CyanogenMod 6.1.1 for B7610 update 01-06 // I am stopping my contributions as my phone died

[Guest erikcas]

attached dmesg and logcat. 1 - after calling off a handsfree, 2 - with the included handsfree. I also noticed if you do not answer the call, then after 4 seconds - the vibration and melody stop, although the call continues to go. I would be very grateful if you help with this disaster

from dmesg I should say that the headset is active. This must have to do with the keyboardslide.

Does this also happen when you do not apply the keyboard-slide.

Elbee and I exprimented with the headset a while ago. We cannot find the gpio (which triggers the on/off/reset for that part of hardware) for it. I suspect the headset-trigger is somehow connected to the keyboard.

I have to think how to solve this. For now, forget about the update we were talking about.

First I ask you to try without the keyboard-slide fix for calls

[Guest pentarick]

I did, thank you.

But the hardware keys remain useless. I just replaced the updated zImage in 'My Storage' with the original zImage (from 'froyo_beta3.zip'), leaving the frb3update.tar.gz in the root of My Storage.

This gives me some usable keys, but I guess that zImage leaves the frb3update alone?

Putting back the updated zImage from Kernel-B7610-FroyoREV75.zip cripples the keyboard again. Only the 'Call' key is useful, which acts as a backspace now.

Gosh, this is annoying. Yesterday I was playing happily with Beta1, then I looked here if there was any progress made, excited that I was when I found the beta 3, and now the fun is over. :(

I'm not a noob at all with this kind of stuff, but I just can't figure out why everybody got it working except for me. :S

Please tell me, what information do you need to get this fixed?

Or, can I update qwerty.kl manually somehow?

Edited October 31, 2011 by pentarick

[Guest Isascaboy89]

Just a little issue:

for my with physical keyboard it's impossible to create special characthers, such "?() and similar..

But this still remain a very very very good job!!!!!

[Guest erikcas]

Just a little issue:

for my with physical keyboard it's impossible to create special characthers, such "?() and similar..

But this still remain a very very very good job!!!!!

qwerty?

For me they are all working

Edited October 31, 2011 by erikcas

[Guest erikcas]

I did, thank you.

But the hardware keys remain useless. I just replaced the updated zImage in 'My Storage' with the original zImage (from 'froyo_beta3.zip'), leaving the frb3update.tar.gz in the root of My Storage.

This gives me some usable keys, but I guess that zImage leaves the frb3update alone?

Putting back the updated zImage from Kernel-B7610-FroyoREV75.zip cripples the keyboard again. Only the 'Call' key is useful, which acts as a backspace now.

Gosh, this is annoying. Yesterday I was playing happily with Beta1, then I looked here if there was any progress made, excited that I was when I found the beta 3, and now the fun is over. :(

I'm not a noob at all with this kind of stuff, but I just can't figure out why everybody got it working except for me. :S

Please tell me, what information do you need to get this fixed?

Or, can I update qwerty.kl manually somehow?

Can you tel EXACTLY how you install and how you apply the update?

I compiled the update and tested 20 fresh installs before publishing. I have absolute NO problems

When I read your post I tried to reproduce your problem but I cannot reproduce.

please post every exact steps you do from scratch untill you booted after update. And offcourse rebooted after applying the update.

I want to help you but for that I have to know the exact steps you take

Edited October 31, 2011 by erikcas

[Guest pentarick]

Well, I tried some different things already, each time giving the same result. But this is what I did last time:

First, I remove all the partitions from the SD-card using MiniTool Partition Wizard, no cardreader, but phone functioning as mass storage device. SD-card is 2GB.

Then, I delete all the files from 'My storage' ('Mijn Opslag' in my case, dutch language), I did delete the partition once, but last time I just deleted the files.

After that, I copy all the files from 'froyo_beta3.zip' to 'My Storage' from my computer, using Mass Storage mode.

Then I disconnect the phone from the computer, and I start 'androidinstall.exe' in o2beta folder. Install runs fine, as far I can tell, phone reboots.

Ok, I reconnect the phone to the computer, and I copy 2 files (zImage and frb3update.tar.gz') from 'Kernel-B7610-Froyo-REV75.zip' to the root of 'My Storage', noticing a file 'startup.txt' has appeared in the root.

