Re: [Announce] kernel-power stable v53 in Extras Testing
 

Well well well ... now my phone doesn't want to start ... but actually I'm not even sure if it's related to the KP update.

Here is the story :

Yesterday, I discovered there was a new CSSU-thumb update. And because I have also seen the KP update, I thank i could do a full update.
So I started by the CSSU update, and then rebooted the phone. It seemed to work, at the exception of the camera. The switch didn't activate the camera-menu, and all of the applications which use it show a black display. But I thought I will be able to check this later.

After that, I tried to perform the KP update, with backupmenu. I have also U-Boot installed, so I did it in the following way :

1. in first, I removed U-boot, and the KP52 u-boot image
2. then I updated backupmenu
3. next, I updated KP
4. after I reinstalled U-Boot
5. I installed the KP53 U-Boot image
6. and finally, I rebooted my phone, but it failed.

Now, when I try to start my phone, I have the U-Boot menu, and if I try to launch the attached kernel, I can see that I launch the zImage-2.6.28-omap1 kernel. After there is only a black screen.

Does somebody has an idea for fixing my phone ?

Re: [Announce] kernel-power stable v53 in Extras Testing
 

Why you removed uboot at first step? And why do not boot kp53 from uboot instead attached kernel?

Re: [Announce] kernel-power stable v53 in Extras Testing
 

In u-boot you see other options?

Else a solution could be to flash the kp53 kernel only (flash-only=kernel). Then you will loose u-boot but phone should start...

Re: [Announce] kernel-power stable v53 in Extras Testing
 

Hey, I installed yesterday the last CSSU version 21.2011.38.
Before that I used the stable version.
Suddenly I didn't have wifi connection and the device couldn't see any wlan networks.
I use usually the device without sim. I booted the device several times and took the battery off. I couldn't see any wlan.
I also changed wlan settings and took the batterysaving off, no result.
Then I put the sim card in the device and after that the device worked and could find wlan networks.

I havn't installed the scripts.

After removing the simcard the device couldn't see any networks

Re: [Announce] kernel-power stable v53 in Extras Testing
 

It looks like regulatory settings... Maemo system set wifi regulatory zone from cellular information.

Re: [Announce] kernel-power stable v53 in Extras Testing
 

Why the stable version worked? I didn't have simcard and everything worked.

Re: [Announce] kernel-power stable v53 in Extras Testing
 

hmm... maybe something other... do not know

Re: [Announce] kernel-power stable v53 in Extras Testing
 

Because when I install a new kernel, it will flash the nand and there will not be u-boot during the start of my phone. So I have to reinstall it.

In the U-Boot menu, the only choices I have are "Attached kernel", "Internal eMMC", "External SD Card", "U-Boot boot order", and "U-Boot console". How can I choose the kernel ?

Re: [Announce] kernel-power stable v53 in Extras Testing
 

This means that U-Boot cannot read config file which is stored in MyDocs FAT32 partition. There are some problems with tools (from MS too!) which format FAT partition incorrectly.

Quick fix is:
1) Boot kp53 via flasher-3.5 or 0xFFFF without flashing kernel image
2) Backup all data from MyDocs (if is still readable and not corrupted)
3) Format MyDocs (on Maemo or via USB mass storage mode with mkfs.vfat)
4) Copy backup back

Re: [Announce] kernel-power stable v53 in Extras Testing
 

The bootmenu.scr file is empty.

Re: [Announce] kernel-power stable v53 in Extras Testing
 

sudo u-boot-update-bootmenu

Re: [Announce] kernel-power stable v53 in Extras Testing
 

are we sure it is a problem? I mean - does mce in R1.3 behave differently?

anyway, I'll check what stock mce does as soon as I get back to my desktop.

EDIT: code in stock mce and mce in cssu is 100% identical, it must be something else

Re: [Announce] kernel-power stable v53 in Extras Testing
 

I forgot to say: I was running latest CSSU stable (no thumb) and kp52 before the proximity stopped working (no u-boot at all).

Only packages updated were kp and kp settings and after that proximity stopped working.

I don't know how can I help to debug this, in case tell me useful informations I could provide (logs, cat, etc..)

Re: [Announce] kernel-power stable v53 in Extras Testing
 

I also have no wifi w/o sim card

Re: [Announce] kernel-power stable v53 in Extras Testing
 

I was unable to insmod wl1251_spi.ko due to Unknown symbol crc7

Re: [Announce] kernel-power stable v53 in Extras Testing
 

see first post where is link to packet injection load script

Re: [Announce] kernel-power stable v53 in Extras Testing
 

Let me guess. Your WiFi is on channel 13, right?

Re: [Announce] kernel-power stable v53 in Extras Testing
 

I have tried the script, but possibly older version. I also have the same error using interface switcher applet(can not remember exact name now).
channel doesn't matter because I can not load all driver modules so no interface to interact with channel.

