[Q] Getting IOMX interface - Galaxy Tab 10.1 Q&A, Help & Troubleshooting

Hi,
I am trying to get access to the IOMX interface for use in a video decoding app.
I realize the API's I am using are non public and subject to change but the public API are not good enough for me.
I need access to the compressed and decoded data and so this seems to the way to go.
I found some sample code on the net that shows how to get the IOMX pointer.
(tried to post the link but it won't let me since, apparently I am a noob :cyclops
So I tried running that and I get a crash.
This is the code:
Code:
sp<IServiceManager> sm = defaultServiceManager();
Vector<String16> vec = sm->listServices();
sp<IBinder> binder = sm->getService(String16("media.player"));
sp<IMediaPlayerService> service = interface_cast<IMediaPlayerService>(binder);
[COLOR="Red"]sp<IOMX> mOMX = service->getOMX();[/COLOR]
And the crash occurrs on the last line (red) - start of dump below.
All I want is to be able to enumerate the OMX components for now. When that works I expect I can continue.
Should I be initializing anything before running the above code? OMX? Stagefright?
Thank you in advance,
Tomer
06-25 10:02:40.300: I/DEBUG(13769): *** *** *** *** *** *** *** *** *** *** *** *** *** *** *** ***
06-25 10:02:40.300: I/DEBUG(13769): Build fingerprint: 'samsung/GT-P7500/GT-P7500:3.2/HTJ85B/JJKK3:user/release-keys'
06-25 10:02:40.300: I/DEBUG(13769): pid: 15347, tid: 15359 >>> com.ai.playcast <<<
06-25 10:02:40.300: I/DEBUG(13769): signal 11 (SIGSEGV), code 1 (SEGV_MAPERR), fault addr ae684e99
06-25 10:02:40.300: I/DEBUG(13769): r0 ae684e99 r1 ae684e99 r2 00000001 r3 00000008
06-25 10:02:40.300: I/DEBUG(13769): r4 5c55e914 r5 ae684e99 r6 5c55e9d8 r7 00268910
06-25 10:02:40.300: I/DEBUG(13769): r8 a90388f5 r9 ae684e99 10 00268048 fp fffffe90
06-25 10:02:40.300: I/DEBUG(13769): ip a81233dc sp 5c55e8c0 lr a811aafd pc aff14aec cpsr 00000010

Questions go in the Q&A section

I have some updates on this that may make this less opaque.
I looked into the implementation of getOMX().
It seems straight forward enough so I copied it to my own code instead of calling getOMX(). This time it didn't crash - this is becoming a little voodoo.
However I got the handle to the IOMX object.
So then I tried enumerating the components but the list comes back empty.
Anyone has any idea why?
Here is my original crashing code (crashing on the red line):
Code:
android::ProcessState::self()->startThreadPool();
sp<IServiceManager> sm = defaultServiceManager();
sp<IBinder> binder = sm->getService(String16("media.player"));
sp<IMediaPlayerService> service = interface_cast<IMediaPlayerService>(binder);
[COLOR="Red"]sp<IOMX> mOMX = service->getOMX();[/COLOR]
This is my non-crashing code that returns an empty component list:
Code:
android::ProcessState::self()->startThreadPool();
sp<IServiceManager> sm = defaultServiceManager();
sp<IBinder> binder = sm->getService(String16("media.player"));
sp<IMediaPlayerService> service = interface_cast<IMediaPlayerService>(binder);
sp<IOMX> mOMX;
{
Parcel data, reply;
data.writeInterfaceToken(service->getInterfaceDescriptor());
binder->transact(6, data, &reply);
mOMX = interface_cast<IOMX>(reply.readStrongBinder());
}
List<IOMX::ComponentInfo> componenInfos;
status_t err = mOMX->listNodes(&componenInfos);
if (componenInfos.size() == 0)
[COLOR="red"]log_err("no components found %d", err);[/COLOR]
I've also tried calling OMX_Init() before all this but it didn't help.
Please, if anyone has any idea what's going on I would greatly appreciate it.
Thank you in advance

