About make installer_img command error for x86

143 views
Skip to first unread message

Skuld

unread,
Apr 9, 2009, 6:50:41 AM4/9/09
to android-porting
Hi :

I download new source code on 2009.04.07 and I add three project (alsa-
lib,alsa_sound,eeepc) in local_manifest.xml.
Than I execute $TARGET_ARCH=x86 TARGET_PRODUCT=eee_701
DISABLE_DEXPREOPT=true make installer_img.

But it happen following error message.How can I fix this?

===================Error=========================================

Copy: out/target/product/eee_701/installer/root/system/xbin/strace
Copy: out/target/product/eee_701/installer/root/system/bin/netcfg
Copying installer base files...
Copy: out/target/product/eee_701/obj/lib/libdiskconfig.so -> out/
target/product/eee_701/installer/root/system/lib/libdiskconfig.so
Copy: ->
dirname: missing operand
Try `dirname --help' for more information.
mkdir: missing operand
Try `mkdir --help' for more information.
cp: missing file operand
Try `cp --help' for more information.
Copy: ->
dirname: missing operand
Try `dirname --help' for more information.
mkdir: missing operand
Try `mkdir --help' for more information.
cp: missing file operand
Try `cp --help' for more information.
Copy: ->
dirname: missing operand
Try `dirname --help' for more information.
mkdir: missing operand
Try `mkdir --help' for more information.
cp: missing file operand
Try `cp --help' for more information.
Copy: ->
dirname: missing operand
Try `dirname --help' for more information.
mkdir: missing operand
Try `mkdir --help' for more information.
cp: missing file operand
Try `cp --help' for more information.
Copy: ->
dirname: missing operand
Try `dirname --help' for more information.
mkdir: missing operand
Try `mkdir --help' for more information.
cp: missing file operand
Try `cp --help' for more information.
Copy: ->
dirname: missing operand
Try `dirname --help' for more information.
mkdir: missing operand
Try `mkdir --help' for more information.
cp: missing file operand
Try `cp --help' for more information.
Copy: ->
dirname: missing operand
Try `dirname --help' for more information.
mkdir: missing operand
Try `mkdir --help' for more information.
cp: missing file operand
Try `cp --help' for more information.
Copy: ->
dirname: missing operand
Try `dirname --help' for more information.
mkdir: missing operand
Try `mkdir --help' for more information.
cp: missing file operand
Try `cp --help' for more information.
Copy: ->
dirname: missing operand
Try `dirname --help' for more information.
mkdir: missing operand
Try `mkdir --help' for more information.
cp: missing file operand
Try `cp --help' for more information.
Copy: ->
dirname: missing operand
Try `dirname --help' for more information.
mkdir: missing operand
Try `mkdir --help' for more information.
cp: missing file operand
Try `cp --help' for more information.
Copy: ->
dirname: missing operand
Try `dirname --help' for more information.
mkdir: missing operand
Try `mkdir --help' for more information.
cp: missing file operand
Try `cp --help' for more information.
make: *** [out/target/product/eee_701/installer/ramdisk-installer.img]
Error 1


Luca Belluccini

unread,
Apr 9, 2009, 9:29:18 AM4/9/09
to android-porting
Caused by build/tools/findleaves.sh @ line 86

# Print out all files that match, as long as the path isn't explicitly
# pruned. This will print out extraneous results from directories
whose
# parents have a match. These are filtered out by the awk script
below.
find -L "${@:1:$nargs-1}" $findargs -type f -name "$filename" -print
2>/dev/null |

Luca Belluccini

unread,
Apr 9, 2009, 9:35:57 AM4/9/09
to android-porting
No sorry...
They are the 11 files generated by
external/e2fsprogs/Android.mk
Include them for x86 arch - they're necessary for installer. Place an
ifeq or something similar. Or just uncomment.

choo75

unread,
Apr 9, 2009, 12:54:37 PM4/9/09
to android-porting
hi,

Maybe you can also try following workaround I've read inanother thread
(info from Chen)

...you can modify under
external/e2fsprogs/Android.mk:
--- a/Android.mk
+++ b/Android.mk
@@ -1,3 +1,3 @@
ifneq ($(TARGET_SIMULATOR),true)
-# include $(call all-subdir-makefiles)
+ include $(call all-subdir-makefiles)
endif

just add the include....i was successful in building afterwars.

