Active Topics

 


Reply
Thread Tools
Posts: 3,074 | Thanked: 12,960 times | Joined on Mar 2010 @ Sofia,Bulgaria
#31
Originally Posted by Android_808 View Post
funny enough, i spent some of yesterday looking into it. as a warning, 1st time attempting RE, so i may be way off here.

i was expecting to find xprot settings string in strings section of disassembly for use with string comparison to identify sections within config file. so far no luck in -common or -music.
Code:
.rodata:000485A8 0000001B C x-maemo.xprot.displacement    
.rodata:00048388 0000001E C x-maemo.xprot.parameters.left 
.rodata:000483A8 0000001F C x-maemo.xprot.parameters.right
.rodata:000485C4 0000001A C x-maemo.xprot.temperature


EDIT: sorry, it is in voice
__________________
Never fear. I is here.

720p video support on N900,SmartReflex on N900,Keyboard and mouse support on N900
Nothing is impossible - Stable thumb2 on n900

Community SSU developer
kernel-power developer and maintainer

 

The Following 5 Users Say Thank You to freemangordon For This Useful Post:
Posts: 2 | Thanked: 9 times | Joined on Sep 2014 @ Minsk
#32
I compiled and packaged Enlightement DR19 for armhf debian unstable. Compiled with systemd support.
From E19 version 0.19.0 alpha 1 mobile profile are deprecated. But default profile is usable.
Trick to add Nokia N900 keyboard support in Enlightenment. After first run (after initial configs created), run in terminal (enlightenment session need to be stopped!):
Code:
eet -d .e/e/config/standart/e.cfg config e.txt
Now, edit e.txt
Find:
Code:
        group "E_Config_XKB_Layout" struct {
            value "name" string: "us";
            value "model" string: "default";
            value "variant" string: "basic";
        }
Replace with:
Code:
        group "E_Config_XKB_Layout" struct {
            value "name" string: "us";
            value "model" string: "nokiarx51";
            value "variant" string: "basic";
        }
After this, run in terminal:
Code:
eet -e .e/e/config/standart/e.cfg config e.txt 1
All E19 packages:
https://www.dropbox.com/sh/ql5b60dyd...w53vFds4a?dl=0
 

The Following 8 Users Say Thank You to Niko2040 For This Useful Post:
Posts: 915 | Thanked: 3,209 times | Joined on Jan 2011 @ Germany
#33
RE: critical speaker frequency:
According to this [1] 480Hz is either the critical frequency or well on the safe side. (I'm no audio expert)
So muting everything below that frequency in alsamixer should make audio safe.
Who want's to play the guinea pig?

On another note, I finally got around to install Debian on real hardware, installed LXDE and it works mostly fine.

I found some small issues:

1. After a reboot when I select Maemo (CSSU thumb kernel sitting on eMMC) the first attempt to boot will always fail at the 5 dots of dread, the second one will always be fine. No idea if it's a Debian, a u-boot or a CSSU thumb thing.
2. In LXDE tapping the screen will always trigger the right mouse button which makes it hard to control the desktop. I guess that's manageable via evdev. Reading man evdev also revealed that tap-hold for right click can be configured via evdev and therefore desktop-independent.
3. I noticed that the CLI and GUI keymaps differ (e.g. no Tab via Shift+Space in GUI). Where is the CLI keymap defined for comparison?
4. LXDE battery monitor may not work. I say "may" because I only tried with power cable connected.

Other than that I'm quite impressed.
btw. is there an easy way to overclock the CPU within Debian, probably the same way KP does it in Maemo?
I didn't try phone functions so far, but I'll try soon. Unfortunately the fso-deviced package isn't installable on i386/amd64 atm. Not sure if it's any better on armhf. I guess it's due to changes in connection with systemd introduction (but right now it doesn't work with sysvinit either).


[1] http://213.128.137.28/showthread.php?p=1160889
 

The Following 7 Users Say Thank You to sulu For This Useful Post:
Posts: 915 | Thanked: 3,209 times | Joined on Jan 2011 @ Germany
#34
I played a bit:
Originally Posted by sulu View Post
2. In LXDE tapping the screen will always trigger the right mouse button which makes it hard to control the desktop. I guess that's manageable via evdev. Reading man evdev also revealed that tap-hold for right click can be configured via evdev and therefore desktop-independent.
How do I configure evdev in DebiaN900?
I need to figure out which is the correct device. Usualy it's done like this (output from my PC):
Code:
$ find /dev/input/by-id/ -name "*event-mouse"
/dev/input/by-id/usb-Logitech_USB_Mouse-event-mouse
But on the N900 the whole by-id directory isn't there:
Code:
$ ls /dev/input/
by-path  event0  event1  event2  event3  mice  mouse0
In short: I have no idea how to change the mouse button mapping now.

