Related
Hi everyone,
There has been a few questions on how to compile native android code (for exploits and such). Easy enough.
Go to http://source.android.com/download. You will need to be running Linux. Ubuntu is easiest. Follow the directions to get the source code for android downloaded and compiled.
Run this command
export PATH=/path/to/android/source/prebuilt/linux-x86/toolchain/arm-eabi-4.2.1/bin
Run arm-aebi-gcc or arm-aebi-g++ (depending on the language, c or c++) followed by
-o (OUTPUT) (INPUT)
So, for example, test.c would be:
arm-aebi-gcc -o test test.c
And test.cpp would be
arm-aebi-g++ -o test test.cpp
Just a note, this will make STATICALLY linked files. Meaning any headers will be included INSIDE the executable. Simply put, this means the files will be HUGE for large projects. There is a program, named agcc, which fixes this and can be found here:
http://plausible.org/andy/agcc
Put it in /bin by:
Code:
cd /bin
sudo wget http://plausible.org/andy/agcc
chmod 755 agcc
chmod +x agcc
Run agcc -o (OUTFILE) (INPUT) to compile. Be warned though, if a header is in the file that isn't in bionic (android's smaller libc) it won't compile.
Hope this helps!
+1
Awesome
............(stuipid mistake >>was<< here)..............
Thanks man...
now i can break out my bootable 50 meg linux disc and play around.
love that thing used to use it to crack windows passwords
should see the guys face when you crack his 20 char password in 5 mins without ever needing to use it.
well not really cracking but changing it. used to work at the pentagon. this one guys who used to be support for one dept. thought it would be funny to change all the admin passwords in his office. so when one of the pc's was beyond his repair. i showed up and he was like give me 10 min and i'll log you in. well 5 mins later i was fixing the machine while he was screwing off. boy was he pissed.
how big's gcc? cause i'll need to compile it for my linux.
rigamrts said:
how big's gcc? cause i'll need to compile it for my linux.
Click to expand...
Click to collapse
Massive. You're definitely better off using the prebuilt toolchain found in the Android (N|S)DK.
I would like to add something to this.
libc is essentially derived from the kernel. So, if you take agcc, and make changes to use the libc directory (I don't remember it atm) and NOT bionic, the app won't compile. Simple enough, libc is based of the kernel, so bionic is based of the android kernel. Things missing in bionic that are in libc WON"T work simply because certain kernel calls in glibc DON'T exist on the Android platform.
My exploit relied on the fact that I would be able to compile exploits using glibc, instead of using bionic. So it failed. Thats what I've figured out so far anyways.
zifnab06 said:
My exploit relied on the fact that I would be able to compile exploits using glibc, instead of using bionic. So it failed. Thats what I've figured out so far anyways.
Click to expand...
Click to collapse
Would you be willing to share the code for that exploit, even if it doesn't work? (sorry if you already have, I didn't see it anywhere)
I may be able to help.
Look up anything in our old thread, especially when we were talking about "sys/personality.h". The one I was working with exploited a hole that was patched (min_map_addr).
This blog post
honeypod.blogspot.com/2007/12/dynamically-linked-hello-world-for.html
(Sorry, my account isn't allowed to post links yet.)
gives a minimalist approach to using dynamically linked executables. (In particular, see steps #2 and #3 for the sources for hello.c and start.c) I gave it a try, and it seemed to work without agcc, e.g. with a makefile like the following (and with the arm-eabi- executables in the PATH of the user invoking the make) :
Code:
AR = arm-eabi-ar
AS = arm-eabi-as
CC = arm-eabi-gcc
CXX = arm-eabi-c++
LD = arm-eabi-ld
NDK_KIT = /opt/android/android-ndk-1.5_r1
PLATF_KIT = build/platforms/android-1.5
ARM_INC = $(NDK_KIT)/$(PLATF_KIT)/arch-arm/usr/include
ARM_LIB = $(NDK_KIT)/$(PLATF_KIT)/arch-arm/usr/lib
PLATF_INC = $(NDK_KIT)/$(PLATF_KIT)/common/include
OBJS = hello.o start.o
EXES = hello
hello: hello.o start.o
$(LD) \
--entry=_start \
--dynamic-linker /system/bin/linker -nostdlib \
-rpath /system/lib -rpath $(ARM_LIB) \
-L $(ARM_LIB) -lc -o hello hello.o start.o
hello.o: hello.c
$(CC) -I $(ARM_INC) -I $(PLATF_INC) -c hello.c
start.o: start.c
$(CC) -I $(ARM_INC) -I $(PLATF_INC) -c start.c
clean:
rm -f $(OBJS) $(EXES)
HTH
bftb0
Just curious, but I'm trying to get some native code that I've compiled to run on the Incredible. I've followed the instructions to download the the arm gcc, compiled my C code, and adb push'ed the executable over to /sdcard but I get a "permission denied" error when I try running it from my phone and adb shell. Does the phone have to be rooted in order to run native C compiled executables?