regards,
choo

Yi Sun

unread,
Apr 9, 2009, 1:01:21 PM4/9/09
to android...@googlegroups.com
Guys,
I'm verifying my final patch for all the issues we are having on x86
recently. I will send it out as soon as I finished test.
Yi

Luca Belluccini

unread,
Apr 9, 2009, 1:29:55 PM4/9/09
to android-porting
I posted in "Instructions for last codebase" some other files to be
patched/conditioned for building for eee_701 target.

Yi Sun

unread,
Apr 9, 2009, 2:10:33 PM4/9/09
to android...@googlegroups.com
Yes, I saw that and I may have missed something you mentioned
(especially for the WIFI part). I will send out the first run of the
patch and then you can tell which part I have missed and I can add them
in.
Yi

Luca Belluccini

unread,
Apr 9, 2009, 2:31:33 PM4/9/09
to android-porting
Need someone to help me to bring up wifi...
I have a weird SIGIO... error...
Contact me in some way.

Yi Sun

unread,
Apr 9, 2009, 2:49:45 PM4/9/09
to android...@googlegroups.com
Could you post the steps so that we can reproduce the problem you are
having?

Luca Belluccini

unread,
Apr 9, 2009, 7:26:56 PM4/9/09
to android-porting
First of all... wpa_supplicant.conf was modified at ctrl_interface
line:

ctrl_interface=DIR=/data/system/wpa_supplicant/ GROUP=system

Is it typed correctly?

Second: in init.rc file I'm giving system system as owner of:

- /sys/class/net/wlan0
- /etc/wifi/wpa_supplicant.conf
- /data/misc/wifi/wpa_supplicant.conf
- /data/misc/wifi/socket/*

Then... I try to launch, using console at ALT+F1:
/system/wpa_supplicant -Dwext -iwlan0 -c/data/misc/wifi/
wpa_supplicant.conf

ioctl[SIOCSIWAUTH]: Operation not supported on transport endpoint.
WEXT auth param 4 value 0x0 - CTRL-EVENT-STATE-CHANGE id=-1 state=2
CTRL-EVENT-SCAN-RESULTS Ready
CTRL-EVENT-STATE-CHANGE id=-1 state=1

Then the console hangs (the keyboard doesn't allow to CTRL+C... Anyone
knows how to enable it?)... But Android is working...

I'm on a EeePc 701, atk5 as built-in in kernel 2.6.27...
It hangs after some message lines

Yi Sun

unread,
Apr 10, 2009, 1:51:55 AM4/10/09
to android...@googlegroups.com
Ok, I have done with all the patches I have collected from this list for
building x86. I tested them with a fresh check out of cupcake on a
Virtualbox and it is ok. I will start to work on rebuild the drivers and
kernel for the real EeePC device over the weekend and hope I could get
it done before next Monday. After that I will post the binaries and try
to check in the code.

I have posted the patch set in
http://code.google.com/p/patch-hosting-for-android-x86-support/
I also posted a wiki page for how to build a boot-able image with latest
cupcake tree. It is base on the instructions posted by Lim and Chen Yang
in the list. Please let me know if I missed anything and made anything
wrong.

I hope that we can have a central place to capture all the known issues
and solutions/patches so that people can feel easier on their life. The
mailing list should be used to discuss the new issues not to repeat the
same problem again and again.
Because I can only do this in my private time. So bare with me, if I
missed something or slow.If you want to join this project, please let me
know, I can always add you to member list.


