android on omap

android on omap

[twebb]

I'm using linux-2.6.26 from the l-o git on an omap3 based target, and
have patched it with android-0.9 patch derived as described in various
forums about porting android to real hardware.  I'm also using the android
filesystem and binaries as extracted from 0.9 release.  I can start target and
all seems OK - to a point.  Kernel starts OK, android init starts OK,
all processes appear to start OK.  However, when process
'system_server' starts, screen changes from "ANDROID_" to blank, and
about 45-60 seconds later 'system_server' goes away and process
'zygote' restarts.  When started with strace, I see the error output
below.

Can anyone make any sense of this error/exception?  By the way, the
target does not have network connectivity via USB, but does have it
via ethernet.  Might this be an issue?

Thanks,
twebb

java.lang.RuntimeException: ANDROID_SOCKET_zygote unset or invalid
       at com.android.internal.os.ZygoteInit.registerZygoteSocket(ZygoteInit.java:156)
       at com.android.internal.os.ZygoteInit.main(ZygoteInit.java:461)
       at dalvik.system.NativeStart.main(Native Method)
Caused by: java.lang.NumberFormatException: unable to parse 'null' as integer
       at java.lang.Integer.parseInt(Integer.java:362)
       at java.lang.Integer.parseInt(Integer.java:341)
       at com.android.internal.os.ZygoteInit.registerZygoteSocket(ZygoteInit.java:154)

RE: android on omap

[Deacon, Keith]

There is too little information to say much.  Try running logcat (or just cat /dev/log/main) to get more info.  Since dalvik started, it probably has something to do with the input config files or runtime database.  File permissions can also be and issue.  Cleaning out the /data and /system/data directories can also help.

...Kd


> -----Original Message-----
> From: linux-omap-owner@vger.kernel.org [mailto:linux-omap-
> owner@vger.kernel.org] On Behalf Of twebb
> Sent: Friday, October 03, 2008 9:51 AM
> To: linux-omap@vger.kernel.org Mailing List
> Subject: android on omap
>
> I'm using linux-2.6.26 from the l-o git on an omap3 based target, and
> have patched it with android-0.9 patch derived as described in various
> forums about porting android to real hardware.  I'm also using the android
> filesystem and binaries as extracted from 0.9 release.  I can start target
> and
> all seems OK - to a point.  Kernel starts OK, android init starts OK,
> all processes appear to start OK.  However, when process
> 'system_server' starts, screen changes from "ANDROID_" to blank, and
> about 45-60 seconds later 'system_server' goes away and process
> 'zygote' restarts.  When started with strace, I see the error output
> below.
>
> Can anyone make any sense of this error/exception?  By the way, the
> target does not have network connectivity via USB, but does have it
> via ethernet.  Might this be an issue?
>
> Thanks,
> twebb
>
> java.lang.RuntimeException: ANDROID_SOCKET_zygote unset or invalid
>        at
> com.android.internal.os.ZygoteInit.registerZygoteSocket(ZygoteInit.java:15
> 6)
>        at com.android.internal.os.ZygoteInit.main(ZygoteInit.java:461)
>        at dalvik.system.NativeStart.main(Native Method)
> Caused by: java.lang.NumberFormatException: unable to parse 'null' as
> integer
>        at java.lang.Integer.parseInt(Integer.java:362)
>        at java.lang.Integer.parseInt(Integer.java:341)
>        at
> com.android.internal.os.ZygoteInit.registerZygoteSocket(ZygoteInit.java:15
> 4)
> --
> To unsubscribe from this list: send the line "unsubscribe linux-omap" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: android on omap

[twebb]

>
> Had problems with using NFS because Android tries to set specific permission and ownership of files - and has a nasty habit of testing those permissions frequently.  If you have any local storage options - that would help.  Also, check your init.rc file for things like mount trying to remount the filesystem readonly - Android assumes a very specific partition layout - which can be consolidated into one partition as long as init.rc is adjusted to match.
>

> Does the root file system reside on NFS? Have you enabled "Networking" while
> configuring the kernel?
>

I've now moved filesystem to target (jffs2).  Things are a little
better, and different.  Using logcat, I'm now seeing dalvikvm errors
and warnings as shown below.  Particularly odd since I wasn't seeing
this error when rootfs was mounted via NFS.  Any ideas?

I really appreciate your input.