Thanks!
zebdor44 said:
Just curious, but I'm trying to get some native code that I've compiled to run on the Incredible. I've followed the instructions to download the the arm gcc, compiled my C code, and adb push'ed the executable over to /sdcard but I get a "permission denied" error when I try running it from my phone and adb shell. Does the phone have to be rooted in order to run native C compiled executables?
Thanks!
Click to expand...
Click to collapse
On an unrooted phone, push your code to /system/local or another place that you can write to and chmod it to be executable. By default the sdcard is mounted no execute. You will either need to add the directory you put it in to the path or execute it implicitly by specifying it is in the local directory.
for example ./myprogram
I hope that helps. I re-read it and it doesn't make much sense unless you have a firm grasp of the things that happen between the lines.
best of luck.
Thanks. Good stuff.
rigamrts said:
now i can break out my bootable 50 meg linux disc and play around.
love that thing used to use it to crack windows passwords
should see the guys face when you crack his 20 char password in 5 mins without ever needing to use it.
well not really cracking but changing it. used to work at the pentagon. this one guys who used to be support for one dept. thought it would be funny to change all the admin passwords in his office. so when one of the pc's was beyond his repair. i showed up and he was like give me 10 min and i'll log you in. well 5 mins later i was fixing the machine while he was screwing off. boy was he pissed.
how big's gcc? cause i'll need to compile it for my linux.
Click to expand...
Click to collapse
This is one of the most unbelievable stories ive read in a while. The fact that the pentagon had an administration department without policies or security in place to prevent such a widely known method, is comical.
btw, such a linux cd is no secret. Its called pnordahl.
Useful information
I'm surprised to see that many of you don't use the Forum's search function and simply start new topics over and over again. What funny is that the info you put here is old and useless.
I've posted an article on how to compile native C code for Android months ago, with several examples and tools:
http://forum.xda-developers.com/showthread.php?t=514803
or direct link herE:
http://www.pocketmagic.net/?p=682
However this technique is now too old.
The best approach is to simply use the NDK and build a custom Makefile for Cygwin's make under windows or easier under linux, see:
http://betelco.blogspot.com/2010/01/buildingdebugging-android-native-c.html
radhoo said:
I'm surprised to see that many of you don't use the Forum's search function and simply start new topics over and over again. What funny is that the info you put here is old and useless.
I've posted an article on how to compile native C code for Android months ago, with several examples and tools:
http://forum.xda-developers.com/showthread.php?t=514803
or direct link herE:
http://www.pocketmagic.net/?p=682
However this technique is now too old.
The best approach is to simply use the NDK and build a custom Makefile for Cygwin's make under windows or easier under linux, see:
http://betelco.blogspot.com/2010/01/buildingdebugging-android-native-c.html
Click to expand...
Click to collapse
Thank you for your very informative links. I'll take a look tonight, since I find this very interesting and would love to compile a few things for Android.
"What funny is that the info you put here is old and useless." - This was posted almost 10 months ago. That's a long time in smart phone years. By the same token, if it were January 2010 and I was looking for this info, I would assume that a post from May 2009 would be dated too.
"I'm surprised to see that many of you don't use the Forum's search function and simply start new topics over and over again." - Maybe zifnab did search and find your post and deemed your technique too old for Jan. 2010, so he created a new post with newer information. Or maybe he wanted to show a different way to do the same thing. Or maybe he figured that many users only have/take the time to look in their phone-specific forum. No one knows other than zifnab.
Personally, I welcome multiple posts by different people on the same topic. Everyone is different and often have different takes on the same thing. I find it easier to understand many techniques/topics if I get multiple perspectives.
Again, thank you for your contribution.
I need to compile the library with some modifications. how to do it as simple as possible? what will it take?
vlad072 said:
I need to compile the <library> with some modifications. how to do it as simple as possible? what will it take?
Click to expand...
Click to collapse
Did you get any thing in this regard ? Even i want to compile a part for library for the Android 5.1.1 device but not able to find any resource. Help will be appreciated.
