Android on Freerunner image and some instructions
Sean McNeil
of Android for people to try out. You can find my musings at:
http://wiki.openmoko.org/wiki/User:Seanmcneil3
This will take time to get down right, but hopefully now you will all be
able to begin playing with Android.
Cheers,
Sean
Peter Neubauer
Thanks for the great work all Android porters!
I am tryingt o boot up my newly android-flashed FreeRunner, but the
boot screen animation seems to be out of scale, the android text and
non-animated droid is not in the center but in the lower left quadrant
of the screen, the animation in the vertical center near the right
edge of the screen, bouncing away.
After a while, the bouncing just stops and nothing happens. Is there
any way to see what's wrong? Is there a Windows driver for connecting
the USB port to my desktop in order to get adb discovering the device?
Thanks for any hints!
/peter
http://www.oredev.se - Be there or be gone.
GTalk: neubauer.peter
Skype peter.neubauer
ICQ 18762544
Phone +46704 106975
LinkedIn http://www.linkedin.com/in/neubauer
Twitter http://twitter.com/peterneubauer
http://www.neo4j.org - New Energy for Data - the Graph Database.
http://www.ops4j.org - New Energy for OSS Communities - Open
Participation Software.
http://www.qi4j.org - New Energy for Java - Domain Driven Development.
Sean McNeil
The animator currently doesn't scale to the screen and assumes QVGA. I'm
sure that will be fixed pretty quick.
I've put information as well as a link to adb on my wiki page at
Openmoko. What you want to do is attach the phone via. usb to your host
machine with ethernet-over-usb setup. The phone will come up as the
normal "neo" hostname (aka 192.168.0.202). You can then issue the command:
ADBHOST=neo ./adb logcat
This will show you what is going on. Most likely, you do not have your
sdcard divided in half as required. (see Openmoko wiki page).
Sean
Peter Neubauer
On Wed, Nov 5, 2008 at 10:19 AM, Sean McNeil <seanm...@gmail.com> wrote:
>
> Hi Peter,
>
> The animator currently doesn't scale to the screen and assumes QVGA. I'm
> sure that will be fixed pretty quick.
Ok, great.
> This will show you what is going on. Most likely, you do not have your
> sdcard divided in half as required. (see Openmoko wiki page).
Ahh, is that a requirement? I have a 8GB card, so I need to split it
up? On my FR Debian, there is no mkfs.vfat (need to add that
probably), formatting the vfat partition, even if you are not going to
put user data in there, is a requirement too so it can be mounted at
least, I guess?
Thanks!
/peter
Brian Swetland
>
> Hi Peter,
>
> The animator currently doesn't scale to the screen and assumes QVGA. I'm
> sure that will be fixed pretty quick.
>
> I've put information as well as a link to adb on my wiki page at
> Openmoko. What you want to do is attach the phone via. usb to your host
> machine with ethernet-over-usb setup. The phone will come up as the
> normal "neo" hostname (aka 192.168.0.202). You can then issue the command:
>
> ADBHOST=neo ./adb logcat
>
> This will show you what is going on. Most likely, you do not have your
> sdcard divided in half as required. (see Openmoko wiki page).
If you want to get adb running directly, you could give the adb gadget
driver a shot:
This should allow adb to operate over the usb transport directly (as we
do on dream/g1) instead of tunneled over ethernet.
Sean McNeil
Yes, but it is currently a requirement. Also, I've never used an SDHC
card on my freerunner, so I don't even know if the kernel I've provided
supports it. I'm pretty sure the newer kernels do at least.
Sean
Neng-Yu Tu (Tony Tu)
>> This will show you what is going on. Most likely, you do not have your
>> sdcard divided in half as required. (see Openmoko wiki page).
> Ahh, is that a requirement? I have a 8GB card, so I need to split it
> up? On my FR Debian, there is no mkfs.vfat (need to add that
> probably), formatting the vfat partition, even if you are not going to
> put user data in there, is a requirement too so it can be mounted at
> least, I guess?
