[WIP][Thinktank][DEV]Kernel 3.4.x - Xperia Play Android Development

Hello,
As most threads are full of thanks posts or non development related posts, I am making a new one.
A note to everyone: This is a DEV thread, every post not related to this project will be removed!
What is the current status?
Currently the kernel is booting, but touchscreen isn't working.
What's working:
Booting into Android
Screen (Showing things, not touch)
Sound
What's not working or not tested because of touchscreen:
Everything else
Sources:
https://github.com/nickholtus/msm7x30-3.4.x-nAa
Credits:
NobodyAtAll for the initial source
MikeNG for a lot of fixes
DooMLoRD For giving me the device and a lot of work on the source
Mikeioannina A lot of work on the sources
Again: Please keep this thread development related!
If I helped you, please press thanks or consider donating to me!

---

Current status on touchscreen:
Still investigating for the problem, can't really find anything yet.

---

Threads with some more info:
http://forum.xda-developers.com/showthread.php?t=2563508

---

Is the driver crashing? Post dmesg
Sent from my C6902 using xda app-developers app

---

DooMLoRD said:
Is the driver crashing? Post dmesg
Sent from my C6902 using xda app-developers app
Click to expand...
Click to collapse
Will post this evening, am at work right now.

---

Bakisha posted a dmesg here http://forum.xda-developers.com/showpost.php?p=48869177&postcount=20
Although it is not from the most recently updated test kernel.

---

I've posted even full bugreport from zImage(4), wich i think is latest test version. I'll copy paste dmesg from there and attach it here

---

Bakisha said:
I've posted even full bugreport from zImage(4), wich i think is latest test version. I'll copy paste dmesg from there and attach it here
Click to expand...
Click to collapse
ok so here is what i see from the dmesg:
Code:
<6>[ 5.784790] synaptics_touchpad: product 1, firmware revision 13, bootloader date 2011-02-16
<6>[ 5.785827] input: synaptics_touchpad as /devices/virtual/input/input4
<6>[ 5.786254] synaptics_touchpad 4-0020: synaptics_touchpad_set_power: powered 1, activated 0, users 0, standby 0
<7>[ 5.786834] CY8CTMA300_TOUCH: V0.9b-mt built Jan 14 2014 22:15:21
<7>[ 5.787200] CY8CTMA300_TOUCH: SPI setup (requesting 1000000Hz) OK
...
<6>[ 5.949859] CY8CTMA300_TOUCH: TP_REG_FW 0x00
<6>[ 6.109924] CY8CTMA300_TOUCH: Bootloader FW Ver 0.0
<6>[ 6.129882] CY8CTMA300_TOUCH: Application FW Ver 0.0
<6>[ 6.149902] CY8CTMA300_TOUCH: Silicon Revision E
<6>[ 6.169891] CY8CTMA300_TOUCH: Customer ID 0
<6>[ 6.189880] CY8CTMA300_TOUCH: old FW version, applying reset workarounds
<6>[ 6.330108] input: cy8ctma300_touch as /devices/platform/spi_qsd.0/spi_master/spi0/spi0.0/input/input5
<6>[ 6.334136] CY8CTMA300_TOUCH: invalid data?
then...
Code:
<6>[ 33.252838] CY8CTMA300_TOUCH: invalid data?
<6>[ 33.651550] CY8CTMA300_TOUCH: invalid data?
<6>[ 34.024230] synaptics_touchpad 4-0020: synaptics_touchpad_set_power: powered 0, activated 0, users 1, standby 0
...
<6>[ 34.026611] synaptics_touchpad 4-0020: synaptics_touchpad_set_power: powered 0, activated 0, users 0, standby 0
<6>[ 34.027099] synaptics_touchpad 4-0020: synaptics_touchpad_set_power: powered 0, activated 0, users 1, standby 0
<6>[ 34.027130] synaptics_touchpad 4-0020: synaptics_touchpad_set_power: powered 0, activated 0, users 0, standby 0
<6>[ 34.050384] CY8CTMA300_TOUCH: invalid data?
<6>[ 34.491516] CY8CTMA300_TOUCH: invalid data?
<6>[ 34.890441] CY8CTMA300_TOUCH: invalid data?
see that the "set_power" is powered 0... something is powering down the touchpad, may be due to "invalid data?"
can u give me exact time stamp from bootup when the touchscreen stops working? post the corresponding dmesg...