I don't have a working build environment and I'm not sure how to do this. I want to edit the file from source. The source code path is:
msm/drivers/power/ds2784_battery.c
Which belongs in the file system:
/sys/devices/platform/ds2784-battery
I specifically want to edit one line of code and change the following line 380:
(di->status.current_uA <= 80000) &&
And change it to
(di->status.current_uA <= 10000) &&
I don't know enough, but is it possible to make a small edit like this and recompile just this individual file, and then for me to push this edited file to my phone(rooted obviously)? Is this feasible or am I just in way over my head? If anyone could guide me, or better yet make this edit for me I would be very greatful and willing to test it out on my phone. Thanks.
Hi,
I build the file for you from synced tree about 20min ago from cyanogen's repository.
Unfortunately, u can't push any files to /sys even when you have root because those files are automaticaly generated by kernel as same as files in /dev.
So i build the whole kernel for nexus one with that modification you want.
single file is there http://www.multiupload.com/WJA1GXEANV
and whole kernel is there http://www.multiupload.com/BS3YR6P2K1
Just to notice you, you wanted to lower the value from 80000 to 10000 but there is just 1024 instead of 80000 in the cyanogen's source so i raised it to 10000 as you want.
Hope this helps. Let me know if this works..
(and sorry for my english )
Thank you. I'm not sure I understand. Are you saying that the cyanogen version of this file ds2784_battery.c was not set to 80000 like the stock source code showed? You say that you had to raise that value from 1024. I don't understand why it would be different than AOSP. Thank you, I'll try it when I get home.
Yes, you understand right, cyanogen's source have that value on 1024 but don't know why.
Yes I am looking at the cyanogen code now, and It is different than AOSP file. However running cyanogen ROM the file still behaves as if 80000 was there. Hope I can find the correct value and I will post back. Thanks
I just looked at the source again and maybe i changed something else because line 380 is not same as you wrote.
Look yourself here http://github.com/CyanogenMod/cm-kernel/blob/android-msm-2.6.34/drivers/power/ds2784_battery.c
Yes the code is different, so I am looking for the Correct line to change. Thank I'll post back when I find it.
I've looked and looked thru as much as i could, and i cannot locate the identified code i found in the earlier version, where it speifies mA < 80000 (80 mA). the strange thing is that my current installation of cyanogen RC3 is still functioning as if 80 mA were set. the charger shuts off still at 80 mA. maybe someone else can guide me to see what i'm missing.
ok i have the new code i'd like to try out if you dont mind. i'm running cm rc3, and wanted to edit the ds2784_battery.c file with the following line 450:
original:
PHP:
if (di->status.status_reg & 0x80) {
di->status.battery_full = 1;
charge_mode = CHARGE_BATT_DISABLE;
} else
di->status.battery_full = 0;
new
PHP:
if ((di->status.status_reg & 0x80) && (di->status.current_uA <=20000)) {
di->status.battery_full = 1;
charge_mode = CHARGE_BATT_DISABLE;
} else
di->status.battery_full = 0;
only one change was made to line 450. could you make a flashable file? thanks so much if you have the time
Rog, just in case you were waiting for a notification here:
http://forum.xda-developers.com/showpost.php?p=7819793&postcount=261
Current version: !IMEIme 2.2.0.4
Bug Fix
Fixed bug in use previous patch that could result in variable used before declared error.
Changed processing order when custom patches were to be used
The program will now process custom patches prior to editing framework.jar and build.prop edits. With new kernel patches requiring a new build.prop users would lose build.prop edits if the kernel was included in custom patches, the program will now patch any user modifications, then process IMEI generation and build.prop edits.
Updated to work with ROMs that do not include GSMPhone.smali
Recently, many ROMs are not including GSM phone utilities in framework.jar. I have added testing for missing GSMPhone.smali and patching via TelephonyManager.smali if necessary.
UPDATED FILES UPLOADED
MANY of the support files have been updated to the newer versions (smali, baksmali, adb and components).
I encourage you to delete all files in your existing IMEI Generator folder and use the new !IMEIMe.exe to generate the files necessary.
The devices.dat file if you've used the previous version has several issues that prevents the device model from being correctly patched on many of the devices. This has been fixed here and in the device list thread.
There is a known issue with the GUI when your screen settings are set at 125% in Control Panel - Appearance and Personalization - Display... I will work on fixing that in the next release.
Bug reporting thread for !IMEIme
Device list thread
New features:
Will patch GSMPhone.smali if present in framework... patches TelephonyManager.smali otherwise.
I chose this method since more ROMs are coming out for wifi tablets that do not have GSM phone information included in framework.jar. I was playing with CM10.1 and discovered GSMPhone.smali is not present, thus I was getting unable to patch GSMPhone.smali error, and there was no patching for an IMEI. In all honesty... this should be irrelevent, since IMEI is only utilized in cellular communications on GSM phones... however... some applications MAY (xda free does) require an IMEI to work, even on wifi only devices.