Finally, I disconnect the phone again and I run HaRet.exe in the root of 'My storage'. Everything seems to go well, after welcome thing and setup the device asks me which desktop to run, I click one. And then I'm screwed ;)

Writing this down one thing comes to mind, I've read several times in all install-guides: 'After first boot, DO NOT DO ANYTHING but reboot immediately'. I'm sure I did this once, but certainly not the last attempt. Oh well, at least I have something else to try. ;)

About the hardware keys, I'm not sure if the device uses an X and Y grid for keymapping, but mine seems to be switched. Pressing A and Shift gives me 1 and 2, respectively. And W-S-Z gives me 8-9-0. I get one column of Q-W-E and another of P-[-], after that it's all garbled, single keys giving me 'Auouo' and '=%<!q' or something.

[Guest erikcas]

Well, I tried some different things already, each time giving the same result. But this is what I did last time:

First, I remove all the partitions from the SD-card using MiniTool Partition Wizard, no cardreader, but phone functioning as mass storage device. SD-card is 2GB.

Then, I delete all the files from 'My storage' ('Mijn Opslag' in my case, dutch language), I did delete the partition once, but last time I just deleted the files.

After that, I copy all the files from 'froyo_beta3.zip' to 'My Storage' from my computer, using Mass Storage mode.

Then I disconnect the phone from the computer, and I start 'androidinstall.exe' in o2beta folder. Install runs fine, as far I can tell, phone reboots.

Ok, I reconnect the phone to the computer, and I copy 2 files (zImage and frb3update.tar.gz') from 'Kernel-B7610-Froyo-REV75.zip' to the root of 'My Storage', noticing a file 'startup.txt' has appeared in the root.

Finally, I disconnect the phone again and I run HaRet.exe in the root of 'My storage'. Everything seems to go well, after welcome thing and setup the device asks me which desktop to run, I click one. And then I'm screwed ;)

Writing this down one thing comes to mind, I've read several times in all install-guides: 'After first boot, DO NOT DO ANYTHING but reboot immediately'. I'm sure I did this once, but certainly not the last attempt. Oh well, at least I have something else to try. ;)

About the hardware keys, I'm not sure if the device uses an X and Y grid for keymapping, but mine seems to be switched. Pressing A and Shift gives me 1 and 2, respectively. And W-S-Z gives me 8-9-0. I get one column of Q-W-E and another of P-[-], after that it's all garbled, single keys giving me 'Auouo' and '=%<!q' or something.

it seems indeed android is using the old qwerty.kl

the new one is included in the update and should work after second boot (after update)

[Guest pentarick]

Well, I've resetted my device three times now, keys are still not working so I cannot reboot using 'End Call' button. I've still haven't set up anything in Android, so I get the 'Touch the Android to begin' screen when I boot up.

Sensors are not working as well, is this correct? Last thing I can think of is that Android is not looking in the correct folder for updates. Is there any reference to the correct folder (or device, as that is how Unix works) that I can modify? Or am I looking in the wrong direction?

[Guest erikcas]

Well, I've resetted my device three times now, keys are still not working so I cannot reboot using 'End Call' button. I've still haven't set up anything in Android, so I get the 'Touch the Android to begin' screen when I boot up.

Sensors are not working as well, is this correct? Last thing I can think of is that Android is not looking in the correct folder for updates. Is there any reference to the correct folder (or device, as that is how Unix works) that I can modify? Or am I looking in the wrong direction?

Android is not looking for an update, it is a linux script @ boot.

Sensors are incorrect, that is a fact. They have to be corrected for B7610, it is a mathematical riddle.

Did the frb3update.tar.gz disappear from your my storage after boot?

[Guest pentarick]

No, accelerometer is not available at all, according to AndroSensor.

So the reported values are still incorrect, maybe I can help with that, I've solved a few 3d geometric puzzles in the past.

The tarball is still in My Storage.

[Guest erikcas]

No, accelerometer is not available at all, according to AndroSensor.

So the reported values are still incorrect, maybe I can help with that, I've solved a few 3d geometric puzzles in the past.

The tarball is still in My Storage.

Rename the tarball to frb2update.tar.gz and see if it disappears.

Reason it is not working for you, the tarball isn' t extracted by the update-script so nothing is updated in your android filesystem. This is essential for the kernel (zImage) to work correctly.

Think your install is from the extracted beta2 image. So the update script is looking for frb2update.tar.gz instead of frb3update.tar.gz

Did you download the ext4.tar.gz seperately? If yes, delete everything you downloaded before.

Only download the automatic installer package and unzip to your my storage. Don't use any separate downloaded ext4.tar.gz

Install and after that appy update.

