maemo.org - Talk

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)

knobby 2011-09-22 16:53

Re: [Announce] kernel-power v49
 
yup...not in repository till now...

knobby 2011-09-22 16:55

Re: [Announce] kernel-power v49
 
hey guys...showing some error with cleven...?

dignified 2011-09-22 17:22

Re: [Announce] kernel-power v49
 
1 Attachment(s)
pls help me about this

fahadj2003 2011-09-22 17:28

Re: [Announce] kernel-power v49
 
Quote:

Originally Posted by dignified (Post 1094041)
pls help me about this

ure worser then a noob.
i refuse to help u.

poogash 2011-09-22 17:28

Re: [Announce] kernel-power v49
 
Quote:

Originally Posted by dignified (Post 1094041)
pls help me about this

sudo gainroot

knobby 2011-09-22 17:37

Re: [Announce] kernel-power v49
 
if u have rootsh package installed, then enter command sudo gainroot
then proceed further...

trisha02 2011-09-22 17:40

Re: [Announce] kernel-power v49
 
Quote:

Originally Posted by fahadj2003 (Post 1094043)
ure worser then a noob.
i refuse to help u.

lol..hahahahaha. at least i am far away from that stage

dignified 2011-09-22 17:58

Re: [Announce] kernel-power v49
 
Quote:

Originally Posted by fahadj2003 (Post 1094043)
ure worser then a noob.
i refuse to help u.

thanks..f... u

dignified 2011-09-22 18:00

Re: [Announce] kernel-power v49
 
how can i erase kernel48 in multiboot?thanks

Eminem 2011-09-22 18:19

Re: [Announce] kernel-power v49
 
Installed, rebooted and no errors yet. :)

cloud596 2011-09-22 18:21

Re: [Announce] kernel-power v49
 
open file box with root access go to /etc/multiboot.d/

check a document with kernel48

maacruz 2011-09-22 18:26

Re: [Announce] kernel-power v49
 
Quote:

Originally Posted by don_falcone (Post 1093902)
@pali: pre-release kp49 flashed, working, ramzswap working, testing wl1251 now.

@all: use modprobe from now on (it's better/safer than insmod anyway), procedure:
Code:

modprobe ramzswap num_devices=1 disksize_kb=98304
swapon /dev/ramzswap0
swapoff /dev/mmcblk0p3
swapoff /dev/mmcblk1p2

(with second partition on SD being a swap partition)

EDIT: wl1251 also confirmed working (used fAircrack for checking)

EDIT 2: too early... after another reboot, modprobe with disksize_kb parameter supplied results ALWAYS in instant reboot. I've seen this behavior with colin.stephane's kp48 builds too. Some ideas where to check? I've sysklogd installed.

So, it works the first reboot after kernel flashing, but never after that. Every time a spontaneous reboot during which the custom kernel settings are not loaded.

EDIT 3: Confirmed. Compcache works only if:
- i freshly dpkg'd -i kernel, modules, and flasher (in that sequence, kernel-settings were retained from kp47)
AND
- performed a reboot.
And only then it works after this first reboot. From each consecutive reboot onwards. i observe the instant reboots after ramzswap0 initializing with disksize_kb. If i solely insert the module, no reboot occurs, but of course i cannot initialize it separately.

Mmmm, those reboots look like a clear symptom that the compcache modules are buggy.
Unfortunately I haven't setup scratchbox for N900 yet, so I can't help.

knobby 2011-09-22 18:40

Re: [Announce] kernel-power v49
 
open filebox
go to /etc/multiboot.d and delete kernel power 48

Nokia 5700 2011-09-22 18:52

Re: [Announce] kernel-power v49
 
So... Is Kernel-power v48 considered stable (in a kernel-power way) now? Are there any regressions or bugs?

fahadj2003 2011-09-22 18:55

Re: [Announce] kernel-power v49
 
Quote:

Originally Posted by dignified (Post 1094064)
thanks..f... u

thanks to you too kind sir.

Daneel 2011-09-22 19:17

Re: [Announce] kernel-power v49
 
Working stable for 12 hours so far although I haven't tried compcache yet.

Alfred 2011-09-22 19:49

Re: [Announce] kernel-power v49
 
Me stable too

Gusse 2011-09-22 19:54

Re: [Announce] kernel-power v49
 
I think v49 is pretty stable except compcache when disksize is defined to be so big that it does not use eMMC swap (mmcblk0p3) disk at all.