ODEX files still in the works
odex file support... I think this solution will work on odex file systems as long as the patching is done on the ROM prior to flashing to device (anyone using odexed system please let us know) and I am working on in place patching on odexed systems... however, I am not completely comfortable since there is a lot of work done by the device itself during odexing of the modified files... I am very hesitant since any mistake could render a bricked device and I don't have a system to test with prior to release.
Previous Important Changes
The new version of the IMEI Generator will no longer overwrite your existing devices.dat file with the current. To use new devices.dat file, delete the old one prior to running the program, or download the new one and unzip it in the IMEI Generator directory.
Device Communications not necessary in certain situations
If you select to Update ROM, using Serial Number based IMEI and do not select Encrypt IMEI, the program will no longer need to communicate with the device when performing its tasks. The framework.jar patch will not hard patch the IMEI in this situation as before. This is useful for patching a ROM for distribution to multiple people, since they will all maintain unique IMEI's. This is accomplished with the following change in the framework.jar
Code:
/com/android/internal/telephony/gsm/GSMPhone.smali
.method public getDeviceId()Ljava/lang/String;
[b]changed[/b] iget-object v0, p0, Lcom/android/internal/telephony/gsm/GSMPhone;->mImei:Ljava/lang/String;
[b]to[/b] sget-object v1, Landroid/os/Build;->SERIAL:Ljava/lang/String;
prior to patching in code to prepend "0"
.method public getDeviceSvn()Ljava/lang/String;
[b]changed[/b] iget-object v0, p0, Lcom/android/internal/telephony/gsm/GSMPhone;->mImeiSv:Ljava/lang/String;
[b]to[/b] sget-object v1, Landroid/os/Build;->SERIAL:Ljava/lang/String;
prior to patching in code to prepend "0"
To try to explain the above a little...
The above is always changed, no matter what IMEI generation method you select...
If you select Serial Number and New Type IMEI and not Encrypt: no other patching is done for the IMEI... this can be implemented on many devices, since each will have a unique serial number.
If you select Serial Number and do not select New Type: additional code is added to format the IMEI to the old standard ("00-" and "-"s)... this can be implemented on many devices for same reason.
If you select MAC Address or Encrypt (or both): additional code is added that results in the IMEI being hard coded, this makes it very much device specific.
If you select MAC Address or Encrypt (or both) and do not select New Type: additional code is added that results in the IMEI being hard coded as well as code to format the IMEI, this makes it very much device specific.
Use Custom Patch NOTE: This is only used when patching a ROM
This is going to take some major explanation, since I ran into so many possible scenarios...
One thing of note... the only additional lines added to updater-script will be for files in the base directory
The order of processing is:
1. Original ROM updater-script and files
2. Custom Patch zip file
3. Custom Patch folder
The program will utilize folders (from Patch zip file or Patch folder itself) named modboot, modsys, or system (not case sensitive in windows) as well as files in the base folder
Any files in modboot will be moved to the root of the **ROM**-IMEI.zip file and lines added to updater-script as needed
Any files in modsys will be moved to the system directory of the **ROM**-IMEI.zip file
If Custom Patch is checked...
/META-INF/com/google/android/updater-script is extracted from the ROM
the program will ask you to select the Custom Patch Folder
If there is a zip file present in the folder the program will ask if you want to use it
You have 3 options, "Yes", "No" or "Cancel"
Yes = Use the zip file
No = Don't use it, select another
Cancel = Don't use a zip file
If you use a zip file, it will extract the zip file and process the updater-script in it for any additional lines needed
After the above, any non-zip files and modboot, modsys and system directories in the Patch Folder will be processed
I chose this order so you can have a "go to" patch zip file, and test other additions by using the file, folder options prior to including them in the zip.
Example here:
I have my custom patches in folder /CM7/UserMods with these contents:
/META-INF
/modboot
/modsys
patch.zip
The program processes patch.zip first, then overwrites any files with the files in modboot and modsys
It also processes /META-INF/com/google/android/updater-script for any lines extracting files to /boot and adds them to the original ROM updater-script if not already there.
It then adds lines for any files originally in /modboot to updater-script to extract them to /boot
"New IMEI Type" of IMEI which no longer has the "-"s in it, but maintain backward compatibility for those who already have IMEI's generated or prefer the old style. When the new type is selected in the GUI:
NOTE: Per the IMEI standards... Using a single 0 prepended to the IMEI indicates a TEST IMEI for a country with 3 digit international code... while it should have no implications to us since we are not on a cell... it may provide potential country validity issues... I will monitor this and resort to 00 prefix in the new type of IMEI if necessary.