Related

Boot loop with CM 9.0 and 9.1 - SOLVED

:victory:
Hi everyone,
I love this phone! Unfortunately I am stuck at this time with CM 7.2, because when I tried 9.0 and 9.1 I get this boot loop:
Code:
E/libgenlock( 1005): genlock_create_lock: open genlock device failed (err=No such file or directory)
E/msm7x30.gralloc( 1005): alloc_impl: genlock_create_lock failed
E/msm7x30.gralloc( 1005): terminateBuffer: unmapping a non pmem/ashmem buffer flags = 0x0
E/memalloc( 1005): getAllocator: Invalid flags passed: 0x0
E/msm7x30.gralloc( 1005): Could not unmap memory at address 0x21
E/memalloc( 1005): getAllocator: Invalid flags passed: 0x0
F/libc ( 1005): Fatal signal 11 (SIGSEGV) at 0x00000000 (code=1)
I/DEBUG ( 121): *** *** *** *** *** *** *** *** *** *** *** *** *** *** *** ***
I/DEBUG ( 121): Build fingerprint: 'SEMC/SK17i_1249-7380/SK17i:4.0.4/4.1.B.0.431/UL5_3w:user/release-keys'
I/DEBUG ( 121): pid: 1005, tid: 1018 >>> system_server <<<
I/DEBUG ( 121): signal 11 (SIGSEGV), code 1 (SEGV_MAPERR), fault addr 00000000
Any ideas what's going on? I've also gotten the same thing with an Xperia Active. I must be doing something wrong I think?
Thank you!
Edit: This is related to a few bug reports:
http://code.google.com/p/cyanogenmod/issues/detail?id=5504
http://code.google.com/p/cyanogenmod/issues/detail?id=5586
http://code.google.com/p/cyanogenmod/issues/detail?id=5522
I also found this thread: http://forum.xda-developers.com/showthread.php?t=1737703 which indicates that the problem could be that I need to reformat with some other partition type? I tried the SuperWipe file they suggested but it won't run on the SonyEricsson. I'm pretty good with ADB and I know how to use Clockwork Mod and all that but I don't know how to mess around with partitions (yet).
These bug reports say it is fixed but it is not, at least on my two devices here. Anything I can do to help figure this out, I will try!
Follow this http://forum.xda-developers.com/showthread.php?p=30732959
Sent from my WT19i using xda app-developers app
rachit.rc96 said:
Follow this http://forum.xda-developers.com/showthread.php?p=30732959
Click to expand...
Click to collapse
Thanks for the suggestion. No luck on that. I formatted /system, wiped the cache, /data, and Dalvik cache. Nothing changed. Anything else I should do? Surely I'm not the only one getting this?
UPDATE: SUCCESS!!!
The magic is that I need to install the new boot.img that comes with CM 9.1 ICS. The way to do this is:
adb reboot bootloader
Then unzip the cm-9.1-mango.zip file, find the boot.img file, and install it using fastboot:
./platform-tools/fastboot flash boot /tmp/f/boot.img
Then everything will work! My problem is that I had previously installed CM 7.2 and the install script for CM 9.0 doesn't update the boot partition from that. Anyway, of course it's user error, and now I'm glad I have it working. Now I get to play around with ICS on my favorite phone!
Toughdroid said:
Thanks for the suggestion. No luck on that. I formatted /system, wiped the cache, /data, and Dalvik cache. Nothing changed. Anything else I should do? Surely I'm not the only one getting this?
Thanks
Click to expand...
Click to collapse
follow the same procedure but do not format system, it will work then

[Q] Build Kernel MT6577 - Can't compile