>
I have tried using my old 128 MB SD card, and format into 8MB (vfat,
first partition, primary) and 120 (ext3, 2nd partition, primary) and works.
seems 2 partitions is required, and 2nd one must use ext3.
As I saw, most system data was storage in ext3 (2nd partition), 1
partition seems used for media content only.
Tony Tu (Neng-Yu Tu)
Vincenzo Ampolo
> I have tried using my old 128 MB SD card, and format into 8MB (vfat,
> first partition, primary) and 120 (ext3, 2nd partition, primary) and works.
>
> seems 2 partitions is required, and 2nd one must use ext3.
Good!
I'm going to make a android.img to be flashed to a 512+ MB SD card to install
android easily with dd.
thanks for your work guys :)
--
Vincenzo Ampolo
http://vincenzo-ampolo.net
cedric.berger
On 5 nov, 10:32, Brian Swetland <swetl...@google.com> wrote:
> [Sean McNeil <seanmcne...@gmail.com>]
> > I've put information as well as a link to adb on my wiki page at
> > Openmoko. What you want to do is attach the phone via. usb to your host
> > machine with ethernet-over-usb setup. The phone will come up as the
> > normal "neo" hostname (aka 192.168.0.202). You can then issue the command:
>
> > ADBHOST=neo ./adb logcat
>
> driver a shot:
>
> This should allow adb to operate over the usb transport directly (as we
> do on dream/g1) instead of tunneled over ethernet.
I tried to install adb usb driver from android SDK. I edited the .inf
(I do not really know what I do there...) so that device ID is
recognized as Neo Freerunner, so it was "correctly" installed in XP,
and it detects android when plugged in.
But adb does not detect device...
Toggy
Can I have some info for text input.
Tnx
Nikolaus
on my Freerunner that soon...
Only one thing, if I run Android with my SIM-card inserted it won't
accept my PIN code.
Has anyone else had that issue? It's a german o2 SIM card.
Cheers,
Nik
Mike Lockwood
#define DRIVER_VENDOR_ID 0x18D1
#define DRIVER_PRODUCT_ID 0x0001
If you have a USB inspecting utility like USBView or lsusb, you could check to make sure your device appears on the USB bus with these values for the VID and PID. It should have a single interface with two bulk endpoints. Note that the android_adb driver differs from how the G1 appears on the USB bus.
--
Mike Lockwood
Google android team
Peter Neubauer
I have tried several times now with several cards etc, but all I get
is the same as I described, and the AUX LED flashing fast red when the
booting stops with the Android icon animation standing still. Can that
give any hint on what could be wrong?
/peter
http://www.oredev.se - Be there or be gone.
GTalk: neubauer.peter
Skype peter.neubauer
ICQ 18762544
Phone +46704 106975
LinkedIn http://www.linkedin.com/in/neubauer
Twitter http://twitter.com/peterneubauer
http://www.neo4j.org - New Energy for Data - the Graph Database.
http://www.ops4j.org - New Energy for OSS Communities - Open
Participation Software.
http://www.qi4j.org - New Energy for Java - Domain Driven Development.
natr0x
Thanks for the tipp!
But i can't get the driver running on windows xp sp2 (i tryed the
following driver from the android website: http://dl.google.com/android/android_usb_windows.zip)
i also tryed to change the vendor and the product id, but no chance!
any idea? or a running driver ;)
greets Daniel
On 5 Nov., 19:56, "Mike Lockwood" <lockw...@android.com> wrote:
Mike Lockwood
That is the latest driver we have. When you connect your device, does Windows put up the Hardware Wizard dialog? If not, then there is probably something wrong on the device side. If you do get the hardware wizard, then you need to choose the manual install option and choose the directory containing the driver files.
You might have better luck debugging this with a Linux machine, since no driver is necessary and you can get useful diagnostics from dmesg and lsusb.