Credit:
* Dima Zavin<di...@android.com>, the original guy ported the Android to
EeePC 701
* Lim <lee...@gmail.com>, Wrote original "HOWTO Build Android-X86 Full
Source" and "How to use x86 android platform on virtualbox "
* Chen Yang<sunse...@gmail.com> for the instructions on how to make
boot-able image and how to build Android for VMware/VirtualBox, and
investigations on almost all of the build breaks and run time
issues
* Jackie Wu<jack...@intel.com> for the fix in alarm.c
* Androidphan<niels....@tass.nl> for the investigation on the path
issue in the battery service jni code
* Luca Belluccini<lucabel...@gmail.com> for the first patch to fix
the path issue in the battery service jni cod and other information

Please let me know if I missed someone or something.
Yi
On Thu, 2009-04-09 at 11:31 -0700, Luca Belluccini wrote:

Yi Sun

unread,
Apr 10, 2009, 2:18:59 AM4/10/09
to android...@googlegroups.com
FYI-- The powermanagement service is still not working correctly on
VirtualBox. In certain cases, it tells me that the screen is off.
Yi

Skuld

unread,
Apr 10, 2009, 6:12:54 AM4/10/09
to android-porting
Dear Yi:

Thank you for your e2fsprogs.patch.
It's work and produce installer_img.
I install it on x86 PC and boot it, but it can't boot android system
(show android logo).
I use console mode and check all process.
I discover zygote service don't run.

How to fix it?
My kernel can work on android in 2009/03/24
> > * Dima Zavin<d...@android.com>, the original guy ported the Android to
> > EeePC 701
> > * Lim <lee...@gmail.com>, Wrote original "HOWTO Build Android-X86 Full
> > Source" and "How to use x86 android platform on virtualbox "
> > * Chen Yang<sunsety...@gmail.com> for the instructions on how to make
> > boot-able image and how to build Android for VMware/VirtualBox, and
> > investigations on almost all of the build breaks and run time
> > issues
> > * Jackie Wu<jackie...@intel.com> for the fix in alarm.c
> > * Androidphan<niels.kee...@tass.nl> for the investigation on the path
> > issue in the battery service jni code
> > * Luca Belluccini<lucabellucc...@gmail.com> for the first patch to fix

Yi Sun

unread,
Apr 10, 2009, 11:48:55 AM4/10/09
to android...@googlegroups.com
What kind of x86 PC are you using? The current building steps posted on
the web site is only verified for VirtualBOX.
Also, you need to use all the patches otherwise, you will hit on the
per-load unsupport class issue, battery file path issue and the crash
caused by wrong init.rc.
Please also send the logcat output if you can.
Thanks
Yi

Skuld

unread,
Apr 14, 2009, 10:51:34 PM4/14/09
to android-porting
Dear Yi :

I get logcat message.It look like some code not found for preloading.
If I patch framework.patch, there aren't error and it can work.
Why it have preloading issue ?