---

DooMLoRD said:
see that the "set_power" is powered 0... something is powering down the touchpad, may be due to "invalid data?"
can u give me exact time stamp from bootup when the touchscreen stops working? post the corresponding dmesg...
Click to expand...
Click to collapse
I;ll attach dmesg that i just made, but i don't think that lines have something to do with touchscreen because line with that "powered 0" is even in dmesg logs of GB and ICS kernels for Play (i've checked on Lupus kernels for GB and ICS and on those touchpads and touchscreen works normally)
But i noticed difference in next lines:
Code:
<7>[ 5.525909] CY8CTMA300_TOUCH: V0.9b-mt built Jan 14 2014 22:15:21
<7>[ 5.526306] CY8CTMA300_TOUCH: SPI setup (requesting 1000000Hz) OK
<6>[ 5.584747] Charging with 0mA current
<6>[ 5.584960] entering into lpm
<6>[ 5.585144] msm_otg_suspend: Turn on phy comparators
<6>[ 5.585449] msm_otg_suspend: usb in low power mode
<6>[ 5.685119] CY8CTMA300_TOUCH: TP_REG_FW 0x00
<6>[ 5.845153] CY8CTMA300_TOUCH: Bootloader FW Ver 0.0
<6>[ 5.865142] CY8CTMA300_TOUCH: Application FW Ver 0.0
<6>[ 5.885192] CY8CTMA300_TOUCH: Silicon Revision E
<6>[ 5.905181] CY8CTMA300_TOUCH: Customer ID 0
<6>[ 5.925170] CY8CTMA300_TOUCH: old FW version, applying reset workarounds
<6>[ 6.065368] input: cy8ctma300_touch as /devices/platform/spi_qsd.0/spi_master/spi0/spi0.0/input/input5
<6>[ 6.071472] CY8CTMA300_TOUCH: invalid data?
<6>[ 6.084808] CY8CTMA300_TOUCH: Device registered OK
And in other kernels (even in Wedgess's 2.6 kernel for his test release JB4.2) it is like:
Code:
<7>CY8CTMA300_TOUCH: V0.9b-mt built Feb 5 2014 22:14:40
<7>CY8CTMA300_TOUCH: SPI setup (requesting 1000000Hz) OK
<6>CY8CTMA300_TOUCH: TP_REG_FW 0x00
<6>CY8CTMA300_TOUCH: Bootloader FW Ver 0.1
<6>CY8CTMA300_TOUCH: Application FW Ver 0.114
<6>CY8CTMA300_TOUCH: Silicon Revision E
<6>CY8CTMA300_TOUCH: Customer ID 5
<6>CY8CTMA300_TOUCH: SSD firmware [Project ID 7]
<3>CY8CTMA300_TOUCH: firmware versions match.
<6>input: cy8ctma300_touch as /devices/platform/spi_qsd.0/spi0.0/input/input3
<6>CY8CTMA300_TOUCH: Device registered OK
<6>GPIO Input Driver: Start gpio inputs for keypad-zeus in interrupt mode
<6>GPIO Input Driver: Start gpio inputs for keypad-zeus in interrupt mode
<6>GPIO Input Driver: Start gpio inputs for keypad-zeus in interrupt mode
<6>GPIO Input Driver: Start gpio inputs for keypad-zeus in interrupt mode
<6>GPIO Input Driver: Start gpio inputs for keypad-zeus in interrupt mode
<6>input: keypad-zeus as /devices/virtual/input/input4

---

i think probably is touchpad related also..seems like kernel is powering down the touchpads and touchscreen. @nickholtus and if you try kernel like NEOL as the board is the same, but kernel wont look for touchpads..and maybe we can flash on play without touchpads support..just a test,we can flash neol kernel and vice versa..only touchpads and camera inverted..but maybe its possible to make it work.

---

Personally I think the problem is here:
<6>[ 5.845153] CY8CTMA300_TOUCH: Bootloader FW Ver 0.0
<6>[ 5.865142] CY8CTMA300_TOUCH: Application FW Ver 0.0
<6>[ 5.885192] CY8CTMA300_TOUCH: Silicon Revision E
<6>[ 5.905181] CY8CTMA300_TOUCH: Customer ID 0
<6>[ 5.925170] CY8CTMA300_TOUCH: old FW version, applying reset workarounds
Click to expand...
Click to collapse
Because it can't get the right fw versions, it says old FW version, applying reset workarrounds.
In other kernel this isn't the case.