Thanks,
twebb

D/dalvikvm(  483): DexOpt: --- BEGIN 'core.jar' (bootstrap=1) ---
E/dalvikvm(  501): unable to mmap DEX cache: Invalid argument
E/dalvikvm(  501): Optimization failed
D/dalvikvm(  483): DexOpt: waiting for verify+opt, pid=501
W/dalvikvm(  483): DexOpt: --- END 'core.jar' --- status=0xff00, process failed
E/dalvikvm(  483): Unable to extract+optimize DEX from
'/system/framework/core.jar'
D/dalvikvm(  483): Failed on '/system/framework/core.jar' (boot=1)
D/dalvikvm(  483): VM cleaning up
D/dalvikvm(  483): LinearAlloc 0x0 used 4100 of 4194304 (0%)
W/dalvikvm(  483): JNI_CreateJavaVM failed
E/AndroidRuntime(  483): JNI_CreateJavaVM failed

RE: android on omap

[Deacon, Keith]

Haven't seen that error - sorry.

Unable to mmap stands out.  I know there was a lot of updates to the shared memory subsystem and there was some Android share memory driver introduced with the Android patch set.  Seems like it's trying to mmap some shared memory and failing.  The rest just seems to be fallout because it can't do what it wants to now.

...Kd


> -----Original Message-----
> From: twebb [mailto:taliaferro62@gmail.com]
> Sent: Monday, October 06, 2008 4:46 PM
> To: Deacon, Keith; linux-omap@vger.kernel.org Mailing List
> Subject: Re: android on omap
>
> >
> > Had problems with using NFS because Android tries to set specific
> permission and ownership of files - and has a nasty habit of testing those
> permissions frequently.  If you have any local storage options - that
> would help.  Also, check your init.rc file for things like mount trying to
> remount the filesystem readonly - Android assumes a very specific
> partition layout - which can be consolidated into one partition as long as
> init.rc is adjusted to match.
> >
>
> > Does the root file system reside on NFS? Have you enabled "Networking"
> while
> > configuring the kernel?
> >
>
> I've now moved filesystem to target (jffs2).  Things are a little
> better, and different.  Using logcat, I'm now seeing dalvikvm errors
> and warnings as shown below.  Particularly odd since I wasn't seeing
> this error when rootfs was mounted via NFS.  Any ideas?
>
> I really appreciate your input.
>
> Thanks,
> twebb
>
> D/dalvikvm(  483): DexOpt: --- BEGIN 'core.jar' (bootstrap=1) ---
> E/dalvikvm(  501): unable to mmap DEX cache: Invalid argument
> E/dalvikvm(  501): Optimization failed
> D/dalvikvm(  483): DexOpt: waiting for verify+opt, pid=501
> W/dalvikvm(  483): DexOpt: --- END 'core.jar' --- status=0xff00, process
> failed
> E/dalvikvm(  483): Unable to extract+optimize DEX from
> '/system/framework/core.jar'
> D/dalvikvm(  483): Failed on '/system/framework/core.jar' (boot=1)
> D/dalvikvm(  483): VM cleaning up
> D/dalvikvm(  483): LinearAlloc 0x0 used 4100 of 4194304 (0%)
> W/dalvikvm(  483): JNI_CreateJavaVM failed
> E/AndroidRuntime(  483): JNI_CreateJavaVM failed

Re: android on omap

[Adrian Hunter]

Deacon, Keith wrote:
> Haven't seen that error - sorry.
> 
> Unable to mmap stands out.  I know there was a lot of updates to the shared memory subsystem and there was some Android share memory driver introduced with the Android patch set.  Seems like it's trying to mmap some shared memory and failing.  The rest just seems to be fallout because it can't do what it wants to now.

JFFS2 supports only read-only mmap.  If that is a problem, you could try UBIFS.

