Android running on BBB with Linux 3.8

[Andrew Henderson]

Hello all.  I posted this information over at the Google+ group for BeagleBoard.org, but I wanted to post it here as well.  I have Android running on the BBB using the Jellybean (4.2.2) source from arowboat.org and the Linux 3.8.13 kernel from my BeagleSNES project (www.beaglesnes.org).  I just added the Android drivers (Drivers -> Staging Drivers -> Android) into the same kernel configuration that I use for BeagleSNES.  This kernel is really just a stripped down configuration (a 1.9 meg zImage) using Robert Nelson's 3.8.x kernel source tree.  Instead of using the GCC cross-compiler that comes in Rowboat's current prebuilt tools, I use the 3.7 GCC that comes with Robert's kernel tree to build the entire system.  I also updated the u-boot to v2013.04.

Here is a video of it running: https://www.youtube.com/watch?v=yeSt1Y7FhEs
Here is the download of a working Android image that can be written onto a 4GB microSD card using "dd": http://icculus.org/~hendersa/BBB_JB_Android_3_8_13.img.bz2

There is no hardware accelerated video in this version, since SGX support isn't ready for the 3.8 kernel.  To shut off hardware accelerated video, I pass "qemu=1" as a kernel command line argument in uEnv.txt and I add "debug.egl.hw=0" to the /system/build.prop file in the root file system.  To get rid of that blinking cursor on the framebuffer (which I accidentally left on in the image that I made), add this to the uEnv.txt kernel command line arguments: "vt.global_cursor_default=0"

This build will be of interest to Android experimenters who want to use HDMI for audio and video output (rather than using an LCD cape) and who want to experiment with using the device tree.  My kernel is pretty stripped down, so feel free to fetch Robert's latest kernel source for the 3.8.x branch and rebuild your own kernel to drop into the image that I have supplied.  That way, you can include whatever drivers you would like in your kernel.  Use a USB mouse (and a keyboard, too, if you want) to interact with the system.

Thanks, and enjoy!

Andrew

[r2d...@gmail.com]

This is awesome, how would I go about adding custom apps or ones from app stores? 

[Andrew Henderson]

On Tuesday, June 18, 2013 11:57:24 AM UTC-4, r2d...@gmail.com wrote:

This is awesome, how would I go about adding custom apps or ones from app stores?

I forget the exact file path off the top of my head, but you can copy your .apk file to somewhere convenient on the /data partition and then browse to it via Android and "launch" the .apk file by double-clicking it.  As long as you can browse to the .apk file, you're good.  Under Settings -> Applications, make sure that the "Unknown sources" box is checked.  The .apk file will be treated as an installer of sorts by the OS when launched.  To get the .apk onto the microSD card in the first place, you can use the wired ethernet connection and the web browser to download the .apk.  You can also just mount the /data partition under Linux and copy it over that way.

[Venkat Bommakanti]

Hi Andrew,

Thanks for the info.
When possible, could you pl. write up a "step by step" howto for newbie's like me. Don't understand all the steps listed above.
Or, could you point to links that basically lists the steps for generic android builds that reflect your steps above.

thanks,
/venkat

[Ryan Ramchandar]

Thanks for sharing Andrew, This looks like a very interesting project.

Can you elaborate further on how you were able to boot the Android 4.2.2 image on the BBB. From what I've been reading the arowboat guys have not yet got it working on the BBB.

Thanks, 

Ryan R.

[Robert Nelson]

On Tue, Jun 18, 2013 at 12:42 PM, Ryan Ramchandar <ryan...@gmail.com> wrote:
> Thanks for sharing Andrew, This looks like a very interesting project.
>
> Can you elaborate further on how you were able to boot the Android 4.2.2
> image on the BBB. From what I've been reading the arowboat guys have not yet
> got it working on the BBB.

It's actually pretty cool what Andrew just did.. So in the last few
kernel releases there has been renewed interest in merging the android
kernel stuff in mainline.. So Andrew took the kernel source we are
using for Angstrom/Ubuntu and rebuilt it with more of the Android
config options enabled.. Essentially ignoring the 'arowboat' kernel..
Tweaked a few things, and bam, Android works. ;)

Regards,

--
Robert Nelson
http://www.rcn-ee.com/

[Andrew Henderson]

On Tuesday, June 18, 2013 1:42:30 PM UTC-4, Ryan Ramchandar wrote:

Thanks for sharing Andrew, This looks like a very interesting project.

Can you elaborate further on how you were able to boot the Android 4.2.2 image on the BBB. From what I've been reading the arowboat guys have not yet got it working on the BBB.

I am still working on writing up some step-by-step instructions for building the complete Android image from scratch.  My efforts here are due to my needing a working Android image for the BBB for a class of graduate students in a "Smartphone Technology" at Syracuse University this summer.

An Android kernel is a Linux kernel with about 25k of extra Android-specific code in it.  As of Linux 3.3, Android support is part of the mainline Linux kernel, so if you turn on the extra Android features in the kernel configuration you have an Android kernel.  You can actually run an Android-enabled Linux kernel on your Linux system, so there isn't any harm in turning on the additional features.  So really, if you have a Linux kernel and system that runs well and stable on the BBB with the drivers that you need, that kernel is ready for an Android system.

Rowboat does have a BBB build, but it is quite new.  Last I checked, the wiki documentation had not yet caught up to it.  The first I saw of it was when Manish Shakya posted a picture of a 4.1.2 Android build with a 3.2 kernel running on BBB on 6 June to the Google+ BeagleBoard.org discussion area.  Once Manish had done the hard work of getting a BBB build into the Rowboat build scripts, swapping out the kernel for 3.8 and upgrading the bootloader were fairly straight-forward.  I had to make sure that the kernel data structures that were exposed to the Android system apps during building were correct, so the 3.8 kernel source headers had to be placed into the Rowboat source tree prior to building Rowboat.  I actually built the kernel and the bootloader without Rowboat.  Only building the root Android file system was done using the Rowboat makefile.  I also modified the top-level Rowboat makefile to use the same compiler that I used to build the kernel (to keep everything "simple").

All of the pieces were there: BBB Rowboat support, Robert's kernel tree, and a new u-boot that supports device tree kernels.  It just took some (very time-consuming) effort to get them all built, in-sync with each other, and configured properly.  Luckily, my experience with my own BBB projects has given me the know-how to troubleshoot all of these little system hiccups and cobble together a working Android image.  It is my hope that having a "working" reference platform for the 3.8 kernel to start from will assist people that want to try their own Android-enabled kernels, play with the uEnv.txt settings, use a "real" Android system for their application development, etc.

Andrew

[Ryan Ramchandar]

Thanks again Andrew for your work. I really liked the last sentence of your post!

I am new to building Android for the BBB and your step-by-step instructions will be a great resource. I am looking forward to it.

Ryan R.

--
For more options, visit http://beagleboard.org/discuss
---
You received this message because you are subscribed to a topic in the Google Groups "BeagleBoard" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/beagleboard/Yc8CdtU5MEs/unsubscribe.
To unsubscribe from this group and all its topics, send an email to beagleboard...@googlegroups.com.
For more options, visit https://groups.google.com/groups/opt_out.
 
 

[Aaron Buckner]

Awesome! I can't wait to try this out! I am getting my BBB today finally after fighting between it being in stock and me having the $ 

[a.jain]

Will you let us know when the instructions will be out? Was it hard to create? Overall it runs well, glad that Android can be supported on the black! Thanks

[Vishveshwar]

Great work Andrew!

Thanks for taking this effort and sharing your results.

-Vishveshwar

[Andrew Henderson]

I have put together some directions for building Android for the BBB with the 3.8 kernel.  Take a look: http://icculus.org/~hendersa/android

Andrew

[a.jain]

Thanks for the tutorial! I had one problem:

When I started up the Bone, it went to a screen that said 'Automatically Logging In' with the beagleboard logo in the right handside. Android didn't boot up, and I don't know why. Did I do something wrong, or should I just wait?

[Andrew Henderson]

On Wednesday, June 19, 2013 10:29:50 PM UTC-4, a.jain wrote:

Thanks for the tutorial! I had one problem:

When I started up the Bone, it went to a screen that said 'Automatically Logging In' with the beagleboard logo in the right handside. Android didn't boot up, and I don't know why. Did I do something wrong, or should I just wait?

What does the debug output over your FTDI cable show?  Was the microSD card detected?  If so, is there an error message from the bootloader saying that the kernel wasn't found?  Is there an error message about the .dtb file missing? Do you have four partitions on your microSD card (boot, rootfs, data, usrdata)? Offhand, it sounds like you're booting off of the eMMC, not the microSD card.  Is this a fairly new BBB? Did you follow step #8, here: http://beagleboard.org/Getting%20Started ?  We'll need more information to be able to help you.  That's definitely not Android that you are seeing.

Andrew

[pe...@icecavern.com]

Andrew,

Great work! This sounds like a big step.

One question, does the 3.8 kernel under Android bring with it Cape support for Capes supported currently? Most interested in the LCD4 which I have, will Android be displayed on that like X does on the standard kernel and allow us to use touch screen? :)

[Andrew Henderson]

The short answer is "if it is supported under 3.8 Linux, it is supported under Android because it is the same kernel".  Now, that being said, I have no cape hardware at all, so I can't tell anyone if it definitely works because I have not seen it with my own eyes.  As a PhD student, I just can't afford to buy all of the various capes for testing.  Looking at the am335x-bone-common.dtsi device tree file in the 3.8 kernel, I see support for Adafruit's 1.8" TFT display (BB-BONE-TFT-01) and the LCD3 cape (BB-BONE-LCD3-01).  Looking in my patches, I see the 0167-capes-add-LCD7-A3.patch and 0168-capes-add-basic-support-for-LCD4-capes.patch.  So, offhand, I would say support for it is in there.  I'm not sure about how loading cape firmware under Android will work, as I have not tried it, but I suspect that it would be built into the kernel as a firmware blob at compile time (Device drivers -> Generic Driver Options -> Include in-kernel firmware blobs in kernel binary).  In the stripped-down kernel that I provided for the sample Android image, capes are enabled (since HDMI is considered a built-in cape), but I don't believe I included the firmware in the boot partition.  I didn't not turn on support for the Geiger or Nixie capes in the kernel configuration, though, so I hope no one was hoping to use my test Android image to make a Geiger counter.

Android is running on the Linux framebuffer, so I suspect that the cape will show an image similar to how it displays X under Linux: the top-left portion of the screen will be mapped to the physical resolution of the LCD.  After all... at its lowest level, X is drawn onto the framebuffer for display.  From the cape standpoint, the cape hardware is just pulling its graphics from the framebuffer.  Since the DVI cape replicates the display that is shown on the HDMI video output, I would expect LCD capes to act in a similar fashion.

I'm sure that with the hardware and a little time to hack on it, I could cobble something together and provide you with some steps to get everything working.  But, unfortunately, I don't have that knowledge for you right now.  Sorry.