Mike
c_c
I have android running on my freerunner - and its amazing to see so
much progress so soon. I do have the following observations :-
a) The touch screen seems a little jittery. I need to press and
hold an icon to make it launch.
b) If I let the phone suspend - and call it - i can hear the phone
ringing. But the phone does not respond and can't come out of suspend.
c) I can receive SMS's.
d) I can make calls. The volume is a little low - but is usable for
me. There seems to be a little echo - but less than that on
Qtextended / OM 2008.9. Maybe because the volume is lower.
e) I received one call while charging - don't know what I pressed /
did. Other than that (and I have not been able to replicate this) I
can't receive calls. Though the phone rings - but I can't pick up the
call.
f) If I press the power button for ~ 2sec, the screen blanks. I can
recover from that if I press the aux button fast enough. But once the
screen has blanked for > 2sec or so - the phone needs to be rebooted.
g) The font sizes and icons seems a little small for the freerunner.
Probably need to be scaled a bit ~ 125%.
h) There are people who have ported android onto other devices and
written basic keyboards. Eg http://it029000.massey.ac.nz/vogue/. Maybe
they could help.
i) Suspend does not resume - that one has been said many times
now :-)
Don't mean to nitpick - just some things off the top of my head that
need work. Am going to try the kernel from the Koolu site to see if
that is any better.
Thanks and Kodos for a great job!
Sean McNeil
crashed. This could be because of the SDHC support I mentioned was never
tested by me with that kernel, or some other issue with hardware or the
flashed kernel image. You might want to try the sdcard that was provided
with the phone and to check your kernel image size then re-flash.
Sean McNeil
> On Wed, Nov 5, 2008 at 5:53 AM, cedric.berger
> <cedric....@gmail.com <mailto:cedric....@gmail.com>> wrote:
>
>
>
> On 5 nov, 10:32, Brian Swetland <swetl...@google.com
> > [Sean McNeil <seanmcne...@gmail.com <mailto:seanmcne...@gmail.com>>]
> >
> > > I've put information as well as a link to adb on my wiki page at
> > > Openmoko. What you want to do is attach the phone via. usb to
> your host
> > > machine with ethernet-over-usb setup. The phone will come up
> as the
> > > normal "neo" hostname (aka 192.168.0.202
> >
> > > ADBHOST=neo ./adb logcat
> >
> > If you want to get adb running directly, you could give the adb
> gadget
> > driver a shot:
> >
> >
> http://android.git.kernel.org/?p=kernel/common.git;a=commit;h=fc20ff5...
> >
> > This should allow adb to operate over the usb transport directly
> (as we
> > do on dream/g1) instead of tunneled over ethernet.
>
> Is it possible to connect via adb from Windows ?
> I tried to install adb usb driver from android SDK. I edited the .inf
> (I do not really know what I do there...) so that device ID is
> recognized as Neo Freerunner, so it was "correctly" installed in XP,
> and it detects android when plugged in.
> But adb does not detect device...
>
>
> The Windows driver should work as long as you edit the VID and PID in
> the .inf file to match the values in android_adb.c:
>
> #define DRIVER_VENDOR_ID 0x18D1
> #define DRIVER_PRODUCT_ID 0x0001
>
> If you have a USB inspecting utility like USBView or lsusb, you could
> check to make sure your device appears on the USB bus with these
> values for the VID and PID. It should have a single interface with
> two bulk endpoints. Note that the android_adb driver differs from how
> the G1 appears on the USB bus.
The Android ADB USB gadget isn't in this kernel. I use the
ethernet-over-USB gadget as I debug by chrooting Android from a minimal
NAND filesystem.
Sean McNeil
Yes, it has been reported by someone else as well. A logcat dump would
be very useful in tracking this problem down.