> 
> ...Kd
> 
> 
>> -----Original Message-----
>> From: twebb [mailto:taliaferro62@gmail.com]
>> Sent: Monday, October 06, 2008 4:46 PM
>> To: Deacon, Keith; linux-omap@vger.kernel.org Mailing List
>> Subject: Re: android on omap
>>
>>> Had problems with using NFS because Android tries to set specific
>> permission and ownership of files - and has a nasty habit of testing those
>> permissions frequently.  If you have any local storage options - that
>> would help.  Also, check your init.rc file for things like mount trying to
>> remount the filesystem readonly - Android assumes a very specific
>> partition layout - which can be consolidated into one partition as long as
>> init.rc is adjusted to match.
>>> Does the root file system reside on NFS? Have you enabled "Networking"
>> while
>>> configuring the kernel?
>>>
>> I've now moved filesystem to target (jffs2).  Things are a little
>> better, and different.  Using logcat, I'm now seeing dalvikvm errors
>> and warnings as shown below.  Particularly odd since I wasn't seeing
>> this error when rootfs was mounted via NFS.  Any ideas?
>>
>> I really appreciate your input.
>>
>> Thanks,
>> twebb
>>
>> D/dalvikvm(  483): DexOpt: --- BEGIN 'core.jar' (bootstrap=1) ---
>> E/dalvikvm(  501): unable to mmap DEX cache: Invalid argument
>> E/dalvikvm(  501): Optimization failed
>> D/dalvikvm(  483): DexOpt: waiting for verify+opt, pid=501
>> W/dalvikvm(  483): DexOpt: --- END 'core.jar' --- status=0xff00, process
>> failed
>> E/dalvikvm(  483): Unable to extract+optimize DEX from
>> '/system/framework/core.jar'
>> D/dalvikvm(  483): Failed on '/system/framework/core.jar' (boot=1)
>> D/dalvikvm(  483): VM cleaning up
>> D/dalvikvm(  483): LinearAlloc 0x0 used 4100 of 4194304 (0%)
>> W/dalvikvm(  483): JNI_CreateJavaVM failed
>> E/AndroidRuntime(  483): JNI_CreateJavaVM failed
> 
> --
> To unsubscribe from this list: send the line "unsubscribe linux-omap" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
> 

Re: android on omap

[twebb]

>
> JFFS2 supports only read-only mmap.  If that is a problem, you could try
> UBIFS.
>

Now when mounting rootfs as yaffs2, the servicemanager dies/exits
almost immediately on android startup, and logcat output is shown
below.  However, when mounting same rootfs via NFS, I don't see this
problem (though I do have other issues).  Does the error

E/dalvikvm-gc(  483): Could not create 176128-byte ashmem mark stack

indicate a specific problem?  Might there be a way to dump more
specific/verbose output at startup to get a better understanding of
what's wrong?

Thanks,
twebb


I/DEBUG   (  481): debuggerd: Aug 14 2008 18:58:32
E/mountd  (  480): could not read initial mass storage state
D/AndroidRuntime(  483):
D/AndroidRuntime(  483): >>>>>>>>>>>>>> AndroidRuntime START <<<<<<<<<<<<<<
D/AndroidRuntime(  483): CheckJNI is ON
D/AndroidRuntime(  483): --- registering native functions ---
I/Zygote  (  483): Preloading classes...
E/dalvikvm-gc(  483): Could not create 176128-byte ashmem mark stack

RE: android on omap

[Deacon, Keith]

Ashmem is the Android shared memory subsystem.  You can find it in the kernel under mm/ashmem.c.  There is a config option to turn it on and off.  The error indicates that it's having trouble with the driver - although it's not specific as to exactly what.  Make sure it's enabled, the device is present (/dev/ashmem), etc..

...Kd


> -----Original Message-----
> From: twebb [mailto:taliaferro62@gmail.com]
> Sent: Thursday, October 09, 2008 10:09 AM
> To: Adrian Hunter
> Cc: Deacon, Keith; linux-omap@vger.kernel.org Mailing List;
> anilvpr@gmail.com; Mat Johnson
> Subject: Re: android on omap
>
> >
> > JFFS2 supports only read-only mmap.  If that is a problem, you could try
> > UBIFS.
> >
>
> Now when mounting rootfs as yaffs2, the servicemanager dies/exits
> almost immediately on android startup, and logcat output is shown
> below.  However, when mounting same rootfs via NFS, I don't see this
> problem (though I do have other issues).  Does the error
>
> E/dalvikvm-gc(  483): Could not create 176128-byte ashmem mark stack
>
> indicate a specific problem?  Might there be a way to dump more
> specific/verbose output at startup to get a better understanding of
> what's wrong?
>
> Thanks,
> twebb
>
>
> I/DEBUG   (  481): debuggerd: Aug 14 2008 18:58:32
> E/mountd  (  480): could not read initial mass storage state
> D/AndroidRuntime(  483):
> D/AndroidRuntime(  483): >>>>>>>>>>>>>> AndroidRuntime START
> <<<<<<<<<<<<<<
> D/AndroidRuntime(  483): CheckJNI is ON
> D/AndroidRuntime(  483): --- registering native functions ---
> I/Zygote  (  483): Preloading classes...
> E/dalvikvm-gc(  483): Could not create 176128-byte ashmem mark stack