Andrew

[pe...@icecavern.com]

OK I'll see if I can find the time this evening to run up your image on my board with the LCD4 installed and see what happens :)

Will report back once I've given it a go.

Pete

[Andrew Henderson]

On Thursday, June 20, 2013 9:15:38 AM UTC-4, pe...@icecavern.com wrote:

OK I'll see if I can find the time this evening to run up your image on my board with the LCD4 installed and see what happens :)

Will report back once I've given it a go.

If it doesn't work with the kernel in my image, try rebuilding the kernel with all of the features in it.  That instructional link that I included gives the information on how to fetch and build everything, but you only need to worry about the kernel.  I may have removed some kernel feature that your cape depends upon (since that kernel came from a different project of mine), so don't get discouraged if you use my image and it doesn't work right away. Rebuilding the kernel will also give you all of the firmware blobs, which you might need to copy over to the boot partition to get your cape working.

Andrew

[a.jain]

I realized that I didn't use step 8, so I did that. However, nothing happened. None of the User LEDs turned on, and I am not sure what is going on

[Andrew Henderson]

On Thursday, June 20, 2013 5:07:54 PM UTC-4, a.jain wrote:

I realized that I didn't use step 8, so I did that. However, nothing happened. None of the User LEDs turned on, and I am not sure what is going on

Well, without more information, I don't think that I can provide much assistance to you.  The FTDI output that I asked for will at least let me know if you're booting off the eMMC, if you're trying to boot off the microSD card and failing (i.e. bad kernel name in the uEnv.txt), and whether the microSD card is even detected by your BBB.  Also, can you at least mount the microSD card on a Linux system and see all four of the partitions?

Andrew

[Andrew Henderson]

I don't like to leave threads unresolved (in case someone is searching the archives when experiencing a similar problem), so here is what we found out.  The problem was that the Android image was dd'd to the microSD card while it was still bzip2 compressed.  Once it was bunzip2'd and dd'd to the card, everything was fine.  The compressed image has a .img.bz2 extension and is around 700 megs in size, and the decompressed image has a .img extension and is closer to 4 GB in size.  Under Linux, you can use bunzip2 to decompress the image. Under Windows, you can use a utility like 7zip (http://www.7-zip.org/download.html).

Andrew

[pe...@icecavern.com]

Sorry I didn't get a chance to try this last night, family things took over yesterday. I will try over the weekend though.

Pete

[Ajay Jain]

For those who are wondering about how to install it on Windows, here's
the link to the wiki:

http://circuitco.com/support/index.php?title=Android

[uros...@gmail.com]

Used Keka inplace of 7zip on mac and dd'd to sd card....for Mac folk.  Worked like a charm.  Watch out for the prompt to use ".img.bz2 inplace of .bz2  for proper extension."  You will end up with a .img.img.bz2 file. I bit at that.......but just once.....and in a weak moment.  So just .bz2 and Bob's your frickin uncle.

[Andrew Henderson]

I wanted to let everyone know that I uploaded a new copy of the Android image.  I took some measurements during the build process so that I could document the total time required to build Android from source (about 6 hours) and the hard drive space needed (about 35 GB).  The download link for the image is still the same: http://icculus.org/~hendersa/BBB_JB_Android_3_8_13.img.bz2

In this image, I built Robert Nelson's kernel without disabling anything, which should allow people with supported cape boards and USB/Bluetooth peripherals to use them with Android.  I got rid of that blinking framebuffer cursor with the proper setting in uEnv.txt.  I also copied all of the built kernel modules and firmware into the Android filesystem.  My previous build had no modules or firmware in it, so any capes that were dependent upon firmware definitely wouldn't work.

This image is a "first run" image, meaning that it will be like turning your Android system on for the first time.  It will take about 2.5 minutes to start up and get to the "welcome to your device" screen.  On subsequent runs, it will only take about a minute or so to get started.

I used this build as a test of my build instructions from scratch, and I did find a few trivial details that I neglected to mention.  These items have been addressed and added into my instructions at: http://icculus.org/~hendersa/android

Andrew

[Robert Nelson]

On Fri, Jun 21, 2013 at 11:36 AM, Andrew Henderson <hend...@icculus.org> wrote:
> I wanted to let everyone know that I uploaded a new copy of the Android
> image. I took some measurements during the build process so that I could
> document the total time required to build Android from source (about 6
> hours) and the hard drive space needed (about 35 GB). The download link for
> the image is still the same:
> http://icculus.org/~hendersa/BBB_JB_Android_3_8_13.img.bz2
>
> In this image, I built Robert Nelson's kernel without disabling anything,
> which should allow people with supported cape boards and USB/Bluetooth
> peripherals to use them with Android. I got rid of that blinking
> framebuffer cursor with the proper setting in uEnv.txt. I also copied all
> of the built kernel modules and firmware into the Android filesystem. My
> previous build had no modules or firmware in it, so any capes that were
> dependent upon firmware definitely wouldn't work.

btw: for 'some' of the capes to work fully, you need the patched
device-tree-compiler 'dtc', since I didn't see this mentioned in your
directions, i figured i'd better point it out, just in case..

Here's a quick script to build the patched dtc, till the '@' option
hits mainline dtc:
https://github.com/RobertCNelson/tools/blob/master/pkgs/dtc.sh

wget -c https://raw.github.com/RobertCNelson/tools/master/pkgs/dtc.sh
chmod +x dtc.sh
./dtc.sh

[a.jain]

I have two other questions: 

One: How can I get Wifi on the Black when it is running Android? I am using this Wifi dongle: http://www.adafruit.com/products/814, and a tutorial to set it up with a BeagleBone White: http://learn.adafruit.com/beaglebone/wifi.

Two: How can I get the BeagleBone Black to recognize the rest of the space on the microSD card? I want to use the rest of the space on the microSD card for storage for apps and games for Android to run. 

[Andrew Henderson]

On Friday, June 21, 2013 12:49:14 PM UTC-4, RobertCNelson wrote:

btw: for 'some' of the capes to work fully, you need the patched
device-tree-compiler 'dtc', since I didn't see this mentioned in your
directions, i figured i'd better point it out, just in case..

Here's a quick script to build the patched dtc, till the '@' option
hits mainline dtc:
https://github.com/RobertCNelson/tools/blob/master/pkgs/dtc.sh

wget -c https://raw.github.com/RobertCNelson/tools/master/pkgs/dtc.sh
chmod +x dtc.sh
./dtc.sh

You know, I actually followed these steps when I built my Android image, and in my directions I told people to follow the steps in the "Linux Kernel" section of your wiki.  Since the instructions for installing the modified dtc were in that section, I figured that people would just do EVERYTHING listed in that section and install the dtc also.  But, I suppose that isn't very clear, and some folks might jump the gun and head straight to grabbing the kernel source.  I didn't want to copy all of the steps verbatim from your wiki, in case you decided to make a minor change here or there, but I might as well just include every step in my instructions to keep people from getting confused.

Andrew

[cody...@gmail.com]

Hey, so i got the img running on my BBB just fine. It seems all the system functionality is working fine but i am having a problem getting apps onto the SD card. Whenever i try downloading anything from the web browser, it either crashes before the download or just completely ignores the download. I tried putting an APK on a thumbdrive and plugged it into the BBB and under storage/usb1 the folder is there for it but the folder is completely empty.

any suggestions on getting an apk on the device maybe a different method or maybe i'm doing something wrong?

Thanks in advance

[rhkr...@gmail.com]

I have tried in vain to get the Android image to boot from uSD card on BBB  I followed your directions exactly, tried two different SD cards, tried two different image writers, tried two different downloads, but the results are the same.  when I hold down the boot button and power up the board either the LEDs don't come on at all or the board boots to Angstrom which makes me think the card isn't being recognized.   It was even stranger last night - after multiple attempts Android did boot from the SD card but it was not when I held down the button.  I have not been able to repeat that even though I've tried all the above.

The SD card is formatted as FAT 32, it looks like all the files are on the card...except I do not have Linux so can't tell what all the partitions are.   Is there a way to boot to Angstrom and mount the SD card?   Could it be that I have a faulty board (Rev A5B)?  Any suggestions would be much appreciated.  I am out of ideas.

Thanks

Roy

[Gerald Coley]

If the board boots the original image form the eMMC, then the boards is fine. Check that first. Then work with the uSD image for the Android.

Gerald

--

For more options, visit http://beagleboard.org/discuss
---

You received this message because you are subscribed to the Google Groups "BeagleBoard" group.
To unsubscribe from this group and stop receiving emails from it, send an email to beagleboard...@googlegroups.com.

For more options, visit https://groups.google.com/groups/opt_out.
 
 

--

Gerald

 

ger...@beagleboard.org

g-co...@ti.com

http://beagleboard.org/

http://circuitco.com/support/

[rhkr...@gmail.com]

The board boots to Angstrom from the eMMC with no problems.  

But, still no luck w/ booting from the uSD card.  I tried using a Kingston 8GB SDHC card;  downloaded the image again, used 7Zip to decompress it wrote the .img file to the SD card using WIn32 Disk imager, all just like the instructions said.  Mounted the card under Ubuntu and it recognized 3 partitions, boot, rootfs and usrdata.  boot and rootfs had what I expected.  usr data was empty except for lost+found.

Same thing, though, when I installed the card and tried to boot to Android.  Held button down, applied power, continued to hold button down for about 10 seconds.  Nothing - no LEDs lit, no Android boot.  don't hold the button and apply power and Angstrom starts up.   Is there any way to tell what's happening during booting?  I can't figure out how to get the console working - until the system partially boots COM3:  (the com port assigned to the BBB) is not recognized.

I read somewhere that if the MLO file on the eMMC card is deleted than the board will be forced to boot from the SD Card.  Is that something to try?   If so, rather than deleting the file can I just rename it to something different will that accomplish the same thing?  At least then I'd know whether the SD Card was being recognized. 

I am out of ideas.  Today I tried a right-out-of-the-package 4GB SDHC card but the image is too big to load  (Win32 imager said not enough room on the card).  What else can I try?  I'm starting to re-think that there is something wrong with the board or with the pushbutton.

Thanks

Roy

[Ryan Ramchandar]

Hey Roy, I am having the exact same problems as you. I followed the same steps and I am getting the same result. If it helps at all, I tried with the HDMI out, and the LCD3 shield. I also played with the resolutions to no avail.

You received this message because you are subscribed to a topic in the Google Groups "BeagleBoard" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/beagleboard/Yc8CdtU5MEs/unsubscribe.
To unsubscribe from this group and all its topics, send an email to beagleboard...@googlegroups.com.

[Wilfredo Nieves]

Sounds to me like the boot button works, otherwise pressing or not pressing the button would get you the same result. Have you looked at the sd socket to make sure all the pins are making contact? Have you tried booting anything else from sd? Have tried to see if you can mount the sd from uboot or at least see if it is visible? Have you formatted sd like the BBB wiki states?

I know that these are a lot of questions but they should help to isolate the problem.