/dalvikvm( 2077): Added shared lib /system/lib/libwebcore.so 0x0
D/dalvikvm( 2077): GC freed 459 objects / 26328 bytes in 9ms
D/dalvikvm( 2077): GC freed 475 objects / 38936 bytes in 9ms
E/Zygote ( 2077): Class not found for preloading:
com.android.internal.policy.impl.PhoneLayoutInflater
E/Zygote ( 2077): Class not found for preloading:
com.android.internal.policy.impl.PhoneWindow
E/Zygote ( 2077): Class not found for preloading:
com.android.internal.policy.impl.PhoneWindow$1
E/Zygote ( 2077): Class not found for preloading:
com.android.internal.policy.impl.PhoneWindow$ContextMenuCallback
E/Zygote ( 2077): Class not found for preloading:
com.android.internal.policy.impl.PhoneWindow$DecorView
E/Zygote ( 2077): Class not found for preloading:
com.android.internal.policy.impl.PhoneWindow$PanelFeatureState
E/Zygote ( 2077): Class not found for preloading:
com.android.internal.policy.impl.PhoneWindow$PanelFeatureState
$SavedState
E/Zygote ( 2077): Class not found for preloading:
com.android.internal.policy.impl.PhoneWindow$PanelFeatureState
$SavedState$1
D/dalvikvm( 2077): GC freed 1027 objects / 77360 bytes in 9ms
D/dalvikvm( 2077): GC freed 292 objects / 22136 bytes in 9ms
D/dalvikvm( 2077): GC freed 314 objects / 18936 bytes in 9ms
D/dalvikvm( 2077): GC freed 419 objects / 20736 bytes in 10ms
D/dalvikvm( 2077): GC freed 14221 objects / 476944 bytes in 13ms
D/dalvikvm( 2077): GC freed 11164 objects / 455488 bytes in 14ms
D/dalvikvm( 2077): GC freed 7567 objects / 316288 bytes in 13ms
D/dalvikvm( 2077): GC freed 747 objects / 42024 bytes in 12ms
D/dalvikvm( 2077): GC freed 603 objects / 31128 bytes in 12ms
D/dalvikvm( 2077): GC freed 442 objects / 27712 bytes in 12ms
E/Zygote ( 2077): Zygote died with exception
E/Zygote ( 2077): java.lang.IllegalStateException: Missing class(es)
for preloading, update preloaded-classes
[com.android.internal.policy.impl.PhoneLayoutInflater
com.android.internal.policy.impl.PhoneWindow
com.android.internal.policy.impl.PhoneWindow$1
com.android.internal.policy.impl.PhoneWindow$ContextMenuCallback
com.android.internal.policy.impl.PhoneWindow$DecorView
com.android.internal.policy.impl.PhoneWindow$PanelFeatureState
com.android.internal.policy.impl.PhoneWindow$PanelFeatureState
$SavedState com.android.internal.policy.impl.PhoneWindow
$PanelFeatureState$SavedState$1]
E/Zygote ( 2077): at
com.android.internal.os.ZygoteInit.preloadClasses(ZygoteInit.java:299)
E/Zygote ( 2077): at com.android.internal.os.ZygoteInit.main
(ZygoteInit.java:512)
E/Zygote ( 2077): at dalvik.system.NativeStart.main(Native Method)
D/AndroidRuntime( 2077): Shutting down VM
W/dalvikvm( 2077): threadid=3: thread exiting with uncaught exception
(group=0xa8007df8)
D/dalvikvm( 2077): DestroyJavaVM waiting for non-daemon threads to
exit
D/dalvikvm( 2077): DestroyJavaVM shutting VM down
D/dalvikvm( 2077): VM cleaning up
D/dalvikvm( 2077): LinearAlloc 0x0 used 2070036 of 4194304 (49%)
D/skia ( 2077): purging 6K from font cache [1 entries]
D/AndroidRuntime( 2078):
D/AndroidRuntime( 2078): >>>>>>>>>>>>>> AndroidRuntime START
<<<<<<<<<<<<<<
D/AndroidRuntime( 2078): CheckJNI is ON
D/AndroidRuntime( 2078): --- registering native functions ---
I/Zygote ( 2078): Preloading classes...
D/dalvikvm( 2078): GC freed 763 objects / 40584 bytes in 0ms
D/dalvikvm( 2078): GC freed 290 objects / 17288 bytes in 0ms
D/dalvikvm( 2078): GC freed 234 objects / 13968 bytes in 1ms
D/dalvikvm( 2078): GC freed 3834 objects / 182480 bytes in 3ms
D/dalvikvm( 2078): GC freed 12196 objects / 423408 bytes in 5ms
D/dalvikvm( 2078): GC freed 11296 objects / 455376 bytes in 6ms
D/dalvikvm( 2078): GC freed 9417 objects / 384240 bytes in 6ms
D/dalvikvm( 2078): GC freed 15439 objects / 527680 bytes in 8ms
D/dalvikvm( 2078): GC freed 10650 objects / 438760 bytes in 8ms
D/dalvikvm( 2078): GC freed 5527 objects / 225616 bytes in 7ms
D/dalvikvm( 2078): GC freed 348 objects / 36544 bytes in 6ms
D/dalvikvm( 2078): GC freed 778 objects / 46624 bytes in 9ms
D/dalvikvm( 2078): GC freed 321 objects / 36128 bytes in 9ms
D/dalvikvm( 2078): GC freed 490 objects / 29456 bytes in 9ms
D/dalvikvm( 2078): Trying to load lib /system/lib/libwebcore.so 0x0
D/dalvikvm( 2078): Added shared lib /system/lib/libwebcore.so 0x0
D/dalvikvm( 2078): GC freed 459 objects / 26328 bytes in 9ms
D/dalvikvm( 2078): GC freed 475 objects / 38936 bytes in 9ms
E/Zygote ( 2078): Class not found for preloading:
com.android.internal.policy.impl.PhoneLayoutInflater
E/Zygote ( 2078): Class not found for preloading:
com.android.internal.policy.impl.PhoneWindow
E/Zygote ( 2078): Class not found for preloading:
com.android.internal.policy.impl.PhoneWindow$1
E/Zygote ( 2078): Class not found for preloading:
com.android.internal.policy.impl.PhoneWindow$ContextMenuCallback
E/Zygote ( 2078): Class not found for preloading:
com.android.internal.policy.impl.PhoneWindow$DecorView
E/Zygote ( 2078): Class not found for preloading:
com.android.internal.policy.impl.PhoneWindow$PanelFeatureState
E/Zygote ( 2078): Class not found for preloading:
com.android.internal.policy.impl.PhoneWindow$PanelFeatureState
$SavedState
E/Zygote ( 2078): Class not found for preloading:
com.android.internal.policy.impl.PhoneWindow$PanelFeatureState
$SavedState$1

