[Announce] kernel-power stable v49 in Extras-Testing - Page 147 - maemo.org

Active Topics

The N900 is now 15 years old! (9)
to Nokia N900 by Arno_11 - 4 hrs, 57 mins ago
Sailfish OS for the Motorola Moto G7 Power (XT1955-5) - (ocean) (4)
to SailfishOS by Maemish - 5 days, 17 hrs ago
more...

Page 147 of 300

147

Thread Tools

trisha02	2011-09-01 , 17:45
Posts: 915 \| Thanked: 57 times \| Joined on Jun 2011	#1461

Originally Posted by Gusse

Download following files to your N900
kernel-power_2.6.28-10power47
kernel-power-modules_2.6.28-10power47
kernel-power-flasher_2.6.28-10power47

Then install those manually.
Code:
sudo gainroot
dpkg -i kernel-power_2.6.28-10power47_armel.deb
dpkg -i kernel-power-modules_2.6.28-10power47_armel.deb
dpkg -i kernel-power-flasher_2.6.28-10power47_armel.deb
Reboot and that's it.

Reason is most probably that you have not included extras-devel repository, but please do not add it. I think you have had enough problems already.

i installed those.checked it by uname -r.fyn.but i wanted to install karam's task swapper through xtermianl and i got this?

BusyBox v1.10.2 (Debian 3:1.10.2.legal-1osso30+0m5) built-in shell (ash)
Enter 'help' for a list of built-in commands.

~ $ sudo gainroot
Root shell enabled

BusyBox v1.10.2 (Debian 3:1.10.2.legal-1osso30+0m5) built-in shell (ash)
Enter 'help' for a list of built-in commands.

/home/user # apt-get install task-swapper
Reading package lists... Done
Building dependency tree
Reading state information... Done
You might want to run `apt-get -f install' to correct these:
The following packages have unmet dependencies:
kernel-power-bootimg: Depends: kernel-power-modules (= 1:2.6.28-10power48) but 1:2.6.28-10power47 is to be installed
E: Unmet dependencies. Try 'apt-get -f install' with no packages (or specify a solution).
/home/user #

Quote & Reply |

evan	2011-09-01 , 18:59
Posts: 136 \| Thanked: 115 times \| Joined on Jul 2010 @ Greece	#1462

about leafpad. I won't rollback to a previous kp to see if it's a new behaviour for v48 although I doubt it, must be a leafpad bug.

n900 will instantly reboot when you try save as and leafpad started
from command prompt no matter if it is user or root and both of the following are true
1) it starts with something as a parameter no matter if it is a file that exists or not
2)current directory is not Mydocs
wierd ehh?

edit:
~ $ cd MyDocs/
~/Mydocs $ leafpad whatever
save as -> ok
exit
~/Mydocs $ cd
~ $ leafpad whatever2
save as -> reboot

same as root
no reboots when it starts with just leafpad, root or user
previously I checked with no filename as parameter so I had not spotted this.

Last edited by evan; 2011-09-01 at 19:14.

Quote & Reply |

The Following 2 Users Say Thank You to evan For This Useful Post:
jpala16, reinob

	Estel	2011-09-01 , 20:12
	Posts: 5,028 \| Thanked: 8,613 times \| Joined on Mar 2011	#1463

Yea, but it was all "discovered" a while ago (still, thank for summing it up). What is even more strange, it fluctuate for some users, i.e. I can save from user no matter if I'm in MyDocs or not.

I hope mtd2 provided will help freemangordon to solve this issue. If You got this problems also (it seems so) post also your mtd2 as per freeman's request.

__________________
N900's aluminum backcover / body replacement
-
N900's HDMI-Out
-
Camera cover MOD
-
Measure battery's real capacity on-device
-
TrueCrypt 7.1 | ereswap | bnf
-
Hardware's mods research is costly. To support my work, please consider donating. Thank You!

Quote & Reply |

vi_	2011-09-01 , 20:21
Posts: 1,680 \| Thanked: 3,685 times \| Joined on Jan 2011	#1464

Leafpad is gay, use vim. problem solved.

__________________

N900: One of God's own prototypes. A high-powered mutant of some kind never even considered for mass production. Too weird to live, and too rare to die.

Quote & Reply |

	Estel	2011-09-01 , 20:22
	Posts: 5,028 \| Thanked: 8,613 times \| Joined on Mar 2011	#1465

No, it is not solved. Other people reported it to happen also in fmms, for which You don't have much alternatives.

Also, if it worked in kp47, and crashes system to oblivion in kp48, there is something worth investigating.