Re: android on omap

[twebb]

> Ashmem is the Android shared memory subsystem.  You can find it in the kernel under mm/ashmem.c.  There is a config option to turn it on and off.  The error indicates that it's having trouble with the driver - although it's not specific as to exactly what.  Make sure it's enabled, the device is present (/dev/ashmem), etc..
>
> ...Kd
>

I have ashmem properly configured and it works when the FS is mounted
via NFS.  However as pointed out previously that seemed to be causing
other issues so I moved the FS into JFFS2.  That too had it's issues
(where mmap was failing under certain conditions) that Adrian and Anil
pointed out were due to JFFS2.  So I moved it to YAFFS2.  Now I'm
seeing this "ashmem" issue.  Up to this point, ashmem has worked fine
so I don't believe it's really the culprit.  I feel like it has to do
with permissions or the filesystem.  One difference is that the
emulator uses three different yaffs2 partitions (userdata, system, and
cache) while I currently have all of these on a single partition.

Any other ideas as to what might cause the ...

E/dalvikvm-gc(  483): Could not create 176128-byte ashmem mark stack

message?

Thanks,
twebb

RE: android on omap

[Deacon, Keith]

Running out of ideas here.  It just opens the ashmem device and makes two ioctl calls - one to set a label and the other allocate the size.  If either fail, you get the error.  Device permissions could be an issue or there isn't enough memory to successfully allocate.  Since it worked before, I would guess device permissions.  The init.rc script does remount / as read only.  I would test /dev/ out by getting into the console and seeing if I can read, write, create a device - preferably as a non-root user - since dalik will most likely be running as a non-prived user.

...Kd


> -----Original Message-----
> From: twebb [mailto:taliaferro62@gmail.com]
> Sent: Thursday, October 09, 2008 5:00 PM
> To: Deacon, Keith
> Cc: Adrian Hunter; linux-omap@vger.kernel.org Mailing List;
> anilvpr@gmail.com; Mat Johnson
> Subject: Re: android on omap
>
> > Ashmem is the Android shared memory subsystem.  You can find it in the
> kernel under mm/ashmem.c.  There is a config option to turn it on and off.
> The error indicates that it's having trouble with the driver - although
> it's not specific as to exactly what.  Make sure it's enabled, the device
> is present (/dev/ashmem), etc..
> >
> > ...Kd
> >
>
> I have ashmem properly configured and it works when the FS is mounted
> via NFS.  However as pointed out previously that seemed to be causing
> other issues so I moved the FS into JFFS2.  That too had it's issues
> (where mmap was failing under certain conditions) that Adrian and Anil
> pointed out were due to JFFS2.  So I moved it to YAFFS2.  Now I'm
> seeing this "ashmem" issue.  Up to this point, ashmem has worked fine
> so I don't believe it's really the culprit.  I feel like it has to do
> with permissions or the filesystem.  One difference is that the
> emulator uses three different yaffs2 partitions (userdata, system, and
> cache) while I currently have all of these on a single partition.
>
> Any other ideas as to what might cause the ...
>
> E/dalvikvm-gc(  483): Could not create 176128-byte ashmem mark stack
>
> message?
>
> Thanks,
> twebb

Re: android on omap

[Anil Sasidharan]

twebb,

           As far as I know Android root file system cannot reside on
jffs2 partition as jffs2 does not have support for memory mapped IO.
One workaround may be to create an EXT2 image for the Android root FS
and place it in the jffs2 partition and try mounting it as loop
device. A brief procedure can be as shown here:

1. create EXT2/EXT3 image for android root FS
2. store the image under the jffs2 partition in NAND (booting out of
NFS, mount MTD partition and place the image file)
3. Boot your device using NFS or ramdisk
4. After boot completes, mount the EXT2/EXT3 image as loop device
5. chroot into that mounted directory and run android.