ADDITIONAL NOTE: Per the IMEI standards... For devices without an IMEI, they are to provide a unique serial number to be used... This program modifies framework.jar to allow this.
I am now patching framework.jar in the /com/android/internal/telephony/gsm/GSMPhone.smali file instead of /android/telephony/TelephonyManager.smali (this change is what allows the information to display in the about tablet information)
I am renaming and patching 2 functions... getDeviceID() and getDeviceSvn()
By patching the two functions in this file... the IMEI now shows in Settings... About Tablet... Status... no longer have to use external program or dial *#06# to verify the device is patched.
getDeviceID() shows it in IMEI
getDeviceSvn() shows it in IMEI SVN
You can rename or copy !IMEIme.ini to IMEIme.ini and the program will work.... useful for *nix users and probably mac users... since they have issues with special char actors (!)... While I like to use it in windows to keep the executable and ini file at the top of the file list in windows explorer... anyway...
The program looks for IMEIme.ini first and uses it if present... if it is not... it then looks for !IMEIme.ini (which will be there... because the program installs the generic !IMEIme.ini if it isn't ) This also provides a good way to keep your ini.. and see the new settings in the compiled in ini.
GUI selection and related ini setting
GUI: New IMEI Type
INI Setting:
New_Type =
; If 0 then the old type of "00-XXXXXX-YYYYYY-ZZZ" will be used
; If 1 then the new type of "00XXXXXXYYYYYYZZZ" will be used
BUG FIX
No known or reported bugs to work out.
!IMEIme.ini file default settings and explanation:
Code:
;The setting options are 1 (use the option) or 0 (don't use the option)
;WiFi IP Address can be set to your Nook's IP address here to a default to use
;IMEI can be set to a default here... you can also set the seed you use for generation
;Setting Device_Manufacturer to anything will result in an edit to build.prop setting the entered manufacturer
;IF Device_Manufacturer is NOT blank then:
;Setting Manufacturer_Device to anything will result in an edit to build.prop setting the entered device
;
;NOTE: ONLY Device_Manufacturer is necessary for this edit... there have been no software that appears to
; require a device edit
;
;Setting LCD_Density will result in build.prop edit for this setting regardless of Device_Manufacturer setting
;
;Set all options in [Settings] section at the bottom
[Settings_Explained]
Use_In_Place = 1
; If 0 Disable In Place patching... useful for those who always update AOSP ROM files and never patches on device framework.jar
; If 1 Enables In Place patching if ADB is working
Use_Previous_Patch = 0
; If 0 Ignore IMEI.fix
; If 1 AND IMEI.fix exists... use it for patching
Use_Serial_Number = 1
; If 0 then do not base IMEI off of Device Serial Number
; If 1 then base IMEI off of Device Serial Number
; NOTE: This takes priority over Use_MAC_Address
Use_MAC_Address = 0
; If 0 then do not base IMEI off of Device MAC Address
; If 1 then base IMEI off of of DeOvice MAC Address (last 5 hex words) (2 bytes = 1 hex word)
; 0A is converted to 010, FF is converted to 255 etc.
; NOTE: Use_Serial_Number takes priority
Use_Manual_Input = 1
; If 0 then Manual Input disabled
; If 1 then Manual Input enabled
Encrypt_IMEI = 1
; If 0 then uses actual data for IMEI... i.e. Serial Number (last 15 digits) or MAC Address (last 5 hex words) is actual IMEI
; If 1 then program encrypts data for IMEI generation... hiding actual Device data
New_Type = 1
; If 0 then the old type of "00-XXXXXX-YYYYYY-ZZZ" will be used
; If 1 then the new type of "00XXXXXXYYYYYYZZZ" will be used
Use_ADB = 1
; If 0 then ADB is disabled... this will prevent In-Place updating from working all together
; If 1 then ADB is enabled... In-Place will work... IF adb is working on your device
; NOTE: This takes priority over Use_ADB(usb) and Use_ADB(WiFi)
Use_ADB(usb) = 1
; If 0 then ADB via USB connection is disabled... I use this since some ROM's have Debug Mode issues
; If 1 then ADB via USB is enabled and attempted first
; NOTE: Use_ADB takes priority over Use_ADB(usb) and Use_ADB(WiFi)
Use_ADB(WiFi) = 1
; If 0 then ADB via WiFi connection is disabled
; If 1 then ADB via WiFi is enabled... I use this since some ROM's have Debug Mode issues
; NOTE: Use_ADB takes priority over Use_ADB(usb) and Use_ADB(WiFi)
Clean_Up = 1
; If 0 then the program will leave all support files when cleaning up and exiting
; If 1 then the program will delete all support files when cleaning up and exiting if none of them
; existed at program start
Include_Patch = 0
; If 0 then custom patches is disabled
; If 1 then the program will prompt for custom patches to include
Device_Manufacturer =
; If blank then the program will not edit build.prop
; If anything other than blank the program will edit build.prop to include manufacturer
Manufacturer_Device =
; If blank then the program will not include device in build.prop edit
; IF anything other than blank the program will include device in build.prop edit
; NOTE: No build.prop edit will occur if Device_Manufacturer is blank
Device_Model =
; If blank then the program will not include model in build.prop edit
; IF anything other than blank the program will include model in build.prop edit
; NOTE: No build.prop edit will occur if Device_Manufacturer is blank
Build_Fingerprint =
; If blank then the program will not include Build Fingerprint in build.prop edit
; IF anything other than blank the program will include Build Fingerprint in build.prop edit
; NOTE: This edit will occur even if Device_Manufacturer is blank
LCD_Density =
; If blank then the program will not include LCD Density in build.prop edit
; IF anything other than blank the program will include LCD Density in build.prop edit
; NOTE: This edit will occur even if Device_Manufacturer is blank
WiFi_IP_Address =
; You can enter the default Device IP address here... especially useful if you are only using this on one device...