Edited October 31, 2011 by erikcas

[Guest pentarick]

No luck after renaming the file either.

I downloaded the files from the sources you mentioned on page 16, no separate ext4.tar.gz used:

Download installer from http://o2droid.phj.hu/index_en.php

B7610 update available from http://code.google.com/p/omnia2droid in the downloadsection

Do you know where the script is located, is it in zImage? I can't modify that, can I?

Edited October 31, 2011 by pentarick

[Guest erikcas]

So the reported values are still incorrect, maybe I can help with that, I've solved a few 3d geometric puzzles in the past.

oke.. This part of the puzzle,in this file, I think has to be solved:

(compare it to REV75 and not to the update I PM' ed you because I did some trial and error in it! In REV75 sensors work but inverted)

	acc_data.x = (x - 2080) / div_val; 

	acc_data.y = (y - 2080) / div_val;

	acc_data.z = (z - 2080) / div_val;



	if (swap == 1) {

		acc_data.x = (y - 2080) / div_val; 

		acc_data.y = (x - 2080) / div_val;

	}	


	if (swap == 2) {

		acc_data.x = (z - 2080) / div_val; 

		acc_data.z = (x - 2080) / div_val;

	}	


	if (swap == 3) {

		acc_data.y = (z - 2080) / div_val;

		acc_data.z = (y - 2080) / div_val;

	}	



	if ( change_sign & 1) acc_data.x *= -1;

	if ( change_sign & 2) acc_data.y *= -1;

	if ( change_sign & 4) acc_data.z *= -1;

[Guest erikcas]

No luck after renaming the file either.

I downloaded the files from the sources you mentioned on page 16, no separate ext4.tar.gz used:

Do you know where the script is located, is it in zImage? I can't modify that, can I?

no it is within the ext4.tar.gz filesystem.

strange. it is in loadmod

if [ -f /storage/frb3update.tar.gz ];

then

# tar xzvf /storage/frb3update.tar.gz

 cat /storage/frb3update.tar.gz | tar xzvf -

 rm /storage/frb3update.tar.gz

fi

if [ -f /storage/frb3update.sh ];

then

 chmod +x /storage/frb3update.sh

 /storage/frb3update.sh > /storage/frb3update.txt

 rm /storage/frb3update.sh

fi

I will prepare you an ext4.tar.gz and zImage pre-updated with rev 75

Will pm you the link. It will be ready in half an hour or so.

Download and replace the ext4 and zImage from installer with my version. It will work I think.

But I am curious what will happen with future updates on your phone

[Guest erikcas]

No luck after renaming the file either.

I downloaded the files from the sources you mentioned on page 16, no separate ext4.tar.gz used:

Do you know where the script is located, is it in zImage? I can't modify that, can I?

PM' ed you a link. Currently uploading 148 mb can take some 15 minutes

[Guest pentarick]

Ah, thank you very very much, hardware keys are now functional.

The accelerometer seems quite messed up, yes. ;)

Well, going to bed now, I'll look into it later this week.

[Guest geniussbk]

from dmesg I should say that the headset is active. This must have to do with the keyboardslide.

Does this also happen when you do not apply the keyboard-slide.

Elbee and I exprimented with the headset a while ago. We cannot find the gpio (which triggers the on/off/reset for that part of hardware) for it. I suspect the headset-trigger is somehow connected to the keyboard.

I have to think how to solve this. For now, forget about the update we were talking about.

First I ask you to try without the keyboard-slide fix for calls

Thank you for taking the time, I try without keyboard-slide fix. The result is the same. Perhaps this is due to the fact that the headset connects via microusb, as there is no 3.5 mm mini jack. Talk through the headset I'm not going, so I'll try to just work the microphone and loudspeaker during a call.

event responsible for a headset at all would be removed. Do you happen to know this event?

Edited November 1, 2011 by geniussbk

[Guest erikcas]

Thank you for taking the time, I try without keyboard-slide fix. The result is the same. Perhaps this is due to the fact that the headset connects via microusb, as there is no 3.5 mm mini jack. Talk through the headset I'm not going, so I'll try to just work the microphone and loudspeaker during a call.

event responsible for a headset at all would be removed. Do you happen to know this event?

I will search for that. But it seems we have a hardware difference here. Could you take a look at the hardware specs wiki page at the google project page and collect as much data about the B7620 as you can?

