Re: [Announce] kernel-power stable v50 in Extras-Devel
 

hi everyone,

first big thanks for this great work :)

I have tryed to get my logitech rumblepad 2 usb to run with h-e-n
but it fails cause of the missing hid-logitech.ko kernel drivers.

So i compiled my own kernel based on kp50 with CONFIG_HID_LOGITECH=m
and after setting up hal with a custom fdi file the gamepad works fine.

Because im not really used to modify maemo/linux kernels
i would like to ask if there is an easy way
to get thinks like this driver to work with the stock kp50 kernel
and if not if it would be possible
to get this option enabled in some further release.

greetings from germany
Hoga

Re: [Announce] kernel-power stable v50 in Extras-Devel
 

Would love to see that too, wasn't it already enabled on older versions? I vaguely remember that I was using my logitech dual action gamepad, but I am not sure though. Too busy with other things to game with N900 :(

Re: [Announce] kernel-power stable v50 in Extras-Devel
 

hello guys
I have 2.6.28-omap1 kernel and I wanna know more about modified kernels and what is its positive and negative aspects from your experience.
Actually I've always been afraid from updating kernel because the last time I tried to update it the device stopped booting and I had to make a reflash.
So how safe it's to get KP ? and does stable version means that it's totally safe?

Re: [Announce] kernel-power stable v50 in Extras-Devel
 

Many thanks to the KP team for this incredible and ongoing effort.

I am upgrading from KP 49. I was using KP 49 @ 950MHz but found that videoconferencing stopped working on KP 50. Connection would fail when trying to use video, the red light would stay on afterwards with high CPU and battery use - no further skyping (audio or video) until a reboot. I changed frequency down a notch to 900 and I have video again (current config below).

Cheers :)

current kernel configuration:
current frequency: 500
supported frequencies: 125 250 500 550 600 720 805 850 900 950 1000 1100 1150
min. frequency: 125
max. frequency: 900
avoid frequencies: 125 250
active frequencies: 500:33,360 550:40,400 600:42,430 720:48,520 805:52,600 850:54,600 900:56,600 950:56,600 1000:60,600 1100:72,600 1150:72,600
SmartReflex VDD1=1, VDD2=1
governor ondemand: ignore nice load= 0, up threshold= 95, sampling rate= 300000, powersave bias= 0

Re: [Announce] kernel-power stable v50 in Extras-Devel
 

Hi ,

i have a quick question. Are power49 and 50 incompatible with the hacked fmtxd? I remember reading something to that effect but can't find it now. If it is incompatible, which is the newest power kernel that still supports it? I need to use it for a while. I can't get it to work with 49.

Thx

Re: [Announce] kernel-power stable v50 in Extras-Devel
 

Well if you compile it with the =m option you should have a .ko module file that you can insmod on the 'normal' kernel power so you do not have to flash your own kernel... but yes, you will have to compile ;)

Re: [Announce] kernel-power stable v50 in Extras-Devel
 

@Pali, what do you think for yet another package, called kernel-power-modules-extra, where all that rarely used drivers (see above posts) to live?

Re: [Announce] kernel-power stable v50 in Extras-Devel
 

If such thing was to happen do not forget this:
http://talk.maemo.org/showpost.php?p...0&postcount=54

Re: [Announce] kernel-power stable v50 in Extras-Devel
 

Sounds good. Now we have more extra modules in -module package and also extra usb drivers (for host mode). So I think that extra usb modules should also go to -modules-extra package.

But big problem is again maemo.org package interface...

Re: [Announce] kernel-power stable v50 in Extras-Devel
 

I use jacekowski's modified fmtx and it's working fine with kp50.

