I did it the hard way before i read your answer. I deinstalled the kernel power which reverse back to stock kernel then download all of the above deb files and install like your exact order but i skipped the headers file. is it necessary to install the header file? since it seem to reboot and work just fine without it.

I have updated my original post here http://talk.maemo.org/showpost.php?p...&postcount=204 with new version of DSP patch, this one should not break any multimedia functionality.

Can this be installed on v 47 ? And how ?

It is a patch (code difference), not an installable package, wait until pali include it in some next version of KP.

Edit: and btw do not expect that with this patched kernel your n900 to be able to play or record 720p. It will be a little miracle if it still works as before this patch, having in mind size of code changed.

Headers file doesn't need to be installed unless you intend on compiling kernel-linked programs on-device, as I vaguely understand it.

That was what I thought it was....since I saw that package header wasn't a dependency of any other packages. Thanks for clearing that up for me.

Flashing is definitely faster solution if you got correct backups, but it is worse solution if You want to track down source of problems

You're probably much more experienced in overclocking N900 than me, but i just can't agree with that. Have You traced (by simple logging script or whatever) CPU usage before and after overclocking, in tasks that you feel are snappier than before? I did, and unless You force N900 to stay at higher frequency all the time, it just DON'T jump into higher frequency, because it DON'T even need to - ondemand governor need set CPU load (for example, 90 %, you can change it) before it can request higher frequency. So, unless there is a bug in maemo at all (which can't be called improvement in any condition), this must be placebo effect. Unless You or someone can show hard proofs - I. E. "scientific" (logged) data with explanations "why it is faster". Of course, if people just feel their devices faster, even it is placebo, and like to overclock it to 1000 or 1150 just because of that - I'm ok with this. It just not my business

Also, i CAN'T agree with statement that shorter lifetime is just theory. I got ~10 years of experience in overclocking desktop machines, from things like basic homemade watercooling, through rreone-based devices, to Dry Ice things. And i just can't stand spreading misinformation about how overclocking passive cooled device have no impact on lifetime (which is main reason of this post at all, by the way). Higher electro-migration occur even when CPU isn't stressed enough, but got higher voltage (higher than producer advise, not the stock settings, that ARE overclocked even in brand new N900) and operate in higher temperature. As far as i remember, it is explained in details in Wiki, so I'm not going to repeat all hard data here.

Keep in mind that I'm also sceptical about bad/damage effects on Your N900 if you "overclock with brain". That's why i use overclocking - but people tend to misunderstand basic fact about N900 design. It is EVERYTHING but not cooling-friendly. As every Cortex, it is practically not cooled at all, also case is thermal isolator. Theoretically You can achieve miracles by assembling Your own case from scratch, using galvanized cooper (or in fact even galvanized aluminum case would be great improvement over plastic-like design in terms of thermo-conductivity) - i even though about doing it with my N900 myself - but this is pure hobbyist thing, without major real-life advantage, not to mention price/gain factor.

Making this long story short, overclocking with proper on-demand governor AND not too high AND without long-time stressing N900 in times of ambient temperature ~30C/bad around-case ventilation should be relatively safe. Overclocking to stable max and increasing voltage too much, on the other hand, can and probably WILL (yea, i know about how ridiculous in terms of grammar this sentence is :P ) produce bad results, and i presume that some people are going to feel it in 2-3 years from now. If not felt it already - I remember there was a topic from guy who was overclocking his device to 1000-1150. He started receiving N900 total unusualness due to effects similar to CPU end-of-life. He send his N900 to Nokia Care and even thought about clearing kernel-flashing records before that, but dropped this idea cause it's probably impossible without hard-bricking device [unflashable] - unfortunately, he didn't write how this story end.

It's all about operating temperatures, I'm afraid, and temperature is rising by n*n factor with every overclocking/overvoltage.

Estel

Ps.

Now i REALLY feel spamming this topic - doesn't all this discussion belongs to "Overclocking N900" discussion?

sorry but i dont know as if i am using the older version of power kernel ,i tried to check it out ,in conky i saw it was shown as:
linux 2.6.28.10power46 on armv7I
i love n900 but i really dont know about programming etc ,so just wanted to know as if i am using older version or ?

Yep, 46 is the previous version, crafted by Titan.

thanks but i think i have updated my n900,hopefully it shows now that it is v47 ..still updating ))