[Project] Porting 3.X kernel to Defy(+)

[Blechd0se]

so that's the fixed version?
http://www.blechdose-live.de/download/kernel/3.0.x/

Yes, i added only XZ compression and applied a patch for the lowmemorykiller.

If some other devs see this; the low performance is because of the SGX driver (2D and 3D) or did i miss something?

 

[drunk_ryder24]

kernel compiling error

tried to compile kernel 3.0.8 for defy but i get and gcc error of fiq not set using ndk

 

[Kayant]

tried to compile kernel 3.0.8 for defy but i get and gcc error of fiq not set using ndk

ndk

Last time I compiled 3.0 it works fine with the 4.5.4 linaro toolchain and the 4.6 one from google which built the kernel in my compiled build of CM10.1 two days ago......

Are you sure you set the CROSS_COMPILE path correctly because there should not be any problems apart from maybe the kernel failing to compile because of Quarx latest changes but ndk errors that's just weird

 

[Boss442]

what about audio/ril/bluetooth/etc. drivers? :silly:

 

[drunk_ryder24]

ndk

Last time I compiled 3.0 it works fine with the 4.5.4 linaro toolchain and the 4.6 one from google which built the kernel in my compiled build of CM10.1 two days ago......

Are you sure you set the CROSS_COMPILE path correctly because there should not be any problems apart from maybe the kernel failing to compile because of Quarx latest changes but ndk errors that's just weird

for google i set the path correctly,otherwise terminal would show gcc**not found but i used 4.4.3 toolchain?!?,tried codesourcery it gives me tons of scriptmod errors and the linaro toolchain hell knows where it did install

 

[Kayant]

for google i set the path correctly,otherwise terminal would show gcc**not found but i used 4.4.3 toolchain?!?,tried codesourcery it gives me tons of scriptmod errors and the linaro toolchain hell knows where it did install

if it's showing gcc not found then it's not being set properly........

Try this ---->

$ Export ARCH = arm

$ Export CROSS_COMPILE = ~/path to toolchain

e.g Export CROSS_COMPILE = ~/toolchain/4.5.4_linaro/bin/arm-eabi
Must sure it points to bin/arm-eabi that's why gcc errors are showing

$ make mapphone_defconfig

$ make -j<x>
x being number of your cpu threads plus one.....

 

[drunk_ryder24]

if it's showing gcc not found then it's not being set properly........

Try this ---->

$ Export ARCH = arm

$ Export CROSS_COMPILE = ~/path to toolchain

e.g Export CROSS_COMPILE = ~/toolchain/4.5.4_linaro/bin/arm-eabi
Must sure it points to bin/arm-eabi that's why gcc errors are showing

$ make mapphone_defconfig

$ make -j<x>
x being number of your cpu threads plus one.....

did this and allready got 2 different errors,ndk gives me an fiq module error(toolchain is recognized),codesourcery gives me error empty.o mod

 

[Kayant]

did this and allready got 2 different errors,ndk gives me an fiq module error(toolchain is recognized),codesourcery gives me error empty.o mod

If you're still getting errors try another tool chain either the 4.5.4 linaro or 4.6 that comes default in 4.2 and 4.4.3 should work also but why use it when you can get better compler options with the other two
These two toolchains work fine .... anything else am not sure if it works search it up......

 

[drunk_ryder24]

If you're still getting errors try another tool chain either the 4.5.4 linaro or 4.6 that comes default in 4.2 and 4.4.3 should work also but why use it when you can get better compler options with the other two
These two toolchains work fine .... anything else am not sure if it works search it up......

thx for your help,also i uset as kernel source the omap git and the command make mapphone_defconfig cloned from quarxs repo

 

[junesoung]

I'm making jb-3.0 device source..
https://github.com/Docker-J/android_device_moto_jordan-common
https://github.com/Docker-J/android_device_moto_mb526
https://github.com/Docker-J/proprietary_vendor_motorola

 

[Quarx]

I'm making jb-3.0 device source..
https://github.com/Docker-J/android_device_moto_jordan-common
https://github.com/Docker-J/android_device_moto_mb526
https://github.com/Docker-J/proprietary_vendor_motorola

i'm see only fork of my repo

The problem with the flickering display is, that the current power_mode wasn't the expected_mode and so the display gets rebooted every
couple of seconds.
So i set the power_mode to 0x09 and the display flickering stops. I already made a commit for that https://github.com/Blechd0se/jordan-kernel/commit/7e62c65ea22e3089950be038b62470b12efe0412

However i guess this is just a little workaround, i think we need to set the display permanently to the given value, at least the display works correct

I tried it, but commit doesn't change anything for me, same, if display enabled without problem at boot, it works fine all the time.

