[Announce] kernel-power stable v53 in Extras Testing
 

Finally kernel-power version v53 is now in Extras Testing!

Update (from old version):
Use Hildon Application Manager or via apt-get.

Installation:
From Hildon Application Manager install package: Linux kernel for power user
Or via apt-get package: kernel-power-flasher

Overclock support (optional):
From Hildon Application Manager install package: Linux kernel for power user (settings and overclock)
Or via apt-get package: kernel-power-settings

U-Boot support (optional):
From Hildon Application Manager install packages: U-Boot with kernel 2.6.28-omap1 and Linux kernel for power user (boot image for U-Boot)
Or via apt-get packages: u-boot-flasher and kernel-power-bootimg
Make sure you accept flashing U-Boot!

Packet injection wifi driver:
Kernel modules are already included and installed in /opt/packet-injection-modules/2.6.28.10-power53/
To load/unload it use shell script: http://atrey.karlin.mff.cuni.cz/~pal...t-injection.sh
Support for IBSS (Ad-Hoc) mode is now fixed.

BME replacement support:
This kernel has support for BME replacement (open source battery management). More info in this thread.

New USB host mode:
This kernel has support for new usb host mode package USB Mode version 1.0-4. More info in this thread.

Uninstallation:
Overclock support or Bootimg package can be uninstalled via Hildon Application Manager or via apt-get. Uninstalling kernel itself is possible only via menu entry "Nokia Kernel". NEVER UNINSTALL KERNEL POWER PACKAGES!!! IT CAN MAKE YOUR DEVICE UNBOOTABLE.

===

Changelog:

wl1251-maemo (0.3) unstable; urgency=low

* Include IBSS fixes patches

-- Pali Rohár <hidden> Fri, 09 May 2014 19:58:51 +0200

kernel-power-settings (0.18) fremantle; urgency=low

* start upstart script after mounting /home and /opt
* upstart kernel-modules file is service

-- Pali Rohár <hidden> Fri, 09 May 2014 17:58:06 +0200

kernel-power (1:2.6.28-10power53) fremantle; urgency=low

* Backport security fixes from upstream kernel for radio driver in patch radio-bcm2048-security.patch
* fix sign of value in bcm2048_##prop##_read()
* fix missing unlock on error in bcm2048_rds_fifo_receive()
* fix coccinelle warnings: Replace memcpy with struct assignment
* fix unneeded semicolon in bcm2048_fops_read()
* Fix dangling if statement in patch power-supply-no-verbose.diff
* Backport usb ids from upstream kernel for dm9601.ko and asix.ko drivers
* Add patch nokia-av_key.patch for immediately headset button press event as KEY_PROG1
* Remove shm_size=0x400000 from /etc/modprobe.d/bridgedriver because it is not needed
* Update patch overclock_smartreflex_900.diff to improve the stability on some rare devices
* Fix suspend to memory problems, generally only useful in offline mode
use /sys/power/state to enter (not used by Mameo5, just less broken now)
* advance system clock after resume using RTC time
* computer USB port now is able to be detected and charge after resume
* disable WARN_ON in getnstimeofday when timing system suspended

-- Pali Rohár <hidden> Sat, 10 May 2014 22:17:57 +0200

===

For questions about development, bugs, etc... write to thread
[Development] kernel-power on http://talk.maemo.org/showthread.php?t=78371
Here is also information how to compile kernel, send patches, ...

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

Thanks pali.

I am a bit ashamed as I am using this kernel already almost half a year (and did not give any feedback to you :().
But now: everything is fine. Never ran into any kind of problem.

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

I have not done any changes since May, so version in Extras-Devel & Testing should be same, just (re)compiled by Maemo infra.

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

Thanks @pali

I can confirm that apt-get upgrade'ing from kp52 to kp53 works fine. Note that if you use U-boot and you want to keep it, you have to cancel when asked if OK to replace u-boot with kp53!!!

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

And I can confirm that an upgrade through FAM works just as well ;)
I had to reboot manually though to load the new kernel.