; or if you keep seperate folders for each device you use (!IMEIme.exe and !IMEIme.ini must be in each folder)...
; i.e. folder for "sister" containing the program and ini file at minimum.
; If blank the program will prompt you for the IP address of the device to establish ADB WiFi connection
IMEI =
; Enter a base 10 (integer) and it will be used as the IMEI (duplicated until 15 digits is reached)
; Enter your "seed" and the program will generate an IMEI based off of it
; NOTE: If you try to generate the old GENERIC IMEI the program will not do it
[Settings]
Use_In_Place = 0
Use_Previous_Patch = 0
Use_Serial_Number = 1
Use_MAC_Address = 0
Use_Manual_Input = 1
Encrypt_IMEI = 0
New_Type = 1
Use_IMEI(15) = 0
Use_ADB = 1
Use_ADB(usb) = 1
Use_ADB(WiFi) = 1
Clean_Up = 1
Include_Patch = 1
Device_Manufacturer =
Manufacturer_Device =
Device_Model =
Build_Fingerprint =
LCD_Density =
WiFi_IP_Address =
IMEI =
Credits:
mthe0ry: Credit for the original IMEI patches released for us Nookers(TM). His original thread is here...
martian21: Took mthe0ry's work and maintained it for releases of CM7, upeating it for each nightly that needed a new one. Martian21's thread.
HacDan on irc.freenodes.net #nookcolor for helping me figure out patching GSMphone.smali instead of TelephonyManager.smali
Thank you's:
paleh0rse: I believe was the first to download and test this program... I think the first bug report too... helped many users with suggestions regarding their apps.
mr_fosi: Continues testing and reporting despite no need to. Tested a few private beta builds to help iron out a significant issue. Also providing information regarding Phone App *#06# IMEI test.
martian21: Set the wheels turning. Provides invaluable feedback and suggestions. He is an invaluable tester and Q&A guy. Thanks for dangling that bait
mellopete: Provided the very first bug report... prompted me to include necessary files in the program itself.
TheMainCat, 12paq and frankusb: Provided bug reports leading me to look at why some Windows versions didn't run the program initially.
Nayla1977: Bug report regarding a mistyped EndIf in my source.
jdexheimer: Bug report that lead me to find a problem with folders with spaces in them.
LinuxParadigm: Bug report regarding missmatching If - EndIf's.
BitingChaos: first public post to get me back on target.
dillweed, garrisj and many others: for PM's indicating the importance of this solution.
lemdaddy for reporting the bug that we tracked down to the java version and reporting back that it was the java version causing issues.
adusumilli for reporting the bug where IMEI was generated as "00-cat: c-an't o-pen"
topcaser for being persistent enough with the bug causing In-Place to fail in certain situations.
HacDan on IRC for leading me in the right direction to impliment the patching of GSMphone.smali.
We are all adults, if we break our toys... we only have ourselves to blame and we may have to buy new ones... (this will NOT break your Nook... I PROMISE you that! but it may break some of your apps... more on that later in post)
BUG REPORTING:
This program was initially ineteded to generate a unique IMEI based on your device S/N and update Dev's install zip files... it has become so much more, and as such there are many functions involved in this process.