Re: [Announce] kernel-power stable v53 in Extras Testing
 

I resolved uninstalling and installing again packages

- shortcutd
- proximityd (shortcutd alone did not resolve)

I can confirm no event is shown in dmesg, although shortcutd now is working fine;
while triggering proximity during calls results in dmesg showing proximity events.

So my previous post reports normal behaviour, it's not a bug. :o

Re: [Announce] kernel-power stable v53 in Extras Testing
 

For the sake of knowledge, I had same problem reported above upon rebooting.

I found the problem was related to proximityd not starting on boot (shortcutd needs proximityd running).

I solved changing this line into /etc/event.d/proximityd
to this one (I saw shortcutd starts this way)

Now proximityd starts properly on boot and shortcutd works good.
This change made necessary since upgrade to kp53, as reported previously.

Re: [Announce] kernel-power stable v53 in Extras Testing
 

Hello,
I am using this one with U-Boot. It works great, but can I overclock the device without losing U-Boot? As I see, kernel-power-settings has kernel-power-flasher in dependencies. If I will install it, that will reflash NAND and I will lose U-Boot, right? Is there any way to overclock the device without losing bootmenu?
Waiting for the answer, thanks in advance.

Re: [Announce] kernel-power stable v53 in Extras Testing
 

Nope, at install stage you will be asked to reflash NAND or not. And you can cancel reflashing process, so NAND will be untouched.

Re: [Announce] kernel-power stable v53 in Extras Testing
 

It worked like charm. Thanks!
UPD: Well, that really worked. But now I get this error during installation/deinstallation in faster application manager. I tried to do whatever it says, but it proposes me... to flash the NAND, yes!

Re: [Announce] kernel-power stable v53 in Extras Testing
 

Read first post. Use HAM or apt-get.

You can skip and do not accept flashing to NAND (click outside of window).

Re: [Announce] kernel-power stable v53 in Extras Testing
 

Maybe I've descibed it bad, but it works like that:
1) I type "sudo apt-get install kernel-power-settings".
2) It asks me, do I really want install it. I type "y" and press Enter.
3) The message about reflashing shows, I press on the blurred area.
4) ????
5) I can change the frequency of my CPU, but can't install anything from any package manager cause "unmet dependencies: kernel-feature-smartreflex is not installed". After reboot my settings turns back to default [250, 600], even if I save current [250, 900] state as default.
6) I type "sudo apt-get remove kernel-power-settings" when I want install something else.
7) RAGE!

Re: [Announce] kernel-power stable v53 in Extras Testing
 

And are you booting kernel-power? Or not?

Re: [Announce] kernel-power stable v53 in Extras Testing
 

Surely yes.

Re: [Announce] kernel-power stable v53 in Extras Testing
 

Anybody? Guys?

Re: [Announce] kernel-power stable v53 in Extras Testing
 

bumpity-bump-bump

Re: [Announce] kernel-power stable v53 in Extras Testing
 

Please install kernel-power-settings via apt-get (sudo apt-get install kernel-power-settings) and send xterm output here.

Re: [Announce] kernel-power stable v53 in Extras Testing
 

Being back on N900 and hitting against old problems like low memory :) and eternal swap time
Did somebody ever think to compile power_kernel with page_size set to 32K? ARM architecture supports this, and on my uSD this should mean (for me, testing with flashbench) 5x better transfer speed in swapping to a uSD partition...
Crazy idea? In case, where to find kernel-power sources and how to setup scratchbox for Kernel compilation?

Re: [Announce] kernel-power stable v53 in Extras Testing
 

Read first post! In that you find link to development thread :-) where are also steps how to download & compile kernel-power yourself.

http://talk.maemo.org/showthread.php?t=78371

Re: [Announce] kernel-power stable v53 in Extras Testing
 

Thanks! I was missing the source dev repository in my /etc/apt/sources.list and of course apt-get was not installing the sources...
I'll report as soon as I have time to compile/install/test

Re: [Announce] kernel-power stable v53 in Extras Testing
 

Really 5x? Seems like massive improvement, why nobody thought about that before? Or did one?

Re: [Announce] kernel-power stable v53 in Extras Testing
 

Can we have the card name and some absolute speed numbers for 4k and 32k please. On which device did you run the tests .. we need infos ;)

Re: [Announce] kernel-power stable v53 in Extras Testing
 

Results from a Transcend 8M class4 uSD card:
Results from a Kingston 4M class4 uSD card
ARMV7 supports 4k, 64k + 1M, 2M and 1G page sizes with LPAE implementation.
With the second card, moving from 4k to 1Mb is 5x performance, but with the first card to move from 4k to 64k (no LPAE) would mean 10x.
Changing page_size is not an easy task, though... I am dirtying my hand into kernel source and I really feel the task is a bit hard to accomplish and over my capabilities. Specially with so little time to dedicate :(
There is a patch called transparent hugepages floating around, but it will not beneficial since we need to look for and modify how the pages are phisically swapped in and out from the block device.
This is where I am standing right now.