---

nickholtus said:
Personally I think the problem is here:
Because it can't get the right fw versions, it says old FW version, applying reset workarrounds.
In other kernel this isn't the case.
Click to expand...
Click to collapse
To add to this - it also applies a different file after realizing the FW version doesn't match up.
In Wedgess 2.6 Kernel is uses 'input3':
Code:
cy8ctma300_touch as /devices/platform/spi_qsd.0/spi_master/spi0/spi0.0/input/input3
Whereas in this Kernel is uses 'input5'
Code:
cy8ctma300_touch as /devices/platform/spi_qsd.0/spi_master/spi0/spi0.0/input/input5
I doubt it's as simple as this, but if this is referencing different files why not try copying/renaming 'input3' to 'input5' and trying again?
Edit: Yeah, it doesn't look like it's as simple as this. xD
Either that or try changing the FW versions? Since for some reason they are currently '0.00', we'd just need to somehow change them to:
Code:
<6>CY8CTMA300_TOUCH: Bootloader FW Ver 0.1
<6>CY8CTMA300_TOUCH: Application FW Ver 0.114
Another thing to note is that 'input3' is identified as 'pm8xxx-keypad'
Code:
<6>[ 5.400421] input: pm8xxx-keypad as /devices/platform/msm_ssbi.0/pm8058-core/pm8xxx-keypad/input/input3
Is this the same in the 2.6 Kernel?
@Bakisha - Would you be able to post the dmesg for the 2.6 Kernel please? I don't have the means to get it myself unfortunately - I was going to see if there were any other discrepancies between the two.

---

Spizzy01 said:
@Bakisha - Would you be able to post the dmesg for the 2.6 Kernel please? I don't have the means to get it myself unfortunately - I was going to see if there were any other discrepancies between the two.
Click to expand...
Click to collapse
It's little bit old and it don't have timestamp, but untill i can make new one, it will do. Attaching (it is copy-pasted from previous bugreport i posted in other thread)
As for myself, i'll look in differences in source code between Nick's and Wedgess's kernel, to see if there is any differences

---

Hi all,
I hav some news, but it's not directly related to the kernel.
The play I have is going to Achotjan for some time, he is going to develop for it.
We are going to develop together on this kernel. After a while i will get the play back.
So hopefully progress will go a bit faster now we have a new developer.
Regards,
Nick

---

I'm trying to rebuild my Linux development/kernel environment and compile the Kernel in preparation for helping nick/nAa if required.
Has anyone ever ran into this before
Code:
CHK include/generated/utsrelease.h
/home/spizzy/android/msm7x30/scripts/gcc-version.sh: line 25: arm-eabigcc: command not found
/home/spizzy/android/msm7x30/scripts/gcc-version.sh: line 26: arm-eabigcc: command not found
HOSTCC scripts/conmakehash
CC scripts/mod/empty.o
Generating include/generated/mach-types.h
arm-eabigcc: No such file or directory
Is your PATH set correctly?
make[2]: *** [scripts/mod/empty.o] Error 2
make[1]: *** [scripts/mod] Error 2
make[1]: *** Waiting for unfinished jobs....
CC kernel/bounds.s
arm-eabigcc: No such file or directory
Is your PATH set correctly?
make[1]: *** [kernel/bounds.s] Error 2
make: *** [prepare0] Error 2
make: *** Waiting for unfinished jobs....
make: *** [scripts] Error 2
I'm using nAa's sources and the paths are set correctly... I can't understand why it's saying it's not. Unless I'm missing something else? oO
I'll look into it more tonight if I can't work it out soon, but if anyone can help or if anyone is able to talk me through it that'd be great.. (Y)
Edit: Nvm. Fixed. xD

---

Spizzy01 said:
I'm using nAa's sources and the paths are set correctly... I can't understand why it's saying it's not. Unless I'm missing something else? oO
I'll look into it more tonight if I can't work it out soon, but if anyone can help or if anyone is able to talk me through it that'd be great.. (Y)
Click to expand...
Click to collapse
Wich toolchain are you using to build zImage?
Or post link to source that you are trying to compile and i'll look into it too...

---

Bakisha said:
Wich toolchain are you using to build zImage?
Or post link to source that you are trying to compile and i'll look into it too...
Click to expand...
Click to collapse
I fixed that issue now... Had issues with the filesystem/Linux itself lol.
Now I'm getting this though:
Code:
drivers/gpu/msm/adreno.c: In function 'adreno_iommu_setstate':
drivers/gpu/msm/adreno.c:425: warning: the frame size of 1040 bytes is larger than 1024 bytes; see http://go/big_stack_frame
error, forbidden warning: adreno.c:425
make[3]: *** [drivers/gpu/msm/adreno.o] Error 1
make[2]: *** [drivers/gpu/msm] Error 2
make[1]: *** [drivers/gpu] Error 2
make: *** [drivers] Error 2
make: *** Waiting for unfinished jobs....
Hopefully an easy fix.. Working on it now lol.
Oh Linux... Y U DO DIS! O_O
FYI: I'm using this toolchain:
Code:
github.com/AndroidUnited/Android_Toolchains