There is a new dependecy package, fiasco-image-update-ask. Why is it needed and why is it still needed after the reboot? I could not uninstall it, everything depends on it.

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

-flasher package depends on it. and that package cause user will be asked if some SW/script try to flash *other* kernel (non kernel-power) into nand. It is protection against removing u-boot or kernel-power.

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

Hmm, that does not seem to work...
Currently I'm at kp52 and Extras Testing is enabled.

HAM does not show any update available at all and after "apt-get update" "apt-get upgrade" says:
So here I pressed "n", but what could be the reason for this situation?

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

broken dependences. try to use aptitude upgrade

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

Hi guys. Hope U can help me...

Updated to KP v53 via HAM. Worked like a charm. Restarted device...and it won't boot up. It vibrates, then the Nokia logo appears and than it shut itself off. I have removed battery and tried again, got the same results.
Even have tried to flash it but my computer didn't recognised it. It says "Suitable USB device not found, waiting.".
It looks like it strugles to establish connection to a PC or a charger cuz from time to time it vibrates and than yellow led shines until it happens over and over again. Battery is fully charged and usb port is in order. Don't know what to do, please help.

Edit:

I have succeeded in flashing it. Will install everything and try again.

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

it's probably just the fiasco-ask-fiasco :) so just doing "apt-get dist-upgrade" should do it.

Actually, I didn't even know we had aptitude in the repos :)

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

aptitude upgrade succeeded, thanks! :)

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

NO, NEVER CALL dist-upgrade. On maemo it is broken.

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

Slight correction:
Maemo is (and has always been) broken. dist-upgrade works just fine if all dependencies, conflicts and transitions are declared correctly.

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

Also, I have another problem...when I installed KP 52 or 53 my multimedia got broken (images got distorted as well as video that got some crackling sounds when recording). It worked normal on nokia's kernel. Does anybody have an idea what that could be?

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

Not sure if this is related with KP53 or not, but after installing it, shortcutd proximity doesn't work anymore. However, been running for a day and it's been running stable.

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

Via apt-get use the command kernel-power-settings not kernel-power-flasher

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

@pinochio, kernel-power-settings is an independent package that just happened to be updated at the same time. To upgrade the kernel, as pali says and update kernel-power-flasher.

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

I install (Update from old version) from Hildon Application Manager package Linux kernel for power user, but first I had to uninstall Linux kernel for power user (boot image for U-Boot).
Then in the terminal I install kernel-power-settings (v. 18).
Is it right?

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

aptitude works fine in CLI but bombs for me using ncurses. Always has.

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

"computer USB port now is able to be detected and charge after resume"

Can you please clarify "after resume"?

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

https://gitorious.org/community-ssu/...942df84c9eae6f

Maemo system does not use suspend mode, so basically this patch is only for "power" users who manually set n900 to suspend.

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

Please vote for packages so they can be moved to Extras:
http://maemo.org/packages/package_in....28-10power53/
http://maemo.org/packages/package_in....28-10power53/
http://maemo.org/packages/package_in....28-10power53/
http://maemo.org/packages/package_in....28-10power53/
http://maemo.org/packages/package_in....28-10power53/
http://maemo.org/packages/package_in...settings/0.18/

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

Also fiasco-image-update-ask needs votes becase kernel-power depends on it:

http://maemo.org/packages/package_in...pdate-ask/0.4/

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

It did not take long ;)

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

But this one needs one or two more ...

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

Next time, can we try to encourage people to "test" the packages rather than just "vote" for them.

I know it's Pali and there's probably nothing wrong with his packages but there is a QA procedure to follow. Otherwise we might as well do away Extra-testing and allow developers to promote packages directly to Extras.

There is the problem that there aren't many testers nowadays so perhaps we need to organise more testing marathons and/or make more users super-testers.

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

Well said I couldn't agree more.