Hello... I tried compiling kernel from wiko source https://github.com/wiko-sources/cink-king
The problem is when I compile kernel by this command ./makeMtk s9081 r k
I'm getting this errors
/home/anoopkumar/ANDROID/ALPS/kernel/mediatek/kernel/drivers/combo/drv_wlan/mt6620/wlan/os/linux/include/gl_kal.h:1860:1: error: inlining failed in call to always_inline 'kalOidComplete': function body not available
/home/anoopkumar/ANDROID/ALPS/kernel/mediatek/kernel/drivers/combo/drv_wlan/mt6620/wlan/os/linux/include/gl_os.h:
/home/anoopkumar/ANDROID/ALPS/kernel/mediatek/kernel/drivers/combo/drv_wlan/mt6620/wlan/common/wlan_lib.c:2355:35: error: called from here
/home/anoopkumar/ANDROID/ALPS/kernel/mediatek/kernel/drivers/combo/drv_wlan/mt6620/wlan/os/linux/include/gl_kal.h:1860:1: error: inlining failed in call to always_inline 'kalOidComplete': function body not available
/home/anoopkumar/ANDROID/ALPS/kernel/mediatek/kernel/drivers/combo/drv_wlan/mt6620/wlan/os/linux/include/gl_os.h:
/home/anoopkumar/ANDROID/ALPS/kernel/mediatek/kernel/drivers/combo/drv_wlan/mt6620/wlan/common/wlan_lib.c:2274:27: error: called from here
/home/anoopkumar/ANDROID/ALPS/kernel/mediatek/kernel/drivers/combo/drv_wlan/mt6620/wlan/os/linux/include/gl_kal.h:1860:1: error: inlining failed in call to always_inline 'kalOidComplete': function body not available
/home/anoopkumar/ANDROID/ALPS/kernel/mediatek/kernel/drivers/combo/drv_wlan/mt6620/wlan/os/linux/include/gl_os.h:
/home/anoopkumar/ANDROID/ALPS/kernel/mediatek/kernel/drivers/combo/drv_wlan/mt6620/wlan/common/wlan_lib.c:2097:39: error: called from here
/home/anoopkumar/ANDROID/ALPS/kernel/mediatek/kernel/drivers/combo/drv_wlan/mt6620/wlan/os/linux/include/gl_kal.h:1860:1: error: inlining failed in call to always_inline 'kalOidComplete': function body not available
/home/anoopkumar/ANDROID/ALPS/kernel/mediatek/kernel/drivers/combo/drv_wlan/mt6620/wlan/os/linux/include/gl_os.h:
/home/anoopkumar/ANDROID/ALPS/kernel/mediatek/kernel/drivers/combo/drv_wlan/mt6620/wlan/common/wlan_lib.c:2427:35: error: called from here
make[8]: *** [mediatek/kernel/drivers/combo/drv_wlan/mt6620/wlan] Error 2
make[7]: *** [mediatek/kernel/drivers/combo/drv_wlan/mt6620] Error 2
make[6]: *** [mediatek/kernel/drivers/combo/drv_wlan] Error 2
make[5]: *** [mediatek/kernel/drivers/combo] Error 2
make[4]: *** [mediatek/kernel/drivers] Error 2
make[3]: *** [mediatek/kernel] Error 2
make[2]: *** [sub-make] Error 2
My Config:
[OS]: Ubuntu 10.04.4 (64-bit)
[make]: 3.81 (64-bit)
[perl]: 5.10.1 (64-bit)
[python]: 2.6.5 (64-bit)
[arm-linux-androideabi-gcc]: 4.7 (64-bit)
[gcc]: 4.4.3 (64-bit)
[jdk]: 1.6.0_29 (64-bit)
[bison]: 2.4.1 (64-bit)
[flex]: 2.5.35 (64-bit)
[gperf]: 3.0.3 (64-bit)
[mingw]: Installed
[unix2dos/tofrodos]: Installed
Please help me to fix this issue....
Please Anyone! Help Us!!!
open /mediatek/kernel/drivers/combo/drv_wlan/mt6620/wlan/os/linux/include/gl_kal.h
search for inline
then you'll see
Code:
inline VOID
kalOidComplete (
IN P_GLUE_INFO_T prGlueInfo,
IN BOOLEAN fgSetQuery,
IN UINT_32 u4SetQueryInfoLen,
IN WLAN_STATUS rOidStatus
);
remove the inline and save
let like this to be precise
Code:
VOID
kalOidComplete (
IN P_GLUE_INFO_T prGlueInfo,
IN BOOLEAN fgSetQuery,
IN UINT_32 u4SetQueryInfoLen,
IN WLAN_STATUS rOidStatus
);
you'll need to do it on the 6628 too