This is known issue, which was noticed in "COMPCACHE on power-kernel" thread. See this post. It is good to read the whole tread, if interested in about this nice feature.

I hope freemangordon has time to look this issue, because at the beginning compcache is fast and smooth until after some time system start to become unresponsive.

I did try again several disksizes and system settings with v49, but always unresponsiveness came after some usage. It would be a nice addon for N900 to get compacache running smooth.

misterc 2011-09-22 21:17

Re: [Announce] kernel-power v48
 
Quote:

Originally Posted by Carlito (Post 1089103)
When I turn off the phone and charge it with the wall connector it keeps showing me the nokia logo than the 5 dots showing up. They disappear than the logo shows up again and so on . . . very strange. When turning the phone on it doesn't charge at all.

copy that!
but only on one of my two N900s :eek:
unfortunately the primary one :(
guess i will switch & reflash the faulty one & see if it is SW (PK V48) or HW related (i.e. reinstall all the other SWs except PK V48)

maacruz 2011-09-22 22:39

Re: [Announce] kernel-power v49
 
1 Attachment(s)
Quote:

Originally Posted by don_falcone (Post 1093902)
Code:

modprobe ramzswap num_devices=1 disksize_kb=98304
swapon /dev/ramzswap0
swapoff /dev/mmcblk0p3
swapoff /dev/mmcblk1p2



EDIT 3: Confirmed. Compcache works only if:
- i freshly dpkg'd -i kernel, modules, and flasher (in that sequence, kernel-settings were retained from kp47)
AND
- performed a reboot.
And only then it works after this first reboot. From each consecutive reboot onwards. i observe the instant reboots after ramzswap0 initializing with disksize_kb. If i solely insert the module, no reboot occurs, but of course i cannot initialize it separately.

I'm using it right now with disksize_kb=131072 with no issues at all, and no slowness for now.
Unfortunately, this being compcache 0.6.2 means that ramzswap statistics are accessible only from ioctl, so having rzscontrol is a must for debugging and checking its behavior.

EDIT: I have built rzscontrol in my N810 scratchbox, works fine, but the ramzswap module has been built without stats support, so it is useless.
@pali: could you build the module with stats enabled (define CONFIG_RAMZSWAP_STATS), please?

EDIT2: And could you apply the changes I suggested in the compcache thread? they'll increase the memory efficiency.
I have increased the ramzswap size to 192 MB and no stability problems so far, although I have hit a funny memory management problem when applying memory pressure: ramzswap usage has grown to 168 MB, swap partition remains at 0, and the cpu is pegged because the free memory goes up and down like a roller coaster against the used memory. I guess this is some kernel parameter configuration issue that can be easily fixed with a script "a la swapolube".

flashgc 2011-09-23 05:41

Re: [Announce] kernel-power v49
 
Strange. I had a problem with random reboots (3-4 + reboots per day with KP up to v48). Now with KPv49 uptime 15 hours and there was no reboot. Stable.

biobobby89 2011-09-23 05:44

Re: [Announce] kernel-power v49
 
Been testing this for the past 12 hrs or so...No issues reported in KP48 seem to manifest themselves here and its pretty stable as of now...

werks 2011-09-23 06:59

Re: [Announce] kernel-power v49
 
ive had no reebots since i installed kp49,

however, when i did the steps for compcache by don_falcone, it kept saying that im out of memory.

cat /proc/swaps shows that ramzswap0 is in use.

SirSocke 2011-09-23 07:01

Re: [Announce] kernel-power v49
 
KP49 is pretty stable for 16h in normal use now. Great work.

don_falcone 2011-09-23 08:46

Re: [Announce] kernel-power v49
 
Quote:

Originally Posted by DeeGee (Post 1093964)
Were the commands copy-pasted? In that case the swapon command might have came too soon after loading ramzswap. That's the reason you need to use "sleep 1" when enabling ramzswap on a script.

Or, if it already was inserted/enabled.

prankster 2011-09-23 08:48

Re: [Announce] kernel-power v49
 
i hope it arrives soon in repos with stable compcache,probably apt-get update/upgrade will help .

don_falcone 2011-09-23 08:48

Re: [Announce] kernel-power v49
 
Quote:

Originally Posted by werks (Post 1094328)
ive had no reebots since i installed kp49,

however, when i did the steps for compcache by don_falcone, it kept saying that im out of memory.

cat /proc/swaps shows that ramzswap0 is in use.