Originally Posted by sulu View Post
4. LXDE battery monitor may not work. I say "may" because I only tried with power cable connected.
No, it doesn't work. Afair the reason is that it only looks in /sys/class/power_supply/BAT* but it should look in /sys/class/power_supply/bq27200-0
I had a look at the code long ago and I believe this is easily changed. But on the other hand I could have sworn there is a way to configure this without recompilation, which doesn't seem to be true - so no idea how accurate my memory is here.

Originally Posted by sulu View Post
I didn't try phone functions so far, but I'll try soon. Unfortunately the fso-deviced package isn't installable on i386/amd64 atm. Not sure if it's any better on armhf.
It doesn't install on armhf either, with the same error. I filed a bug report:
https://bugs.debian.org/cgi-bin/bugr...cgi?bug=766114
 

The Following 4 Users Say Thank You to sulu For This Useful Post:
wicket's Avatar
Posts: 634 | Thanked: 3,266 times | Joined on May 2010 @ Colombia
#35
Originally Posted by Niko2040 View Post
I compiled and packaged Enlightement DR19 for armhf debian unstable. Compiled with systemd support.
From E19 version 0.19.0 alpha 1 mobile profile are deprecated. But default profile is usable.
Trick to add Nokia N900 keyboard support in Enlightenment. After first run (after initial configs created), run in terminal (enlightenment session need to be stopped!):
Code:
eet -d .e/e/config/standart/e.cfg config e.txt
Now, edit e.txt
Find:
Code:
        group "E_Config_XKB_Layout" struct {
            value "name" string: "us";
            value "model" string: "default";
            value "variant" string: "basic";
        }
Replace with:
Code:
        group "E_Config_XKB_Layout" struct {
            value "name" string: "us";
            value "model" string: "nokiarx51";
            value "variant" string: "basic";
        }
After this, run in terminal:
Code:
eet -e .e/e/config/standart/e.cfg config e.txt 1
All E19 packages:
https://www.dropbox.com/sh/ql5b60dyd...w53vFds4a?dl=0
Thanks! At some stage I'd like to set up a APT repository for additional packages such as these. If anyone can help with hosting please let me know.
__________________
DebiaN900 - Native Debian on the N900. Deprecated in favour of Maemo Leste.

Maemo Leste for N950 and N9 (currently broken).
Devuan for N950 and N9.

Mobile devices with mainline Linux support - Help needed with documentation.

"Those who do not understand Unix are condemned to reinvent it, poorly." - Henry Spencer
 

The Following 2 Users Say Thank You to wicket For This Useful Post:
wicket's Avatar
Posts: 634 | Thanked: 3,266 times | Joined on May 2010 @ Colombia
#36
Originally Posted by sulu View Post
RE: critical speaker frequency:
According to this [1] 480Hz is either the critical frequency or well on the safe side. (I'm no audio expert)
So muting everything below that frequency in alsamixer should make audio safe.
Who want's to play the guinea pig?
You could test it first with headphones plugged in (assuming the audio will automatically be redirected to the headphone jack, I've not tested this myself but I can confirm that the kernel does detect when headphones have been plugged in).

I'd personally prefer a solution that works at driver level if possible. That way it would cover all possible audio configurations.

Originally Posted by sulu View Post
On another note, I finally got around to install Debian on real hardware, installed LXDE and it works mostly fine.

I found some small issues:

1. After a reboot when I select Maemo (CSSU thumb kernel sitting on eMMC) the first attempt to boot will always fail at the 5 dots of dread, the second one will always be fine. No idea if it's a Debian, a u-boot or a CSSU thumb thing.
I experience this sometimes too. I think lifeguard timer is being triggered and is resetting the device. Try activating R&D mode to see if that makes a difference.

Originally Posted by sulu View Post
2. In LXDE tapping the screen will always trigger the right mouse button which makes it hard to control the desktop. I guess that's manageable via evdev. Reading man evdev also revealed that tap-hold for right click can be configured via evdev and therefore desktop-independent.
Originally Posted by sulu View Post
I played a bit:
How do I configure evdev in DebiaN900?
I need to figure out which is the correct device. Usualy it's done like this (output from my PC):
Code:
$ find /dev/input/by-id/ -name "*event-mouse"
/dev/input/by-id/usb-Logitech_USB_Mouse-event-mouse
But on the N900 the whole by-id directory isn't there:
Code:
$ ls /dev/input/
by-path  event0  event1  event2  event3  mice  mouse0
In short: I have no idea how to change the mouse button mapping now.
When did you grab the scripts? On 8th October I merged this pull request from Niko2040. Perhaps the EmulateThirdButtonTimeout value needs tweaking. evdev is normally configured in xorg.conf.