Kernel Compiling Error

Code:
CC net/netfilter/xt_qtaguid.o
CC net/netfilter/xt_quota.o
CC net/netfilter/xt_quota2.o
CC net/netfilter/xt_socket.o
In file included from include/linux/kernel.h:23:0,
from include/linux/cache.h:4,
from include/linux/time.h:7,
from include/linux/stat.h:60,
from include/linux/module.h:10,
from net/netfilter/xt_socket.c:13:
net/netfilter/xt_socket.c: In function 'xt_socket_get4_sk':
include/linux/dynamic_debug.h:75:21: warning: 'protocol' may be used uninitialized in this function [-Wmaybe-uninitialized]
error, forbidden warning: dynamic_debug.h:75
make[2]: *** [net/netfilter/xt_socket.o] Error 1
make[1]: *** [net/netfilter] Error 2
make: *** [net] Error 2
[email protected]:~/kernel$
Original file here.
I'm using ndk gcc 4.8. This error is present also on gcc 4.7. The only way to get rid of this error is to use toolchain 4.6. I need to compile it using a newer toolchain, that's why I request your help.
Any ideas?
Kernel Compilation
srsdani said:
Code:
CC net/netfilter/xt_qtaguid.o
CC net/netfilter/xt_quota.o
CC net/netfilter/xt_quota2.o
CC net/netfilter/xt_socket.o
In file included from include/linux/kernel.h:23:0,
from include/linux/cache.h:4,
from include/linux/time.h:7,
from include/linux/stat.h:60,
from include/linux/module.h:10,
from net/netfilter/xt_socket.c:13:
net/netfilter/xt_socket.c: In function 'xt_socket_get4_sk':
include/linux/dynamic_debug.h:75:21: warning: 'protocol' may be used uninitialized in this function [-Wmaybe-uninitialized]
error, forbidden warning: dynamic_debug.h:75
make[2]: *** [net/netfilter/xt_socket.o] Error 1
make[1]: *** [net/netfilter] Error 2
make: *** [net] Error 2
[email protected]:~/kernel$
Original file here.
I'm using ndk gcc 4.8. This error is present also on gcc 4.7. The only way to get rid of this error is to use toolchain 4.6. I need to compile it using a newer toolchain, that's why I request your help.
Click to expand...
Click to collapse
I guess you are looking on another file.
Code:
error, forbidden warning: dynamic_debug.h:75
I guess there is a problem in dynamic_debug.h in line 75. I guess you need to initialise something.
Here is dynamic_debug.h but I can't find the problem.
I will appreciate any opinion.
Try this method!
srsdani said:
Here is dynamic_debug.h but I can't find the problem.
Click to expand...
Click to collapse
I am not much used to kernel related development but I analysed your dynamic_debug.h and found the problem (maybe a problem):
Normally, a do-while loop syntax is this:
Code:
do
{
Statement1;
Statement2;
...
}while(condition);
But I don't see a semicolon in the end of those loops in your code. This maybe a syntax error.
Also analyse these codes:
Code:
#if defined(CONFIG_DYNAMIC_DEBUG)
extern int ddebug_remove_module(const char *mod_name);
extern __printf(2, 3)
int __dynamic_pr_debug(struct _ddebug *descriptor, const char *fmt, ...);
struct device;
extern __printf(3, 4)
int __dynamic_dev_dbg(struct _ddebug *descriptor, const struct device *dev,
const char *fmt, ...);
struct net_device;
extern __printf(3, 4)
int __dynamic_netdev_dbg(struct _ddebug *descriptor,
const struct net_device *dev,
const char *fmt, ...);
Every intialisation has 4 protocols but the first initialisation contain only 3.
Try the above things and tell me back if it works or not.
So, the problem was in Makefile and .config. I compared and added some flags from Mako (from franco) and it worked. I still have some warning/s, some breakpoints but at least now it's working. Anyway, this thread may be useful for all of us. Thanks a lot for help.
Now, another problem.
Code:
CC fs/proc/loadavg.o
CC fs/proc/meminfo.o
fs/proc/meminfo.c: In function ‘meminfo_proc_show’:
fs/proc/meminfo.c:175:1: internal compiler error: in optimize_sc, at modulo-sched.c:1069
}
^
Please submit a full bug report,
with preprocessed source if appropriate.
See <https://bugs.launchpad.net/gcc-linaro> for instructions.
make[2]: *** [fs/proc/meminfo.o] Error 1
make[1]: *** [fs/proc] Error 2
make: *** [fs] Error 2
[email protected]:~/kernel2$
With linaro 4.8.3 => OK;
linaro 4.8.3 + A15 optimize=> fail, error above. meminfo.c it's exactly like the one from nexus 4.
FIX: compiler bug;
srsdani said:
Code:
CC net/netfilter/xt_qtaguid.o
CC net/netfilter/xt_quota.o
CC net/netfilter/xt_quota2.o
CC net/netfilter/xt_socket.o
In file included from include/linux/kernel.h:23:0,
from include/linux/cache.h:4,
from include/linux/time.h:7,
from include/linux/stat.h:60,
from include/linux/module.h:10,
from net/netfilter/xt_socket.c:13:
net/netfilter/xt_socket.c: In function 'xt_socket_get4_sk':
include/linux/dynamic_debug.h:75:21: warning: 'protocol' may be used uninitialized in this function [-Wmaybe-uninitialized]
error, forbidden warning: dynamic_debug.h:75
make[2]: *** [net/netfilter/xt_socket.o] Error 1
make[1]: *** [net/netfilter] Error 2
make: *** [net] Error 2
[email protected]:~/kernel$
Original file here.
I'm using ndk gcc 4.8. This error is present also on gcc 4.7. The only way to get rid of this error is to use toolchain 4.6. I need to compile it using a newer toolchain, that's why I request your help.
Click to expand...
Click to collapse
Hey sarsdani Apologies that I'm quoting you 12 month later. I am getting an identical error in the same file/address as you had when you posted this question. Im compiling stock HTC m7 Android L kernel with Google tool chain 4.8 - android 5.1 r1
You mentioned later on that you changed the Makefile and .config to fix the [-Wmaybe-uninitialized] error. Can you please detail what those changes were or point me to an example ( if you can remember ) I could just change Wmaybe in Makefile to Wno but turning the flag off altogether seems dangerous.
@Joeisgood99 there are two ways to solve this.
1. using a "hack". So, you need to add in makefile (in root folder from kernel source) at KBUILD_CFLAGS: -Wno-maybe-uninitialized \
2. Open that file with the error and asign a value to protocol.
The first option is not dangerous. Why? Because only the toolchain say this. In gcc 4.6, the reason why you don't get this error is because it assign a value automatically to that "protocol", usualy random value but not always. In gcc 4.7/4.8/4.9(or 4.10) this is threated as warning/error.
/home/austin/cm11/kernel/lge/w3c/net/ipv4/ipconfig.c:264: undefined reference to `msecs_to_jiffies'
/home/austin/cm11/kernel/lge/w3c/Makefile:877: recipe for target '.tmp_vmlinux1' failed
make[2]: *** [.tmp_vmlinux1] Error 1
Makefile:130: recipe for target 'sub-make' failed
make[1]: *** [sub-make] Error 2
make[1]: Leaving directory '/home/austin/cm11/kernel/lge/w3c'
build/core/tasks/kernel.mk:188: recipe for target 'TARGET_KERNEL_BINARIES' failed
make: *** [TARGET_KERNEL_BINARIES] Error 2
Any ideas what this could be? I'm a bit stumped i've never seen an error like this before
It's okay i retired the project as of now since trying to modify the kernel source code isn't do anything but causing more errors. I do appreciate any help when everyone has a chance