-Wil

[Charles Steinkuehler]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On 6/23/2013 9:59 PM, rhkr...@gmail.com wrote:
> The board boots to Angstrom from the eMMC with no problems.
>
> But, still no luck w/ booting from the uSD card. I tried using a
> Kingston 8GB SDHC card; downloaded the image again, used 7Zip to
> decompress it wrote the .img file to the SD card using WIn32 Disk
> imager, all just like the instructions said. Mounted the card
> under Ubuntu and it recognized 3 partitions, boot, rootfs and
> usrdata. boot and rootfs had what I expected. usr data was empty
> except for lost+found.
>
> Same thing, though, when I installed the card and tried to boot to
> Android. Held button down, applied power, continued to hold button
> down for about 10 seconds. Nothing - no LEDs lit, no Android boot.
> don't hold the button and apply power and Angstrom starts up.

This sounds like the 'Bone isn't recognizing your SD card for some
reason. If you have an SD card inserted and boot without pressing the
boot button, uboot is read off the eMMC but the _very_ first thing it
does is look to see if there is an sd card inserted. If there is, the
eMMC uboot will switch to the sd installed uboot (regardless of
whether or not the sd card contains a boot loader or is even formatted).

So...if you're getting Angstrom to boot out of the eMMC memory *WITH*
an SD card installed, there is something very wrong. If you have any
recognized SD card installed, you should either get the boot image off
the SD card, or the 'Bone will be "hung" at the uboot prompt on the
serial console.

If you hold down the pushbutton, you force the 'Bone to boot directly
off the SD card (bypassing the uboot on the eMMC).

That you're getting Angstrom to boot when you have an SD card
installed, and no boot when you hold down the pushbutton all indicates
your SD card is not being recognized by the system. I'd check for any
bent pins, make sure the SD card is fully inserted, etc.

> Is there any way to tell what's happening during booting? I can't
> figure out how to get the console working - until the system
> partially boots COM3: (the com port assigned to the BBB) is not
> recognized.

Use a USB serial cable. IMHO this is required equipment for anyone
working with the BBB.

> I read somewhere that if the MLO file on the eMMC card is deleted
> than the board will be forced to boot from the SD Card. Is that
> something to try? If so, rather than deleting the file can I just
> rename it to something different will that accomplish the same
> thing? At least then I'd know whether the SD Card was being
> recognized.

Your SD card is *NOT* being recognized. You don't have to mess with
the eMMC to know that (see above).

> I am out of ideas. Today I tried a right-out-of-the-package 4GB
> SDHC card but the image is too big to load (Win32 imager said not
> enough room on the card). What else can I try? I'm starting to
> re-think that there is something wrong with the board or with the
> pushbutton.

It sure sounds like there could be a physical problem with the SD card
socket.

- --
Charles Steinkuehler
cha...@steinkuehler.net
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.11 (MingW32)
Comment: Using GnuPG with Thunderbird - http://www.enigmail.net/

iEYEARECAAYFAlHITlUACgkQLywbqEHdNFwLdQCgj0FqeUjr1Ien83yOl9b0xVbN
q0UAoO5ckNgkDezFrPKyGzcXX+0JU+s1
=KkEp
-----END PGP SIGNATURE-----

[Daniel Escasa]

Elsewhere in this group is a similar thread (maybe this one <g>), and a solution was to power the 'bone through the mains, not through the USB port.

I've been having the same problems, tried an AC/DC adapter and still didn't get anywhere. Had the adapter tested, and it was defective. Still haven't had the time to get a new one, but I'm hoping that a new adapter will get me booting from the uSD

[rhkr...@gmail.com]

Trying to get caught up with all the posts...thanks everybody:

o On did I check the SD socket.  I checked for bent pins, poor solder joints, etc.  Nothing showed up under a lighted magnifier.  In fact I could see scratch marks on the SD card where the pins on the socket made contact.  All of the scratches were about in the middle of the fingers on the SD card so I don't think there is anything wrong there.  Solder joints looked good.  

o On booting something else from the SD card.  Well, the first time I tried the board I followed the Getting Started guide and updated to the latest Angstrom.  That worked as far as I can tell.  Curiously, I also got Android to boot once or twice without the button pressed which Charles explained.   That was the first night.  I have tried multiple cards, multiple images etc. with the same result.,.hold the button apply power, nothing happens,.no LEDs lit.  

o Have I tried mounting the card from UBoot?  No...I don't have a USB-Serial cable...yet to check and I don't get far enough to bring up the virtual serial port.

o Have I formatted the SD card according to the BBB Wiki. I think so.  I also tried writing the image from Linux and Windows w/ the same result.  Partition-wise there are three partitions...boot, usrdata and rootfs.  SD card mounts in both Windows and Ubuntu so I think it's formatted right.

o On powering through the mains and not the USB.  Tried both.  Mains power w/ a 5V@1A wall wart that has worked on similar systems in the past.

So, I've concluded that this seems like a hardware problem with this board.  As somebody pointed out the button is hard to press and hold down so I wonder whether I flexed the board and an internal trace or solder joint is damaged.  Requested an RMA from Mouser...after a few exchanges they said OK.  Now I'll need to get a new board and use the same SD card...I'll post again when I have results.

Roy

[Gerald Coley]

RMA requests are not to go to the distributor but to BeagleBoard.org. http://beagleboard.org/support/RMA

http://circuitco.com/support/index.php?title=BeagleBoneBlack#RMA_Support

Gerald

--

[Wilfredo Nieves]

Hey Roy,

Can you create a backup .img of you sd and provide a download link. I would love to test that image on my BBB with several sd cards to see if it works.

-Wil

--

[Ryan Ramchandar]

Okay, I've had a second go at it. I've tried every combinations of the following:

1. 2 microSD cards: SanDisk 4GB HC Speed 2, and Kingston 8GB HC Speed 4
2. 2 power supplies: 5V @ 2amps
3. 2 BeagloneBone Blacks rev A5B

I flashed the image on the microSD cards using Win8 64bit w/ DiskImager per the instruction of the wiki: http://circuitco.com/support/index.php?title=Android

I tried unplugging all other connections (USB, Ethernet, LCD Capes etc.) and many different combinations. I did notice that the ethernet port LEDS blink rapidly. Though the 4 user LEDS did not become solid and nothing happened.

Unfortunately, after all this I was not able to boot the Android image by Andrew. I believe the issue lies in the Android image itself.

You received this message because you are subscribed to a topic in the Google Groups "BeagleBoard" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/beagleboard/Yc8CdtU5MEs/unsubscribe.
To unsubscribe from this group and all its topics, send an email to beagleboard...@googlegroups.com.

[duncan...@gmail.com]

Have you tried using a computer that isn't running Windows 8? 

I know that I've had similar issues when writing images from Windows 8.  Eventually I tried writing the same image to the same card using my Windows 7 laptop it worked.

[Tommy Snyder]

Ryan Ramchandar,

I just wanted to make sure you aren't expecting the Android image to flash to your eMMC.

This sentence,

> Though the 4 user LEDS did not become solid and nothing happened.

made it seem like you are expecting it to flash to the eMMC. Just wanted to make sure that wasn't the case.

I also wanted to say trying with a Windows 7 computer sounds like a better idea as well.

Tommy

[Andrew Henderson]

On Wednesday, June 26, 2013 12:48:51 AM UTC-4, Ryan Ramchandar wrote:

Okay, I've had a second go at it. I've tried every combinations of the following:

1. 2 microSD cards: SanDisk 4GB HC Speed 2, and Kingston 8GB HC Speed 4
2. 2 power supplies: 5V @ 2amps
3. 2 BeagloneBone Blacks rev A5B

I flashed the image on the microSD cards using Win8 64bit w/ DiskImager per the instruction of the wiki: http://circuitco.com/support/index.php?title=Android

I tried unplugging all other connections (USB, Ethernet, LCD Capes etc.) and many different combinations. I did notice that the ethernet port LEDS blink rapidly. Though the 4 user LEDS did not become solid and nothing happened.

Unfortunately, after all this I was not able to boot the Android image by Andrew. I believe the issue lies in the Android image itself.

I'm sorry that you're having such a difficult time with this!  For what it is worth, I have a class of graduate students that are using this Android image for their projects this semester, and they should all be writing this image to a microSD card prior to their lab session this Friday.  I'll be asking everyone how it went and what OS and tools everyone tried using.  I know that we have a mix of Linux, Win7/8, and Macs in the class, so I can get a better idea of what platforms might be a problem by seeing who had trouble and who didn't.  Maybe this information will help you out.  There are also a mix of rev A5A-C BBB's in the class, as well.

I always use "dd" under Linux to write images to my microSD cards.  I have not tried any other OSes or tools.  Prior to uploading the Android image, I zeroed out a microSD card and then dd'd the image to the card.  On my rev A5A BBB, it booted without any issue.

Andrew

[Ryan Ramchandar]

Thanks guys for the feedback, 

Duncan, I will try writing the image with Win 7 64bit tonight.

Tommy, sorry I meant to say 'until the USER LEDs begin to flash' per the instructions. (Though flashing the eMMC would be a great feature)

Andrew, thanks, I look forward to the results with your class (do you plan to make any material public?). I have a virtualbox of Ubuntu that I will try flashing the image using "dd".

--

[Andrew Henderson]

For those that are having difficulty with installing apps on the Android image, there are a few things that I've been fixing in the image.  I'll continue my work and upload a new image with these items addressed (I'll post to this thread when I do), but until then, you can do the following:

1. The mount device listed in the rootfs's /fstab.am335xevm file is incorrect.  The mount device should be /dev/block/mmcblk0p3, not /dev/block/platform/omap/omap_hsmmc.0/mmcblk0p3.  If you change the path in that file, then the third partition of the SD card (usrdata) will mount to the directory /data in rootfs.  This is what is giving people trouble with storing settings and downloading files: these items are saved to that partition.  The mount will occur on the next system restart.

2. Once you have gotten that changed, copy the .apk that you wish to install into the /media/0/Download directory of the usrdata filesystem.  This directory is linked as the "/sdcard/Download" directory of the rootfs filesystem in the running Android system.  To get that file copied, you have three options:

OPTION 1: Mount the SD card's usrdata filesystem on a Linux system and use "sudo" to copy the file into /media/0/Download directory. 

OPTION 2: Download the file via the web browser app while the system is running.

OPTION 3: Use a USB drive to copy the files over while the system is running.  To do this, you need to have a shell into the running system via your FTDI cable to manually mount your USB drive and copy the files off.  Use the following command:

mount -t FS_TYPE /dev/block/sdaX /usbdrive

... where X is the partition number of the USB drive that you wish to mount (usually you just want the first partition, sda1), and FS_TYPE is the type of filesystem (vfat, most likely, if you are using Windows).  Once you have the USB drive mounted, copy the file from /usbdrive/wherever into /sdcard/Download.

Once the file is in place, launch the application "OI File Manager" and navigate to "sdcard -> Download".  You will see any files that you have manually placed in that Download directory, and also any files that you have downloaded via the web browser (this is where they end up).  Just double-click on the .apk file that want to install and it will install and appear with the rest of your apps.  If you get a warning about not having the correct permissions set, make sure that you have checked the "Unknown Sources" checkbox in the Settings -> Applications dialog.