Originally Posted by sulu View Post
3. I noticed that the CLI and GUI keymaps differ (e.g. no Tab via Shift+Space in GUI). Where is the CLI keymap defined for comparison?
X11 uses a separate key map. Both key maps are taken from Arch Linux for the N900 as mentioned in the first post. The X11 key map has some problems that are yet to be fixed that have been inherited from the Arch patches. The console key map gets installed with this command and then ends up under /etc/console/boottime.kmap.gz

Originally Posted by sulu View Post
4. LXDE battery monitor may not work. I say "may" because I only tried with power cable connected.
Originally Posted by sulu View Post
No, it doesn't work. Afair the reason is that it only looks in /sys/class/power_supply/BAT* but it should look in /sys/class/power_supply/bq27200-0
I had a look at the code long ago and I believe this is easily changed. But on the other hand I could have sworn there is a way to configure this without recompilation, which doesn't seem to be true - so no idea how accurate my memory is here.
E17 has the same issue although I've not yet looked into it. If the problem is as you describe, I wonder if it could be worked around by creating symlinks.

Originally Posted by sulu View Post
Other than that I'm quite impressed.
btw. is there an easy way to overclock the CPU within Debian, probably the same way KP does it in Maemo?
I'm personally not interested in overclocking so I haven't looked into that at all. I would assume it would be similar to KP under Maemo but bare in mind that things may have changed in recent kernels.

Originally Posted by sulu View Post
I didn't try phone functions so far, but I'll try soon. Unfortunately the fso-deviced package isn't installable on i386/amd64 atm. Not sure if it's any better on armhf. I guess it's due to changes in connection with systemd introduction (but right now it doesn't work with sysvinit either).]
Originally Posted by sulu View Post
It doesn't install on armhf either, with the same error. I filed a bug report:
https://bugs.debian.org/cgi-bin/bugr...cgi?bug=766114
Note that my scripts revert to good old System V init (at least for now). When Debian changed the default init system to systemd, it broke the N900 power button functionality. I'm loathed to troubleshoot problems with Lennartware so I installed System V init and within an instant all sanity was restored!

oFono is also available for Debian if you want to try an alternative to FSO.
__________________
DebiaN900 - Native Debian on the N900. Deprecated in favour of Maemo Leste.

Maemo Leste for N950 and N9 (currently broken).
Devuan for N950 and N9.

Mobile devices with mainline Linux support - Help needed with documentation.

"Those who do not understand Unix are condemned to reinvent it, poorly." - Henry Spencer
 

The Following 5 Users Say Thank You to wicket For This Useful Post:
Posts: 915 | Thanked: 3,209 times | Joined on Jan 2011 @ Germany
#37
Originally Posted by wicket View Post
I experience this sometimes too. I think lifeguard timer is being triggered and is resetting the device. Try activating R&D mode to see if that makes a difference.
Maybe I will. But it's not much of a problem and since it's predictable I'm fine with this.

Originally Posted by wicket View Post
When did you grab the scripts? On 8th October I merged this pull request from Niko2040. Perhaps the EmulateThirdButtonTimeout value needs tweaking. evdev is normally configured in xorg.conf.
I downloaded the zip you provided in the first post on 17th or 18th and created the SD card on 18th.
I think the problem lies deeper since at least to my understanding /dev/input/by-id/ should be present regardless of xorg.conf. It should be there once xserver-xorg-input-evdev is installed.

Originally Posted by wicket View Post
X11 uses a separate key map. Both key maps are taken from Arch Linux for the N900 as mentioned in the first post. The X11 key map has some problems that are yet to be fixed that have been inherited from the Arch patches. The console key map gets installed with this command and then ends up under /etc/console/boottime.kmap.gz
In Easy Debian I copy the whole /usr/share/X11/xkb directory from Maemo to the image (and then set xkb-data on hold). This is the only way I've found so far to make sure the keymap is correct. I tried different things with setxkbmap (the N900 keymap is in Debian) but can't make it work correctly. According to diff both directories are identical.
I'll try if that works for DebiaN900 too.

Originally Posted by wicket View Post
E17 has the same issue although I've not yet looked into it. If the problem is as you describe, I wonder if it could be worked around by creating symlinks.
Afaik one can't create symlinks in /sys.