[20141217] Unofficial CM12

Use at your own risk - untested
I am not responsible for bricked devices.
Back up before doing anything
Compiled using the following device tree: https://github.com/CyanogenMod/android_device_sony_z3c​
Working:
Nothing, as far as I am concerned which is why you should use at your own risk.​Not Working:
Everything, as far as I am concerned which is why you should use at your own risk.​
The commands I used to make zip package was:
Code:
. build/envsetup.sh
lunch
make bacon
(Brunch and Breakfast are for official devices only, I believe)
Local Manifest includes:
Code:
<project name="CyanogenMod/android_device_sony_z3c" path="device/sony/z3c" remote="github" />
<project name="CyanogenMod/android_device_sony_shinano-common" path="device/sony/shinano-common" remote="github" />
<project name="CyanogenMod/android_kernel_sony_msm8974" path="kernel/sony/msm8974" remote="github" />
<project name="CyanogenMod/android_device_sony_qcom-common" path="device/sony/qcom-common" remote="github" />
<project name="CyanogenMod/android_device_sony_common" path="device/sony/common" remote="github" />
<project name="CyanogenMod/android_device_sony_msm8974-common" path="device/sony/msm8974-common" remote="github" />
<project name="CyanogenMod/android_device_qcom_common" path="device/qcom/common" remote="github" />
File can be found in Downloads Tab
And,
Mirror: AndroidFileHost
XDA:DevDB Information
ROM compilation, ROM for the Sony Xperia Z3 Compact
Contributors
m0d
Thank you to Kali- and chirayudesai for setting up device tree at: android_device_sony_z3c
And, thank you to the people who spared the time in correcting my silly mistakes, and for responding to my even sillier questions.
ROM OS Version: 5.0.x Lollipop
Version Information
Status: Testing
Created 2014-12-15
Last Updated 2014-12-17
No clue how you managed to build, I've been trying for days!
Anyhow, booted it up and it restarts repeatedly at the 'starting system apps' stage. Here's a logcat:
Code:
https://bpaste.net/show/6aaa730e0985
---------- Post added at 07:42 PM ---------- Previous post was at 07:37 PM ----------
Gairtial said:
No clue how you managed to build, I've been trying for days!
Anyhow, booted it up and it restarts repeatedly at the 'starting system apps' stage. Here's a logcat:
Code:
https://bpaste.net/show/6aaa730e0985
Click to expand...
Click to collapse
I should also mention that the installation initially failed. The device calls itself 'aries', not 'd5803'. Had to modify the script in the ZIP before flashing.
Gairtial said:
No clue how you managed to build, I've been trying for days!
Click to expand...
Click to collapse
Magic!
But, there is no magic.
Send me your build error,
And also the error for show commands and I will try to help.
As that is the secret. In programming I come to find more and more that there is no better place to go than an actual person - not all errors are documented and some strange things happen for no apparent reason, or so one thinks.
Gairtial said:
I should also mention that the installation initially failed. The device calls itself 'aries', not 'd5803'. Had to modify the script in the ZIP before flashing.
Click to expand...
Click to collapse
Yes, someone mentioned that before.
Not sure where to look.
I did find this in BoardConfig.mk,
Code:
# Assert
TARGET_OTA_ASSERT_DEVICE := D5803,D5833,z3c
And, in init.sony.usb.rc - this:
Code:
on boot
write /sys/class/android_usb/android0/iSerial ${ro.serialno}
write /sys/class/android_usb/android0/iManufacturer ${ro.product.manufacturer}
write /sys/class/android_usb/android0/iProduct ${ro.product.model}
The problem is sony calls the device aries, the code for the device is d5803, and the device tree by cm12 is z3c.
I like z3c but if anyone knows where, maybe one of the mk files, to make the changes?
Any help appreciated. I am learning as I go and as I said above:
Thank you to the people who take the time to correct my silly mistakes, and for responding to my even sillier questions
Click to expand...
Click to collapse
I only compiled device tree, no changes as of yet, other than small fixes for compilation.
m0d said:
Magic!
But, there is no magic.
Send me your build error,
And also the error for show commands and I will try to help.
As that is the secret. In programming I come to find more and more that there is no better place to go than an actual person - not all errors are documented and some strange things happen for no apparent reason, or so one thinks.
Click to expand...
Click to collapse
The build errors started with libtime_genoff.so being missing from the output directory so I manually copied it over and it continued on fine.
There were some issues with AAPT but I manually fixed those (and submitted patches for review to Gerrit).
Just waiting for the build to complete now. Are you on IRC/Hangouts or something? Mind if I PM you so we can talk a little more freely?
m0d said:
Magic!
But, there is no magic.
Send me your build error,
And also the error for show commands and I will try to help.
As that is the secret. In programming I come to find more and more that there is no better place to go than an actual person - not all errors are documented and some strange things happen for no apparent reason, or so one thinks.
Yes, someone mentioned that before.
Not sure where to look.
I did find this in BoardConfig.mk,
Code:
# Assert
TARGET_OTA_ASSERT_DEVICE := D5803,D5833,z3c
And, in init.sony.usb.rc - this:
Code:
on boot
write /sys/class/android_usb/android0/iSerial ${ro.serialno}
write /sys/class/android_usb/android0/iManufacturer ${ro.product.manufacturer}
write /sys/class/android_usb/android0/iProduct ${ro.product.model}
