1) Probably need to a get a linux kernel working on the device first
before worrying about a whole system such as android, there have been
attempts at that which you can read about online
2) Sources are at android.git.kernel.org
On Aug 19, 3:41 am, "bikcmp(creator and main programmer)"
wrote:
>
gcc is not installed in all its pieces or they are not in your path or
the proper variables have not been pointed to the proper versions by
the config process. Note that the build process likely requires both
a normal installation of gcc to produce executables for your
development machine, and a
Porting the VM itself is probably practical if you want it badly
enough... grab the files, type make, and start working on the errors
Porting the rest of android would likely be painful... there are
mechanisms added to the linux kernel which you would have to replicate
or substitute, on top of th
On Sep 27, 2:28 am, Jay wrote:
> I am looking into to Interfacing an Android Phone with a PIC
> Microcontoller and retrieve data via the USB port. How would I go
> about doing this? What API calls do I need to make to retrieve data
> from the PIC? And will it be easier if I were to do this via
What I can find in the way of specs makes it look a bit limited. But
even in the realm of squeezing a limited sd card based build in, your
questions are going to be a) can you execute arbitrary privileged code
such as the linux kernel you would have to get running on it, and b)
can you figure out
one of these Samsung phones and wanted to
> poke around with it :P
>
> On Feb 28, 9:40 pm, Chris Stratton wrote:
> > What I can find in the way of specs makes it look a bit limited. But
> > even in the realm of squeezing a limited sd card based build in, your
> > questi
The underlying file i/o c library calls are likely to be the common 32
bit signed versions unless someone specifically chose the 64 bit
versions.
On Mar 8, 1:13 pm, Hedge wrote:
> I am trying to play back 3GB videos from an SDHC card in OpenCORE
> v2.05 in Cupcake. The player crashes every time.
ould fail. As chris said below, there
> > > might be a break in i/o lib.
> >
> > > On Wed, Mar 10, 2010 at 2:52 AM, hedwin wrote:
> > > > FAT32 itself is limited to 2GB. If you need to handle files larger than
> > > > 2GB
> > > >
You will amost certainly need a phone you can root and install a
customized kernel on. The adp1, g1, magic and similar devices has a
low voltage (adapyer REQUIRED) serial port on the extended usb
connector using some of the headphone pins, there are plenty of
writeups with the details online. Don
First you should build that driver source for a desktop linux and make
sure it works with the peripheral. You may need to add vid/pid codes
to it, or it might be you have the wrong driver. Figure that out on
the desktop first.
Then you should look at the driver source and see where that no more
If you are working on a typical 1st-generation qualcom based phone
that already runs android (G1 etc) I think you want the kernel from
msm/ rather than common/ but same idea
Ashwin Bihari wrote:
> The Android Kernel is located at
> http://android.git.kernel.org/?p=kernel/common.git;a=summary, bu
This is neither a very practical project nor a good place for someone
unfamiliar with both the linux kernel and android to start. You need
to learn a lot more about working with the linux kernel in embedded
systems and also about using android on a system it already runs on,
such as its stock emul
As with any other embedded linux you want a console for kernel
messages and ideally an interactive shell under the gui. Serial ports
are still preferred for this (htc's phones have one with wierd
voltages on the extusb headphone lines) - less advantageous
possibilities are doing it over usb or eve
No, the logs indicate it did mount the filesystem, then failed later.
Might want to try a different hub / try that hub and storage device on
a desktop linux box while stress testing it somehow.
Srikant wrote:
> Check this from the logs:
> init: cannot find '/system/bin/mountd', disabling 'mountd'
Just stabbing in the dark here, but can you figure out what
usbmodeswitch actually causes the kernel to do, and then make your
modem driver module cause that itself?
Assuming of course that you can figure out how to make the modem
driver module claim the cdrom-mode version of the device, before so
Did you make, manually or in a script, a corresponding device file?
> adb shell ls -l /dev/binder
crw-rw-rw-1 root root 10, 44 Sep 23 08:51 /dev/binder
If you aren't getting something similar you will need to mknod it
On Sep 24, 3:42 am, siva ram krishna wrote:
> I am porting an
On Oct 27, 5:43 am, Vincent wrote:
> I am trying to read out data from an USB audio device (/dev/snd/dsp,
> 14:3) through native C function. Same C code works with Ubuntu 10.04
> but not with Android (execution stucks at read(...) which never
> returns), any tips about it?
It's probably blocking
On Oct 28, 2:45 pm, Semitsoglou-Tsiapos Dimitrios
wrote:
> Then, on the android device, the file needs to be moved to a directory where
> it
> can be exected from; /system/xbin for example:
> mv /mnt/sdcard/hello /system/xbin/hello
>
> Now do the chmod wizardry:
> chmod +x /system/xbin/hello
No
Figure out what parts of the program are suspect and instrument it
with debug output.