> Cheers,
> Nik
> >
>
Cédric Berger
>
> Mike Lockwood wrote:
>> On Wed, Nov 5, 2008 at 5:53 AM, cedric.berger
>> The Windows driver should work as long as you edit the VID and PID in
>> the .inf file to match the values in android_adb.c:
>>
>> #define DRIVER_VENDOR_ID 0x18D1
>> #define DRIVER_PRODUCT_ID 0x0001
>>
>> If you have a USB inspecting utility like USBView or lsusb, you could
>> check to make sure your device appears on the USB bus with these
>> values for the VID and PID. It should have a single interface with
>> two bulk endpoints. Note that the android_adb driver differs from how
>> the G1 appears on the USB bus.
>
> The Android ADB USB gadget isn't in this kernel. I use the
> ethernet-over-USB gadget as I debug by chrooting Android from a minimal
> NAND filesystem.
>
Ok so that is why it does not work.
I indeed changed the
.inf with the VID / PID I picked in windows manager when neo was
plugged : I added
; For XP and later
[Google.NTx86]
; Neo
%USB\VID_1457&PID_5117.DeviceDescRelease%=androidusb.Dev, USB\VID_1457&PID_5117
%USB\VID_1457&PID_5117&Rev_0212.DeviceDescRelease%=androidusb.Dev,
USB\VID_1457&PID_5117&Rev_0212
and
[Strings]
USB\VID_1457&PID_5117.DeviceDescRelease="Neo Android"
USB\VID_1457&PID_5117&Rev_0212.DeviceDescRelease="Neo Freerunner Android"
Then Windows recognizes the device.
adb does not work but at least, now battery is correctly charged when
plugged in PC !
And is it possible to get internet connectivity through USB for android ?
Does it work under linux with this ethernet-over-USB gadget ?
ding...@163.com
I was confused about the android GUI init process.
I don't find some tip on /init.rc about the GUI startup.
I don't know about java. How the java startup the GUI.
and I also want know how to execute the java function from shell.
Peter Neubauer
after examining the SDHC card, it seems the kernel is supporting it,
since I get Android related data in there (dalvik-cache, app, misc
etc). However, is there any log or other output so I could look at to
know what is causing the crash, if it is not the SD card?
/peter
http://www.oredev.se - Be there or be gone.
GTalk: neubauer.peter
Skype peter.neubauer
ICQ 18762544
Phone +46704 106975
LinkedIn http://www.linkedin.com/in/neubauer
Twitter http://twitter.com/peterneubauer
http://www.neo4j.org - New Energy for Data - the Graph Database.
http://www.ops4j.org - New Energy for OSS Communities - Open
Participation Software.
http://www.qi4j.org - New Energy for Java - Domain Driven Development.
Anson
you should check init.rc again, and the code in framework/base/cmds/app_process and service manager , grep the main() function under this directory , you will seen how UI start , UI is start by surfaceFlinger.
Peter Neubauer
investigating further, there might be some problem with my previous
Debian-install leaving strange boot environment settings behind? Can I
from the bootloader prompt start the android kernel in order to get
some output?
In the bootloader, getenv is giving me the following, there might be
some errors regarding the bootargs_base?
Thanks for any hints!
/peter
boot_menu_timeout=300
bootargs_base=rootfstype=jffs2 root=/dev/mtdblock6
console=ttySAC2,115200 console=tty0 loglevel=8 regular_boot
bootcmd=setenv bootargs ${bootargs_base} ${mtdparts} rootfstype=ext2
root=/dev/mmcblk0p2 rootdelay=5; mmcinit; ext2load mmc 1 0x32000000
${sd_image_name}; bootm 0x32000000
bootdelay=1
menu_1=Boot from Flash: setenv bootargs ${bootargs_base} ${mtdparts};
nand read.e 0x32000000 kernel 0x200000; bootm 0x32000000
menu_2=Reboot: reset
menu_3=Power off: neo1973 power-off
menu_4=Set console to USB: setenv stdin usbtty; setenv stdout usbtty;
setenv stderr usbtty
menu_5=Set console to serial: setenv stdin serial; setenv stdout
serial; setenv stderr serial
mtdids=nor0=physmap-flash,nand0=neo1973-nand
mtdparts=mtdparts=physmap-flash:-(nor);neo1973-nand:0x00040000(u-boot),0x00040000(u-boot_env),0x00800000(kernel),0x000a0000(splash),0x00040000(factory),0x0f6a0000(rootfs)
quiet=1
sd_image_name=uImage.bin
splashimage=nand read.e 0x32000000 splash 0x5000; unzip 0x32000000
0x8800000 0x96000
usbtty=cdc_acm
partition=nor0,0
mtddevnum=0
mtddevname=nor
pcb_rev=0x305
pcf50633_int1=0x88
pcf50633_int2=0x02
stdin=usbtty
stdout=usbtty
stderr=usbtty
Environment size: 1172/262140 bytes
http://www.oredev.se - Be there or be gone.
