maemo.org - Talk - [Announce] kernel-power stable v49 in Extras-Testing

maemo.org - Talk (https://talk.maemo.org/index.php)

- Maemo 5 / Fremantle (https://talk.maemo.org/forumdisplay.php?f=40)

- - [Announce] kernel-power stable v49 in Extras-Testing (https://talk.maemo.org/showthread.php?t=71879)

Re: [Announce] kernel-power stable v49 in Extras-Testing

Quote:

Originally Posted by woody14619 (Post 1166367)

One thing I find odd though, why is fmtxd running on your system at all? On my N900, the only time that runs is when FM has been enabled and it's actively transmitting audio. By default it runs at start up but goes away on it's own after a short delay. So the fact that it was even running implies to me that something was trying to activate it?

During my effort the track down battery draining programs and processes on n900 while it's idle I found out the same: the fmtxd was waking up the cpu all the time when it is not needed.
So I tried out the modified binary from here and the problem was solved- the fmtxd is now shut down completely when it is not used.

HTH and HAND

Michael

Re: [Announce] kernel-power stable v49 in Extras-Testing

@Pali, patches 0001-mtd-fix-a-huge-latency-problem-in-the-MTD-CFI-flash-.diff and 0002-mtd-change-struct-flchip_shared-spinlock-locking-int.diff are still enabled in series file, please disable them as it is confirmed those cause reboots/slowdowns.

Also I am unable to use QtMobileHotspot with USB, device reboots in 1-2 seconds after connecting USB cable. Please anyone to test if it is working for him with latest KP. I suspect it might be USB sysfs notify patch, will disable it and will test.

EDIT:
it is not USB sysfs notify which bugs the kernel.

Re: [Announce] kernel-power stable v49 in Extras-Testing

Quote:

Originally Posted by freemangordon (Post 1166571)

Also I am unable to use QtMobileHotspot with USB, device reboots in 1-2 seconds after connecting USB cable. Please anyone to test if it is working for him with latest KP. I suspect it might be USB sysfs notify patch, will disable it and will test.

EDIT:
it is not USB sysfs notify which bugs the kernel.

Running latest KP (including git updates from 2 days ago) and I'm writing this reply from laptop using USB QtMobileHotspot. No problems.

Re: [Announce] kernel-power stable v49 in Extras-Testing

Well, seems I am hitting this. Will prepare a patch and will send it to Pali (if this solves the kernel crash)

EDIT: Bug smashed, patch sent to Pali.

Re: [Announce] kernel-power stable v49 in Extras-Testing

Quote:

Originally Posted by pali (Post 1166340)

This is because that wifi driver is compiled against old kernel-power. For proper working each module must be compiled against actual kernel version. But there was nothing changed in wifi stack, so packet injection wifi driver worked. But now in kernel power are some changes (disabled mesh, ...) and binary wl1251 packet injection driver is incompatible with last kernel-power...

So, someone should build them against KP50

Re: [Announce] kernel-power stable v49 in Extras-Testing

I have tried the latest kp pre-50 and worked no problems for me with the exception of wifi lxps driver so I went back to freemangordons pre-50 and as usual I do a autoremove & autoclean only this time after autoclean a lot of my aps with lots of lib files where deleted resulting in a mess. I could not get any updates as repositories reported errors ex: "sub-process bzip2 returned an error code (1)" and "applet not found" so I had to reflash my device. Not sure if there is a bug by reverting back to a previous KP you can lose files (being deleted) by trying to autoclean your device, can anyone else confirm the above.

Re: [Announce] kernel-power stable v49 in Extras-Testing

Quote:

Originally Posted by magic_doc (Post 1166561)

So I tried out the modified binary from here and the problem was solved- the fmtxd is now shut down completely when it is not used.

I think this may be a regional thing? One would hope that such a setting wouldn't change behavior like this, but all things are possible with closed-source blobs. :(

Re: [Announce] kernel-power stable v49 in Extras-Testing

Quote:

Originally Posted by freemangordon (Post 1166571)

I builded new deb packages. I disabled both patches + included that you sent me.

Re: [Announce] kernel-power stable v49 in Extras-Testing

Be prepared, long post of kernel-power-settings which is driving me crazy. Really kernel-power settings is gone nuts.

In this post I describe exactly what steps I took to get this result, and it's reproducable 10/10.

I already thought kernel power-settings was acting weird on me so I ran:

Code:

apt-get remove --purge kernel-power-settings

rm -r /home/user/.kernel

That last command was to be sure there were no bits left of my personal config files.

This part describes why I am posting this...
After removing kernel-power-settings, I installed kernel-power-settings again by issuing

Code:

sudo gainroot

apt-get install kernel-power-settings

Followed by loading the default config http://www.mediafire.com/imgbnc.php/...36d17c5e6g.jpg

I thought these voltages were quite weird, and yet it becomse even weirder, here the voodoo kicks in, when I recheck the config I just loaded with nano

http://www.mediafire.com/imgbnc.php/...7bab7c656g.jpg

WTF :confused:, other values in the config file than kernel-config show reports... Especially freq 500 and 600 are weird

I did not set any profile as default so why don't we reboot and let it load the original "nokia" config? Easy said easy done:

http://www.mediafire.com/imgbnc.php/...736d348d6g.jpg

Then load the default profile and check again. Seems quite normal, except still that voltage voodoo bug ( 550 and 600)

Let's make the default profile default so it loads at boot
http://www.mediafire.com/imgbnc.php/...8dffdb676g.jpg

Reboot...
Checking again
http://www.mediafire.com/imgbnc.php/...89ef13746g.jpg
Voltages seems normal atleast they are linear increased with frequency increasement, not the weird jumps as when loading it. Though it still makes me wonder where the vuck do these values come from as they are definitely not in the config file(see screenshot 2, the one opened in nano)

Re: [Announce] kernel-power stable v49 in Extras-Testing

@mr_pingu: breathe man, SR is enabled and you're seeing what SR has calculated for each frequency :p