Latest product reviews
ARCHOS 50 Diamond
ARCHOS GamePad2
ARCHOS
Smartphones
ARCHOS
TV Connect
ARCHOS 101 XS

A A A
Avatar

Please consider registering
Guest

Search

— Forum Scope —






— Match —





— Forum Options —





Minimum search word length is 3 characters - maximum search word length is 84 characters

Register Lost password?

ClockworkMod (CWM) Recovery easy install for RK3066 and RK3188 -- TWRP/CWM Flash-Tool and root for RK3288

 Please donate to support OMA and CrewRKTablets firmware work, thank you !

sp_Feed sp_TopicIcon
Updating TrekStor Ventos 10.1 (ST10216-2A) to Android 4.4?
Any way to update this device to a more recent ROM/kernel/Android version?
Avatar
JochenKauz

Moderator

Firmware Guru
Forum Posts: 450
Member Since:
2012/11/19
sp_UserOfflineSmall Offline

Thanks Received: 76
91
2015/03/08 - 15:53
sp_Permalink sp_Print

Hi,

the problem is that there are several ways of turning the tablet.

You have the sensor.so in android and the kernel driver. Normally the sensor.so in Android should not make any rotation by it self, but unfortunately some developer do this. Our sensor.so does not turn in any way, so you should do it in the kernel.

Looking at your kernel source of the driver I think you should experiment with the following method:

STK831x_ReportValue
there you find the reporting of x,y and z values:
input_report_abs(stk->input_dev, ABS_X, (stk->raw_data[1]));
input_report_abs(stk->input_dev, ABS_Y, -(stk->raw_data[2]));
input_report_abs(stk->input_dev, ABS_Z, (stk->raw_data[0]));
 

raw_data[1] is x, raw_data[2] is y and raw_data[0] is z

if you want to swap x and y change the values according to your needs, e.g.

input_report_abs(stk->input_dev, ABS_X, -(stk->raw_data[2]));
input_report_abs(stk->input_dev, ABS_Y, (stk->raw_data[1]));
input_report_abs(stk->input_dev, ABS_Z, (stk->raw_data[0]));
 

Cheers

JochenKauz

The following users say thank you to JochenKauz for this useful post:

awl
Avatar
awl
Regular Member
Forum Posts: 90
Member Since:
2015/02/11
sp_UserOfflineSmall Offline

Thanks Received: 1
92
2015/03/08 - 18:25
sp_Permalink sp_Print

Thanks, Jochen, for the quick reply and the pointer.

It was not that easy as to simply swap [1] and [2], because, as suspected, the developers of sensor.rk30board.so additionally swap the values in the sensor library, so what is called X, Y and Z in the kernel driver won't end up as X, Y and Z after the sensor library and also change the algebraic signs from + to - and the other way round.

So I ended up checking all the possible permutations, and came up with the following, which now indeed works fine:

        input_report_abs(stk->input_dev, ABS_X, -(stk->raw_data[0]));
        input_report_abs(stk->input_dev, ABS_Y, -(stk->raw_data[2]));
        input_report_abs(stk->input_dev, ABS_Z, (stk->raw_data[1]));

So for the rest of today I will now finally look into whether I get the light sensor to work (which really is optional), but everything important seems to work fine now! Cool

I'll then take a little rest (as I also will be quite busy during next week), and then try to look into using a 3.0.36+ kernel (without SELinux) and a KitKat build.

@Oma:
What is the best way to provide my kernel.img, boot.img and "model patch" ZIP to you for the CM 10.1 image? Can I upload it somewhere, or shall I send them to you via PN or mail?

Thanks again for all your kind help during this interesting lesson in looking behind the scenes of how Android kernels and ROMs work behind the scenes... Wink

Avatar
JochenKauz

Moderator

Firmware Guru
Forum Posts: 450
Member Since:
2012/11/19
sp_UserOfflineSmall Offline

Thanks Received: 76
93
2015/03/08 - 20:30
sp_Permalink sp_Print

Glad to hear that it now works.

Perhaps it would be possible for you to upload your kernel source to github, so other people could build an own kernel.

 

Cheers

JochenKauz

The following users say thank you to JochenKauz for this useful post:

awl
Avatar
awl
Regular Member
Forum Posts: 90
Member Since:
2015/02/11
sp_UserOfflineSmall Offline

Thanks Received: 1
94
2015/03/08 - 23:22
sp_Permalink sp_Print sp_EditHistory

Hello Jochen & Oma,

I don't know whether a git will make much sense here, as the kernel source has already been published by Trekstor - unfortunately not managed by git, but as a plain .tar.gz file:

TrekStor 3.0.8+ kernel sources for ST10216-2A

My changes to these kernel sources are really minimal: They mainly consist of an adapted config file, two Makefile/Kconfig fixes, replacing the MT5931 driver v2.07 by v2.14, fixing the EINVAL issue in this driver (see earlier in this thread) and changing the STK8312 source to make rotation work properly.