GTalk: neubauer.peter
Skype peter.neubauer
ICQ 18762544
Phone +46704 106975
LinkedIn http://www.linkedin.com/in/neubauer
Twitter http://twitter.com/peterneubauer
http://www.neo4j.org - New Energy for Data - the Graph Database.
http://www.ops4j.org - New Energy for OSS Communities - Open
Participation Software.
http://www.qi4j.org - New Energy for Java - Domain Driven Development.
AndersHedberg
but the sudden white splash and then the black screen right after
having unpacked the kernel. I've followed every single instruction I
could find it seems. My orginal uSD card has a 8Mb VFAT32 (empty, even
though gparted left an .formatted file (manually removed)) filesystem
+ 300+Mb Ext3 filesystem(empty) (after booting the ext3 partition do
contain some extra folders) (I've used the images on the net this day
(17th of november + lates nand boot flash...).
Any suggestions? Does it work with 3G sim cards? European 1800 GSM
verisons? Why does it boot with a black screen... impossible to
debug...
Best regards
Anders Hedberg
Sean McNeil
Packet Video (there are license issues with the codecs). I've now
deleted this bad rootfs and will place a new one up when I have it fixed.
One way to debug is to look at the logcat output as outlined on the wiki
page:
ADBHOST=neo ./adb logcat
Sean
cedric.berger
> As of 12 hours ago, I broke the rootfs for Freerunner when I disabled
> Packet Video (there are license issues with the codecs). I've now
> deleted this bad rootfs and will place a new one up when I have it fixed.
>
> One way to debug is to look at the logcat output as outlined on the wiki
> page:
>
> ADBHOST=neo ./adb logcat
>
from previous image... (/system/lib/libmedia_jndi.so, ...)
I noticed there was also some corrections already made in this new
image (such as access to GPS).
Talking about licensing issues,
Sean's image does not include :
/system/framework/com.google.android.maps.jar
/system/framework/com.google.android.gtalk.jar
/system/app/Maps.apk
Are they published somewhere ? (even as closed source)
Since I could not find them elsewhere, I extracted them from the
Emulator, and installed it on my Neo.
Are we allowed to do that ?
(and also same question about these video codecs... are they available
somewhere ?)
AndersHedberg
some promises to show Android on the Freerunner this week. Please use
http://www.yousendit.com/ or similar service since it's something like
30 Megs. (I'll send a pouch and headset to the first person who do
this with something I can get up and running with) my email adress is
anders dot hedberg at factorycode dot com
/Anders Hedberg
AndersHedberg
will send a package only to the first one. I managed to get Android up
and running. (unfortunately it started in an endless screensaver
coming, going, coming, going thing, but at least it is working and
I'll have something to show. Thanks again! /Anders H.
On Nov 20, 7:38 am, AndersHedberg <anders.hedb...@factorycode.com>
wrote:
Peter Neubauer
Congrats on that one! If you want, you can bring it to Öredev tomorrow
and show it off to Mike Jennings from the Android team, I will get you
in I promise!
/Peter
http://www.oredev.se - Be there or be gone.