It is the bad mount point that has been giving everyone trouble when installing apps.

Andrew

[Andrew Henderson]

I just updated the uploaded image to address the issue of the usrdata partition not mounting.  Also, I added a "consoleblank=0" to the kernel command line in uEnv.txt to keep the HDMI interface from shutting off after 10 minutes of operation.

Engineering samples of the LCD3, LCD4, and LCD7 capes arrived today, and I am starting to look at those for Android compatibility and integration.

Andrew

[Ajay Jain]

Let us know how Android looks on the LCD7. I want to know if it looks
blurry or not considering the resolution is 800 x 480. Also, does the
BeagleBone Black now support LCD7 cape?

[Ajay Jain]

If/when you look at the LCD7?, can you take a picture and put it on
here? I want to see how it looks...

On Thu, Jun 27, 2013 at 2:57 PM, Andrew Henderson <hend...@icculus.org> wrote:

[Andrew Henderson]

On Thursday, June 27, 2013 4:27:59 PM UTC-4, a.jain wrote:

If/when you look at the LCD7?, can you take a picture and put it on
here? I want to see how it looks...

I can do better than just a picture: https://www.youtube.com/watch?v=2MmWkt_7-y0

My camera is not that great, and there is a great deal of reflection off of the protective film on the screen, but you can see how it looks.  The screen is actually quite sharp.  There is a little unused space to the right of the screen, but the full 800x480 surface is visible.  Touchscreen works OK in the center of the screen, but it gets progressively more and more skewed from its proper position once you touch around a centimeter or two away from the edge of the screen. My FTDI shell debug output showed that the cape was picked up and that the firmware was loaded out of the filesystem, so it should work for others as-is.  You'll only need to change the resolution from the current "video=720x480-16@60" to "video=800x480-16" on the kernel command line in the uEnv.txt. 

As for the other capes, I saw the upper-left corner of the display being shown on the LCD3 cape (not surprising, since it is only 320x240), but I did not receive any touchscreen events from it.  I'll look into the touchscreen, but I might be able to change the DPI for the display in Android to get it running better.  The screen was refreshing at about 0.5 Hz (I could see it changing scanline-by-scanline), so my configuration is clearly wrong for that particular cape.  For the LCD4 cape, the backlight came on and the screen remained solid white.  The touchscreen events went thorough, though.  I was cat'ing out /dev/input/event1 and getting data when I poked the screen.

Andrew

[Ajay Jain]

Thanks man. Resolution looks decent according to your video, so I'll probably get this screen. 

Are you planning to do any SD card partitioning so Android can recognize SD card space, as well as the Wifi tutorial soon? 

--

[Andrew Henderson]

On Thursday, June 27, 2013 8:20:15 PM UTC-4, a.jain wrote:

Thanks man. Resolution looks decent according to your video, so I'll probably get this screen. 

Are you planning to do any SD card partitioning so Android can recognize SD card space, as well as the Wifi tutorial soon? 

Not at the moment.  I squeeze in my efforts with the BBB here and there when I can, and I have a list of things that I need to get working sooner rather than later.  That stuff takes priority at the moment.

Andrew

[Ajay Jain]

Okay, I understand. Let us know when you can get a chance to work on it.

--

[venkat.e...@gmail.com]

Thanks for the latest image, Andrew.

It does boot up fine and recognize my Samsung HP-T5064 720p TV, though the picture is grainy (using a HDMI 1.4 cable that works fine with an LG 1080p full HD TV). Made no changes to the resolution setting in the uEnv.txt file.

Also, in your image, the audio is not coming through - though the same Samsung TV works (video and audio) fine, using the same cable, with the latest Angstrom image. 'am presuming the audio in your youtube video is from an external source and not from the HDMI monitor.

Is anything special needed to get audio working with your latest android image...

Thanks,
/venkat

[levuph...@gmail.com]

Dear . Andrew Henderson 
I have downloaded file android .img in //circuitco.com/support/index.php?title=Androidhttp

I boot failed. But I boot angstrom and ubuntu successfully.

please help me show error.

thank for any idea.

ö<0>ð<0><0><0><0><0><0><0>

U-Boot SPL 2013.04-dirty (Jun 19 2013 - 09:57:14)

musb-hdrc: ConfigData=0xde (UTMI-8, dyn FIFOs, HB-ISO Rx, HB-ISO Tx, SoftConn)

musb-hdrc: MHDRC RTL version 2.0 

musb-hdrc: setup fifo_mode 4

musb-hdrc: 28/31 max ep, 16384/16384 memory

USB Peripheral mode controller at 47401000 using PIO, IRQ 0

musb-hdrc: ConfigData=0xde (UTMI-8, dyn FIFOs, HB-ISO Rx, HB-ISO Tx, SoftConn)

musb-hdrc: MHDRC RTL version 2.0 

musb-hdrc: setup fifo_mode 4

musb-hdrc: 28/31 max ep, 16384/16384 memory

USB Host mode controller at 47401800 using PIO, IRQ 0

OMAP SD/MMC: 0

mmc_send_cmd : timeout: No status update

reading u-boot.img

reading u-boot.img

U-Boot 2013.04-dirty (Jun 19 2013 - 09:57:14)

I2C:   ready

DRAM:  512 MiB

WARNING: Caches not enabled

NAND:  No NAND device found!!!

0 MiB

MMC:   OMAP SD/MMC: 0, OMAP SD/MMC: 1

*** Warning - readenv() failed, using default environment

musb-hdrc: ConfigData=0xde (UTMI-8, dyn FIFOs, HB-ISO Rx, HB-ISO Tx, SoftConn)

musb-hdrc: MHDRC RTL version 2.0 

musb-hdrc: setup fifo_mode 4

musb-hdrc: 28/31 max ep, 16384/16384 memory

USB Peripheral mode controller at 47401000 using PIO, IRQ 0

musb-hdrc: ConfigData=0xde (UTMI-8, dyn FIFOs, HB-ISO Rx, HB-ISO Tx, SoftConn)

musb-hdrc: MHDRC RTL version 2.0 

musb-hdrc: setup fifo_mode 4

musb-hdrc: 28/31 max ep, 16384/16384 memory

USB Host mode controller at 47401800 using PIO, IRQ 0

Net:   <ethaddr> not set. Validating first E-fuse MAC

cpsw, usb_ether

Hit any key to stop autoboot:  0 

gpio: pin 53 (gpio 53) value is 1

mmc0 is current device

micro SD card found

mmc0 is current device

gpio: pin 54 (gpio 54) value is 1

SD/MMC found on device 0

reading uEnv.txt

534 bytes read in 4 ms (129.9 KiB/s)

Loaded environment from uEnv.txt

Importing environment from mmc ...

Running uenvcmd ...

reading zImage

3320040 bytes read in 403 ms (7.9 MiB/s)

reading am335x-boneblack.dtb

24129 bytes read in 9 ms (2.6 MiB/s)

## Flattened Device Tree blob at 815f0000

   Booting using the fdt blob at 0x815f0000

   Using Device Tree in place at 815f0000, end 815f8e40

Starting kernel ...

Uncompressing Linux... done, booting the kernel.

[    0.000000] Booting Linux on physical CPU 0x0

[    0.000000] Initializing cgroup subsys cpu

[    0.000000] Linux version 3.8.13-bone22 (notroot@ubuntu) (gcc version 4.7.3 20130328 (prerelease) (crosstool-NG linaro-1.13.1-4.7-2013.04-20130415 - Linaro GCC 2013.04) ) #1 SMP Thu Jun 20 22:32:44 UTC 2013

[    0.000000] CPU: ARMv7 Processor [413fc082] revision 2 (ARMv7), cr=10c5387d

[    0.000000] CPU: PIPT / VIPT nonaliasing data cache, VIPT aliasing instruction cache

[    0.000000] Machine: Generic AM33XX (Flattened Device Tree), model: TI AM335x BeagleBone

[    0.000000] Memory policy: ECC disabled, Data cache writeback

[    0.000000] AM335X ES1.0 (neon )

[    0.000000] PERCPU: Embedded 9 pages/cpu @c0e2c000 s14080 r8192 d14592 u36864

[    0.000000] Built 1 zonelists in Zone order, mobility grouping on.  Total pages: 129792

[    0.000000] Kernel command line: consoleblank=0 console=ttyO0,115200n8 androidboot.console=ttyO0 mem=512M root=/dev/mmcblk0p2 rw rootfstype=ext4 rootwait init=/init ip=off video=720x480-16@60 qemu=1 vt.global_cursor_default=0

[    0.000000] PID hash table entries: 2048 (order: 1, 8192 bytes)

[    0.000000] Dentry cache hash table entries: 65536 (order: 6, 262144 bytes)

[    0.000000] Inode-cache hash table entries: 32768 (order: 5, 131072 bytes)

[    0.000000] __ex_table already sorted, skipping sort

[    0.000000] allocated 1048576 bytes of page_cgroup

[    0.000000] please try 'cgroup_disable=memory' option if you don't want memory cgroups

[    0.000000] Memory: 511MB = 511MB total

[    0.000000] Memory: 507256k/507256k available, 17032k reserved, 0K highmem

[    0.000000] Virtual kernel memory layout:

[    0.000000]     vector  : 0xffff0000 - 0xffff1000   (   4 kB)

[    0.000000]     fixmap  : 0xfff00000 - 0xfffe0000   ( 896 kB)

[    0.000000]     vmalloc : 0xe0800000 - 0xff000000   ( 488 MB)

[    0.000000]     lowmem  : 0xc0000000 - 0xe0000000   ( 512 MB)

[    0.000000]     pkmap   : 0xbfe00000 - 0xc0000000   (   2 MB)

[    0.000000]     modules : 0xbf000000 - 0xbfe00000   (  14 MB)

[    0.000000]       .text : 0xc0008000 - 0xc08d6580   (9018 kB)

[    0.000000]       .init : 0xc08d7000 - 0xc0920700   ( 294 kB)

[    0.000000]       .data : 0xc0922000 - 0xc09a4e00   ( 524 kB)

[    0.000000]        .bss : 0xc09a4e00 - 0xc0a1c180   ( 477 kB)

[    0.000000] Hierarchical RCU implementation.

[    0.000000] RCU restricting CPUs from NR_CPUS=4 to nr_cpu_ids=1.