---

Spizzy01 said:
FYI: I'm using this toolchain:
Code:
github.com/AndroidUnited/Android_Toolchains
Click to expand...
Click to collapse
Maybe to try with 4.7?
I'm using the one from
https://github.com/DooMLoRD/android_prebuilt_toolchains/tree/master/arm-linux-androideabi-4.7

---

Bakisha said:
Maybe to try with 4.7?
I'm using the one from
https://github.com/DooMLoRD/android_prebuilt_toolchains/tree/master/arm-linux-androideabi-4.7
Click to expand...
Click to collapse
I changed the toolchain as you suggested... It's building currently, I'll let you know if it gets further
Thanks for the replies Bakisha. Much appreciated!
Edit: Sadly it died... And funnily enough, on the touchscreen.
By the way, I haven't modified anything in the source yet. At least, not for this installation. I do however hopefully have a fix which I want to try once I get this working correctly.
Code:
drivers/input/touchscreen/cy8ctma300_touch.c: In function 'cy8ctma300_touch_probe':
drivers/input/touchscreen/cy8ctma300_touch.c:1354:35: error: expected ')' before ';' token
drivers/input/touchscreen/cy8ctma300_touch.c:1382:1: error: expected ')' before '}' token
drivers/input/touchscreen/cy8ctma300_touch.c:1382:1: error: expected ')' before '}' token
drivers/input/touchscreen/cy8ctma300_touch.c:1382:1: error: expected expression before '}' token
drivers/input/touchscreen/cy8ctma300_touch.c:1329:3: error: label 'err_gpio_setup' used but not defined
drivers/input/touchscreen/cy8ctma300_touch.c:1382:1: warning: control reaches end of non-void function [-Wreturn-type]
error, forbidden warning: cy8ctma300_touch.c:1382
make[3]: *** [drivers/input/touchscreen/cy8ctma300_touch.o] Error 1
make[2]: *** [drivers/input/touchscreen] Error 2
make[1]: *** [drivers/input] Error 2
make: *** [drivers] Error 2
make: *** Waiting for unfinished jobs....
Could you by any chance upload your 'cy8ctma300_touch.c' file please? I have a feeling the file I have has been modified somehow... Which makes no sense as it was cloned straight from github... oO
I've found an older file myself, I'll try this one too...
Edit: OK... I'm pretty sure this is the problem...? Can someone try to compile the Kernel again, and select only '2' workers (aka -j2). If I put a large amount of workers then it'll most likely compile as it waits for the rest of the tasks to finish... But in actual fact, the kernel isn't compiling correctly because of the error above...? Unless of course I'm the only one getting this error... But as I said, I've not made any changes yet.
I'm gonna go through the file and see if I can work out what the issue is, most likely it's just a missing expression or something. I'll also try sourcing the file from somewhere else, see if it helps..

---

Spizzy01 said:
By the way, I haven't modified anything in the source yet. At least, not for this installation. I do however hopefully have a fix which I want to try once I get this working correctly.
Code:
drivers/input/touchscreen/cy8ctma300_touch.c: In function 'cy8ctma300_touch_probe':
drivers/input/touchscreen/cy8ctma300_touch.c:1354:35: error: expected ')' before ';' token
drivers/input/touchscreen/cy8ctma300_touch.c:1382:1: error: expected ')' before '}' token
drivers/input/touchscreen/cy8ctma300_touch.c:1382:1: error: expected ')' before '}' token
drivers/input/touchscreen/cy8ctma300_touch.c:1382:1: error: expected expression before '}' token
drivers/input/touchscreen/cy8ctma300_touch.c:1329:3: error: label 'err_gpio_setup' used but not defined
drivers/input/touchscreen/cy8ctma300_touch.c:1382:1: warning: control reaches end of non-void function [-Wreturn-type]
error, forbidden warning: cy8ctma300_touch.c:1382
make[3]: *** [drivers/input/touchscreen/cy8ctma300_touch.o] Error 1
make[2]: *** [drivers/input/touchscreen] Error 2
make[1]: *** [drivers/input] Error 2
make: *** [drivers] Error 2
make: *** Waiting for unfinished jobs....
Click to expand...
Click to collapse
I found a workaround for that.
To me it looks like number of opened "(" doesn't match with closed ")", so I copied code from Lupus kernel:
Change:
Code:
while (((err = reset_device(this);
to
Code:
while (((err = reset_device(this)) < 0) && reset_retries--) {
DEBUG_PRINTK(KERN_DEBUG
"CY8CTMA300_TOUCH: retrying reset_device()\n");
msleep(100);
}
Don't ask why, i don't understand what it does, but for me it fixed problem with compiling (in case few more errors, compare this file with file from lupus jb 2.6 kernel )