Re: [Announce] kernel-power stable v53 in Extras Testing
 

My feeling is that it doesnt improve the situation:

Fragmentation may looses ram:
http://yarchive.net/comp/linux/page_sizes.html

mb/s isnt latency:
If your software needs random read like swap, you cant convert it to sequential by just changing the kernel page size.
(write isnt that important cause you can change the swappiness to swapout ahead of time (at the cost of battery life), the kernel first overwrites code pages which he just rereads from the filesystem (no need to write them, they are already saved) and so most swap operations are random reads (which are slow even compared to sequential(!) writes which you can do when you actually do swap out other data) (But this may depend on your swap policy))

I hope swap gets already paged in and out in multiple Blocks, no need to recompile:
http://www.gluster.org/community/doc..._Kernel_Tuning
"cat /proc/sys/vm/page-cluster"

But maemo + kernel seems so old (and sadly not good at io) that i found sparse, outdated and even contradicting info.

ram via sd would be nice ;)
But not as nice as better software.

But i would like to be proofed wrong. So post real world experience!

Re: [Announce] kernel-power stable v53 in Extras Testing
 

I'm confused. I need a real simple step-by-step to start with :confused:

So this is what I think I need to do. Please correct me if I have it wrong!

1) From Hildon Application Manager install package: Linux kernel for power user.

2) That's it, right? Then I just reboot?

Now, I do have some sort of choices on booting, but I don't think I have U-boot. I think I installed some kind of multiboot. (I was playing with nitdroid, I think. It's been a few years. :o ) So, do I need U-boot? Not need U-boot? I seem to remember U-boot had some issues where you had to uninstall something else. ??

As you can see, confused! If you can help, I'd be super-grateful!

Re: [Announce] kernel-power stable v53 in Extras Testing
 

It's ok man.
A fellow Maemoan wise in what you need explained... will help.
Give it a bit of time for those fellows to notice.

Most explanations here are geared for either a modest or high ( or if you wish to call it "deep") understanding to begin with to fathom, appreciate and execute some desired results...
I do believe in making things as easy as possible in explanation...
as if explaining a process always to someone completely new to our device lines...
It may make such explanation long winded....but everyone understands it.
And god knows I need my own walk-throughs at times..since it is so rare that I do some processes (due to the stability of our devices and platforms) I forget how to do them!!! So I have had to resort to my own walk-throughs many times myself...

I must admit that in some areas here we truly do need "Noob to Pro" written and / or video guides..
It may take some effort to create them ( I know...I've done them... ) ...
but once created...
such walk-throughs are invaluable...
such prepared guides just run all on their own...
like little engines :)
easy to always refer to...
Saving frustration on the side of the ignorant ( not "ignorant" in the derogotory sense...but "ignorant" as in "not-knowing" sense )

and it saves the perpetual explanation and exasperation on the side of those explaining....

and saves both A LOT of time better pursued doing other things...

More and more I believe such guides are necessary as more and more new people come to our device lines with a great desire to understand ...but have a problem with serious technical terms and methods (which many of us here understand and take for granted everyone "must" also understand...but there was always a time when we all "didn't" get it...so compassion and understanding metes out and you will get a suitable explanation I am sure my good sir..) ...

This could be easily remedied by a step by step walk-through written for all levels of understanding to comprehend ...best is both written and video combined...

But unless it was a "given" protocol here..
it is really up to the creator of "...." to decide their approach to explanation...

Re: [Announce] kernel-power stable v53 in Extras Testing
 

You should first tell us what you want to do! :)

When you install a kernel (step (1) you mentioned above) it is placed in a special partition so that the boot loader (NOLO) will run it.

U-boot acts like a kernel, i.e. it is stored on the kernel partition so that NOLO runs it. U-Boot can then run an attached (embedded) kernel (usually the default Nokia kernel) or another kernel of your choice, being installed somewhere else (another partition, Mydocs, sdcard, whatever).

For completeness:
Multiboot is for all I (we) care deprecated. It basically replaces the installed kernel (i.e. in the kernel partition) with whatever kernel you choose when booting. I've never actually used it, and I don't think there's any valid use-case for it. U-boot does that and much more and much more reliably.

Then there's bootmenu and backupmenu. They are not kernels but they run very early (first thing after the kernel has done its initialization). You can use bootmenu and/or backupmenu independently of which kernel you boot (Nokia kernel, kernel power, or U-boot with some kernel).

In short: if you now install kernel-power it will install itself in the kernel partition (which for you is OK because you don't want/need/know-about U-boot). If you have multiboot it will continue to appear but I'd advise you to remove it. For this, please ask again (or make a new thread, or search here). I believe you only need to replace /sbin/preinit but as I said I've never installed/used/uninstalled it.

Good luck.

PS: Before you do anything, make sure you have a working set-up for flashing the N900. That means the firmware, the flasher and a computer than can actually work with that. If in doubt, ask first.