[    0.000000] NR_IRQS:16 nr_irqs:16 16

[    0.000000] IRQ: Found an INTC at 0xfa200000 (revision 5.0) with 128 interrupts

[    0.000000] Total of 128 interrupts on 1 active controller

[    0.000000] OMAP clockevent source: GPTIMER1 at 24000000 Hz

[    0.000000] sched_clock: 32 bits at 24MHz, resolution 41ns, wraps every 178956ms

[    0.000000] OMAP clocksource: GPTIMER2 at 24000000 Hz

[    0.000000] Console: colour dummy device 80x30

[    0.000380] Calibrating delay loop... 363.67 BogoMIPS (lpj=354304)

[    0.017385] pid_max: default: 32768 minimum: 301

[    0.017636] Security Framework initialized

[    0.017735] Mount-cache hash table entries: 512

[    0.027617] Initializing cgroup subsys cpuacct

[    0.027655] Initializing cgroup subsys memory

[    0.027722] Initializing cgroup subsys blkio

[    0.027875] CPU: Testing write buffer coherency: ok

[    0.028468] CPU0: thread -1, cpu 0, socket -1, mpidr 0

[    0.028549] Setting up static identity map for 0x806166f0 - 0x80616748

[    0.030251] Brought up 1 CPUs

[    0.030278] SMP: Total of 1 processors activated (363.67 BogoMIPS).

[    0.031699] devtmpfs: initialized

[    0.097407] pinctrl core: initialized pinctrl subsystem

[    0.097703] rstctl core: initialized rstctl subsystem

[    0.098218] regulator-dummy: no parameters

[    0.098762] NET: Registered protocol family 16

[    0.099925] DMA: preallocated 256 KiB pool for atomic coherent allocations

[    0.110117] pinctrl-single 44e10800.pinmux: 142 pins at pa f9e10800 size 568

[    0.110894] platform 49000000.edma: alias fck already exists

[    0.110925] platform 49000000.edma: alias fck already exists

[    0.110950] platform 49000000.edma: alias fck already exists

[    0.112340] OMAP GPIO hardware version 0.1

[    0.117160] gpio-rctrl rstctl.3: loaded OK

[    0.122361] hw-breakpoint: debug architecture 0x4 unsupported.

[    0.124517] cpsw.0: No hwaddr in dt. Using c8:a0:30:c4:c4:19 from efuse

[    0.124550] cpsw.1: No hwaddr in dt. Using c8:a0:30:c4:c4:1b from efuse

[    0.139272] bio: create slab <bio-0> at 0

[    0.150945] edma-dma-enginstered

[    0.518284] io scheduler cfq registered (default)

[    0.520435] tps65217-bl tps65217-bl: no platform data provided

[    0.520481] tps65217-bl: probe of tps65217-bl failed with error -22

[    0.521332] Serial: 8250/16550 driver, 4 ports, IRQ sharing enabled

[    0.524096] omap_uart 44e09000.serial: did not get pins for uart0 error: -19

[    0.524437] 44e09000.serial: ttyO0 at MMIO 0x44e09000 (irq = 88) is a OMAP UART0

[    1.249582] console [ttyO0] enabled

[    1.254414] [drm] Initialized drm 1.1.0 20060810

[    1.272866] brd: module loaded

[    1.283117] loop: module loaded

[    1.286586] at24 0-0050: 32768 byte 24c256 EEPROM, writable, 1 bytes/write

[    1.293879] at24 1-0054: 32768 byte 24c256 EEPROM, writable, 1 bytes/write

[    1.301156] at24 1-0055: 32768 byte 24c256 EEPROM, writable, 1 bytes/write

[    1.308434] at24 1-0056: 32768 byte 24c256 EEPROM, writable, 1 bytes/write

[    1.315709] at24 1-0057: 32768 byte 24c256 EEPROM, writable, 1 bytes/write

[    1.329971] bone-capemgr bone_capemgr.8: Baseboard: 'A335BNLT,0A5A,1913BBBK2213'

[    1.337763] bone-capemgr bone_capemgr.8: compatible-baseboard=ti,beaglebone-black

[    1.376939] bone-capemgr bone_capemgr.8: slot #0: No cape found

[    1.414047] bone-capemgr bone_capemgr.8: slot #1: No cape found

[    1.451155] bone-capemgr bone_capemgr.8: slot #2: No cape found

[    1.488265] bone-capemgr bone_capemgr.8: slot #3: No cape found

[    1.494515] bone-capemgr bone_capemgr.8: slot #4: specific override

[    1.501115] bone-capemgr bone_capemgr.8: bone: Using override eeprom data at slot 4

[    1.509147] bone-capemgr bone_capemgr.8: slot #4: 'Bone-LT-eMMC-2G,00A0,Texas Instrument,BB-BONE-EMMC-2G'

[    1.519286] bone-capemgr bone_capemgr.8: slot #5: specific override

[    1.525882] bone-capemgr bone_capemgr.8: bone: Using override eeprom data at slot 5

[    1.533918] bone-capemgr bone_capemgr.8: slot #5: 'Bone-Black-HDMI,00A0,Texas Instrument,BB-BONELT-HDMI'

[    1.543965] bone-capemgr bone_capemgr.8: slot #6: specific overri[   67.712532] warning: `zygote' uses 32-bit capabilities (legacy support in use)

[  100.036647] net eth0: initializing cpsw version 1.12 (0)

[  100.044738] net eth0: phy found : id is : 0x7c0f1

[  100.049753] libphy: PHY 4a101000.mdio:01 not found

[  100.054792] net eth0: phy 4a101000.mdio:01 not found on slave 1

[  100.068821] IPv6: ADDRCONF(NETDEV_UP): eth0: link is not ready

[  150.519662] init: sys_prop: permission denied uid:1003  name:service.bootanim.exit

[  153.566758] binder: release 713:939 transaction 9916 out, still active

[  153.910763] binder: 509:525 transaction failed 29189, size 300-0

[  164.146733] binder: 823:942 transaction failed 29189, size 144-8

[  164.153115] binder: send failed reply for transaction 9916, target dead

[Andrew Henderson]

OK, here are a few general tips to help people out with some of the issues that I have been receiving messages about:

1. If you see debug output over your FTDI cable that shows a message like "warning: `zygote' uses 32-bit capabilities (legacy support in use)", you have actually successfully booted Android.  This means that your kernel has completely bootstrapped and you are well into the "init" start-up sequence for the rest of the OS.  If you aren't seeing anything on your screen at that point, it is because you are trying to use a resolution that is not supported by your display.  So don't worry... if you've gotten this far in the system setup, you've written the microSD card image properly and your system IS actually running!

2. There are two lists of resolutions that the kernel pulls from when initializing the framebuffer.  One list is for CEA video modes.  These modes support both audio and video data, and are used by HDMI devices.  The other list are for VGA modes.  These modes are for video-only (DVI monitors, or DVI monitors with an HDMI connection).  The default 720x480 is a CEA mode, which is intended for HDMI displays.  This mode will not work with most DVI/VGA monitors because it is not a standard VGA mode.  If you are using a DVI/VGA monitor (no audio), you can only set a standard VGA mode: 640x480, 800x600, 1024x768, etc.  If you are using HDMI, you can set a CEA mode, like 720x480.

3. You can try completely removing the "video=720x480-16@60" option from the kernel command line in your uEnv.txt.  This option is an override that forces a particular resolution.  If that resolution doesn't work with your particular display, you won't get any video signal.  Your DVI and HDMI devices will report an EDID block to the BBB that tells the kernel what resolutions it supports.  The kernel will, by default, pick the highest resolution mode in the EDID block that matches an entry in its resolution lists.  DVI monitors automatically pick from the non-CEA list (since they don't report any CEA modes), and HDMI devices pick from the CEA list.  If no display is plugged into the BBB when the kernel bootstraps and sets up the video subsystem, or if you are using an exotic adapter or HDMI cable that prevents the EDID block from being sent to the kernel (like an HDMI-to-VGA cable), the kernel will default to a resolution of 1024x768.  The 1024x768 resolution is a non-CEA mode, so it will not give you audio.

4. While there is theoretically audio support for the CEA mode resolutions that support both video and audio, I haven't tested it.  I don't have any audio media loaded onto the system image, so there are no sound effects.  I always stick some background music on my YouTube clips that show the BBB system running, because otherwise they would be far more boring!  If someone wants to test this a bit, try downloading some audio clips and setting them as the sound effects for the various system events.  I think that you should be able to just download something via the web browser and then play it by clicking it in either the completed downloads or the /sdcard/Download directory (browse to it using the OI File Manager app).  I believe the default audio samples are in .ogg format.

As for anyone that has sent me mail, hang in there for a bit and be patient.  I'm working through the backlog to answer everyone as best as I can!

Andrew

[Junqi Zhuo]

that is a really good news! I have already running android on my beaglebone black! Thank you!

在 2013年6月18日星期二UTC+8上午8时00分07秒，Andrew Henderson写道：

Hello all.  I posted this information over at the Google+ group for BeagleBoard.org, but I wanted to post it here as well.  I have Android running on the BBB using the Jellybean (4.2.2) source from arowboat.org and the Linux 3.8.13 kernel from my BeagleSNES project (www.beaglesnes.org).  I just added the Android drivers (Drivers -> Staging Drivers -> Android) into the same kernel configuration that I use for BeagleSNES.  This kernel is really just a stripped down configuration (a 1.9 meg zImage) using Robert Nelson's 3.8.x kernel source tree.  Instead of using the GCC cross-compiler that comes in Rowboat's current prebuilt tools, I use the 3.7 GCC that comes with Robert's kernel tree to build the entire system.  I also updated the u-boot to v2013.04.

Here is a video of it running: https://www.youtube.com/watch?v=yeSt1Y7FhEs
Here is the download of a working Android image that can be written onto a 4GB microSD card using "dd": http://icculus.org/~hendersa/BBB_JB_Android_3_8_13.img.bz2

There is no hardware accelerated video in this version, since SGX support isn't ready for the 3.8 kernel.  To shut off hardware accelerated video, I pass "qemu=1" as a kernel command line argument in uEnv.txt and I add "debug.egl.hw=0" to the /system/build.prop file in the root file system.  To get rid of that blinking cursor on the framebuffer (which I accidentally left on in the image that I made), add this to the uEnv.txt kernel command line arguments: "vt.global_cursor_default=0"

This build will be of interest to Android experimenters who want to use HDMI for audio and video output (rather than using an LCD cape) and who want to experiment with using the device tree.  My kernel is pretty stripped down, so feel free to fetch Robert's latest kernel source for the 3.8.x branch and rebuild your own kernel to drop into the image that I have supplied.  That way, you can include whatever drivers you would like in your kernel.  Use a USB mouse (and a keyboard, too, if you want) to interact with the system.

Thanks, and enjoy!

Andrew

[levuph...@gmail.com]

Ok.thank you Andrew Henderson .

Right .my beaglebone black was boot successfull.I can console it through serial .

 I cannot out HDMI to TV sony 40EX650 with any OS ( angstrom,unbuntu,android) .I checked with TV sharp and TOSHIBA.it is same result.

TV detected HDMI but no picture.I wake up it (click mouse) . when I connect HDMI to board.serial deug doesn't have any info EDID.

I tried with 2 type cable 

http://www.ebay.com/bhp/mini-hdmi-to-hdmi-connector

and

http://www.amazon.com/Fosmon-HDMI-Micro-Cable-Feet/dp/B003UH0Z9Q

I think that driver HDMI cannot load or hardware

how to get EDID in android?

please tell me any idea to test. I don;t have any idea

Thanks.

phuong le

[Daniel Escasa]

Andrew, thanks for this image, got it working after fiddling with the
video resolution. More precisely, I could *see* that it was working
when I found the right res. Here's what I found:

1. Android doesn't ask to be Google-activated. When I got my Android
phone, I couldn't use it until I signed in with my Google account. No
big deal, I don't really intend to lug it around. However...