Lots of real work has been done on systems that don't have debuggers
available, simply by analytic thinking and adding instrumentation on
the things that can go wrong.
On Oct 19, 7:20 am, Junaid Rana wrote:
> H
You may want to try (ab)using the ndk's gcc compiler to build your
program against bionic rather than using codesourcery gcc. A web
search will turn up instructions for how to make an executable rather
than a library.
If you stick with the non-android gcc, first try building a simple
hello world
On Nov 23, 10:32 am, Shachar Shemesh wrote:
> I'm not sure this is the right forum to ask, but here goes. I want to
> write an APK that can be installed as a standard application, but can
> also be placed in the /system/app folder. More importantly, I want the
> program to behave slightly differe
On Nov 23, 5:33 am, Mikkel Christensen
wrote:
> If the LED is already exposed in sysfs you are lucky and it should be
> quite easy. Otherwise create a Kernel device driver that does the
> exposure of the LED to sysfs.
If it's already exposed in sysfs, you don't need to use native code.
You shoul
Can you post the definition of the interface you want to talk to?
Do you have any actual evidence that using this other interface will
be faster?
(Often the way to speed things up when the interface itself is slow is
to hand off larger units of work at a time)
On Nov 24, 2:10 pm, "jon.schell" w
Maybe you should fix vold?
On Nov 27, 7:22 pm, Carlos Balseiro wrote:
> I forgot to say, I'm mounting internal memory on init.rc because new
> vold is unable to mount it.
>
> On 27 nov, 03:14, Carlos Balseiro wrote:
>
> > Hi. Trying to port Froyo to this device. We got SD Card working ok,
> > an
Many of the android phones such as the G1, Mytouch, Hero (and probalby
Nexus 1) can support USB host mode if you root them and install a
recompiled kernel. The original droid was rumored to have this
capability in the shipping firmware. But that's only the start of the
capability you need - a lot
As embedded systems go, android is extremely resource (memory and CPU
speed) hungry, and also harder than other embedded linuxes to extend
to take advantage of unique hardware capabilities, so all in all, I'd
say this probably not a good project idea.
I think you'd be better off running android on
On Nov 28, 9:51 pm, Ajith Kamath wrote:
> I'm going through PTP now. My initial leaning of the protocol indicated
> there is no need of extra drivers.
> And that communication will be over USB transport media.
> Please provide your view on this.
Unless it already supports PTP (haven't heard of a
You might want to try diff-ing the relevant parts of the customized
Android 2.1 against the official version of 2.1 and then applying
something similar to those changes to the official 2.2
Of course this kind of approach doesn't work if there's been a change
in an underlying mechanism between the
No. ~48 Mb of ram is not going to be sufficient for a usable build of
Android.
First generation android devices like the G1 should be available used
at a reasonable price.
On Dec 8, 9:47 am, Kundan wrote:
> Hi all I am new to this so before involving I want to ask whether is
> it possible that
Indeed, that's how it's done. Development devices like the emulator
have ro.secure=0 in /default.prop which means the adb shell runs as
root.
(Development here meaning development of android itself; application
development does not require privileges and can be done on a consumer
device with ro.s
This is really a question about generic uboot (and probably by
extension linux) support for your board which you should take up with
the board vendor or author of the uboot port - you haven't yet gotten
that working to the point where you are asking an android question.
On Dec 9, 6:53 am, PREMRAJ
SDK/NDK applications do not run with sufficient permissions to access
the framebuffer as they are not in the graphics group.
It might be a fun project to modify the platform to add a new
permission for "FRAMEBUFFER" that would put the application in the
graphics group, though some would likely obj
Android phones typically have their own bootloader (or even chain of
them, there are two processors involved) which may be either fastboot
or something else.
Non-phone android devices might use busybox or whatever is customary
for linux on that platform - there's not really anything unusual about
Android phones typically have their own bootloader (or even chain of
them, there are two processors involved) which may be either fastboot
or something else.
Non-phone android devices might use uboot or whatever is customary for
linux on that platform - there's not really anything unusual about th
On Dec 30 2010, 12:19 pm, A Curtis wrote:
> This is basically what I did to access a GPIO from Java land.
>
> echo "128" >/sys/class/gpio/export
> echo "out" > /sys/devices/virtual/gpio/gpio128/direction
> chmod 0666 /sys/devices/virtual/gpio/gpio128/value
Just FYI, that's not accessing sysfs di
On Jan 2, 3:31 pm, A Curtis wrote:
> On second thought there is a difference,
>
> The problem isn't just a sequence of instructions but having
> sufficient permission to execute. The Java JVM does not run as root.