Also, all of these changes do only make sense if you use the adapted boot.img as well as my "model fix" (built upon and extending Oma's first attempt for a model fix) for the Ventos 10.1 ST10216-2A, and you will also need a custom TWRP (or CWM) recovery build that uses either the stock or my adapted kernel in order to apply the model fix.

In my opinion, it makes more sense to provide the binaries for other owners of a Trekstor ST10216-2A tablet through your site (in case you are interested). In addition, I'll be happy to provide my binaries (as well as source changes/diffs) on demand, answer questions about them and collect some feedback here in this thread or via PM... Smile

Best regards,

awl


Avatar
awl
Regular Member
Forum Posts: 90
Member Since:
2015/02/11
sp_UserOfflineSmall Offline

Thanks Received: 1
95
2015/03/08 - 23:50
sp_Permalink sp_Print

BTW: AnTuTu benchmark result for the ST10216-2A with this new kernel/ROM: 17581 - not too bad in terms of price/performance ratio... Cool

Avatar
Oma7144

Moderator

Firmware Guru
Forum Posts: 6315
Member Since:
2012/10/06
sp_UserOfflineSmall Offline

Thanks Received: 1377
96
2015/03/09 - 06:20
sp_Permalink sp_Print

Fine. Pls attach kernel.img, boot.img and model fix to a post here. I will upload the files to our archive and document the tab on the blog.

- Oma -

Avatar
JochenKauz

Moderator

Firmware Guru
Forum Posts: 450
Member Since:
2012/11/19
sp_UserOfflineSmall Offline

Thanks Received: 76
97
2015/03/09 - 09:31
sp_Permalink sp_Print

It would be nice if you publish your changes here, so I could upload the merged source to our gitlab and give the credits to you.

Even if Trekstor published the base source of the kernel, it is important for the GPL to publish the source again if it gets modified.

 

Cheers

JochenKauz

Avatar
awl
Regular Member
Forum Posts: 90
Member Since:
2015/02/11
sp_UserOfflineSmall Offline

Thanks Received: 1
98
2015/03/09 - 18:47
sp_Permalink sp_Print

Hello Oma,

please find attached in the ZIP file

  • my kernel image (with MT5931 v2.14, which from my experience greatly improves WiFi reception/interoperability with e.g. Deutsche Telekom Speedport routers over v2.07),
  • my boot.img (in which an updated init.rk30board.rc is needed to initialize STK8312 and run wpa_supplicant with nl80211), and
  • the "model fix" update ZIP file to be installed via TWRP or CWM recovery.

For completeness, I also include a version of TWRP 10.1 that I have modified for Ventos ST10216-2a ("Ventos 10c") to be used as recovery image (rename to recovery.img). Special credits for this go to user "mitchum67", who has created the TWRP build for ST10216 "Ventos 10a" and ST10216-2 "Ventos 10b") here:

http://www.android-hilfe.de/root-custom-roms-modding-fuer-trekstor-surftab-ventos-10-1/551771-twrp-ventos10-b-teamwinrecovery-fuer-das-ventos10-1-b.html.

 

Hello Jochen,

the attached ZIP file also contains the unified diff of my kernel sources against stock Ventos ST10216-2A kernel as published by Trekstor (see link in post #94).

 

If you have any questions, just let me know...

Thanks & best regards,

awl

Avatar
JochenKauz

Moderator

Firmware Guru
Forum Posts: 450
Member Since:
2012/11/19
sp_UserOfflineSmall Offline

Thanks Received: 76
99
2015/03/10 - 00:14
sp_Permalink sp_Print

Hi awl,

I applied your patches and published your kernel, with credits, at our git:

http://crewrktablets.arctablet.com:8080/crewrktablets/android_kernel_trekstor_ventos10-1.git

 

Thanks for your work!

 

Cheers

JochenKauz

PS. a funny thing that I found is, that I was not able to compile it with my gcc 4.9.2. I will look at it, perhaps it can be done. The gcc 4.9.2 is much faster. If I can manage it, I will provide you a test kernel.

The following users say thank you to JochenKauz for this useful post:

awl
Avatar
JochenKauz

Moderator

Firmware Guru
Forum Posts: 450
Member Since:
2012/11/19
sp_UserOfflineSmall Offline

Thanks Received: 76
100
2015/03/10 - 01:58
sp_Permalink sp_Print sp_EditHistory

okay,

found the problem. In the mt5931 driver there was a wrong inline declaration of a function. The Linaro toolchain ignored the bug until gcc 4.6, but not in higher versions. I corrected it. If you want, give the attached kernel a try, it is compiled with Linaro gnueabihf 4.9.3.

Cheers

JochenKauz

The following users say thank you to JochenKauz for this useful post:

awl
Avatar
awl
Regular Member
Forum Posts: 90
Member Since:
2015/02/11
sp_UserOfflineSmall Offline

Thanks Received: 1
101
2015/03/10 - 11:18
sp_Permalink sp_Print

Hello Jochen,

many thanks for your late night work! Smile

Unfortunately the kernel compiled by Linaro gnueabihf 4.9.3 fails in the exact same way like my previous attempts with the Ubuntu 14.04.01 default compiler version Linaro gcc 4.8.2:

The boot process fails: somehow "hangs" before the boot animation (animated crewrktablets logo) is displayed.