Due to the complexity the program has taken on... far beyond what I initially intended... to report bugs please try to use the following as a template:
Function attempting: i.e. Updating ROM... In Place Upgrade... Update framwork saved on computer... etc.
Error Messages: any error message you receive... or the last message you saw prior to the issue.
End result: i.e. GSMphone.smali updated, ROM not... GSMphone.smali updated framework.jar not... etc....
Environment: ROM in same folder as IMEIme.exe... ROM on same drive as IMEIme.exe... ROM on different drive... etc. (same for framework if updating framework instead)
!IMEIme.ini settings: you can put your entire ini file if you'd like.
If you could take notes of EXACTLY what which selection in the GUI you have selected and any buttons you click on which prompt it would be EXTREMELY helpful...
As I said, this program has taken on functions I initially had not imagined including... the more features added, the more complex testing and tracking bugs becomes... I don't want to include a bunch of messages just for the sake of letting you know where in the code you are... would not be beneficial to you... more buttons to click for no reason, etc.
The more detailed you can be, the quicker I can see what is happening... otherwise I have to try to duplicate what I think you are doing when you get the error.
mr_fosi and martian21 have been very tedious in reporting bugs... I greatly appreciate their testing despite not needing to, and the manner in which they document what is going on....
Everyone should click "Thanks" on their bug report posts... they have been instrumental in getting the program where it is so far.
Background:
Some developers require a unique number that is supposed to be provided by hardware manufacturers that is unique to every device. This unique number (IMEI) is extremely important in devices utilizing cellular communications.
Since B&N has not registered IMEI numbers for the Nooks, the AOS's we are using do not acquire it as they do in other Android devices.
The developers that require a unique IMEI have been less than receptive of our devices and past methods to provide functionality to utilize their apps.
I decided to provide what I believe to be a viable solution to this problem.
What this program is:
It is a method to provide a unique IMEI (with reasonable certainty) for our Nooks.
It IS intended to be a supplement until IMEI is addressed in dev's ROM's.
It IS viable for Froyo... CM7... CM9... CM10...Honeycomb... MIUI.... AOKP... and others.
I can't think of any reason it will not work with ANY ROM you choose to utilize... if you run across one... just let me know and I'll see if I can't fix that.
What this program is not and does not do:
This is not a perfect solution to our Nook specific issues. Let me make it PERFECTLY CLEAR there is NO PERFECT SOLUTION We are generating an IMEI from something else... I use TEST IMEI patterns based off of our device serial number, to ensure apk devs wouldn't come down on us.
It is not targeting any specific AOS.
It is not guaranteed to be accepted by any other developers.
It is not intended to be the end all, beat all solution.
It is not intended to dissuade other developers from providing what they feel is a better method.
It will not cause any programs to show in the market. That has to be dealt with via APK developers and/or build.prop Manufacturer strings.
Potential issues:
There is NO legitimate solution to the IMEI issue we Nookers (TM) face... unless a group desires to register a block of them for our use... thus I am generating TEST IMEI's... ideal... no, but the only method available to us.
While I feel, with significant certainty, there will be no negative consequences from apk devs in general, I cannot speak for them, or their logic. This can easily be disabled by them again. That is on them, not me or us. By the same token, they can decide to stop providing their service for cause, I still have no control over that.
Above, I emphasize “with reasonable certainty” due to the fact that, in theory, you can wind up with an IMEI that 9 other Nooks that use this software has. That can only happen if the other 9 owners use this program and have a serial number within the same 10 as yours. This is even less likely with the New IMEI Type since it is using the right most 16 digits of a device serial number (and we know they all start with 2)
If everyone who has the same beginning 15 digits utilizes this program to generate an IMEI, you will all wind up with the same IMEI. Given the number of Nooks out there compared to the number of user's hacking them.... I find it extremely difficult to believe, with a reasonable certainty, that any 2 (much less 10) devices would ever wind up with the same IMEI generated by this program. This is prevented when using the New IMEI Type
What this program does/is capable of:
It allows you to extract framework.jar from a developers update zip file.
It will allow you to pull framework.jar from your Nook or use an existing framework.jar already stored on your computer.
It will generate an IMEI based on your Nook's serial number (or MAC Address) if adb is working on your system. If you have issues running adb via USB (ADB(USB)), it provides the opportunity to utilize adb via WiFi (ADB(WiFi)) for any computer-device communications.
It will provide you a method to manually input your serial number if you cannot connect to the device via adb. You can also input a “seed” (easy to remember word or phrase) and generate an IMEI based on the ASCII codes of the text you enter.
