email checking).
    * CPU temperature sensor code from ka6sox
    * Explicit cpu and arch selection in gcc in order to optimise code for the Cortex CPU
    * 128KB kernel buffer to capture all dmesg output all the time
    * All IO schedulers enabled, anticipatory selected by default
    * Full range of frequencies up to 800MHz using unixpsycho's code, default boot speed of 800MHz
    * Force IVA2 into suspend
    * PREEMPT kernel
    * Full selection of CPU frequency governors, with ondemand as default
    * Full selection of TCP congestion control modules, default Reno
    * All options enabled for iotop
    * Kernel compiled with debug info for oops analysis

    Kernel is available at www.fop.ns.ca/pre/uImage.141-6.13-deathstar
    Config is available at www.fop.ns.ca/pre/config-141-6.13-deathstar
    Local settings I am using with this at www.fop.ns.ca/pre/setup.141-6.13-deathstar

    I am particularly interested in reports from people testing prior kernels for the resume from suspend bug to see if this finally nails it. dmesg should give useful information in this case (if you can still novaterm in).

    This kernel is not of release quality, and will not be made available in Preware (although I've been running it all day quite happily). If my fixes work as intended, I will be submitting them upstream for inclusion in a future Preware kernel.

    Cheers, Steve

    THIS KERNEL MAY CAUSE UNAVOIDABLE TIGHTENING OF THE PANTS IN MALES AFTER PROLONGED USE. RISK AND AMELIORATION OF THIS IS AT THE USER'S OWN DISCRETION.
  2. #2  
    Any news on the temp readings?
    I had lost my charger and thus could not test the last kernel as I would like, but I just found it, bring it on!

    Btw, the last kernel didn't lockup on resume after running Preware as the stock one did.
  3. #3  
    Yeah!
    temp1_input: 38
    temp1_input_raw: 59
    Battery Monitor reports: 42

  4. #4  
    Quote Originally Posted by NuttyBunny View Post
    Yeah!
    temp1_input: 38
    temp1_input_raw: 59
    Battery Monitor reports: 42

    Ignore the temp1_input_raw value, it's not a degrees reading like the other two.

    -- 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. #5  
    Quote Originally Posted by rwhitby View Post
    Ignore the temp1_input_raw value, it's not a degrees reading like the other two.

    -- Rod
    Excellent, I was worrying a little that the reading was a lot higher than the battery one

    Thanks!

    Btw, it crashed hard opening Asphalt 5, it didn't even respond to novaterm anymore! (scaling 125000-800000). I'll keep testing
  6. kevank#AC's Avatar
    Posts
    67 Posts
    Global Posts
    73 Global Posts
    #6  
    Installed the kernel... working well. Very fast and no lagging so far. I will keep testing and report back.
  7. #7  
    I can't wait to try this new kernel out once it gets to the beta stage. On a tangentially related note, I think that instead of "deathstar" you should name it "scrambles the death kernel"
  8. #8  
    mmm... there's something odd here:

    scaling_max_freq: 800000
    scaling_min_freq: 500000

    stats/time_in_state
    800000 25478
    720000 1394
    600000 2890
    550000 1245
    500000 2635
    250000 2645
    125000 33941

    After trying to launch Asphalt 5 twice (it gets stuck on the loading screen, flickering)

    800000 47477
    720000 3448
    600000 4490
    550000 2670
    500000 4645
    250000 9230
    125000 36059

    Yes, it's jumping to 250000 and 125000

    I have up_threshold on 20.

    I didn't try this with the other kernel, is this normal?

    Well, it took 2 minutes to boot, let's play
  9.    #9  
    Quote Originally Posted by Armchair Commando View Post
    I can't wait to try this new kernel out once it gets to the beta stage. On a tangentially related note, I think that instead of "deathstar" you should name it "scrambles the death kernel"
    Run dmesg in a terminal window after the phone has been running for a while to understand the naming schema I'm using.

    Cheers, Steve
  10. #10  
    Is this different than the 800mhz kernel I'm running now?
  11.    #11  
    Quote Originally Posted by NuttyBunny View Post
    mmm... there's something odd here:

    scaling_max_freq: 800000
    scaling_min_freq: 500000

    stats/time_in_state
    800000 25478
    720000 1394
    600000 2890
    550000 1245
    500000 2635
    250000 2645
    125000 33941

    After trying to launch Asphalt 5 twice (it gets stuck on the loading screen, flickering)

    800000 47477
    720000 3448
    600000 4490
    550000 2670
    500000 4645
    250000 9230
    125000 36059

    Yes, it's jumping to 250000 and 125000

    I have up_threshold on 20.