The problem is sony calls the device aries, the code for the device is d5803, and the device tree by cm12 is z3c.
I like z3c but if anyone knows where, maybe one of the mk files, to make the changes?
Any help appreciated. I am learning as I go and as I said above:
I only compiled device tree, no changes as of yet, other than small fixes for compilation.
Click to expand...
Click to collapse
Check your build.prop in your compiled build folder (/system) and look for the variables below:
"ro.product.device" "ro.build.product"
If you want to change the values, mount your system.img, make your changes and make new .img
(PM me if you need help with this..)
it's not hard to delect the first sentence in updater-script in zip,then we could get it flashed
Okay, did my own build using similar instructions and came up with the following issues:
libtime_genoff.so wasn't copied during the build process, manually copying required manually copying a second time after the build.
DPI was first set to 480 in build.prop. Reset it to 320 through ADB after booting and it works fine
On the first boot the modem appeared to work (connected to my network), on the second it did not. Nothing I can understand in ADB that's relevant.
Wifi appears to work properly
Camera crashes
GPS error message:
Code:
D/LocSvc_utils_cfg( 5088): D/loc_read_conf: using /etc/gps.conf
D/LocSvc_utils_cfg( 5088): D/loc_set_config_entry: PARAM DEBUG_LEVEL = 2
E/gpsone_dmn( 5088): gpsone_glue_pipeget:46] /data/misc/gsiff_ctrl_q, mode = 2
E/gpsone_dmn( 5088): gpsone_glue_pipeget:50] pipe_name: /data/misc/gsiff_ctrl_q failed: 13
E/gsiff_dmn( 5088): W/gsiff_dmn_init: Unable to initialize Message Queue!
E/gsiff_dmn( 5088): W/gsiff_loc_api_close: Improperly closing Loc API Client!
E/gsiff_dmn( 5088): W/gsiff_dmn_destroy: Could not destroy loc api!
E/gsiff_sp_glue( 5088): W/sp_destroy: Sensor provider not initialized. Sensor provider = 0
E/gsiff_dmn( 5088): W/gsiff_dmn_destroy: Could not destroy sensor provider rc = 0!
E/gpsone_dmn( 5088): gpsone_glue_piperemove:92] fd = -1, /data/misc/gsiff_ctrl_q
E/LocSvc_utils_ll( 5088): W/linked_list_destroy: Invalid list parameter!
F/libc ( 5088): Fatal signal 11 (SIGSEGV), code 1, fault addr 0x8 in tid 5088 (gsiff_daemon)
I/DEBUG ( 263): *** *** *** *** *** *** *** *** *** *** *** *** *** *** *** ***
I/DEBUG ( 263): Build fingerprint: 'Sony/D5803/D5803:4.4.4/23.0.A.2.105/lvf_Rw:user/release-keys'
I/DEBUG ( 263): Revision: '0'
I/DEBUG ( 263): ABI: 'arm'
I/DEBUG ( 263): pid: 5088, tid: 5088, name: gsiff_daemon >>> /system/bin/gsiff_daemon <<<
I/DEBUG ( 263): signal 11 (SIGSEGV), code 1 (SEGV_MAPERR), fault addr 0x8
W/NativeCrashListener( 732): Couldn't find ProcessRecord for pid 5088
I/DEBUG ( 263): r0 00000008 r1 93475f55 r2 93475f55 r3 93475f55
E/DEBUG ( 263): AM write failure (32 / Broken pipe)
I/DEBUG ( 263): r4 00000008 r5 b4f539a4 r6 00000000 r7 b4f539a0
I/DEBUG ( 263): r8 00000003 r9 b6f9a098 sl 00000001 fp b6f94da0
I/DEBUG ( 263): ip b6c5dfbc sp be91f9b8 lr b6f116d1 pc b6f115b6 cpsr 600f0030
I/DEBUG ( 263):
I/DEBUG ( 263): backtrace:
I/DEBUG ( 263): #00 pc 000145b6 /system/lib/libc.so (pthread_mutex_trylock+3)
I/DEBUG ( 263): #01 pc 000146cd /system/lib/libc.so (pthread_mutex_destroy+4)
I/DEBUG ( 263): #02 pc 00002157 /system/lib/libgps.utils.so (msg_q_destroy+66)
I/DEBUG ( 263): #03 pc 00001a85 /system/bin/gsiff_daemon
I/DEBUG ( 263): #04 pc 00002af3 /system/bin/gsiff_daemon
I/DEBUG ( 263): #05 pc 0000fbb9 /system/lib/libc.so (__libc_init+44)
I/DEBUG ( 263): #06 pc 00001968 /system/bin/gsiff_daemon
I/DEBUG ( 263):
I/DEBUG ( 263): Tombstone written to: /data/tombstones/tombstone_
Camera messages:
Code:
E/cacao ( 7380): 1459692 cacao_chokoball.cpp (4914) 9002 E [CKB] ckb preCompile fail. keymgr create credmgr context.
W/cacao ( 7380): 1459761 cacao_chokoball.cpp (4923) 9002 W [CKB] Post process preLoad(0xffffff90)
E/mm-camera-sensor( 297): sony_sensor_power_up:1138,error
E/mm-camera-sensor( 297): sony_sensor_power_up:1145,error
I/Adreno-EGL( 7613): <qeglDrvAPI_eglInitialize:410>: EGL 1.4 QUALCOMM build: RGURRAM_AU_LINUX_ANDROID_LNX.LA.3.5.2.2_RB1.04.04.04.087.030+PATCH[ES]_msm8974_LNX.LA.3.5.2.2_RB1__release_ENGG ()
I/Adreno-EGL( 7613): OpenGL ES Shader Compiler Version: E031.24.00.15
I/Adreno-EGL( 7613): Build Date: 08/12/14 Tue
I/Adreno-EGL( 7613): Local Branch:
I/Adreno-EGL( 7613): Remote Branch: quic/LNX.LA.3.5.2.2_rb1
I/Adreno-EGL( 7613): Local Patches: 8b00bd16f3c1d9d35a2fa902df5e679888d2b2e3 Fixes an llvm crash with mini dEQP apk
I/Adreno-EGL( 7613): 38bad22e162dead4e008444520a0144c78a347bd Fixes a potential dEQP crash.
I/Adreno-EGL( 7613): ce345e1c45c2ae2d1fb2cb125c8d2574f1af5f95 Rev
I/OpenGLRenderer( 7613): Initialized EGL, version 1.4
D/OpenGLRenderer( 7613): Enabling debug mode 0
E/mm-camera( 297): cpp_module_set_clock_freq:1015] clk:133330000, dim:0, stream_type:0,
D/mm-camera( 297): mct_pipeline_process_set:command=800000e
E/cammw ( 7380): E: void* cammw_util_msg_cli_open(const char*, int32_t, cammw_util_msg_callbacks_t, uint32_t, cammw_any_t, uint32_t, cammw_any_t): connect err -1
E/cammw ( 7380): E: cammw_link_isp_cli_open: msg_cli_open fail
E/cammw ( 7380): E: cammw_camera_isp_create: isp create fail
E/cammw ( 7380): E: cammw_camera_open: isp socket create error
E/cammw ( 7380): E: cammw_camera_isp_destroy: invalid arg
E/cammw ( 7380): E: cammw_camera_sensor_destroy: invalid arg
E/cammw ( 7380): E: cammw_link_sensor_cli_close: invalid arg handle = 0x0
E/cammw ( 7380): E: cammw_err_t cammw_util_ashmem_free(cammw_buf_t*): Invalid Arg
E/cammw ( 7380): E: cammw_err_t cammw_util_ashmem_free(cammw_buf_t*): Invalid Arg
D/mm-camera( 297): mct_pipeline_process_set:command=800000b