Originally Posted by wicket View Post
Note that my scripts revert to good old System V init (at least for now). When Debian changed the default init system to systemd, it broke the N900 power button functionality.
I know. I also reverted to sysvinit in one of the VMs mentioned in the bug report, but it seems this doesn't have any effect on what apt does when installing packages.
I guess in the long run there's no way around systemd in Debian because sysvinit compatibility just won't be maintained actively anymore and I don't believe that any of the recent attemts (new GRs) will really change that.

Originally Posted by wicket View Post
oFono is also available for Debian if you want to try an alternative to FSO.
phoneuid indirectly depends on fso-deviced and dos1 promised to make sure phoneui will be supported on the Neo900 running Debian. And since I consider DebiaN900 mainly a testbed for the Neo900 I'd like to use the same solution here.
 

The Following 4 Users Say Thank You to sulu For This Useful Post:
Posts: 2 | Thanked: 9 times | Joined on Sep 2014 @ Minsk
#38
Originally Posted by wicket View Post
When Debian changed the default init system to systemd, it broke the N900 power button functionality.
What do you mean by broke the N900 power button functionality? When I press power button on my N900 with E19 + systemd, it's offer me to poweroff. E19 input settings recognize N900 power button as XF86PowerOff, and you can set action on this key at input settings as well as at ACPI settings too.
 

The Following User Says Thank You to Niko2040 For This Useful Post:
Posts: 915 | Thanked: 3,209 times | Joined on Jan 2011 @ Germany
#39
I think I know why the battery monitor doesn't work.
For comparison my netbook:
Code:
$ cat /sys/class/power_supply/BAT0/uevent 
POWER_SUPPLY_NAME=BAT0
POWER_SUPPLY_STATUS=Discharging
POWER_SUPPLY_PRESENT=1
POWER_SUPPLY_TECHNOLOGY=Li-ion
POWER_SUPPLY_CYCLE_COUNT=0
POWER_SUPPLY_VOLTAGE_MIN_DESIGN=8400000
POWER_SUPPLY_VOLTAGE_NOW=7500000
POWER_SUPPLY_CURRENT_NOW=1192000
POWER_SUPPLY_CHARGE_FULL_DESIGN=6580000
POWER_SUPPLY_CHARGE_FULL=4194000
POWER_SUPPLY_CHARGE_NOW=3337000
POWER_SUPPLY_CAPACITY=79
POWER_SUPPLY_CAPACITY_LEVEL=Normal
POWER_SUPPLY_MODEL_NAME=901
POWER_SUPPLY_MANUFACTURER=ASUS
POWER_SUPPLY_SERIAL_NUMBER=
Notice the key POWER_SUPPLY_CHARGE_NOW which is used to determine the current charge of the battery:
Code:
$ cat /sys/class/power_supply/BAT0/charge_now 
3237000
On the N900 (on battery) I get this:
Code:
$ cat /sys/class/power_supply/bq27200-0/uevent 
POWER_SUPPLY_NAME=bq27200-0
POWER_SUPPLY_STATUS=Discharging
POWER_SUPPLY_PRESENT=1
POWER_SUPPLY_VOLTAGE_NOW=4012000
POWER_SUPPLY_CURRENT_NOW=307912
POWER_SUPPLY_CAPACITY_LEVEL=Normal
POWER_SUPPLY_TEMP=311
POWER_SUPPLY_TECHNOLOGY=Li-ion
POWER_SUPPLY_CHARGE_FULL_DESIGN=2056320
POWER_SUPPLY_CYCLE_COUNT=22
POWER_SUPPLY_POWER_AVG=315818
POWER_SUPPLY_CHARGE_NOW is missing so the battery monitor has no way to determine the current charge of the battery:
Code:
$ cat /sys/class/power_supply/bq27200-0/charge_now
cat: /sys/class/power_supply/bq27200-0/charge_now: Keine Daten verfügbar
("No data available")
 

The Following 5 Users Say Thank You to sulu For This Useful Post:
wicket's Avatar
Posts: 634 | Thanked: 3,266 times | Joined on May 2010 @ Colombia
#40
Originally Posted by sulu View Post
I downloaded the zip you provided in the first post on 17th or 18th and created the SD card on 18th.
I think the problem lies deeper since at least to my understanding /dev/input/by-id/ should be present regardless of xorg.conf. It should be there once xserver-xorg-input-evdev is installed.
The zip is automatically generated with the contents of the master branch so it sounds like your copy contains Niko2040's patch. I changed the value of EmulateThirdButtonTimeout to "750" and it's working much better now. I will update the script to make this configurable in the conf file and I'll set a more sensible default.