Yi Sun

unread,
Apr 14, 2009, 10:55:18 PM4/14/09
to android...@googlegroups.com
because the x86 does not have these classes.

Chen Yang

unread,
Apr 15, 2009, 9:36:30 AM4/15/09
to android...@googlegroups.com
yeah, you may need to remove those classes in frameworks/base/preloaded-classes

--
Chen

Yi Sun

unread,
Apr 15, 2009, 2:16:46 PM4/15/09
to android...@googlegroups.com
I could not find out a decent way to handle this. Current patch(the
quick hack made by me) changed preload-classes to preload-classes.arm
and preload-classes.x86 and has makefile to pick up the right one during
the compile. As pointed out by Dianne Hackbom, whether these classes are
supported has nothing to do with the CPU. I'm thinking about to added
some platform based infrastructures to support different platform/arch
inside the same tree. But I did not get time to work on it yet and don't
have a good idea about it either. There are other platform dependent
things as well for example the path issue in the batteryservice (See my
another brain damage patch).
Now I open for suggestions. Please give your thoughts.
Yi

Luca Belluccini

unread,
Apr 15, 2009, 2:24:19 PM4/15/09
to android-porting
It is policy related... I think the "difference" is between
policy_phone and policy_mid.

On 15 Apr, 20:16, Yi Sun <beyo...@gmail.com> wrote:
> I could not find out a decent way to handle this. Current patch(the
> quick hack made by me) changed preload-classes to preload-classes.arm
> and preload-classes.x86 and has makefile to pick up the right one during
> the compile. As pointed out by Dianne Hackbom, whether these classes are
> supported has nothing to do with the CPU. I'm thinking about to added
> some platform based infrastructures to support different platform/arch
> inside the same tree. But I did not get time to work on it yet and don't
> have a good idea about it either. There are other platform dependent
> things as well for example the path issue in the batteryservice (See my
> another brain damage patch).  
> Now I open for suggestions. Please give your thoughts.
> Yi
>
> On Wed, 2009-04-15 at 21:36 +0800, Chen Yang wrote:
> > yeah, you may need to remove those classes in frameworks/base/preloaded-classes
>
> > --
> > Chen
>
> ...
>
> leggi tutto

Yi Sun

unread,
Apr 15, 2009, 2:46:55 PM4/15/09
to android...@googlegroups.com, android-porting
Well, it is ok to run android on a netbook without phone, right?



On Apr 15, 2009, at 11:24 AM, Luca Belluccini

Luca Belluccini

unread,
Apr 15, 2009, 7:22:47 PM4/15/09
to android-porting
I mean... Differences in preloaded classes are due to different
policies...

On Apr 15, 8:46 pm, Yi Sun <beyo...@gmail.com> wrote:
> Well, it is ok to run android on a netbook without phone, right?
>
> On Apr 15, 2009, at 11:24 AM, Luca Belluccini  
>
> ...
>
> read more »

Chen Yang

unread,
Apr 15, 2009, 10:36:32 PM4/15/09
to android...@googlegroups.com
I agree that this is not related to architecture.
I guess google may refactor some policy related implementations and
may make the preload-classes policy dependant. But don't know how long
it will take.
--
Chen
Reply all
Reply to author
Forward
0 new messages