I started trying to push package promotion a while back but it took forever to get each package promoted, I was posting to myself most of the time, so the thread slowly died.
http://talk.maemo.org/showthread.php?t=90981

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

Just what I meant here.
Regarding the kp packages there are at least 5 persons that I would trust blindly (me included ;)), so voting means testing here.

But I also agree the wording 'voting' is totally misleading regarding the original meaning of testing/confirming working state of the package.
Where does this wording come from?

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

Can somebody else do tests of (fixed) packet injection drivers for normal usage (also power usage, etc)? Now Ad-Hoc (IBSS) mode should be fixed. I'm using only packet injection drivers and for more that half of year everything working fine... So now I'm starting thinking about removing original wl1251 drivers from kernel and provides only packet injection drivers in kernel-power-modules package... What do you think about it?

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

What kind of tests you would like us to do?
I know (erm, knew) only about power consumption problem when having injection drivers enabled and leaving range of wireless network.
What about ad-hoc mode? Never tried with that driver.

What would be the benefit to dismiss original driver? (Saving a few bits on rootfs (modules already on /opt), or?)
For sure a few here will run into problems when using cleven, faircrack and so on...

I am open to it and will enable injection capable drivers only to see

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

In kp52 packet injection driver cause kernel panic and reboot when wifi was set to ad-hoc mode... This should be fixed. Power problems could be fixed too. But both needs to be tested by more people if it is really stable...

Benefit is to not have two (or more) drivers for wifi and unloading/loading when you need to switch between them. Currently for using packet injection by default is needed some hack in /etc/modprobe.d/ (which my script in first post is doing). And it would be great to get rid of it...

Userspace tools for using packet injection drivers can be fixed too if there are problems...

But first I need tests and discussion about possible problems...

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

I think I have to report a problem.
Enabled bleeding edge via advanced interface switcher and after leaving house (wifi range) I did not detect any higher power consumption. But when getting home there was no detection of my wlan nor was any other wlan shown. So I restarted icd2 as well as wlancond to no avail. Only stopping and restarting wlan from aifs applet let the wlan search succeed. :confused:

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

@peterleinchen: thanks for testing. can you use also script from first post for enabling packet injection drivers? maybe there can be problem with other applications/applets (as process of replacing drivers is not simple).

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

Yep, I can do so.
But I believe aifs does it already with same mechanism.

Just a few minutes happened again, wil enable now stock drivers and switch with your script.


--edit
here are aifs scripts
wlan_load0 stock
wlan_load1 bleeding
get_wlan_path checks availability

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

That would be great! No more fiddling with switching between drivers to get pocket injection or save power! :)

And doing so would in a way force people using kernel power to use this version of drivers, so bug hunting would be much much faster (more testers!).

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

Hey pali,
just went out (into the cold ;)) to check.
I have done:
script unload
script load
script enable

And same behaviour.
I had to do another
script load
to get my wlan connectivity back :confused:
(and had high power drain after doing so)

--
you may also check aifs scripts (see my earlier post)

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

Same issue here.. tried to reboot several times, disabling and reenabling shortcutd proximity sensor option in Settings, but no way to get it working again..
Any ideas on how to fix it? :confused:

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

Maybe this is related with shortcutd not working: in dmesg I see proximity log only while making a call;

otherwise I can trigger proximity sensor but no log results in dmesg.

p.s. happy new year to every maemo user in TMO!! :) :)

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

Now I looked at this problem... and proximity gpio-switch is for unknown reason disabled in kernel. I looked at diff between kp52 and kp53 and there is nothing which can disable gpio-switch. So problem must be somewhere in userspace. Maybe CSSU or U-Boot? Maybe other application?

This sysfs can enable or disable proximity gpio-switch:
$ cat /sys/devices/platform/gpio-switch/proximity/disable

And this command enable it (until reboot):
$ echo 0 > /sys/devices/platform/gpio-switch/proximity/disable

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

So it is problem of mce which is in cssu, see:
https://gitorious.org/community-ssu/...witches.c#L324

It disable proximity gpio-switch if call is not active or not ringing.

@freemangordon: what to do now?