D/mm-camera-sensor( 297): module_sensor_stop_session:700 session 1
E/cald ( 7380): 1463787 excal_system.cpp (221) 7891 E [EXC] Camera can't open.
E/cald ( 7380): 1463850 excal_system.cpp (447) 7891 E [EXC] system init is failed. General
E/cald ( 7380): 1463930 excal_comp_mgr.cpp (156) 7891 E [EXC] Unregister component: DebugMgr is not found
E/cald ( 7380): 1463984 excal_comp_mgr.cpp (156) 7891 E [EXC] Unregister component: BufMgr is not found
E/cald ( 7380): 1464037 excal_comp_mgr.cpp (156) 7891 E [EXC] Unregister component: DB is not found
I/cacao ( 7380): 1465315 pal.cpp (3383) 7891 I [INF] PAL_Delete
I/cald ( 7380): 1470629 cald_client.c (2051) 7891 I [INF] Cald_Client_Create
E/libcamera( 7380): failed to create Cald instance
E/libcamera( 7380): createInstance: failed startCamera
I/libcamera( 7380): HAL_openCameraHardware: X created hardware=0x0
E/CameraWrapper( 7380): vendor camera open fail
E/CameraClient( 7380): Could not open camera 0: -19
E/CameraClient( 7380): initialize: Camera 0: unable to initialize device: No such device (-19)
I/CameraClient( 7380): Destroying camera 0
D/NuPlayerDriver( 7380): reset(0xb460d040)
D/NuPlayerDriver( 7380): notifyResetComplete(0xb460d040)
D/NuPlayerDriver( 7380): reset(0xb460d1c0)
D/NuPlayerDriver( 7380): notifyResetComplete(0xb460d1c0)
W/CameraManager( 8386): Error querying camera device 0 for listing.
I/CameraService( 7380): getCameraCharacteristics: Switching to HAL1 shim implementation...
I/libcamera( 7380): HAL_getCameraInfo: E cameraId=1
I/libcamera( 7380): HAL_getCameraInfo: X
I/libcamera( 7380): HAL_getCameraInfo: E cameraId=1
I/libcamera( 7380): HAL_getCameraInfo: X
---------- Post added at 06:15 AM ---------- Previous post was at 06:06 AM ----------
Sorry, noticed a few things were out of date. in my tree. Previous post may not be correct. I'll be doing another build tonight after extracting vendor blobs and syncing properly and I'll post with more results.
Uploading new build I made with following:
I added to check for aries as well:
Code:
# Assert
TARGET_OTA_ASSERT_DEVICE := D5803,D5833,z3c,aries
Which updates update script to:
Code:
assert(getprop("ro.product.device") == "D5803" || getprop("ro.build.product") == "D5803" || getprop("ro.product.device") == "D5833" || getprop("ro.build.product") == "D5833" || getprop("ro.product.device") == "z3c" || getprop("ro.build.product") == "z3c" || getprop("ro.product.device") == "aries" || getprop("ro.build.product") == "aries" || abort("This package is for device: D5803,D5833,z3c,aries; this device is " + getprop("ro.product.device") + "."););
nelvinchi said:
Check your build.prop in your compiled build folder (/system) and look for the variables below:
"ro.product.device" "ro.build.product"
If you want to change the values, mount your system.img, make your changes and make new .img
(PM me if you need help with this..)
Click to expand...
Click to collapse
Build prop is not the problem, consistency is.
Code:
ro.product.name=D5803
ro.build.product=z3c
ro.product.device=z3c
CM uses the above. Sony uses aries.
So I guess checking for all three, and the variation(D5833) should work in the end.
jsscx said:
it's not hard to delect the first sentence in updater-script in zip,then we could get it flashed
Click to expand...
Click to collapse
No, but that is not ideal. It should check the for the correct device.
But, the problem is that there is no consistency so I guess checking for all four should do it - D5803,D5833,z3c,aries.
Gairtial said:
Okay, did my own build using similar instructions and came up with the following issues:
libtime_genoff.so wasn't copied during the build process, manually copying required manually copying a second time after the build.
DPI was first set to 480 in build.prop. Reset it to 320 through ADB after booting and it works fine
On the first boot the modem appeared to work (connected to my network), on the second it did not. Nothing I can understand in ADB that's relevant.
Wifi appears to work properly
Camera crashes
---------- Post added at 06:15 AM ---------- Previous post was at 06:06 AM ----------
[/COLOR]Sorry, noticed a few things were out of date. in my tree. Previous post may not be correct. I'll be doing another build tonight after extracting vendor blobs and syncing properly and I'll post with more results.
Click to expand...
Click to collapse
I am not sure if I had any problems with "libtime_genoff.so", I checked the proprietary files list and it is in none of them.
The aapt was fixed in a commit, z3c.mk is now
Code:
PRODUCT_AAPT_CONFIG := normal hdpi xhdpi
PRODUCT_AAPT_PREF_CONFIG := xhdpi
And it compiled fine.
DPI also had a commit to change to 320.
So doing a repo sync should fix those previous issues.
I believe there was an issue with proprietary files as the proprietary-files.txt has been updated.
And now I have no errors when extracting files, so maybe camera files now are getting tranferred.
Review with updated changes just for z3c - http://review.cyanogenmod.org/#/q/status:merged+project:CyanogenMod/android_device_sony_z3c,n,z
Maybe a commit will be needed to fix update script as mentioned above,
Code:
# Assert
TARGET_OTA_ASSERT_DEVICE := D5803,D5833,z3c,aries
I believe there is a pull request with that specific change as well.
Thank you all for responses
Added to OP:
Mirror: AndroidFileHost​
Also,
Please submit feedback.
All issues are welcome
so here's quick test:
boot with animation,when the app starting screen lasts a few seconds,then restart.
https://www.dropbox.com/s/nfexgudos3bni87/1.txt?dl=0
jsscx said:
so here's quick test:
boot with animation,when the app starting screen lasts a few seconds,then restart.
https://www.dropbox.com/s/nfexgudos3bni87/1.txt?dl=0
Click to expand...