2. I did manage to side-load the Google Play app, along with the GMail
app. They won't run however, probably because I hadn't activated the
device. (I'm supposed to supply a one-time device-specific password
but couldn't find anything in the Settings. I tried the Accounts item
under Settings, got email and one other that I don't remember at the
moment. I entered my Google credentials, and Android told me that I
needed a device-specific password (or something like that).

3. The browser hangs now and then and asks me whether I want to wait
or force-close.

4. The entire environment is slow, my mouse laggy. Wondering if my
monitor has anything to do with it. I've posted a photo of the back of
my monitor in another thread, but for your reference: dick smith LCT
TV with DVD Player, MODEL: GE6607. Other markings: DSE11080793, N19
GE6607, for whatever they may be worth. I don't remember at the moment
what res I entered in uEnv.txt

5. Again, I don't know if it's my monitor and/or its resolution, but I
don't get sound off the HDMI cable. For whatever it may be worth, I
don't get sound either from my ThinkPad running FreeBSD 9.x. Never did
have the chance to test it with Windows (Vista Home Basic).

6. Just for reference, I tried it with a lower res (480x???) and parts
of the output didn't make it to the screen. E.g., I don't see the
Back+Home+Task button at the bottom of the screen. Changing it to a
higher res fixed that some. And now it occurs to me that maybe I can
see the hidden parts of the output by scrolling?

I don't expect to spend a lot of time in Android on the 'bone, but
thought you might want to know how things went.

--
Daniel O. Escasa
independent IT consultant and writer
contributor, Free Software Magazine (http://www.freesoftwaremagazine.com)
Beta blog at http://techfrontier.overblog.com/
Twitter page at http://www.twitter.com/silverlokk
Google Plus page at http://gotoplus.me/DanEscasa

If we choose being kind over being right, we will be right every time.

[Ajay Jain]

I believe that sound isn't enabled yet, but I'm not exactly sure.

[nikolay...@gmail.com]

On Tuesday, July 2, 2013 4:46:50 PM UTC+9, Daniel Escasa wrote:

1. Android doesn't ask to be Google-activated. When I got my Android
phone, I couldn't use it until I signed in with my Google account. No
big deal, I don't really intend to lug it around. However...

Obviously the BBB is not a 'Google  experience' device, so Google 

accounts, etc. are not supported. Especially when you load an AOSP 

build.

2. I did manage to side-load the Google Play app, along with the GMail
app.

There is quite a bit of related infrastructure needed to get Google accounts 

and Google Play to work, just the Play Store app is not enough. That said it 

does work, however doing it is not exactly in line with Google's terms of 

service.  

The real problem with this build is that it doesn't seem to support ADB 

debugging. I've tried starting adbd from init.rc, but it looks like the mini 

USB port is not detected at all. I will try to see if there are any hints 

in the console log, but any ideas? The 'vanilla' rowboat build does 

support adbd, so it is definitely doable. 

[Andrew Henderson]

I'm trying to get caught up on all of the mails coming in asking questions, so I've been a bit behind.

On Tuesday, July 2, 2013 11:57:00 AM UTC-4, nikolay...@gmail.com wrote:

The real problem with this build is that it doesn't seem to support ADB 

debugging. I've tried starting adbd from init.rc, but it looks like the mini 

USB port is not detected at all. I will try to see if there are any hints 

in the console log, but any ideas? The 'vanilla' rowboat build does 

support adbd, so it is definitely doable. 

I've used ADB, but not over the mini-USB.  I used the network connection to do it.  This past Friday, I had a group of students pushing their apps over to the BBB via ADB and debugging using Eclipse with the current Android image: http://i.imgur.com/JSJY9zt.jpg

I'll be working on addressing some of these issues over the next few days as I have time.

Andrew

 

[Ajay Jain]

Thanks for the update Andrew

 

--

[levuph...@gmail.com]

I was used HDMI port with adapter to VGA and supply power external 5V for cable VGA.It worked

and now. i need to programing uart 0,get and send data.how to turn off mod debug on uart 0.

I test with this app https://code.google.com/p/android-serialport-api/ , then receive message "You do not have read/write permission to the serial port."

Someone say:" need too root device or set permision 777".

tomorow ,i will set up ADB and debug with eclipse,then i will report .

did anyone use to get and send date with serial 0? 

thanks

phuong le

[Vishveshwar]

On Tuesday, July 2, 2013 1:16:50 PM UTC+5:30, Daniel Escasa wrote:

Andrew, thanks for this image, got it working after fiddling with the
video resolution. More precisely, I could *see* that it was working
when I found the right res. Here's what I found:

1. Android doesn't ask to be Google-activated. When I got my Android
phone, I couldn't use it until I signed in with my Google account. No
big deal, I don't really intend to lug it around. However...

2. I did manage to side-load the Google Play app, along with the GMail
app. They won't run however, probably because I hadn't activated the
device. (I'm supposed to supply a one-time device-specific password
but couldn't find anything in the Settings. I tried the Accounts item
under Settings, got email and one other that I don't remember at the
moment. I entered my Google credentials, and Android told me that I
needed a device-specific password (or something like that).

3. The browser hangs now and then and asks me whether I want to wait
or force-close.

There is no SGX graphics acceleration available so you are seeing the FCs. I plan to check if the present SGX libraries in rowboat can be made to work with 3.8 kernel.

4. The entire environment is slow, my mouse laggy. Wondering if my
monitor has anything to do with it. I've posted a photo of the back of
my monitor in another thread, but for your reference: dick smith LCT
TV with DVD Player, MODEL: GE6607. Other markings: DSE11080793, N19
GE6607, for whatever they may be worth. I don't remember at the moment
what res I entered in uEnv.txt

Again due to lack of SGX acceleration.

5. Again, I don't know if it's my monitor and/or its resolution, but I
don't get sound off the HDMI cable. For whatever it may be worth, I
don't get sound either from my ThinkPad running FreeBSD 9.x. Never did
have the chance to test it with Windows (Vista Home Basic).

I'm not sure if the audio HAL can handle HDMI audio.

6. Just for reference, I tried it with a lower res (480x???) and parts
of the output didn't make it to the screen. E.g., I don't see the
Back+Home+Task button at the bottom of the screen. Changing it to a
higher res fixed that some. And now it occurs to me that maybe I can
see the hidden parts of the output by scrolling?

At lower resolutions (e.g. 800x480 on the LCD7) you may see a black strip on the right side. This is supposed to be the Nav bar but the buttons actually don't show.

-Vishveshwar

[Vishveshwar]

adb over usb and Android Composite Gadget probably needs some additional kernel patches to be picked from the google android kernel tree.

-Vishveshwar

[nikolay...@gmail.com]

On Wednesday, July 3, 2013 3:17:43 PM UTC+9, Vishveshwar wrote:

adb over usb and Android Composite Gadget probably needs some additional kernel patches to be picked from the google android kernel tree.

Thanks. Adding the gadget driver seems to enable it, but getting a bunch of USB 

errors on the host:

 usb 1-4.1: new high-speed USB device number 44 using ehci_hcd

 usb 1-4.1: no configurations

 usb 1-4.1: can't read configurations, error -22

 

[Andrew Henderson]

Hello all.  I wanted to let you all know that I've gotten the LCD3, LCD4, and LCD7 capes all working with Android on the BBB.  Both video and touchscreen input work well.

LCD3: http://i.imgur.com/MKqCrgw.jpg

LCD4: http://i.imgur.com/OCw6IUE.jpg

LCD7: http://i.imgur.com/RH6wOWI.jpg

For anyone interested in using these capes, I've included instructions at the bottom of the Android building page: http://icculus.org/~hendersa/android

Andrew

[Ryan Ramchandar]

Great work Andrew! In other good news, I was able to boot your latest image w/ LCD3 by flashing it with Win8 and a 8gb microSD card after trying several times. For some reason it automatically boots without needing to press the boot button while applying power as mentioned in the instructions. This is where I was tripping up on before.

Next I am planning to get your image working with the following 7inch LCD cape from chipsee: http://www.chipsee.com/product/evm/beagle/bbb-exp-c.html?___store=english&___from_store=default

It looks like they include their own android image and source. Do you have any pointers on getting your image to work with this cape?

--

[Andrew Henderson]

On Friday, July 5, 2013 11:15:34 AM UTC-4, Ryan R wrote:

Great work Andrew! In other good news, I was able to boot your latest image w/ LCD3 by flashing it with Win8 and a 8gb microSD card after trying several times. For some reason it automatically boots without needing to press the boot button while applying power as mentioned in the instructions. This is where I was tripping up on before.

Next I am planning to get your image working with the following 7inch LCD cape from chipsee: http://www.chipsee.com/product/evm/beagle/bbb-exp-c.html?___store=english&___from_store=default

It looks like they include their own android image and source. Do you have any pointers on getting your image to work with this cape?

That might be a bit tricky.  I suspect that they are running an Android that is based off of the Linux 3.2 kernel (and probably based off of Rowboat or the TI EVM reference SDK).  If that is the case, the device tree mechanism that is present in Linux 3.8 won't be there.  The cape features will be "hard-coded", per se, into their kernel.  That makes supporting that cape for 3.8 tricky, since the hard-coded functionality needs to be changed over to device tree format.  While it should be straightforward to set up the device tree structure by following the pin muxing of the hard-coded cape driver in their kernel, you still have to get the cape's firmware built.

While there are a few interface differences between Linux 3.2 and 3.8, the big difference from the BBB's view is that of the device tree, firmware, and capebus support.  Which, unfortunately, is what "cape support" actually is.

Now, that all being said, it should be doable.  It just won't be doable from a "tweak this config file and you're good to go".  It will take some engineering effort.  If I'm wrong about the kernel that they are using and it is actually a 3.8 kernel their Android is based off on, then it should be more of a "tweak the config file" solution to get it working.

Andrew

[nikolay...@gmail.com]

On Thursday, July 4, 2013 3:23:19 AM UTC+9, nikolay...@gmail.com wrote:

On Wednesday, July 3, 2013 3:17:43 PM UTC+9, Vishveshwar wrote:

adb over usb and Android Composite Gadget probably needs some additional kernel patches to be picked from the google android kernel tree.

Thanks. Adding the gadget driver seems to enable it, but getting a bunch of USB 

errors on the host:

 

This is proving to be trickier than I though. Any pointers to getting a shell on the 

serial interface in the mean time? It shows a root prompt, but does not accept 

commands. I can see kernel output though. 

 

[rhkr...@gmail.com]

I got a serial debug cable (which was very helpful) and a different BBB. The serial console with the board I've been having problems with indicated that the SD card wasn't being found.  Taking the same uSD (Kingston 8GB) to a different BBB worked (other than not having to press the boot button).  Trying  a second card (SanDisk 32MB) also worked so I'm convinced my problem was a board problem.  

