![]() |
Re: [Announce] kernel-power v48
it's not "kernel-power show" but "kernel-config show"
Hope that is (was) your only problem! :) If not, please elaborate.. @bitrocky: you beat me :) |
Re: [Announce] kernel-power v48
Yes, yes, tired fingers typed the wrong code. Nonetheless, when I try to delete kernel-power, I get the response as per my original post.
Question: What is the difference between these two files (their function): /etc/default/kernel_power and /etc/pmconfig Still more concerned about not being able to delete. Quote:
|
Re: [Announce] kernel-power v48
(I mean kernel-power - slippery fingers this morning)
Quote:
|
Re: [Announce] kernel-power v48
unexpected reboot..custom kernel config not loaded.... it is okay?
|
Re: [Announce] kernel-power v48
Yes, that's normal behavior if you have not created a 'default' file. However, if you experience reboots, it means the profile you're loading is not playing nicely with your phone. This happens to me with 'ideal', so I load 'lv' and have no more problems. Limits: 500 850 most of the time
|
Re: [Announce] kernel-power v48
everytitime that happen.. i reboot again to load my default
|
Re: [Announce] kernel-power v48
Let me simplify my problem. I am unable to deinstall kernel-power.
Code:
sudo gainroot Quote:
Quote:
|
Re: [Announce] kernel-power v48
Quote:
As far as I know this is the only "supported" way of removing kernel-power (never tried it though). After that (possibly with a reboot involved), you can install it again. But I still don't understand why you need to uninstall it (or reinstall it). What exactly is the problem you have? Quote:
The pmconfig file has parameters that are used by the maemo power management (not sure exactly how these map to kernel parameters). They do overlap a bit, but AFAIK kernel-power is loaded after pmconfig, so it takes precedence over pmconfig. |
Re: [Announce] kernel-power v48
Thanks. No particular reason to deinstall. I just like to make sure that everything is working as it should, and since I mucked around a bit, I wanted to deinstall and reinstall again. Code doesn't work and Nokia Kernel (which has replaced Deinstall Kernel-Power" in the newest download (I think) doesn't work either, so something is actually wrong.
Thanks for the explanation of the two files. I assume, then, that the 'kernel-power' file would contain any 'make default' changes I've made. Is this correct? And what if I delete it entirely? What happens then? Defaults to the original maemo power management file - pmconfig - I presume? Quote:
|
Re: [Announce] kernel-power v48
@efeinblatt,
kernel-config runs at every boot and loads, by default, /etc/default/kernel-power. If you remove it, it will load the one found on /usr/share/kernel-power-settings/default. (on that folder you have the various defaults, such as ideal, lv, etc.) pmconfig has nothing to do with it (at least in this context). kernel-config doesn't use or touch it. |
Re: [Announce] kernel-power v48
Thanks reinob, this at least clarifies the order of events for me. But I seem to remember that the kernel-power file had altered parameters in it. Is it possible that qcpufreq writes to it when 'saving'? Can't figure out how else the changes could have been made.
On the other question, any insight why a deinstall won't work? |
Re: [Announce] kernel-power v48
Not so sure about this kernel-power file. It says do not edit. Make a copy if you want to customize it. So, where do the original values come from? And if you want to create a new file with, for instance, smart reflex enabled, what name would you use to save it as default?
|
Re: [Announce] kernel-power v48
Quote:
Btw, I assume you *know* what you are doing. |
Re: [Announce] kernel-power v48
Thanks maacruz, but if you read through my earlier posts today you'll see that I've done that and you'll see what message I get back:
Reinstallation of kernel is not possible, it cannot be downloaded. Reinstallation of kernel-flasher is not possible, it cannot be downloaded |
Re: [Announce] kernel-power v48
Kind of know what I'm doing, but I have to admit that I don't know what dpkg does. Can you run me through this? Thanks.
|
Re: [Announce] kernel-power v48
Quote:
You could always disregard the warning and edit them, but it doesn't hurt just copying the one closest to what you want, rename it to whatever you like (e.g. "optimized" :), and edit it as you like. The best is if you then put it in the /etc/default/.. folder, so it will be loaded automatically without having to touch any other script. I'm not sure where/how qcpufreq stores its defaults. I suggest removing that package, as normally you set your parameters once and then forget about it.. BTW, no idea about deinstalling. I guess the bulletproof method is to flash the kernel (and *only* the kernel). But this will not help with reinstalling, as the package system will still think that KP48 is installed.. |
Re: [Announce] kernel-power v48
hey i have a question..im install android.. when i reboot theres have 2 option.. maemo omap and nitdroid.. when i choose maemo omap its set to default the question is how to set again to kernel 48.. beacause i instaled kernel 48 but it didnt load..
|
Re: [Announce] kernel-power v48
has anyone had an issue where the min limit goes back to 125 after a phone call?