This can happen if you switch off swap on swap partition, and ramzswap0 device has not enough memory to receive the contents of said swap partition / the already paged to swap memory contents. Try to close applications before.

don_falcone 2011-09-23 08:50

Re: [Announce] kernel-power v49
 
Quote:

Originally Posted by dignified (Post 1094041)
pls help me about this

As i told you, you obviously know jack sh*t about the basics of the system and there fore if we would go through this "step by step" (man i hate this expression, the same as "tutorial"!) we would bulge under the sh*tload of help requests that would come from you afterwards. No, no chance, at least not from me.

don_falcone 2011-09-23 08:51

Re: [Announce] kernel-power v49
 
Quote:

Originally Posted by trisha02 (Post 1094053)
lol..hahahahaha. at least i am far away from that stage

Yeah, "far away" but still in the wrong direction. Sheesh.

don_falcone 2011-09-23 09:05

Re: [Announce] kernel-power v49
 
Quote:

Originally Posted by don_falcone (Post 1093902)
EDIT 2: too early... after another reboot, modprobe with disksize_kb parameter supplied results ALWAYS in instant reboot. I've seen this behavior with colin.stephane's kp48 builds too. Some ideas where to check? I've sysklogd installed.

So, it works the first reboot after kernel flashing, but never after that. Every time a spontaneous reboot during which the custom kernel settings are not loaded.

EDIT 3: Confirmed. Compcache works only if:
- i freshly dpkg'd -i kernel, modules, and flasher (in that sequence, kernel-settings were retained from kp47)
AND
- performed a reboot.
And only then it works after this first reboot. From each consecutive reboot onwards. i observe the instant reboots after ramzswap0 initializing with disksize_kb. If i solely insert the module, no reboot occurs, but of course i cannot initialize it separately.

Question: how is the initializing handled in compcache version integrated into kernel-power?

As i seem to have quoted issue stemming from the device init, i looked around & found that bug report:
https://bugs.launchpad.net/ubuntu/+s...ls/+bug/600782

Quotes:
Quote:

ramzswap's interface has changed and now requires an 'rzscontrol' tool to initialise devices. I don't believe that's packaged right now. However:

* It's changed yet again upstream, is now called 'zram', and its disk size is controlled by writing to files in sysfs.
* rzscontrol is a trivial ioctl caller.
* compcache's licence is compatible with that of initramfs-tools.
Quote:

This bug was fixed in the package initramfs-tools - 0.98ubuntu2

---------------
initramfs-tools (0.98ubuntu2) maverick; urgency=low

* The ramzswap device changed its interface so that the disk size needs to
be given using an ioctl rather than as a module parameter. Incorporate
the rzscontrol tool from the compcache project
(http://code.google.com/p/compcache/) so that we can control current
ramzswap devices (LP: #600782). This is a strictly temporary measure,
as recent versions of compcache change the interface (and module name)
again to involve writing to sysfs files, which won't require any special
tools.
- which version of compcache was included into kp?

Btw: i cannot retrieve anything relevant from the syslog other than (standard logging configuration after install of sysklogd):

- exiting on signal 15 (==SIGTERM, but only happened once)
- DSME: BOOTSTATE: "USER"

WhiteWolf 2011-09-23 09:08

Re: [Announce] kernel-power v49
 
I have had blackouts at night, either with "kp v48" as with "kp v49". Only ocuure me when I leave loading at night. In the mornings I find online, yes fully charged.

What if I have seen is an improvement in battery life with "kp V49" and batterypatch version 1.9-7

stickymick 2011-09-23 09:13

Re: [Announce] kernel-power v49
 
Battery life was a big improvement in v48 for me. Gonna see how long I get with v49.

I did notice though, the handset didn't get quite so warm while I was watching videos over Wi-Fi with the charger plugged in after the update to v49.
Running at 1.1GHz I did notice video was slightly laggy under v48, whereas v49 was faultless.

Straycat 2011-09-23 10:21

Re: [Announce] kernel-power v49
 
Quote:

Originally Posted by Nokia 5700 (Post 1094095)
So... Is Kernel-power v48 considered stable (in a kernel-power way) now? Are there any regressions or bugs?

Well, yes for a lot fo people.