NB: i've tried the above steps for an SD card formatted as FAT but
never tried this for NAND/jffs2 combination.

Warm Regards,
Anil

On Tue, Oct 7, 2008 at 3:11 AM, twebb <taliaferro62@gmail.com> wrote:
>
> > Does the root file system reside on NFS? Have you enabled "Networking" while
> > configuring the kernel?
> >
>
> Anil,
> I've now moved filesystem to target (jffs2).  Things are a little
> better, and different.  Using logcat, I'm now seeing dalvikvm errors
> and warnings as shown below.  It's particularly odd since I wasn't
> seeing this error when rootfs mounted via NFS.  Any ideas?
>
> I really appreciate your input.
>
> Thanks,
> twebb
>
> D/dalvikvm(  483): DexOpt: --- BEGIN 'core.jar' (bootstrap=1)
> ---
> E/dalvikvm(  501): unable to mmap DEX cache: Invalid
> argument
> E/dalvikvm(  501): Optimization
> failed
> D/dalvikvm(  483): DexOpt: waiting for verify+opt,
> pid=501
> W/dalvikvm(  483): DexOpt: --- END 'core.jar' --- status=0xff00,
> process failed
> E/dalvikvm(  483): Unable to extract+optimize DEX from '/system/
> framework/
> core.jar'
> D/dalvikvm(  483): Failed on '/system/framework/
> core.jar' (boot=1)
> D/dalvikvm(  483): VM cleaning
> up
> D/dalvikvm(  483): LinearAlloc 0x0 used 4100 of 4194304
> (0%)
> W/dalvikvm(  483): JNI_CreateJavaVM
> failed
> E/AndroidRuntime(  483): JNI_CreateJavaVM
> failed
>

Re: Android on OMAP

[barban]

Hi Dirk & all you OMAP experts,
I have reached this mailing list by investigating the very interesting
results you achieved concerning Android on OMAP.

>> The whole Filesystem is about 28MBytes in bz2 format.
>> I don't know where can I upload it. I can not find some place where  
>>   can hold such big file for a long time.
>> I am glad to upload it , if someone has the network storage resources.

I don't know if you have already found a good place for collecting the
stuff related to your work: if not, I would be glad to offer you
network storage as well as other facilities you can need at
www.OpenHandsetDevelopers.com

At the moment, you can find in my site a directory which collects
applications for Android to be used on the emulator and discussion
boards for a community of developers interested in the Android
platform and Android devices.

Do not hesitate to contact me, just in case it can be useful to create  
a specific category/section for your work.

Best regards and enjoy with Android!

Andrea, owner of OpenHandsetDevelopers.com




> skyflame wrote:
>>>> Hi
>>>>
>>>> I move to m5-r14 with both kernel driver and filesystem grabed    
>>>> from the sdk.
>>>> Now, I can enter the homescreen , and the keyboard is also working 8 )
>>>>
>>>> I don't do anything extra than what I am doing on the m3-r14 release.
>>>
>>>
>>> Great! Congratulations! :)
>>>
>>> Do you think you can make the kernel patch (against 2.6.23-omap1?)  
>>>   and the filesystem available somewhere? Then we can link wiki   
>>> page  to it.
>>>
>>> Dirk
>>>
>>>
>>
>> Hi Dirk
>>
>> I was working on a 2.6.21 kernel. For I don't got a working 2.6.23   
>>  kernel on hand.
>> The git tree kernel don't work fine with my omap2430SDP V5.0
>> The PMIC seems not working in the right way, thus the LCD don't run up.
>> So I can not make the diff over 2.6.23 omap1  : (
>
> Okay. I sent my 2.6.23 omap1 Android patch yesterday (seems that it
> didn't reach the list cause of > 100k). If you didn't get it I can
> send it again in private mail. Would be quite nice if you could check
> if there are the same files in it you used.
>
> Anything like
>
> cat my_patch | grep +++ > file1
> cat your_patch | grep +++ > file2
> kdiff3 file1 file2
>
>> The whole Filesystem is about 28MBytes in bz2 format.
>> I don't know where can I upload it. I can not find some place where  
>>   can hold such big file for a long time.
>> I am glad to upload it , if someone has the network storage resources.
>
> As already mentioned yesterday: Same to me ;)
>
> Anybody with ~30-40 MB permanent public storage?
>
> Many thanks
>
> Dirk