/Estel

Re: [Announce] kernel-power stable v50 in Extras-Devel
 

A few observations from me:
- after initial install, at not so fresh but well maintained N900 i had a massive battery drain. Or better if I wrote MASSIVE - six hours from 100% load to powered off phone without any activity and in offline mode (overnight). Uninstalled kp50.
- I`ve wanted to give it another try so flashed my device and installed it again on pure&clean CSSU testing (updated from PR1.3.1). From there I`ve started to install my applications one by one, watching how system performs battery wise. Right now I installed all of them (the same like when that huge battery drain appeared) except one - hildon desktop lockable in portrait (http://talk.maemo.org/showthread.php...sktop+supports).
- SUCCESS! Running with "dsp" (VDD1&2=1) profile as default, with whole bunch of my apps installed, third day at one charge. Yes, THIRD DAY. About 20% juice left.
I`m guessing that there are some compatibility issues between kp50 and that specific h-d version. So if any of You were struck by battery drain after kp50 install AND are using that h-d, try to remove it and return to CSSU h-d version. It will prove my guess right or wrong.
I`m completely, totally and absolutely blown away with this battery efficiency: it`s almost three times more than before (1 day to charge vs 2,5~3 days to charge). Massive improvements are especially shown when device is on standby (idle) - battery graph is close to flat. Writing "thank You pali" is not enough for this, but nevertheless: thank You pali!

PS - closing thought: please include proper version of h-d with portrait lock in next CSSU.

Re: [Announce] kernel-power stable v50 in Extras-Devel
 

I also had (have) a massive power drain shortly after installing. But I also installed latest CSSU testing around that time, so no idea which one... Unless the hildon-desktop you mention is included in CSSU-testing (2.1) - which I don't think so - it may be a problem elsewhere....

Re: [Announce] kernel-power stable v50 in Extras-Devel
 

Re: [Announce] kernel-power stable v50 in Extras-Devel
 

Well, I also feel like I need to report something (not so impressive but..)

For a long time I've been using KP46, then KP47 w/o VDD enabled. My average long-term power consumption was around 16mA, which gave me about 3 days per full charge.

Exactly a week ago I've installed KP50, enabled the default dsp profile (but avoided some frequencies) and, although the change is not so impressive, my average is now around 14mA, which is like half-a-day more (~ 10-15% longer time-between-charging).

Other than that. everything works fine (+ obviously no reboots to report).

Re: [Announce] kernel-power stable v50 in Extras-Devel
 

Really? I am connected to 2G, n900 locked and I am at 6mA with 250-600MHz.

Re: [Announce] kernel-power stable v50 in Extras-Devel
 

@all - guys, please, some of you to update the kernel-power wiki with the modificatons/achievments since KP46.

Re: [Announce] kernel-power stable v50 in Extras-Devel
 

Here is a plot of my SR values against frequency.

The X axis is frequency.

The Y axis is calculated power. Power is calculated from P=C*F*V^2. Where P is power in Watts, C is capacitance in Farads, V^2 is volts squared measured in Volts.

The voltage is calculated by the base voltage plus the number of voltage steps(0.0125V per step). So step 20 is (0.0125V*20) +0.6V=0.85V

Unfortunately I have no idea what the gate capacitance of the chip is. I have set it as 0.1F which is INSANELY high, however it is just a multiplication factor so it simply scales the output of the equation. So the power values are total nonsense however the shape of the graph is what is important.


http://i1174.photobucket.com/albums/r607/vipp2/powa.png


As you can see the voltage steps increase linearly, the power required increases exponentially. From this I would conclude that the 'most bang for buck' is in fact at 600MHz.

//EDIT: Real Conclusions

You can see from the graph that the difference in power use between 125-250MHz is negligible. When you consider the performance hit you take by activating 125MHz it makes sense to avoid it. Here doubling the frequency takes a tiny amount of power.

after 250MHz the increase in power use is almost linear up to 600MHz. Above 600 the power use line changes to a much higher gradient. This means increases in speed come at a much higher power cost.

meh, TBC.

Re: [Announce] kernel-power stable v50 in Extras-Devel
 

I'm talking *average*. Idle is obviously much less, but I actually use my N900 :)

Clearly I'm not a "heavy" user (I rarely use 3G, 2G is plenty for me) and do mostly modest, xterm and a bit of browsing.

Re: [Announce] kernel-power stable v50 in Extras-Devel
 

Re: [Announce] kernel-power stable v50 in Extras-Devel
 

@vi_ - seems the whole "race to idle" thingy is a little bit overestimated.

Re: [Announce] kernel-power stable v50 in Extras-Devel
 

Finally :p

Re: [Announce] kernel-power stable v50 in Extras-Devel
 

Anyone else getting MicroB to crash randomly when a dialog box shows up, for example when saving an image? Didn't happen before as much. Using 250 720 with SR1/2 enabled.