im using the ondemand govenor and have 125 and 250 in the avoid-frequencies file. i dont think it happened on kp47. |
Re: [Announce] kernel-power v48
Quote:
the telephone app is closed-source and broken. After a phone call it sets the maximum to 600 Mhz and the minimum to 250 MHz (or 125 MHz if available) irrespective of what you have set before. The are some workarounds in the Kernel-power kernel. |
Re: [Announce] kernel-power v48
Quote:
Can you "prove" that your CPU is actually going down to 125Mhz, or is that merely the output of "kernel-config show". If the frequency is being avoided, it doesn't matter if the minimum is set to that frequency. Check with e.g. conky and see how low the CPU goes with nothing else running (or use powertop if you know how to). AFAIK most people avoid 125Mhz but allow 250Mhz, which doesn't seem to harm. Why do you need to avoid both 125 and 250? |
Re: [Announce] kernel-power v48
Using 500 mhz as minimum is more "productive" in terms of powersaving and avoiding wear on CPU (although, the second effect is probably very little). When You're using 500 mhz as min, Your CPU spend much more time @ 0 mhz during idle (every second, it's sitting ~2x longer at 0 mhz, after doing task @ 500 mhz), so you got additional benefit of saving on power leakage.
so, probably it's not about harm, but additional benefit ;) It's also possible, that he doesn't have idea of power leakage etc, just like to have 500 mhz as starting point ;) |
Re: [Announce] kernel-power v48
Quote:
Quote:
|
Re: [Announce] kernel-power v48
Quote:
One could also argue that fixing the frequency to, say, 600Mhz (or 900Mhz or whatever) would also provide a benefit (race-to-idle and so), so the phone would only alternate between 0 and 600. On the other hand, there may be background tasks (e-mail sync, etc.) which hopefully can manage to do their thing without forcing the CPU to jump to higher frequencies (e.g. do the e-mail check at 250Mhz), in which case you would save a bit of power, especially in combination with IGNORE_NICE_LOAD. When I have some time (lol) I will test with powertop what happens when I set the minimum to 600Mhz instead of 250Mhz and keep the phone idle with WLAN (or EDGE, I normally don't use 3G) and Nokia Messaging running (that's my usual "idle" situation). |
Re: [Announce] kernel-power v48
Nice idea, reinob (tests). As for locking to (for example) 900 mhz, I would not recommend this, due it being "overclocked" (500 mhz is standard frequency). It may be just placebo, but I would not like to "stress" CPU with higher voltage and frequency for 100% of running time.
As for things that doesn't force frequency jump, it's a matter of fact, that on 250 mhz it take 2x more time than on 500 mhz (usually - we're talking about things that doesn't involve high I/O from disk etc). so you benefit from more idle time = no leakage during 0mhz period. At the same time, een on 500 mhz CPu isn't overclocked (600 mhz is stock max frequency, but it *is* factory overclocked, according to CPU specifications). /Estel |
Re: [Announce] kernel-power v48
installed v49, all is fine
no need to pin |
Re: [Announce] kernel-power v48
Quote:
|
Re: [Announce] kernel-power v48
@Pali - when do you plan to release new KP version as was agreed?
|
Re: [Announce] kernel-power v48
UNTESTED PRE v49 is on http://atrey.karlin.mff.cuni.cz/~pali/
|
Re: [Announce] kernel-power v48
Quote:
thanks man. |
Re: [Announce] kernel-power v48
Quote:
I saw it and install directly, no need thinking |
Re: [Announce] kernel-power v48
Quote:
|
Re: [Announce] kernel-power v48
Conky also indicates that the frequency has changed from my 500mhz lower limits, because after a phone call it cycles between 250 and 805
|
Re: [Announce] kernel-power v48
Quote:
|
Re: [Announce] kernel-power v48
Quote:
recovery-boot: http://talk.maemo.org/showthread.php?t=72726 it has nothing with kernel-power v49 |
Re: [Announce] kernel-power v48
@pali
bq2700 no working anymore? |
Re: [Announce] kernel-power v48
Quote:
Quote:
apt is a high level CLI to package management, while dpkg is the low level interface used by apt, akin to rpm. But since you don't know what dpkg is, I warn you that using dpkg not knowing *exactly* what you are doing is dangerous. |
Re: [Announce] kernel-power v48
Quote:
I've had mine set to 500 as the base for some time. There were also other cute tricks they tried to "trick" the phone app (changing the kernel speed to 605 vs 600, etc). Sadly, the only real "solution" seemed to be to re-load kernel config after a call hang-up. There was a hack put in to Titan's kernels to do that as I recall, but I don't know if it's still there? I'd assumed Pali just took the bundle from Titan's stuff as a work base, but that may not be a good assumption. |
Re: [Announce] kernel-power v48
Quote:
|
Re: [Announce] kernel-power v49
kp49 has ramzswap, nice :D
Let's see how that works @pali: Just a question, have you modified the swap notify patch due to the extensive modifications Nokia did to the swap code, or is it the original swap notify patch. |
Re: [Announce] kernel-power v49
Quote:
On this part, the patch come from freemangordon and it is 'patch_swap_notify_core_support_2_2.6.28.diff' from here : Compcache Post #1 Hope it help ... A++ |
All times are GMT. The time now is 04:11. |
vBulletin® Version 3.8.8