GTalk: neubauer.peter
Skype peter.neubauer
ICQ 18762544
Phone +46704 106975
LinkedIn http://www.linkedin.com/in/neubauer
Twitter http://twitter.com/peterneubauer
http://www.neo4j.org - New Energy for Data - the Graph Database.
http://www.ops4j.org - New Energy for OSS Communities - Open
Participation Software.
http://www.qi4j.org - New Energy for Java - Domain Driven Development.
AndersHedberg
On 6 Nov, 09:51, "Cédric Berger" <cedric.berge...@gmail.com> wrote:
>
> > Mike Lockwood wrote:
> >> On Wed, Nov 5, 2008 at 5:53 AM, cedric.berger
> >> The Windows driver should work as long as you edit the VID and PID in
> >> the .inf file to match the values in android_adb.c:
>
> >> #define DRIVER_VENDOR_ID 0x18D1
> >> #define DRIVER_PRODUCT_ID 0x0001
> .inf with the VID / PID I picked in windows manager when neo was
> plugged : I added
>
> ; For XP and later
> [Google.NTx86]
>
> ; Neo
> %USB\VID_1457&PID_5117.DeviceDescRelease%=androidusb.Dev, USB\VID_1457&PID_5117
> %USB\VID_1457&PID_5117&Rev_0212.DeviceDescRelease%=androidusb.Dev,
> USB\VID_1457&PID_5117&Rev_0212
>
windows to use a VID of 1457 to recognize the USB device, the
Freerunner is apparently telling the windows PC it is a VID 1457 USB
device. If the 0x18D1 is hardcoded into adb.exe someone need to
recompile it with the Freerunners VID (or change the VID used by the
Freerunner, but that is unwise as VIDs should be unique...). To
summarize - the lines
#define DRIVER_VENDOR_ID 0x18D1
#define DRIVER_PRODUCT_ID 0x0001
to let adb work with the Freerunner under windows.
Best regards,
Anders Hedberg
AndersHedberg
in several locations (in both partitions, don't know which place was
correct, but it do show up with all other applications), rebooted and
now I have a virtual keyboard on the freerunner! - see
http://groups.google.com/group/android-developers/browse_thread/thread/c743b2b8537be560/3443ce3b4d63b5a8?show_docid=3443ce3b4d63b5a8
On 21 Nov, 01:31, AndersHedberg <anders.hedb...@factorycode.com>
wrote:
AndersHedberg
file there and you have a virtual keyboard! (running debian on vmware
on my windows xp machine, I can mount the SD card, but not connect
directly to the Freerunner)
On 21 Nov, 02:18, AndersHedberg <anders.hedb...@factorycode.com>
wrote:
> in several locations (in both partitions, don't know which place was
> correct, but it do show up with all other applications), rebooted and
Peter Neubauer
this was actually a near-hardware issue. Now, we were able to show off
Sean's and the others great work to Mike Jennings from the Android
team right now mentioning the FreeRunner port it in his talk on
Android @Oredev!
Keep it up!
/peter
http://www.oredev.se - Be there or be gone.
GTalk: neubauer.peter
Skype peter.neubauer
ICQ 18762544
Phone +46704 106975
LinkedIn http://www.linkedin.com/in/neubauer
Twitter http://twitter.com/peterneubauer
http://www.neo4j.org - New Energy for Data - the Graph Database.
http://www.ops4j.org - New Energy for OSS Communities - Open
Participation Software.
http://www.qi4j.org - New Energy for Java - Domain Driven Development.
Peter Neubauer
/peter
http://www.oredev.se - Be there or be gone.
GTalk: neubauer.peter
Skype peter.neubauer
ICQ 18762544
Phone +46704 106975
LinkedIn http://www.linkedin.com/in/neubauer
Twitter http://twitter.com/peterneubauer
http://www.neo4j.org - New Energy for Data - the Graph Database.
http://www.ops4j.org - New Energy for OSS Communities - Open
Participation Software.
http://www.qi4j.org - New Energy for Java - Domain Driven Development.