Maybe we have to look into the gpio's after all. I am not having too much time this week. Maybe spend some time on it this weekend or next week. I will PM you tonight about how to find gpio's

[Guest Isascaboy89]

Guys i see that lot of uers have problem with it...so, hoping that i wasn't the only lucky boy, i post what i've done step by step

1) Download kernel, installer and startup file

2) Connect phone to pc

3) Copy the unzipped conten of the autoinstaller on the root of my storage

4) Disconnect phone from PC

5) Open o2beta folder and run androidinstaller.exe

6) Wait till the end of the black screen with "falling"text

7) The device will reboot in more or less 30 second

8) Connect phone to PC

9) Copy the file "startup.txt" on the root of my storage

10) Run "haret.exe" from the root of my storage

11) Again wait till the end of the black screen with "falling" text

12) Android will boot (it takes lot of time at the first time)

13) Once in android, don't do anything, just reboot

14) Connect phone to PC

15) Copy the unzipped content of "kernel rev 75..." on the root of my storage,overwriting if necessary

16) Run again "haret.exe" from the root of my storage

17) Android will boot

18) I suggest to reboot again once or two time just to be sure

That's is what i've done and (for now) my android is working great.

Hope this helps.

[Guest geniussbk]

I will search for that. But it seems we have a hardware difference here. Could you take a look at the hardware specs wiki page at the google project page and collect as much data about the B7620 as you can?

Maybe we have to look into the gpio's after all. I am not having too much time this week. Maybe spend some time on it this weekend or next week. I will PM you tonight about how to find gpio's

ok, I will wait PM. I try plug the headset into microusb. Appeared sound in games, the player and during calls, but through the headset. If you remove the headset, the sound is always lost. Microphone in the headset also works. Before I did keyboard slide fix the headset is not working .

Edited November 1, 2011 by geniussbk

[Guest matriX1218]

@erik

forgot to mention 2 bugs...

1. battery not charging till 100% stays @99% and still charging...

2. battery is getting very hot while charging... already mentioned at I8000 forum...

[Guest erikcas]

@erik

forgot to mention 2 bugs...

1. battery not charging till 100% stays @99% and still charging...

2. battery is getting very hot while charging... already mentioned at I8000 forum...

Bug 1 is also mentioned by users on I8000. For bug 2 I will make a testversion which you can try. I don't think it will solve it but we can try.

I changed the batterydriver and tested it on my device. It didn't charge in android. But that could be caused by my phone using a white label battery instead of the original.

So I will compile a kernel for B7610 with original devs battery driver tonight and PM you link. Can you test if battery charges with it and test if your battery become less hot?

Note: my phone is charging to 95% one time, next time it charges up to 100% I think this has to do with environmental settings like temp and moisture.

My phone/battery never heats up during charging. But we have to test original battery with original driver anyway.

[Guest crazyhorse78]

Guys i see that lot of uers have problem with it...so, hoping that i wasn't the only lucky boy, i post what i've done step by step

1) Download kernel, installer and startup file

2) Connect phone to pc

3) Copy the unzipped conten of the autoinstaller on the root of my storage

4) Disconnect phone from PC

5) Open o2beta folder and run androidinstaller.exe

6) Wait till the end of the black screen with "falling"text

7) The device will reboot in more or less 30 second

8) Connect phone to PC

9) Copy the file "startup.txt" on the root of my storage

10) Run "haret.exe" from the root of my storage

11) Again wait till the end of the black screen with "falling" text

12) Android will boot (it takes lot of time at the first time)

13) Once in android, don't do anything, just reboot

14) Connect phone to PC

15) Copy the unzipped content of "kernel rev 75..." on the root of my storage,overwriting if necessary

16) Run again "haret.exe" from the root of my storage

17) Android will boot

18) I suggest to reboot again once or two time just to be sure

That's is what i've done and (for now) my android is working great.

Hope this helps.

Mine stoped at "Android on Omnia II Loading..."