Using adb, a device is detected, but any attempt to connect to this device fails, because not even the root device has been mounted properly in read-write mode:

c:\adb>adb shell
- exec '/system/bin/sh' failed: No such file or directory (2) -

The interesting thing with this subtle gcc bug is that, by changing some option values the kernel compile configuration, with some luck, you may find a particular configuration in which the kernel boots fine even with the new gcc, but when activating all those config options that are needed for the Ventos, the kernel builds with at least gcc 4.8 onwards simply do not work. I lost a couple of days during this process before I found out that it is nothing but recent gcc versions to blame for the fact that some of my compiled kernels booted fine and some simply hung... Yell

In case you have any connections to the gcc/Linaro teams, it would be really interesting and helpful to try and sort out with them what goes wrong there or whether there might already be some newly introduced known bugs in code creation (that have not been there in 4.4.x), but I fear that in order to do so, a serial debug console would be needed, which will break the sealed case of my tablet (even Trekstor does not seem to repair them, but only replace...), so I clearly won't volunteer... Frown

Sorry to not have any better news...

Best regards,

awl

 

P.S.:

Note that when you checked in your patch to the MT5931 header file to fix the "inline" issue, you accidentally checked in the kernel and some other unrelated file (both being build results), which makes git slow and unable to display the patch via the web GUI...

Avatar
Oma7144

Moderator

Firmware Guru
Forum Posts: 6315
Member Since:
2012/10/06
sp_UserOfflineSmall Offline

Thanks Received: 1377
102
2015/03/10 - 20:09
sp_Permalink sp_Print

awl said
BTW: AnTuTu benchmark result for the ST10216-2A with this new kernel/ROM: 17581 - not too bad in terms of price/performance ratio... Cool

Update confirmed installation list: http://crewrktablets.arctablet.com/?p=1913

It was a pleasure working with you!

- Oma -

The following users say thank you to Oma7144 for this useful post:

awl
Avatar
awl
Regular Member
Forum Posts: 90
Member Since:
2015/02/11
sp_UserOfflineSmall Offline

Thanks Received: 1
103
2015/03/10 - 20:23
sp_Permalink sp_Print

Thank you, Oma,

the same applies to you (and also to Jochen!) as well: Your kind support and super-fast replies even at night or over the weekend were just great! Smile

And the journey might even go on:

As stated, I am tempted to go back and try again with the 3.0.36+ kernel and a KitKat ROM in some weeks when I have more spare time (remember that I already had a compile-clean kernel with most of the HW support needed, I need to re-check whether gcc was the root cause of the "hang" issues here as well...). And I promise to make another donation if we can even succeed to get this to run... Wink

Will get back to you on this in some weeks time... Cool

Thanks & best wishes,

awl

Avatar
JochenKauz

Moderator

Firmware Guru
Forum Posts: 450
Member Since:
2012/11/19
sp_UserOfflineSmall Offline

Thanks Received: 76
104
2015/03/10 - 22:34
sp_Permalink sp_Print sp_EditHistory

awl said
Note that when you checked in your patch to the MT5931 header file to fix the "inline" issue, you accidentally checked in the kernel and some other unrelated file (both being build results), which makes git slow and unable to display the patch via the web GUI...

this is not a problem of the files.. it is a problem of gitlab itself.

I will delete them, but browse functions has a bug in this gitlab version. But.. what other files you mean? I can only see the kernel.img.

I think I will talk with some people regarding the bug of the compiler, perhaps they have an idea.

 

Cheers

JochenKauz

The following users say thank you to JochenKauz for this useful post:

awl
Avatar
JochenKauz

Moderator

Firmware Guru
Forum Posts: 450
Member Since:
2012/11/19
sp_UserOfflineSmall Offline

Thanks Received: 76
105
2015/03/10 - 23:48
sp_Permalink sp_Print

commit fixed.

The following users say thank you to JochenKauz for this useful post:

awl
Forum Timezone: Europe/Paris

Most Users Ever Online: 749

Currently Online:
82 Guest(s)

Currently Browsing this Page:
1 Guest(s)


Devices in use: Desktop (79), Phone (3)

Top Posters:

finless: 604

DarthJabba: 551

maikal: 394

mussonero1: 350

alex: 252

damo: 243

DanielVd: 237

Mark06: 222

Newest Members:

niteaura

radolini

fede32

Ty3uK

pipirska

telo000788

Forum Stats:

Groups: 10

Forums: 185

Topics: 6037

Posts: 60500

 

Member Stats:

Guest Posters: 43

Members: 262194

Moderators: 5

Admins: 1

Administrators: admin

Moderators: globula_neagra, exelletor, JochenKauz, Oma7144, cracktech


CrewRKTablets moderators: JochenKauz and Astralix