I ended up using Ubuntu running under VMWare w/ a USB uSD card reader to create the images.  I haven't gone back to try to do the same with the Win32 imager.

Thanks for all the support.

Roy

On Tuesday, June 25, 2013 6:02:10 AM UTC-7, Wilfredo Nieves wrote:

Hey Roy,

Can you create a backup .img of you sd and provide a download link. I would love to test that image on my BBB with several sd cards to see if it works.

-Wil

On Jun 25, 2013 6:47 AM, <rhkr...@gmail.com> wrote:

Trying to get caught up with all the posts...thanks everybody:

o On did I check the SD socket.  I checked for bent pins, poor solder joints, etc.  Nothing showed up under a lighted magnifier.  In fact I could see scratch marks on the SD card where the pins on the socket made contact.  All of the scratches were about in the middle of the fingers on the SD card so I don't think there is anything wrong there.  Solder joints looked good.  

o On booting something else from the SD card.  Well, the first time I tried the board I followed the Getting Started guide and updated to the latest Angstrom.  That worked as far as I can tell.  Curiously, I also got Android to boot once or twice without the button pressed which Charles explained.   That was the first night.  I have tried multiple cards, multiple images etc. with the same result.,.hold the button apply power, nothing happens,.no LEDs lit.  

o Have I tried mounting the card from UBoot?  No...I don't have a USB-Serial cable...yet to check and I don't get far enough to bring up the virtual serial port.

o Have I formatted the SD card according to the BBB Wiki. I think so.  I also tried writing the image from Linux and Windows w/ the same result.  Partition-wise there are three partitions...boot, usrdata and rootfs.  SD card mounts in both Windows and Ubuntu so I think it's formatted right.

o On powering through the mains and not the USB.  Tried both.  Mains power w/ a 5V@1A wall wart that has worked on similar systems in the past.

So, I've concluded that this seems like a hardware problem with this board.  As somebody pointed out the button is hard to press and hold down so I wonder whether I flexed the board and an internal trace or solder joint is damaged.  Requested an RMA from Mouser...after a few exchanges they said OK.  Now I'll need to get a new board and use the same SD card...I'll post again when I have results.

Roy

On Monday, June 24, 2013 6:42:07 PM UTC-7, Daniel Escasa wrote:

Elsewhere in this group is a similar thread (maybe this one <g>), and a solution was to power the 'bone through the mains, not through the USB port.

I've been having the same problems, tried an AC/DC adapter and still didn't get anywhere. Had the adapter tested, and it was defective. Still haven't had the time to get a new one, but I'm hoping that a new adapter will get me booting from the uSD

On Monday, June 24, 2013 10:59:19 AM UTC+8, rhkr...@gmail.com wrote:

The board boots to Angstrom from the eMMC with no problems.  

But, still no luck w/ booting from the uSD card.  I tried using a Kingston 8GB SDHC card;  downloaded the image again, used 7Zip to decompress it wrote the .img file to the SD card using WIn32 Disk imager, all just like the instructions said.  Mounted the card under Ubuntu and it recognized 3 partitions, boot, rootfs and usrdata.  boot and rootfs had what I expected.  usr data was empty except for lost+found.

Same thing, though, when I installed the card and tried to boot to Android.  Held button down, applied power, continued to hold button down for about 10 seconds.  Nothing - no LEDs lit, no Android boot.  don't hold the button and apply power and Angstrom starts up.   Is there any way to tell what's happening during booting?  I can't figure out how to get the console working - until the system partially boots COM3:  (the com port assigned to the BBB) is not recognized.

--
For more options, visit http://beagleboard.org/discuss
---

You received this message because you are subscribed to the Google Groups "BeagleBoard" group.
To unsubscribe from this group and stop receiving emails from it, send an email to beagleboard...@googlegroups.com.

[rhkr...@gmail.com]

I saw the post on the LCD3 and removed just the video=720x480-16@60 part of that line in uenv.txt.   That left the qemu=1 and the vt.global_cursor_default=0 portion of the line.  Is that correct?

The reason I ask is that with an LCD on my (now sucessfully booting Android new BBB) I can see the clock in the right corner and it looks like it's scaled correctly for the display, but I don't see the rest of the desktop and the panel seems unresponsive to touch.  

Thanks

Roy

[Andrew Henderson]

On Saturday, July 6, 2013 8:24:49 PM UTC-4, rhkr...@gmail.com wrote:

I saw the post on the LCD3 and removed just the video=720x480-16@60 part of that line in uenv.txt.   That left the qemu=1 and the vt.global_cursor_default=0 portion of the line.  Is that correct?

The reason I ask is that with an LCD on my (now sucessfully booting Android new BBB) I can see the clock in the right corner and it looks like it's scaled correctly for the display, but I don't see the rest of the desktop and the panel seems unresponsive to touch.  

It sounds like you have everything set up correctly.  The "qemu=1" disables hardware graphics acceleration (SGX) because it is not yet supported in the 3.8 kernel.  "vt.global_cursor_default=0" gets rid of the annoying blinking cursor on the framebuffer. The touchscreen should work, though it isn't all that responsive around the edges.  Try testing it out with a USB mouse and making sure that it looks and acts like what you would expect.  If you move the mouse and don't see the cursor, then let us know.  Also, do you see any thrown exceptions from apps via the FTDI interface?  If you accidentally inserted a line break into the bootargs line and moved the "qemu=1" off the bootargs line, then apps will be crashing left and right, leading to the missing desktop.

Is the Android loading animation (the "Android" text with the reflection moving across it) pretty fast?  If it really, really slow, then the resolution setting isn't quite right and touchscreen won't work.  This shouldn't be a problem if you completely remove the "video=" option from the kernel bootargs, though.  The only time that I've seen touchscreen not work is when I've explicitly specified a resolution with "video=" in the bootargs.

Andrew   

[rhkr...@gmail.com]

Thanks Andrew...you were right on.  I did insert a line break (more ignorance than accident when I edited the file.  Fixed that and now my screen looks like yours.

Roy

[nikolay...@gmail.com]

No luck with the merged Android gadget driver, but it mostly works using the  

  Function Filesystem driver as described below. The only problem I have is 

that the serial number shows as '????????????' which Eclipse doesn't like. 

https://plus.google.com/111524780435806926688/posts/AaEccFjKNHE

Also add something like this to init.usb.rc, etc.:

on fs
    mkdir /dev/usb-ffs 0770 shell shell
    mkdir /dev/usb-ffs/adb 0770 shell shell
    mount functionfs adb /dev/usb-ffs/adb uid=2000,gid=2000

[nikolay...@gmail.com]

On Monday, July 8, 2013 2:09:00 AM UTC+9, nikolay...@gmail.com wrote:

 

No luck with the merged Android gadget driver, but it mostly works using the  

  Function Filesystem driver as described below. The only problem I have is 

that the serial number shows as '????????????' which Eclipse doesn't like. 

Adding 'g_ffs.iSerialNumber=<serial>' to uEnv.txt fixes that.

[pe...@icecavern.com]

Great stuff. Just tried the latest image with my LCD4 and it's all working nicely.

One thing I can't seem to do is get back to the home screen with just the screen. With a mouse plugged in I can click on the furthest right pixel and then it goes to the home screen. But I can't touch this far line on the touch screen it appears. Could the buttons on the board be reconfigured somehow?

[ajav...@gmail.com]

I apologize in advance for being a complete newbie to Linux, but I read the instructions provided, and does this mean that if we just use the pre-built image at the link below, we don't have to follow the rest of the instructions on the page for building Android yourself? Thanks : )

On Friday, June 21, 2013 12:36:00 PM UTC-4, Andrew Henderson wrote:

I wanted to let everyone know that I uploaded a new copy of the Android image.  I took some measurements during the build process so that I could document the total time required to build Android from source (about 6 hours) and the hard drive space needed (about 35 GB).  The download link for the image is still the same: http://icculus.org/~hendersa/BBB_JB_Android_3_8_13.img.bz2

In this image, I built Robert Nelson's kernel without disabling anything, which should allow people with supported cape boards and USB/Bluetooth peripherals to use them with Android.  I got rid of that blinking framebuffer cursor with the proper setting in uEnv.txt.  I also copied all of the built kernel modules and firmware into the Android filesystem.  My previous build had no modules or firmware in it, so any capes that were dependent upon firmware definitely wouldn't work.

This image is a "first run" image, meaning that it will be like turning your Android system on for the first time.  It will take about 2.5 minutes to start up and get to the "welcome to your device" screen.  On subsequent runs, it will only take about a minute or so to get started.

I used this build as a test of my build instructions from scratch, and I did find a few trivial details that I neglected to mention.  These items have been addressed and added into my instructions at: http://icculus.org/~hendersa/android

Andrew

[Andrew Henderson]

On Tuesday, August 6, 2013 10:08:51 AM UTC-4, ajav...@gmail.com wrote:

I apologize in advance for being a complete newbie to Linux, but I read the instructions provided, and does this mean that if we just use the pre-built image at the link below, we don't have to follow the rest of the instructions on the page for building Android yourself? Thanks : )

On Friday, June 21, 2013 12:36:00 PM UTC-4, Andrew Henderson wrote:

I wanted to let everyone know that I uploaded a new copy of the Android image.  I took some measurements during the build process so that I could document the total time required to build Android from source (about 6 hours) and the hard drive space needed (about 35 GB).  The download link for the image is still the same: http://icculus.org/~hendersa/BBB_JB_Android_3_8_13.img.bz2

That is correct.  The pre-built image is there to help you get started using Android right away.  You do not need to follow all of my build instructions unless you are interested in building your own Android image for the BBB.  Rather than looking at the build instructions and getting confused, try following the instructions on the CircuitCo wiki:

http://circuitco.com/support/index.php?title=Android

Then, if you decide later that you'd like to try making a complete Android image yourself from scratch, you can follow my build instructions.

Andrew 

[ami...@gmail.com]

Hi Nelson ,

Thanks for all your efforts in adding support for BBB.

I ported your Android Image On My BBB which i used for Device driver Testing.But  with this build my mini usb is not detected with lsusb command thus  i not able to use adb on usb. 

Can you please help me out finding way to enable Mini Usb on BBB.

Thanks 

Amit   

On Tuesday, June 18, 2013 11:31:37 PM UTC+5:30, RobertCNelson wrote:

On Tue, Jun 18, 2013 at 12:42 PM, Ryan Ramchandar <ryan...@gmail.com> wrote:
> Thanks for sharing Andrew, This looks like a very interesting project.
>
> Can you elaborate further on how you were able to boot the Android 4.2.2
> image on the BBB. From what I've been reading the arowboat guys have not yet
> got it working on the BBB.