[Guest pentarick]

oke.. This part of the puzzle,in this file, I think has to be solved:

(compare it to REV75 and not to the update I PM' ed you because I did some trial and error in it! In REV75 sensors work but inverted)

	acc_data.x = (x - 2080) / div_val; 

	acc_data.y = (y - 2080) / div_val;

	acc_data.z = (z - 2080) / div_val;



	if (swap == 1) {

		acc_data.x = (y - 2080) / div_val; 

		acc_data.y = (x - 2080) / div_val;

	}	


	if (swap == 2) {

		acc_data.x = (z - 2080) / div_val; 

		acc_data.z = (x - 2080) / div_val;

	}	


	if (swap == 3) {

		acc_data.y = (z - 2080) / div_val;

		acc_data.z = (y - 2080) / div_val;

	}	



	if ( change_sign & 1) acc_data.x *= -1;

	if ( change_sign & 2) acc_data.y *= -1;

	if ( change_sign & 4) acc_data.z *= -1;

@erikcas, what do you use to check the accelerometer? I checked it with AndroSensor, and I see that the accelerometer isn't reporting any values most of the time. It has something to do with the compass, because if I rotate the phone while keeping the screen level (facing the ceiling), I get readings at some orientations. Probably a coding error which quits the readout subroutine. When I do get readings, all seem to be reverse BUT consistent which is a good thing, I don't know if it's just your zImage that works that way, or if it's the same with REV75. The crippled readout is in all versions, I've noticed. I would change this

    x = kxsd9_get_valid_value(&buf_read[0]);

    y = kxsd9_get_valid_value(&buf_read[2]);

    z = kxsd9_get_valid_value(&buf_read[4]);

to this:

    x = -kxsd9_get_valid_value(&buf_read[0]);

    y = -kxsd9_get_valid_value(&buf_read[2]);

    z = -kxsd9_get_valid_value(&buf_read[4]);

That way, you'll affect all if-cases. The following is also stupid, I think:

        acc_data.x = (x - 2080) / div_val; 

        acc_data.y = (y - 2080) / div_val;

        acc_data.z = (z - 2080) / div_val;


        if (swap == 1) {

                acc_data.x = (y - 2080) / div_val; 

                acc_data.y = (x - 2080) / div_val;

        }       


        if (swap == 2) {

                acc_data.x = (z - 2080) / div_val; 

                acc_data.z = (x - 2080) / div_val;

        }       


        if (swap == 3) {

                acc_data.y = (z - 2080) / div_val;

                acc_data.z = (y - 2080) / div_val;

        }       

Three values are calculated in the top three lines. But if 'swap' is between 1 and 3, two values get recalculated, which costs extra processor time. Little, but still. More efficient would be:

        switch (swap) {

            case 1:

                acc_data.x = (y - 2080) / div_val; 

                acc_data.y = (x - 2080) / div_val;

                acc_data.z = (z - 2080) / div_val;

                break;


            case 2:

                acc_data.x = (z - 2080) / div_val; 

                acc_data.y = (y - 2080) / div_val;

                acc_data.z = (x - 2080) / div_val;

                break;


            case 3:

                acc_data.x = (x - 2080) / div_val; 

                acc_data.y = (z - 2080) / div_val;

                acc_data.z = (y - 2080) / div_val;

                break;


            default:

                acc_data.x = (x - 2080) / div_val; 

                acc_data.y = (y - 2080) / div_val;

                acc_data.z = (z - 2080) / div_val;

                break;

        }       

This way, a value never gets calculated twice, which is important, as divide operations are processor-intensitive, relatively. The 'default' case could probably also be at the top as 'case 0', but I can't test it and this way it behaves exactly as the original code. No time today to contribute anything further, we have to find out why accelerometer-events are not always happening. AndroSensor could be faulty, with the 'Waiting for event' message, but in Froyo beta1 it happened much less, and was unrelated to the compass. Oh, and on line 264 there is bad coding, this gives division by zero errors in the code above, try replacing it with 10000 or something:

              case 30:

                        div_val = 0;

                        break;

Please tell me, do I have to use linux to compile the zImage, or can it be done in Windows? Or should I just Google? I'm really new to Android ;)

[Guest erikcas]

I use gps status. You are absolutely right, every now and then it stops reporting. I think it has to do with a reset of the chip that is wrongly set. It is also reported on I8000.

Also reported on I8000 is batterydrain caused by sensordriver, so maybe I am right about reset function in driver. The devs are looking into this.

You can compile on windows, but I think it is hard. Google on 'windows cross compile eabi' and take a look at codesourcery.com

Download not latest it is (in Ubuntu) not compiling well.

downloads for windows https://sourcery.mentor.com/sgpp/lite/arm/portal/release1294

Tonight I will recompile with your settings. Also I will try to recompile androids sensor libs to our kernel if possible. In the current update I included spica's sensor libs. The original ones from installer package do not work on B7610

Recompiling android libs take several hours. After a crash I had to fresh import the source so I can't use cached compile.

Kernel compilation (zImage) takes some twenty minutes

Edited November 1, 2011 by erikcas