Quote & Reply |

The Following User Says Thank You to Estel For This Useful Post:
bitrocky

freemangordon	2011-09-01 , 20:32
Posts: 3,074 \| Thanked: 12,960 times \| Joined on Mar 2010 @ Sofia,Bulgaria	#1466

I think I found what the bug is, just a little patience

Quote & Reply |

The Following 10 Users Say Thank You to freemangordon For This Useful Post:
bitrocky, cheve, Estel, fw190, ivyking, jpala16, jurop88, moepda, reinob, woody14619

Riddick	2011-09-01 , 20:40
Posts: 6 \| Thanked: 0 times \| Joined on Sep 2011	#1467

after installing v48 my battery lasts for about 8 hours, with v47 it was more than a day.. what could be the problem?

Quote & Reply |

colin.stephane	2011-09-01 , 23:43
Posts: 458 \| Thanked: 783 times \| Joined on Jan 2010 @ France	#1468

Originally Posted by freemangordon

You can get two patches needed for compcache from here

and if you make compcache sources (from the same post) to be part of the kernel (as a patch) I will owe you a beer .

BTW compcache is called ramzswap at kernel.org, will be good to follow naming conventions from there.

Hi freemangordon,

It's time for me to go bed right now ...

Tomorrow I upload something new for you :-)

Code:

[sbox-FREMANTLE_ARMEL: ~/tmp/Building/kernel-power] > dpkg-deb -c kernel-power-modules_2.6.28-10power48_armel.deb | grep ramz
-rw-r--r-- root/root     23312 2011-09-02 01:06:32 ./lib/modules/2.6.28.10-power48/ramzswap.ko

So, stay tunned and ready to test it ...

A++

EDIT !!!

Download from here :

This one contain the ramzswap.ko module directly, no more need to compile yourself under scratchbox or me to provide as external module ...

Changelog :

Latests commits by Pali before he have school obligations (see post #) :

Commits '7ab052e2cf27b79aa8eea32b62f8cde752a87373' & '55fc617b021546d885dc413e15d99690488b3906' :

Enable patch board-rx51-peripherals.diff again, need for I2C_BOARD_INFO bq27x00_battery module. Now disable autoloading bq27x00_battery module by blacklist. Rename blacklist file to kernel-power.blacklist.
Support for binding and unbinding console drivers (CONFIG_VT_HW_CONSOLE_BINDING=y).

New changes by myself :

Applied 'atop' kernel patches (Patch 1: atopcnt & Patch 2: atopacct) :

Patch 01patch-version_atopcnt takes care that statistical counters are added to the process-administration (task_struct) and that these counters are incremented per process. The additional counters are shown via the file /proc/pid/stat as a second line of counters. When this patch is installed, atop automatically shows another layout for the generic screen (including disk- and network-transfers for active processes). However for terminated process these additional counters are not shown, because the conventional layout of the process-accounting record has not been adapted.
Patch 02patch-version_atopacct takes care that the additional statistical counters are added to the process-accounting record (yes, this means an incompatible layout of this record). Notice that this patch is an extension of the patch 01patch-version-atopcnt, i.e. it can not be installed on its own. When this patch is installed, atop automatically shows the additional counters for terminated processes as well. Since the pid of processes is also stored in the modified process-accounting record, matching this info with the formerly running process is less time-consuming compared to the conventional process-accounting record.
Applied 'patch_swap_notify_core_support_2_2.6.28.diff' patch from freemangordon as requested by himself on post #.
Applied 'update_mmu_cache_arm.diff' patch from freemangordon as requested by himself on post #.
Applied 'update_mmu_cache_arm.diff' patch from freemangordon as requested by himself on post #.
Backported kernel module 'ramzswap' integration for 2.6.28 to build it inline during kernel compilation as requested by freemangordon in this post #.

You can get two patches needed for compcache from here

and if you make compcache sources (from the same post) to be part of the kernel (as a patch) I will owe you a beer.

BTW compcache is called ramzswap at kernel.org, will be good to follow naming conventions from there.

Done !!!

A++

PS : freemangordon, can you please test it and repport if it work ?

Last edited by colin.stephane; 2011-09-02 at 10:40.

Quote & Reply |

The Following 13 Users Say Thank You to colin.stephane For This Useful Post:
aanckar, ahmadamaj, Estel, freemangordon, fw190, Gusse, ivyking, jurop88, reinob, Stonik, Trestry, vi_, ysss

freemangordon	2011-09-02 , 19:31
Posts: 3,074 \| Thanked: 12,960 times \| Joined on Mar 2010 @ Sofia,Bulgaria	#1469

@colin.stephane: Thanks but I am a little busy with CSSU, cannot test it right now. Anyway I am sure there will be volunteers to do that for me

.