But in my personal case, KP48 screwed up all the video system on my device: unable to watch videos saved, unable to record any video, unable to take any photos...

In adition, my battery life was worse than on KP47, where I'm for now.

But I have followed the KP48 threads and I'm the only one with this kind of troubles.

Regards.

flashgc 2011-09-23 11:28

Re: [Announce] kernel-power v49
 
Quote:

Originally Posted by Straycat (Post 1094408)
Well, yes for a lot fo people.

But in my personal case, KP48 screwed up all the video system on my device: unable to watch videos saved, unable to record any video, unable to take any photos...

In adition, my battery life was worse than on KP47, where I'm for now.

But I have followed the KP48 threads and I'm the only one with this kind of troubles.

Regards.

N900 strange device. From one to anoother its own glitches. They are unique))

maacruz 2011-09-23 12:01

Re: [Announce] kernel-power v49
 
Quote:

Originally Posted by don_falcone (Post 1094372)
Question: how is the initializing handled in compcache version integrated into kernel-power?

As i seem to have quoted issue stemming from the device init, i looked around & found that bug report:
https://bugs.launchpad.net/ubuntu/+s...ls/+bug/600782

Quotes:




- which version of compcache was included into kp?

Btw: i cannot retrieve anything relevant from the syslog other than (standard logging configuration after install of sysklogd):

- exiting on signal 15 (==SIGTERM, but only happened once)
- DSME: BOOTSTATE: "USER"

You should read my posts instead of going in a noob bashing rampage :D
http://talk.maemo.org/showpost.php?p...postcount=1700
And as any other module, initialization is performed at insertion. Just read the module source code, it is quite easy to follow.

don_falcone 2011-09-23 12:34

Re: [Announce] kernel-power v49
 
Quote:

Originally Posted by maacruz (Post 1094467)
You should read my posts instead of going in a noob bashing rampage :D
http://talk.maemo.org/showpost.php?p...postcount=1700
And as any other module, initialization is performed at insertion. Just read the module source code, it is quite easy to follow.

I haven't seen your post, i'm sorry! Thanks btw.

Morku 2011-09-23 12:49

Re: [Announce] kernel-power v49
 
Quote:

Originally Posted by Straycat (Post 1094408)

But in my personal case, KP48 screwed up all the video system on my device: unable to watch videos saved, unable to record any video, unable to take any photos...

I had this issue, when I loaded ulv. The low voltages didn't reboot my device, but made video playback corrput.

freemangordon 2011-09-23 15:39

Re: [Announce] kernel-power v49
 
Quote:

Originally Posted by maacruz (Post 1094467)
You should read my posts instead of going in a noob bashing rampage :D
http://talk.maemo.org/showpost.php?p...postcount=1700
And as any other module, initialization is performed at insertion. Just read the module source code, it is quite easy to follow.

It was me who insist compcache statistics to be disabled because of performance reasons (there are several spinlocks involved as you may see in source code which COULD result in performance loss). Instead of enabling stats for general public, I think it is better to just rebuild module from compcache thread with stats enabled and test with it. And if you find which exactly kernel parameter causes swap trashing when RAM/compcache is full I think all of us will say BIG thank you :D . For changing compress threshold - maybe it will have slight difference in memory consumption, but I think our major problem is swap trashing, not few megs of RAM more.

EDIT:
And maybe it is better to continue compcache conversation on its dedicated thread.

EDIT2:

For those of you who are unable to play videos with KP48-49 - I am sure it is because of your overclocking profile, DSP subsystem is totally new and voltages which were OK with KP47, are slightly lower than needed in KP48-49. Try to increase your voltages by 1 and try again video playing(after reboot).

Straycat 2011-09-23 16:00

Re: [Announce] kernel-power v49
 
Quote:

Originally Posted by Morku (Post 1094490)
I had this issue, when I loaded ulv. The low voltages didn't reboot my device, but made video playback corrput.

Thanks, Morku. What you say makes sense.

How can I check my voltage profile? 'kernel-config show' do not show this.

maacruz 2011-09-23 16:42

Re: [Announce] kernel-power v49
 
Quote:

Originally Posted by freemangordon (Post 1094596)

EDIT:
And maybe it is better to continue compcache conversation on its dedicated thread.

I agree with that, too much noise in this thread.


All times are GMT. The time now is 10:07.

vBulletin® Version 3.8.8