    I didn't try this with the other kernel, is this normal?

    Well, it took 2 minutes to boot, let's play
    OK that's a new one on me. Can you try rebooting and run the game as the very first thing you do? The only 3D app I have is the glyder2 demo; that locks the freq to 500 once it's done regardless of what I do, so there's something deep in the kernel playing with frequency settings that somehow is overriding the ondemand scheduler.

    Any complaints in dmesg after running the game? It's certainly not normal... Ideally it'd be clocking up to 800MHz to play the game. What happens if you set scaling_(max|min)_frequency the same?

    Cheers, Steve
  12.    #12  
    Quote Originally Posted by Cantaffordit View Post
    Is this different than the 800mhz kernel I'm running now?
    Probably.

    Cheers, Steve
  13. #13  
    Quote Originally Posted by sbromwich View Post
    OK that's a new one on me. Can you try rebooting and run the game as the very first thing you do? The only 3D app I have is the glyder2 demo; that locks the freq to 500 once it's done regardless of what I do, so there's something deep in the kernel playing with frequency settings that somehow is overriding the ondemand scheduler.

    Any complaints in dmesg after running the game? It's certainly not normal... Ideally it'd be clocking up to 800MHz to play the game. What happens if you set scaling_(max|min)_frequency the same?

    Cheers, Steve

    Well, I got several of these, but I don't think they are related to the scaling
    [ 1475.100000] ##### DISPC_IRQSTATUS_GFXFIFOUNDERFLOW
    [ 1475.100000] DISPLAY: ***** Sync Lost LCD (DISPC_IRQSTATUS 30e2) *****
    [ 1475.800000] ##### DISPC_IRQSTATUS_GFXFIFOUNDERFLOW
    [ 1475.800000] DISPLAY: ***** Sync Lost LCD (DISPC_IRQSTATUS 1060) *****
    [ 1518.850000] ##### DISPC_IRQSTATUS_GFXFIFOUNDERFLOW
    [ 1518.850000] DISPLAY: ***** Sync Lost LCD (DISPC_IRQSTATUS 30e2) *****
    [ 1521.000000] ##### DISPC_IRQSTATUS_GFXFIFOUNDERFLOW
    [ 1521.000000] DISPLAY: ***** Sync Lost LCD (DISPC_IRQSTATUS 1060) *****
    [ 1725.210000] ##### DISPC_IRQSTATUS_GFXFIFOUNDERFLOW
    [ 1725.210000] DISPLAY: ***** Sync Lost LCD (DISPC_IRQSTATUS 30e2) *****
    [ 1726.000000] ##### DISPC_IRQSTATUS_GFXFIFOUNDERFLOW
    [ 1726.000000] DISPLAY: ***** Sync Lost LCD (DISPC_IRQSTATUS 1060) *****

    Will try now locking it at 800000, but using ondemand and up_threshold of 20, running it as the first thing after boot.
  14. #14  
    It's the app maybe... but it crashed again (took the Pre down)


    Locked on at 800000
    While on the loading screen, it kept jumping again and flickering.
    [ 403.960000] ##### DISPC_IRQSTATUS_GFXFIFOUNDERFLOW
    [ 403.960000] DISPLAY: ***** Sync Lost LCD (DISPC_IRQSTATUS 30e2) *****
    [ 404.710000] CODEC ref count = 1
    [ 404.730000] CODEC ref count = 2
    [ 404.840000] CODEC ref count = 1
    [ 405.800000] ##### DISPC_IRQSTATUS_GFXFIFOUNDERFLOW
    [ 405.800000] DISPLAY: ***** Sync Lost LCD (DISPC_IRQSTATUS 1060) *****
    [ 412.420000] Waited 5 ms for DSP hibernation