> So,you need something that runs as root to change permissions to
> enable Java acc
On Jan 2, 3:31 pm, A Curtis wrote:
> On second thought there is a difference,
>
> The problem isn't just a sequence of instructions but having
> sufficient permission to execute. The Java JVM does not run as root.
echo isn't setuid root either
> So,you need something that runs as root to change
On Feb 4, 6:58 am, keerthi wrote:
> hey guys im new to android
> im building a new device (tablet)
> i need to know how to make android run on my device.. please help
1) Figure out if your project is plausible, for example does android
or at least an embedded linux already run on something with s
On Mar 10, 11:07 pm, "Eric Wong (hdmp4.com)"
wrote:
> I am not a programmer but I believe many propertiary components, such
> as video codec drivers would not be included in Telechip's public
> source code.
> That means even if you suceed in booting your own linux kernel, and
> even port your own
Most any decent IDE should be able to be configured to invoke build
commands.
Large software projects intended to be easily maintained and
compatible with distributed development tend to avoid the lock-in of
an IDE-driven build process and instead provide solid build scripts,
which can be triggere
On Apr 28, 12:09 am, CW wrote:
> In windows, it has the concept of data segment and code segment in
> memory, an application can't access the code segment in the X86 CPU.
> Do you know if the android have the same feature support?
Not quite sure what you mean by can't access the code segment - ma
On May 2, 6:11 am, larlin wrote:
> I'm trying to sample low frequency signals on the headset microphone
> AD.
> The wolfson driver defines the 3 different filter options available on
> the chip, eg:
> static const char *adc_hpf_text[] = {"Hi-Fi","Voice 1","Voice 2",
> "Voice 3",};
I hadn't reali
For a site-specific change it seems you can add it to $HOME/.android/
adb_usb.ini
For a proper, lasting change you would edit it into core.git adb/
usb_vendors.c and rebuild.
And if you like spending a day figuring out why it stopped working
after an upgrade, there's always the "hex edit the bina
On Monday, May 9, 2011 4:07:16 AM UTC-4, Itai wrote:
> I wrote and built my own libgps.so in my directory on hardware/
> /gps
> Afterwards, I re-built the entire android project. The make completed
> successfully and libgps was placed in system/lib
>
> However, when I load the emulator, with :
On Friday, May 27, 2011 1:48:38 AM UTC-4, Gopi wrote:
>i m using devkit800 and i used to insert the sdcard once before
> the android boot animation comes...
What Dianne said still holds true for the external storage (partition?) of
the sdcard. It's not mounted immediately.
Your ser
On Friday, May 27, 2011 1:17:03 PM UTC-4, Azam Saleem wrote:
>
> I need to ask that I have a project of solar tracking system using
> microcontroller ... I was asking that could I control the hardware
> through android phone such as stepper motor.
There are various methods with various advantag
On Wednesday, June 1, 2011 7:46:47 AM UTC-4, ramya mohan wrote:
I have a scenario , (i.e) If i am using an android based device(need
> not be a phone) ,it is connected to server. I am using around (say
> playing an video / playing a game ) If a critical message(information
> - can be in any for
On Monday, May 30, 2011 2:03:32 AM UTC-4, Gopi wrote:
>
> Hi Chris,
> Is the path '/data' correct to log to the sdcard???
>
No. Well, it depends on what you mean by the sdcard.
If you mean the free-for-all user storage area which has been the
traditional role of sdcards on android dev
On Monday, August 8, 2011 2:10:16 PM UTC-4, Anil wrote:
>
> I work for an OEM and am trying to implement an idea - to log all
> inputs (keystrokes,
> text field input, browser tests including Javascript, plugging/
> unplugging SD card etc.) by the testers to have a set of trails that
> would be
On Monday, August 8, 2011 7:13:25 PM UTC-4, T0N1 wrote:
>
> 1. tried to compile with CodeSourcery and
>
Don't do that. It's rarely the right answer.
> 2. tried to compile with NDK and it resulted a executable file!!!
> It ran ok and has a EABI version 4. But when I try to connect with
On Wednesday, August 17, 2011 11:16:36 PM UTC-4, arif wrote:
>
> I am porting one application to android, there are few depending
> binaries and config files, is it possible to include/build binaries in
> apk ( using NDK or by other method ).
If you just mean data, sure - see the writeup on as
On Tuesday, August 23, 2011 8:12:19 AM UTC-4, Adrien wrote:
>
> I'd like to use updatedd in Android. Updatedd is a tool allowing to publish
> IP address changes to ovh, noip, dyndns, ...