It's actually pretty cool what Andrew just did.. So in the last few
kernel releases there has been renewed interest in merging the android
kernel stuff in mainline..  So Andrew took the kernel source we are
using for Angstrom/Ubuntu and rebuilt it with more of the Android
config options enabled..  Essentially ignoring the 'arowboat' kernel..
Tweaked a few things, and bam, Android works. ;)

Regards,

--
Robert Nelson
http://www.rcn-ee.com/

[Andrew Henderson]

On Thursday, August 8, 2013 3:41:42 AM UTC-4, ami...@gmail.com wrote:

Hi Nelson ,

Thanks for all your efforts in adding support for BBB.

I ported your Android Image On My BBB which i used for Device driver Testing.But  with this build my mini usb is not detected with lsusb command thus  i not able to use adb on usb. 

Can you please help me out finding way to enable Mini Usb on BBB.

Thanks 

Amit   

 

Nikolay Elenkov was experimenting with this Android image and he made some observations on using FunctionFS, rather than the USB gadget driver, to get ADB working: http://nelenkov.blogspot.com/2013/07/building-wireless-android-device.html

Search for "ADB support" in his blog post for more information.

Andrew

[a.jain]

Has anyone gotten wifi working on the BeagleBone Black running Android yet?

On Monday, June 17, 2013 7:00:07 PM UTC-5, Andrew Henderson wrote:

Hello all.  I posted this information over at the Google+ group for BeagleBoard.org, but I wanted to post it here as well.  I have Android running on the BBB using the Jellybean (4.2.2) source from arowboat.org and the Linux 3.8.13 kernel from my BeagleSNES project (www.beaglesnes.org).  I just added the Android drivers (Drivers -> Staging Drivers -> Android) into the same kernel configuration that I use for BeagleSNES.  This kernel is really just a stripped down configuration (a 1.9 meg zImage) using Robert Nelson's 3.8.x kernel source tree.  Instead of using the GCC cross-compiler that comes in Rowboat's current prebuilt tools, I use the 3.7 GCC that comes with Robert's kernel tree to build the entire system.  I also updated the u-boot to v2013.04.

Here is a video of it running: https://www.youtube.com/watch?v=yeSt1Y7FhEs
Here is the download of a working Android image that can be written onto a 4GB microSD card using "dd": http://icculus.org/~hendersa/BBB_JB_Android_3_8_13.img.bz2

There is no hardware accelerated video in this version, since SGX support isn't ready for the 3.8 kernel.  To shut off hardware accelerated video, I pass "qemu=1" as a kernel command line argument in uEnv.txt and I add "debug.egl.hw=0" to the /system/build.prop file in the root file system.  To get rid of that blinking cursor on the framebuffer (which I accidentally left on in the image that I made), add this to the uEnv.txt kernel command line arguments: "vt.global_cursor_default=0"

This build will be of interest to Android experimenters who want to use HDMI for audio and video output (rather than using an LCD cape) and who want to experiment with using the device tree.  My kernel is pretty stripped down, so feel free to fetch Robert's latest kernel source for the 3.8.x branch and rebuild your own kernel to drop into the image that I have supplied.  That way, you can include whatever drivers you would like in your kernel.  Use a USB mouse (and a keyboard, too, if you want) to interact with the system.

Thanks, and enjoy!

Andrew

[pic.m...@gmail.com]

a.jain , 

I was searching too for a wifi working. Andrew Henderson setup is very useful the only thing missing is wifi (for me).

Did you find a solution?

thx

[Ajay Jain]

I have not found a solution yet :(. Hopefully there will be one soon!

[pic.m...@gmail.com]

It seems that Nikolay worked out the details to get wifi running. It would be nice if we can get the binaries to try 

http://nelenkov.blogspot.com/2013/07/building-wireless-android-device.html

[Ajay Jain]

Didn't see this link yet, I'll check it out. Thanks!

[mukel...@gmail.com]

I had a look at his page and I can't seem to get the ADB to work with FunctionFS. Did it work for you?

Also has anyone got sshd to work?

[david.m...@gmail.com]

I'm also having problems starting adbd with kernel 3.8 and I think I found the problem. I tried to load the kernel module using this command:

 insmod /system/lib/modules/3.8.13-bone28/kernel/drivers/usb/gadget/g_ffs.ko idVendor=0x18d1 idProduct=0x4e26 iSerialNumber=0123456789ABCDEF functions=adb

The command generates the following errors:

[  943.915225] g_ffs: Unknown symbol usb_string_ids_n (err 0)

[  943.921212] g_ffs: Unknown symbol usb_interface_id (err 0)

[  943.927004] g_ffs: Unknown symbol usb_assign_descriptors (err 0)

[  943.933337] g_ffs: Unknown symbol usb_string_ids_tab (err 0)

[  943.939334] g_ffs: Unknown symbol usb_composite_unregister (err 0)

[  943.945824] g_ffs: Unknown symbol config_ep_by_speed (err 0)

[  943.951798] g_ffs: Unknown symbol usb_add_function (err 0)

[  943.957555] g_ffs: Unknown symbol usb_composite_probe (err 0)

[  943.963588] g_ffs: Unknown symbol usb_composite_overwrite_options (err 0)

[  943.970761] g_ffs: Unknown symbol usb_ep_autoconfig (err 0)

[  943.976810] g_ffs: Unknown symbol usb_add_config (err 0)

[  943.982402] g_ffs: Unknown symbol usb_free_all_descriptors (err 0)

insmod: can't insert '/system/lib/modules/3.8.13-bone28/kernel/drivers/usb/gadget/g_ffs.ko': unknown symbol in module or invalid parameter

How can I solve this problem?

[Nikolay Elenkov]

On Wed, Oct 2, 2013 at 5:24 AM, <david.m...@gmail.com> wrote:
>> I'm also having problems starting adbd with kernel 3.8 and I think I found
>> the problem. I tried to load the kernel module using this command:
>
> insmod /system/lib/modules/3.8.13-bone28/kernel/drivers/usb/gadget/g_ffs.ko
> idVendor=0x18d1 idProduct=0x4e26 iSerialNumber=0123456789ABCDEF
> functions=adb
>
> The command generates the following errors:
>
> [ 943.915225] g_ffs: Unknown symbol usb_string_ids_n (err 0)

Android doesn't load dependent modules for you, check g_ffs description for
what those are and preload them. And really, why build as a module?
Build this into the kernel.

[david.m...@gmail.com]

I've now built the kernel with g_ffs built directly into the kernel and now adb works!

[sultan.ar...@gmail.com]

hello!

please tell me the steps to make ADB work.
as i have set CONFIG_USB_FUNCTIONFS=y
but when i write ./adb devices
my device not appear in the list
however i check the usb debugging in my BBB.

Best Regards,
Sultan


[vega...@gmail.com]

Andrew:

Thank you for this post.  I have this image working on my BBB using 4D Systems 4DCAPE-43T Cape by editing the uEnv.txt as: video=480x272-16@56 .  I am able to SSH to the device using Android SSH Server on the device and putty on my PC.  However, I need to know the root password for this image.  I like to avoid using 'adb shell' to get to the root.  I really want to root this device.  I use Xamarin as a development tool and it works well.  However, when I try to access, via code, the "/sys/class/leds/beaglebone:green:usr0"

to set the trigger to "none" I get Access Denied.  I need root access via my code...I am pulling my hair on that one.  any help is appreciated.

Kindly

Rich

[Mahendra Gunawardena]

Hello Rich,

I am also trying to boot BBB with 4DCAPE-43T. 

1. Can you please share your uEnv.txt?
2. Did you used a pre-built image from TI-Andriod or something different?
3. Did you boot off the uSD card?

Appreciate your help 

Mahendra

On Monday, December 30, 2013 11:40:53 PM UTC-5, vega...@gmail.com wrote:

Andrew:

Thank you for this post.  I have this image working on my BBB using 4D Systems 4DCAPE-43T Cape by editing the uEnv.txt as: video=48...@56 .  I am able to SSH to the device using Android SSH Server on the device and putty on my PC.  However, I need to know the root password for this image.  I like to avoid using 'adb shell' to get to the root.  I really want to root this device.  I use Xamarin as a development tool and it works well.  However, when I try to access, via code, the "/sys/class/leds/beaglebone:green:usr0"

[Mahendra Gunawardena]

Rich

I got  Andrew Henderson JB image working, 

Thanks

--

[pru....@gmail.com]

On Sunday, July 7, 2013 4:09:26 AM UTC+2, Andrew Henderson wrote:

[...]

Try testing it out with a USB mouse and making sure that it looks and acts like what you would expect.  If you move the mouse and don't see the cursor, then let us know.  [...]

Andrew   

 

Hi Andrew,
I run the 3.8.13 image on BBB with LCD7 cape and I do not have a mouse pointer. Plugging-in mouse I can see in the log it is detected. The usb keyboard is working fine.
Any hints?

Thanks
Pru

[pru....@gmail.com]

I have just tested with hdmi connection (so no LCD7) and it's the same - no mouse.

[Suzanne Lenox]

Hi,

I have Andrew Henderson's image running with LCD4 cape but now I am trying to communicate with the serial ports.

I have tried modifying the init.rc to change the UART permissions as recommended in another post but that did not help.

Also, I cannot access via abd. Any suggestions?

Do I need to compile my own image?

Many thanks for any help.

Suzanne

[paccer]

Strangely, I've been unable to make touchscreen work with "andrews jb" on 4D systems 4DCAPE_70T

I've tried with the prebuilt image, and also tried building from source, using the instructions on Andrews webpage, using latest 3.8.x kernel sources.

Touch screen works fine on TI_Android_JB_4.2.2_DevKit_4.1.1 image, so it's definitely not an hardware issue.

I'm currently baffled, since its reported that LCD7 works, and the 4DCAPE_70T should be similar, it appears.

Any help/suggestions highly appriciated - maybe there's some patches available i need?

I did manage to get GApps (google play etc.) + SuperSU working on the TI_Android_JB image, and i'd gladly provide details if requested.

Thanks

paccer

[Vishveshwar]

On Thursday, February 27, 2014 6:09:34 AM UTC+5:30, paccer wrote:

Strangely, I've been unable to make touchscreen work with "andrews jb" on 4D systems 4DCAPE_70T

I've tried with the prebuilt image, and also tried building from source, using the instructions on Andrews webpage, using latest 3.8.x kernel sources.

Touch screen works fine on TI_Android_JB_4.2.2_DevKit_4.1.1 image, so it's definitely not an hardware issue.

I'm currently baffled, since its reported that LCD7 works, and the 4DCAPE_70T should be similar, it appears.

Any help/suggestions highly appriciated - maybe there's some patches available i need?

There are two things w.r.t touchscreen:
- The touch operation itself - is the driver loaded? are you getting touch interrupts?
- Calibration - mapping the usable area so that the touch interacts with the correct display element (icon etc)

Some information is at the AOSP page:
http://source.android.com/devices/tech/input/touch-devices.html

FYI, In the TI DevKit kernel (3.2) the touch calibration data is part of the platform data for the touch driver.
See my post in android-porting for more information about this:

https://groups.google.com/d/msg/android-porting/CpP4xCk0U_g/DaCoAmqj6FwJ

-Vishveshwar
 

[Daniel Delgado Terán]

Any chances you can please share that image?