Originally Posted by sulu View Post
In Easy Debian I copy the whole /usr/share/X11/xkb directory from Maemo to the image (and then set xkb-data on hold). This is the only way I've found so far to make sure the keymap is correct. I tried different things with setxkbmap (the N900 keymap is in Debian) but can't make it work correctly. According to diff both directories are identical.
I'll try if that works for DebiaN900 too.
These are the XKB patches that are downloaded and applied to /usr/share/X11/xkb/symbols/nokia_vndr/rx-51. If you want to restore the original file just run apt-get --reinstall install xkb-data.

Originally Posted by sulu View Post
Afaik one can't create symlinks in /sys.
Yes, this seems to be the case.

Originally Posted by sulu View Post
I guess in the long run there's no way around systemd in Debian because sysvinit compatibility just won't be maintained actively anymore and I don't believe that any of the recent attemts (new GRs) will really change that.
There's been a lot of FUD spread about the latest GR. I'm quite optimistic about the situation. Most of the opposing responses are concerned that the proposal has come so close to the Jessie freeze rather than opposing it all together. As I understand it, the GR is not about changing Debian policy but rather about enforcing it due to some misunderstandings. Debian calls itself "The Universal Operating System" and the Social Contract states "We will be guided by the needs of our users and the free software community. We will place their interests first in our priorities. We will support the needs of our users for operation in many different kinds of computing environments." so I don't really see anything changing. Should it change and should systemd become the only supported init system, it's bye bye Debian for me but I very much doubt that it will come to that.

Originally Posted by sulu View Post
I think I know why the battery monitor doesn't work.
For comparison my netbook:
Code:
$ cat /sys/class/power_supply/BAT0/uevent 
POWER_SUPPLY_NAME=BAT0
POWER_SUPPLY_STATUS=Discharging
POWER_SUPPLY_PRESENT=1
POWER_SUPPLY_TECHNOLOGY=Li-ion
POWER_SUPPLY_CYCLE_COUNT=0
POWER_SUPPLY_VOLTAGE_MIN_DESIGN=8400000
POWER_SUPPLY_VOLTAGE_NOW=7500000
POWER_SUPPLY_CURRENT_NOW=1192000
POWER_SUPPLY_CHARGE_FULL_DESIGN=6580000
POWER_SUPPLY_CHARGE_FULL=4194000
POWER_SUPPLY_CHARGE_NOW=3337000
POWER_SUPPLY_CAPACITY=79
POWER_SUPPLY_CAPACITY_LEVEL=Normal
POWER_SUPPLY_MODEL_NAME=901
POWER_SUPPLY_MANUFACTURER=ASUS
POWER_SUPPLY_SERIAL_NUMBER=
Notice the key POWER_SUPPLY_CHARGE_NOW which is used to determine the current charge of the battery:
Code:
$ cat /sys/class/power_supply/BAT0/charge_now 
3237000
On the N900 (on battery) I get this:
Code:
$ cat /sys/class/power_supply/bq27200-0/uevent 
POWER_SUPPLY_NAME=bq27200-0
POWER_SUPPLY_STATUS=Discharging
POWER_SUPPLY_PRESENT=1
POWER_SUPPLY_VOLTAGE_NOW=4012000
POWER_SUPPLY_CURRENT_NOW=307912
POWER_SUPPLY_CAPACITY_LEVEL=Normal
POWER_SUPPLY_TEMP=311
POWER_SUPPLY_TECHNOLOGY=Li-ion
POWER_SUPPLY_CHARGE_FULL_DESIGN=2056320
POWER_SUPPLY_CYCLE_COUNT=22
POWER_SUPPLY_POWER_AVG=315818
POWER_SUPPLY_CHARGE_NOW is missing so the battery monitor has no way to determine the current charge of the battery:
Code:
$ cat /sys/class/power_supply/bq27200-0/charge_now
cat: /sys/class/power_supply/bq27200-0/charge_now: Keine Daten verfügbar
("No data available")
If that's what's causing it, I think we're looking at a driver patch to fix it.
__________________
DebiaN900 - Native Debian on the N900. Deprecated in favour of Maemo Leste.

Maemo Leste for N950 and N9 (currently broken).
Devuan for N950 and N9.

Mobile devices with mainline Linux support - Help needed with documentation.

"Those who do not understand Unix are condemned to reinvent it, poorly." - Henry Spencer

Last edited by wicket; 2015-03-23 at 20:25.
 

The Following 3 Users Say Thank You to wicket For This Useful Post:
Reply

Tags
debian, debian900, devuan, maemo 7


 
Forum Jump


All times are GMT. The time now is 13:58.