Re: [Announce] kernel-power stable v50 in Extras-Devel
 

I would like to report some problems with KP50.
to clarify- my N900 is sensitive for OC, and SR wasn't working stable on it. But with KP49 and DSP I was able to work on the phone without any problems. Now with KP50 and DSP I get suden restarts. To fix this and retain the SR I scaled down the max from 805 to 600 on DSP and now it works ok. No reboots.

As for the battery- I have my since day one 2,5 yera and I have about 1,5 or 2 days of usage so I think that this is ok ;)

Re: [Announce] kernel-power stable v50 in Extras-Devel
 

All KP50 packages appeared in the package interface! :O

Re: [Announce] kernel-power stable v50 in Extras-Devel
 

It'd be a good moment to promote v49 to Extras and v50 to Testing.

Re: [Announce] kernel-power stable v50 in Extras-Devel
 

Hello guys
I have recently downloaded the latest update for CSSU-Testing and I have also the stock kernel
(2.6.28-omap1).

so does anybody know which version of kernel is most comatible to my software?

and i wanna know how safe/risky is it to get a new KP
because last time I tried to update the kernel the device went into a COMA and boot loops.and since then I have kernelo-phobia;)

If needed at all, is there a step-by-step guide to have a safe download of kenel-power in a comprehensive way to avoid bricking OS?
thanks to all in advance!!!:)

Re: [Announce] kernel-power stable v50 in Extras-Devel
 

I'd rather promote 50 to testing and then quickly to extras...

Re: [Announce] kernel-power stable v50 in Extras-Devel
 

Agree, KP50 is better in terms of stability, battery life and compatibility with omap1

Re: [Announce] kernel-power stable v50 in Extras-Devel
 

not in performance tbh
omap1 is smoother than kp /not sure but that's what i see\

Re: [Announce] kernel-power stable v50 in Extras-Devel
 

For people experiencing a little higher power drain with kp50 having exactly the same applications as in kp49, I noticed for example in my case that Advanced Power was draining the battery more than in kp49 (i.e. a lot of hildon-desktop-menu CPU), so after I uninstalled AP everything went back to normal. So this means some previous applications might not work as expected in both kp49 and now in kp50.

On the other hand, Advanced Power was needed because default battery status is reporting wrong percentage and signal low battery level earlier than bq27200 module. So I will manage for now without AP or default battery status icon and display the bq27200 data with QBW on desktop only.

Re: [Announce] kernel-power stable v50 in Extras-Devel
 

I wonder if it's related to gangbang X-Fade was getting in "council election" thread ;)

// Edit

Nah, it's related to:
http://talk.maemo.org/showpost.php?p...4&postcount=45

So, kudos to freemangordon for constructively enforcing fix for this problem.

Re: [Announce] kernel-power stable v50 in Extras-Devel
 

Hey firstly thanks for updated kernel-power to v50.
Secondly, I'm still haven't been convinced to overclock, and I didn't measure batterry use objectively, but my subjective opinion is that kp50 is a notch better at batterry usage than previous version.
Thirdly I installed a bunch of apps but I believe none of them were realted to video codecs.

Then a real point: on kp50 it seems the video is EXTREMELY LAGGED on SOME codecs/backends. On others it's PERFECTLY OKAY. So I smell a bug in some driver included with kp50. I didn't have this problem prior to installing kp50. extremely lagged means: video goes 3-4 times slower than sound track (which on my player resulted in 1:20 minutes of 0:40 and sound stop properly). perfectly okay means: there wasn't even a 100 milisecond lag while playing the video, I can't seem to differentiate the video of man playing the instrument and the sound itself. before somebody call me a noob I'll say the same test with same videos was done before. If pali or any of bosses may want making me sorry I accent "I smell the bug in kernel driver" but "I may smell in wrong direction" (which may mean changes in kernel uncovered old bug in player/backend/codec which was supposed to never happen basing on assumptions how hardware and system works, but assumptions become false).