@all - there is a bug (buffer overrun) in dspbridge.ko module which is triggered by a bug in some userspace application/library. That is the reason for Save As reboots with kp48. While it was easy to find the bug in kernel code, it will take a bit longer to find which userland library is buggy. TBH what I think it is maybe the same bug which causes modest/conversations/etc. crashes out of the blue. Regardless if this is true or not I will try to find it (and fix it if it is an opensourced library). But it will take time.

So, shall I release a patch fixing kernel oops ASAP(lets say by Sunday) or you can live few more days without pressing Save As.

Quote & Reply |

The Following 7 Users Say Thank You to freemangordon For This Useful Post:
Estel, fw190, jurop88, kolos, reinob, Stonik, woody14619

colin.stephane	2011-09-02 , 20:40
Posts: 458 \| Thanked: 783 times \| Joined on Jan 2010 @ France	#1470

Originally Posted by freemangordon

@colin.stephane: Thanks but I am a little busy with CSSU, cannot test it right now. Anyway I am sure there will be volunteers to do that for me .

@all - there is a bug (buffer overrun) in dspbridge.ko module which is triggered by a bug in some userspace application/library. That is the reason for Save As reboots with kp48. While it was easy to find the bug in kernel code, it will take a bit longer to find which userland library is buggy. TBH what I think it is maybe the same bug which causes modest/conversations/etc. crashes out of the blue. Regardless if this is true or not I will try to find it (and fix it if it is an opensourced library). But it will take time.

So, shall I release a patch fixing kernel oops ASAP(lets say by Sunday) or you can live few more days without pressing Save As.

No problem on my side ...

I have to find why the camera doesn't work anymore in my build ...

Today, I'm worked on migrating array of our corporate mail server.

I like to take photo of material before and after migration, and trying to use the camera with my build ... Black screen :-(

And here is the syslog message :

Code:

Sep  2 20:10:30 Nokia-N900 kernel: [ 1810.305114] cam_shutter (GPIO 110) is now open
Sep  2 20:10:30 Nokia-N900 camera-ui[1565]: GLIB WARNING ** Gtk - Invalid icon size 216 
Sep  2 20:10:30 Nokia-N900 camera-ui[1565]: GLIB WARNING ** GStreamer - Name v4l2camsrc0 is not unique in bin camerabin0, not adding
Sep  2 20:10:30 Nokia-N900 camera-ui[1565]: GLIB WARNING ** GStreamer - Name camerabinimage0 is not unique in bin camerabin0, not adding
Sep  2 20:10:30 Nokia-N900 camera-ui[1565]: GLIB CRITICAL ** GStreamer - Padname outputselector2:src2 does not belong to element outputselector3 when removing
Sep  2 20:10:30 Nokia-N900 camera-ui[1565]: GLIB CRITICAL ** GStreamer - Padname outputselector2:src1 does not belong to element outputselector3 when removing
Sep  2 20:10:30 Nokia-N900 camera-ui[1565]: GLIB CRITICAL ** GLib-GObject - g_object_get: assertion `G_IS_OBJECT (object)' failed
Sep  2 20:10:30 Nokia-N900 camera-ui[1565]: GLIB CRITICAL ** GStreamer - gst_element_get_pad: assertion `GST_IS_ELEMENT (element)' failed
Sep  2 20:10:30 Nokia-N900 camera-ui[1565]: GLIB CRITICAL ** GStreamer - gst_pad_add_event_probe_full: assertion `GST_IS_PAD (pad)' failed
Sep  2 20:10:30 Nokia-N900 camera-ui[1565]: GLIB CRITICAL ** GStreamer - gst_object_unref: assertion `object != NULL' failed
Sep  2 20:10:30 Nokia-N900 camera-ui[1565]: GLIB CRITICAL ** GStreamer - gst_object_unref: assertion `object != NULL' failed
Sep  2 20:10:30 Nokia-N900 camera-ui[1565]: GLIB WARNING ** default - Error  open viewfinder  Value: Failed: impossible to set the playing state in the pipeline 
Sep  2 20:10:40 Nokia-N900 kernel: [ 1820.609863] cam_shutter (GPIO 110) is now closed

So, I have to find where this s**t occured, and decided to 1st recompile v48 from pali without any modification from you and me to see if I have something working properly at this step ...

Then after that I'm gona push new patches one by one, build, check, then add another one, build, check ... etc ...

A++