>
> I managed to build it using the following:
>
> export PATH=/bin
>
> ./configure --host=arm-none-linux-gnuea
On Thursday, August 25, 2011 5:59:32 AM UTC-4, Shanth wrote:
>
> how to unlock the screen from adb shell command, without user having to tap
> and slide it on the screen... needed to automate some test cases, so
> required it.
>
How about disabling the screen lock?
--
unsubscribe: android-port
On Thursday, August 25, 2011 5:00:39 AM UTC-4, ramon wrote:
>
> Hey,
> I working on AT91SAM9M10 evaluation board currently running android
> 2.1.
> during kernel device development, I'm working with buildroot, for
> easier module manipulation.
> I finished writing my char device, and it's work
On Thursday, March 1, 2012 12:51:08 PM UTC-5, Upendra wrote:
>
> HI,
> I tried to give the shell commands from android java using exec and it
> works fine for ls and mkdir commands but it is not working for mount
> and chmod commands. Even I tried to run scripts from java but that is
> also not
You probably also want to give it special oom killer values to protect it.
On Friday, March 2, 2012 8:07:39 PM UTC-5, Tim wrote:
>
> I'm working with a client who needs to have a persistent server run on
> a device running Android. The device will be wall-powered, so battery
> life is not an i
On Thursday, March 29, 2012 10:02:20 PM UTC-4, NYL wrote:
>
> I would like to create an Android application that communicates with a
> PC Windows app using the USB interface. May I know if this is
> possible? Do I need to modify the Android OS, add in drivers or can
> this be done at the applica
On Monday, April 9, 2012 8:20:40 AM UTC-4, Vikas KM wrote:
>
> I want to execute a shell script in android
>
> *** SHELL SCRIPT
>
> if [ ! -d "/sys/devices/platform/ehci-omap.0/usb1/1-2/1-2.4" ]; then
> if [ -d "/sys/devices/platform/ehci-omap.0/usb1/1-2" ]; th
On Friday, April 20, 2012 3:57:16 AM UTC-4, Vikas KM wrote:
>
> Hi,
>
> Works with busybox..
>
By itself what you have posted accomplishes nothing, so the fact that it
runs is of little importance
Any other way of doing this without buybox..
>
As I said before, tell us what you are trying to ac
On Tuesday, April 17, 2012 10:57:11 AM UTC-4, Glenn Kasten wrote:
>
> One solution I've seen for this kind of problem is to have a separate
> 'factory' filesystem
> which is mounted read/write during factory provisioning, but then is
> mounted read-only
> during normal operation. This is also
Android apps - even system apps - don't run as root, so any actual root
ioctl's are probably being sent from code that is part of the platform
rather than part of any app, but perhaps triggered by system app with
special privileges in the eyes of the platform.
Your best bet for doing this in a
On Tuesday, April 24, 2012 8:46:39 PM UTC-4, HV wrote:
>
> Hi All,
>
> I'm trying to initiate a script as a service from init.rc (running ICS),
> but it doesn't seem to work. Here is what I have:
>
>
> service my_sh /system/bin/sh ./system/etc/my.sh
>
Why are you staring what looks
On Wednesday, April 25, 2012 2:59:01 PM UTC-4, HV wrote:
> With this, I do see the service kicking in. The next challenge is how do I
> get the script to work. Say if I want to cd /mnt/sdcard & this should show
> up in the shell prompt. Is that doable? In other words, it should work
> exactly
On Thursday, April 26, 2012 3:28:57 PM UTC-4, HV wrote:
>
> I wanted the commands that I'm running from within the script get
> translated on to the console. But, I guess that's not possible
>
That depends on what you mean by "translated" and "console" - neither term
matches very well with the p
On Tuesday, April 24, 2012 10:14:16 AM UTC-4, NARAYANA wrote:
>
> After successful installation of SDK on windows machine, and after
> launching SDK Manager.exe, throughs me an error.
>
No one can help you since you did not state what the error message is.
However, you _should not_ do so as pro
On Apr 29, 11:58 am, Rainius wrote:
> I try to build Email in eclipse as an individual application. However,
> Email depends on some resouce in the framework:
> - framework/opt/mailcommon/res
> - frameworks/ex/chips/res
>
> I've try to copy these two folders to Email's folder, and modify the
> And
On Jun 4, 4:31 pm, Ronnyek wrote:
> Edit the init.rc file which may be in your init ramdisk or just in the root
> of the boot partition. You can disable a service by adding a line beneath
> last configuration option for the service called disabled. This will
> temporarily disable the service. or r
67 matches
Mail list logo