Thinking about my experience, there's no dissonance between people saying "It's lagging video a hell I'll go back to KP49" and people saying "Here IT WORKS perfectly", and it isn't matter of being lover/hater noob/pro fail_of_user/fail_of_package_maintainer, It's just a matter people test in different enviroment while claiming it's the same. Problem doesn't seem the software would be unavailble to work properly (since in some cases it work perfectly) but rather seem a trivial bug on time calculation which makes video on some technical cases (codecs/backends) run constant times slower just because of miscalculation. To be sure I'll test it again on BFS kernel (which is based on KP49) since I have pretty multiboot configuration. I'll back soon.

Re: [Announce] kernel-power stable v50 in Extras-Devl
 

...works for me.

seriously wtf are you on about?

What player?
what videos?
what 'CPU settings'

Need more data!

Re: [Announce] kernel-power stable v50 in Extras-Devl
 

answered inside quote. I have still to check some things, but not done yet as my phone is now under intensive task (NOTE: when I played the film there was almost nothing in background tasks)

Try playing films in different resolutions (my ones doesn't perfectly fit the screen - scalling occur - but all are much smaller than 720p) and different formats (mpeg4, flv, matroska H3623, etc.) and different backends (mafw, mplayer, and perhaps others). Note that MediaBox is only a player GUI and uses already installed backends. I have codecs for almost all common video/music formats on the phone. I may try checking which backend and which codec causes problem, but now I run optyfing script (pkgscan.sh) which involves many disk use, and I definately don't want make a system crash (script copies files then symlinks them, I run it because I have much stuffed rootfs, and I have to check if packages are able to optify so I must answer it questions). So i will make any new tests/checks later.

EDIT: After reboot (I managed to not brick my device by optifing some boottime essential) under same KP50 i cannot reproduce my problem (all works properly), but I believe it was always-reproducible in the short-term recheck. So it has to be some rare device state which triggerred the bug. I'm perfectly hapy now it works, but I'm curious what was the culprit.
The things happen during the try which catched the bug: Batterry was about 3% full (almost empty), It was long time since last reboot (before it used the 97% it was constantly under charger), I was connected to the network (GPRS i believe i remember), and openNTTPD was running (network clock syncing). I don't remember any other not normal things which happened to the device while I was for short time able to reproduce the bug. It may be some as mysterious bug as race condition. It could be also triggered by some VDD related function which was at not normal state (triggered by almost empty battery, and thus weird battery voltage). Ah one more I remember - after battery died (or player crashed i don't remember) and I recharged It couldn't boot up properly at all. First i thought I bricked the device, but removing the battery and puting back after 3 minutes helped. So indeed device was at weird state.
Since I can't reproduce problem anymore, You may forget it, and I will shut up about it, until it would happen once more. If somebody other claim it lags the video in any way, ask him what he/she was doing at the device while it happened, and ask him for check if reboot helps. Mysterious one...

Re: [Announce] kernel-power stable v50 in Extras-Devel
 

hello guys
After I downloaded KP50 I faced charging problem.
Sometimes the charger start charging after 20 seconds after I plug the charger in the device
and sometimes it doesn't charge at all and the same thing happens when using USB.
SO what the problem could be?
Plz Help!

Re: [Announce] kernel-power stable v50 in Extras-Devel
 

wish mine charged in 20sec too:D

Re: [Announce] kernel-power stable v50 in Extras-Devl
 

The only format that is guaranteed to play well on the N900 is H.264 Baseline Level 3@800x480.

That means:
No CABAC
No B-Frames
No Weighted P-Frames
8 bit video (10-bit anime will need to be re-encoded)
No Partitions
No Deblocking (?)

If you are using Handbrake, I'd recommend following these settings, plus maxing out the ME, Subpixel ME and Trellis options, to get better video quality at the cost of having ridiculously slow encode times.

Sometimes the voltage of the DSP is too low, requiring a reboot or a battery pull.

Potential USB port failure. Refer to this thread.

Re: [Announce] kernel-power stable v50 in Extras-Devel
 

@ majaczek, checked dmesg when ui/video lagging?

Re: [Announce] kernel-power stable v50 in Extras-Devel
 

I would also like to add that Theora is working fine on N900, and current implementations produce awesome results @ small file sizes. It's currently my codec of choice for all videos stored on N900.

/Estel

Re: [Announce] kernel-power stable v50 in Extras-Devel
 

how do I downgrade to KP49 when I'm using multiboot and KP50 boot image? :)