    And just after it crashed:
    [ 537.410000] BUG: scheduling while atomic: swapper/0/0x40000105
    [ 537.410000]
    [ 537.410000] Pid: 0, comm: swapper
    [ 537.410000] CPU: 0 Not tainted (2.6.24-palm-joplin-3430 #1)
    [ 537.410000] pc : [<a0046c74>] lr : [<a005dfe8>] psr: 40000053
    [ 537.410000] sp : a040ded0 ip : a040dea8 fp : a040df84
    [ 537.410000] r10: 00000001 r9 : a044cda0 r8 : d8005000
    [ 537.410000] r7 : 0003ffff r6 : 00000001 r5 : 0010c389 r4 : 00000000
    [ 537.410000] r3 : 00000000 r2 : a040dea8 r1 : 000003d8 r0 : 00000001
    [ 537.410000] Flags: nZcv IRQs on FIQs off Mode SVC_32 ISA ARM Segment kernel
    [ 537.410000] Control: 00c5387f Table: 85718018 DAC: 00000017
    [ 537.410000] Function entered at [<a0034088>] from [<a005d498>]
    [ 537.410000] r5:00000002 r4:a040de88
    [ 537.410000] Function entered at [<a005d44c>] from [<a0387434>]
    [ 537.410000] r5:00000002 r4:a040c000


    Maybe it helps?

    Will boot it again and disable scaling to see if it helps
  15. kevank#AC's Avatar
    Posts
    67 Posts
    Global Posts
    73 Global Posts
    #15  
    Anyone notice that if you have scaling enabled and you launch the camera, the CPU will never drop back down past 500000? It will scale but never drop down to 250000 or lower. I have tried this on a couple of kernels and I get the same results. Anyone else able to recreate this?
  16. #16  
    Quote Originally Posted by sbromwich View Post
    Probably.

    Cheers, Steve
    Thanks, wise guy. Would you or anyone else like to elaborate on WHAT is different, and if there is benefit to using this rather than the 800mhz kernel from caj2008?

    Thanks!
  17. #17  
    Quote Originally Posted by Cantaffordit View Post
    Thanks, wise guy. Would you or anyone else like to elaborate on WHAT is different, and if there is benefit to using this rather than the 800mhz kernel from caj2008?

    Thanks!
    Well, almost everything that's on the first post
    I think the only thing that's the same is the frequency range.

    This kernel has been compiled after several days (or weeks) of testing several of the options compiled in.

    As sbromwich said, it is not for public release yet, as I can confirm (keeps crashing with Asphalt 5, and the ondemand governor keeps jumping frequencies even if min and max freqs are the same!)
    I just need one more test, I was about to do it but I restarted on my original Kernel to see if my Pre was still working correctly!!
  18. #18  
    Quote Originally Posted by Cantaffordit View Post
    Thanks, wise guy. Would you or anyone else like to elaborate on WHAT is different, and if there is benefit to using this rather than the 800mhz kernel from caj2008?

    Thanks!
    If Steve linked to the source patch, then that would be obvious

    Actually, I think it's all in the config, so just diff that against the stock palm config to see what is different.

    -- Rod
  19. #19  
    Ok, so, not much to report.
    On my Pre scaling seems to be buggy, at least with this Kernel
    It doesn't matter as I love to run at full speed all the time (800). Set scaling_governor to userspace, set the freq to 800,000 and it runs in all its glory.
    Tested Asphalt 5 and it ran great!

    Now, for the lockups.
    I did what I always do to lock Luna and have it restart, I open App Catalog, AppScoop and Preware. Update feeds in Preware (seems to be the trigger). Maybe install something, exit preware, suspend, and on resume most of the time it crashes.

    Still no locks. Will go to bed and keep it running at full speed, hoping for a Luna lockup

    --- well, so much for the fun... I unplugged my Pre from the computer, try to turn it on and poof! Luna lock and restart!
    I'm sending you an email with my /var/log/messages file just after the restart.

    Hope it helps!
    Last edited by NuttyBunny; 04/21/2010 at 12:48 AM. Reason: Luna Lockup! :D
  20. #20  
    Quote Originally Posted by caj2008 View Post
    Well see and please all stop comparing Steves kernels to me. This seems too personal.
    Honestly, I think the people who are comparing the different kernels are doing so because they genuinely don't understand the differences.

    To many, they don't get a single bit of how it works, but 800MHz is 800MHz to them. The rest is all gibberish.

    Now, a reasonable person might say if you don't understand the difference, you probably shouldn't be continuing past the first few lines anyway, but like that's ever stopped anyone before.
Page 1 of 2 12 LastLast

Posting Permissions