About gcc: 4.4.3/4.5/4.6 should works fine

TODO: need fix wifi and get adb debug via network

 

[Blechd0se]

AW: [Project] Porting 3.X kernel to Defy(+)

i'm see only fork of my repo

I tried it, but commit doesn't change anything for me, same, if display enabled without problem at boot, it works fine all the time.

About gcc: 4.4.3/4.5/4.6 should works fine

TODO: need fix wifi and get adb debug via network

I had the problem that the display gets rebooted, so this patch should work when its the case, if power_mode = expceted_mode everything is fine if not my patch solves it at least for me. its not a huge deal, so it wont cost that much implementing it

Sent from my MB526 using xda app-developers app

 

[Kayant]

I had the problem that the display gets rebooted, so this patch should work when its the case, if power_mode = expceted_mode everything is fine if not my patch solves it at least for me. its not a huge deal, so it wont cost that much implementing it

Sent from my MB526 using xda app-developers app

Btw the patch also It works for me also I wonder why it works for me and you but not Quarx??

 

[Quarx]

How this problem can be reproduced? After suspend?

 

[oSandmaNo]

Btw the patch also It works for me also I wonder why it works for me and you but not Quarx??

I think quarx isnt encountering the problem

Sent from my MB526 using xda app-developers app

 

[hDusan]

3.0 kernel working on MB525 green lens ?

Küldve az én GT-I9100-emről

 

[Kayant]

I think quarx isnt encountering the problem

Sent from my MB526 using xda app-developers app

How this problem can be reproduced? After suspend?

Yh it looks like it

I can't say of anyway to reproduce it but here is somethings I have noticed ---->
My test build is compiled with latest changes before your recent commit here and the patch by Blechd0se

The flickering starts from the bootanimation then continues to the startup when it continues randomly then stops for a bit before it continues usually if you start touching the screen. This is the behaviour I experienced with the test build you uploaded on the 23th which also happens on the test build I compiled when I reverted the zImage back to the one from your build with Blechd0se patch

With Blechd0se patch from the bootanimation everything is fine not a single flicker........

If there is anything I can do to produce some logs plz assist me and tell me because it's difficult to pull anything atm since you never know when it will actually boot

 

[Quarx]

Hmm this problem happened for me in first revisions of kernel, but after some changes, i didn't see problem anymore.

 

[Blechd0se]

Hmm this problem happened for me in first revisions of kernel, but after some changes, i didn't see problem anymore.

Cant quite tell if its still present, however i checked the dmesg output on my last commits and i see that
the display rebooted frequently without my changes. So yes, its reproducable without my commit
I guess Kayant has a Defy+, I have the same, could that be the problem,?

Does anyone have this issues with the testkernel? Quarx, you can add this to your source, if its really only affecting some people
(like me) then its a fallback and fixes it.
The problem is, the display gets rebooted at least one time (you wont notice it, but you can see it in the logs). So my idea was to put it
somewhere else.

If some one ca provide a dmesg output, it would be nice.
If there is somthing like:

Power mode in incorrect state, mode = 0x01f, expected = 0x09c

Then you have the same problem like me and need the patch, if you don't have this line, well everything is okay :cyclops:

 

[Kayant]

Cant quite tell if its still present, however i checked the dmesg output on my last commits and i see that
the display rebooted frequently without my changes. So yes, its reproducable without my commit
I guess Kayant has a Defy+, I have the same, could that be the problem,?

Does anyone have this issues with the testkernel? Quarx, you can add this to your source, if its really only affecting some people
(like me) then its a fallback and fixes it.
The problem is, the display gets rebooted at least one time (you wont notice it, but you can see it in the logs). So my idea was to put it
somewhere else.

If some one ca provide a dmesg output, it would be nice.
If there is somthing like:

Power mode in incorrect state, mode = 0x01f, expected = 0x09c

Then you have the same problem like me and need the patch, if you don't have this line, well everything is okay :cyclops:

Yh I think it could be a model issue I guess like you said since I do indeed have a defy+ like you and in my dmesg output I have the same lines....

<4>Uv:910100, Set vsel: 25
<3>mapphone-panel display0: Power mode in incorrect state, mode = 0x1c, expected = 0x9c
<3>mapphone-panel display0: ESD: performing LCD reset
<6>ESD: mapphone_panel_power_off.
<6>ESD: mdelay 20ms.
<6>ESD: mapphone_panel_power_on.
<4>AUO displays require a different setting: 0x09: 0x1f
<6>done EDISCO CTRL ENABLE
<6>Mapphone Display is ENABLE