Click to collapse
The name of the zip package you used will be helpful, i.e. 20141215 or 20141219
19
jsscx said:
so here's quick test:
boot with animation,when the app starting screen lasts a few seconds,then restart.
https://www.dropbox.com/s/nfexgudos3bni87/1.txt?dl=0
Click to expand...
Click to collapse
Maybe @jerpelea could help as more than one person has been encountering this issue.
jsscx said:
19
Click to expand...
Click to collapse
Uploaded new build only on AFH,
Mirror: AndroidFileHost​The repo sync includes these commits,
http://review.cyanogenmod.org/#/c/82475/
http://review.cyanogenmod.org/#/c/82465/
I can confirm the issue with the start loop.
Get a starting apps message, then reverts back to CM bootanimation, does this in a loop, hard to hard shut down phone (Volume up and Power) to get it to turn off.
EDIT
- Also had to delete first line of updater script to get it to install
- The built in recovery still didn't let me install using 'apply' update, i had to use CWM 6.0.5.0 from AndroPlus Kernel to flash it
- Updater script recognized my device as 'Amami' when i don't have any reference to it in my build.prop, all value were D5803
Regards
spida_singh said:
I can confirm the issue with the start loop.
Get a starting apps message, then reverts back to CM bootanimation, does this in a loop, hard to hard shut down phone (Volume up and Power) to get it to turn off.
Regards
Click to expand...
Click to collapse
I had the same problem and I could not test the rom
spida_singh said:
EDIT
- Also had to delete first line of updater script to get it to install
- The built in recovery still didn't let me install using 'apply' update, i had to use CWM 6.0.5.0 from AndroPlus Kernel to flash it
- Updater script recognized my device as 'Amami' when i don't have any reference to it in my build.prop, all value were D5803
Regards
Click to expand...
Click to collapse
Not sure why it is recognizing your device as Amami.
I added 'aries' to the assert so it should work.
I am creating zip packages using "make bacon", by default CM uses CWM so it is not surprising that you must use CWM to flash it.
As for built-in recovery, I am not sure, as it gets created automatically with the build.
Maybe changes are needed.
spida_singh said:
I can confirm the issue with the start loop.
Get a starting apps message, then reverts back to CM bootanimation, does this in a loop, hard to hard shut down phone (Volume up and Power) to get it to turn off.
Click to expand...
Click to collapse
Burakki said:
I had the same problem and I could not test the rom
Click to expand...
Click to collapse
I may have found why 'com.qualcomm.location.apk' was crashing,
It seems devs are trying to move it from 'app' to 'priv-app':
Code:
-app/com.qualcomm.location.apk:priv-app/com.qualcomm.location/com.qualcomm.location.apk
Changing 'proprietary-files.txt' from the above to
Code:
-app/com.qualcomm.location.apk
might fix that issue.
Compiling with change to see if it works.
m0d said:
It seems devs are trying to move it from 'app' to 'priv-app'
Click to expand...
Click to collapse
Any reason behind why this is trying to be done? Hasn't it always been inside app?
someone755 said:
Any reason behind why this is trying to be done? Hasn't it always been inside app?
Click to expand...
Click to collapse
No, not sure.
On Stock ROM it is located in app; however, somewhere along the line CM devs are moving it to priv-app.
The change I made does not seem to make a difference and I can not seem to find where are they moving it to priv-app
Android.mk in 'vendor/sony/shinano-common'
Code:
include $(CLEAR_VARS)
LOCAL_MODULE := com.qualcomm.location
LOCAL_MODULE_OWNER := sony
[COLOR="SeaGreen"]LOCAL_SRC_FILES := proprietary/priv-app/com.qualcomm.location/com.qualcomm.location.apk[/COLOR]
LOCAL_MODULE_TAGS := optional
LOCAL_MODULE_SUFFIX := $(COMMON_ANDROID_PACKAGE_SUFFIX)
LOCAL_MODULE_CLASS := APPS
LOCAL_CERTIFICATE := platform
LOCAL_PRIVILEGED_MODULE := true
include $(BUILD_PREBUILT)
And, it is being copied to: out/target/product/z3c/system/priv-app/com.qualcomm.location
I am not sure if this is of any concern, but from the logs com.qualcomm.location.apk looks like the root of the issue as everything dies after it encounters:
Code:
E/installd( 269): DexInv: --- END '/system/priv-app/com.qualcomm.location/com.qualcomm.location.apk' --- status=0x0100, process failed
D/AndroidRuntime( 4634): Shutting down VM
E/AndroidRuntime( 4634): *** FATAL EXCEPTION IN SYSTEM PROCESS: main
I tested it's working great..
Just a heads up! Cyanogenmod now has an official nightly for the Sony Xperia Z3 compact.
You can download their nighty builds from here:
http://download.cyanogenmod.org/?device=z3c

[Q] Must specify VARIANT_DEFCONFIG

I am trying to build my own kernel using the Android NDK r5, but whenever I try to
Code:
make msm8974_defconfig
(I have Galaxy s5)
It just gives me this:
Code:
[email protected]:~/linux$ make msm8974_defconfig
sound/soc/codecs/audience/Kconfig:40:warning: type of 'SND_SOC_ES_SLIM' redefined from 'boolean' to 'tristate'
sound/soc/codecs/audience/Kconfig:43:warning: type of 'SND_SOC_ES_I2C' redefined from 'boolean' to 'tristate'
sound/soc/codecs/audience/Kconfig:44:warning: choice value used outside its choice group
sound/soc/codecs/audience/Kconfig:41:warning: choice value used outside its choice group
KCONFIG_SELINUX((null))
KCONFIG_LOG_SELINUX((null))
KCONFIG_TIMA((null))
KCONFIG_VARIANT((null))
***
*** You must specify VARIANT_DEFCONFIG !
***
make[1]: *** [msm8974_defconfig] Error 1
make: *** [msm8974_defconfig] Error 2
Google didn't help me out this time
Does someone have the answer?
Having this problem too.. any ideas?

Categories

Resources