It will edit /com/android/internal/telephony/gsm/GSMPhone.smali to rename any existing getDeviceId() and getDeviceSvn() function to getDeviceId2() getDeviceSvn2() and append the patch to end of that file. NOTE: When the program "smali's" the resulting GSMphone.smali... it relocates the appended function to be before the renamed function.
It will save the patch as IMEI.fix, thus allowing you to utilize it for subsequent runs of the program. A caveat to this is... if you run it from the same folder on a friend's Nook... it will overwrite your original one if it is in the same folder or they will have the same IMEI as you do if you use Previous Run.
It will offer to push the patched framework.jar to your Nook... IF you opted to pull framework.jar from your Nook AND adb successfully worked to do that. This facilates in place upgrading.
It will backup the existing developers zip file appending “-IMEI” to it, distinguishing it is one this program has been used on. It will update this file, not the original developers file.
If there are issues with file names that become duplicate in a case insensitive OS such that windows is, it will warn you of this case and not remove the updated framework.jar to facilitate manual updating of the zip file.
Caveats:
This program is known to work on Java version 1.6.0_23 and known NOT to work on version 1.6.0_17 or earlier. If your system seems to work fine... but the nook does not give you an IMEI number... check your java version by typing this in a DOS window (start-run and type in cmd):
java -version
this will tell you the version of java you are running.
Java must be on your system. It must be in your system's path statement, or this program must be in the java/bin folder. It is possible that you must have java 32 bit version, this is being researched.
It will very likely break your swype, or any other app that utilizes IMEI for validation and you have used previous methods to circumvent their validation process.
It will likely break the same software if/when developers include a fix to the Nook IMEI situation in their AOS. Unless you opt to use this method again on their AOS to ensure you maintain the IMEI you used my program to generate.
Since I have opted to utilize test formed IMEI's to prevent duplicating someone's “real device” IMEI, software developers can easily shut us down again. That is their option. I am trying to provide a solution that is acceptable to both sides of the fence.
Closing statement:
As I desire to make this program as beneficial as possible... PLEASE provide any feedback and/or bug reports... just don't continue to push your ideals once it has been discussed... beating dead horses gets tiresome and just wastes precious time.
112 downloads of 2.2.0.3 with bug when pervious fix was selected
1686 downloads of 2.2.0.2 with no bugs reported
141 downloads of 2.2.0.1 with CM10 in place bug that would cause BBSOB and never boot
197 downloads of 2.2.0.0 (that actually appeared to be 2.1.0.4 in the zip) with a few minor bugs... mostly in custom patching
648 downloads of 2.1.0.3 with known GT for GameLoft issues
1123 downloads of 2.1 with no known bugs
182 downloads of 2.0a with a Generic IMEI bug
1919 downloads of 1.9 with no bug reports
3131 downloads of 1.8 with all bug reports being for non-nook devices
80 downloads of 1.7 with no bug reports
600 downloads of 1.6 with a couple of reports of In-Place update bug
880 downloads of 1.5a with 0 bug reports
148 downloads of 1.5 with a bug that could result in IMEI being generated without being properly formed.
36 downloads of 1.4 with a bug that could result in IMEI of "cat: can't open".
258 downloads of 1.3 with 0 bug reports... time to move on with next feature.
1618 downloads of 1.1 and the only bug noted has been tracked to the user's Java version.
12,758 downloads prior to the current version.
Bug reporting thread for !IMEIme
Device list thread
Looks like I have something new to mess with tomorrow night... thanks for working this, we owe ya!
Been looking forward to this! Thanks for your hard work DizzyDen.
Tested it out however it isn't finding 7zip. I've tried both the 64-bit and the 32-bit version (on 64-bit Windows 7). I'm probably doing something wrong if so please feel free to enlighten me
Martian21
martian21 said:
Been looking forward to this! Thanks for your hard work DizzyDen.
Tested it out however it isn't finding 7zip. I've tried both the 64-bit and the 32-bit version (on 64-bit Windows 7). I'm probably doing something wrong if so please feel free to enlighten me
Martian21
Click to expand...
Click to collapse
It wasn't you... there's something weird with the API to the fileopendialog that changes the working directory... a TEMPORARY work around is to copy the zip file to the folder you are running the program from.
Updating to beta 2 to auto extract support files on run.
Beta 2 is up... OP updated... note the bold text... for now the zip file must be in the same folder as IMEIme.exe
That will be fixed shortly.
Updated to beta 3. OP updated.
Fixed file browse for update file.
Improved cleanup behind itself before exiting...
removes helper files
removes framework.jar
removes classes.dex
removes out folder
removes system folder (the one used to add framework.jar to the zip file)
Still debating ability to allow manual input of the IMEI or a serial number... but those that want to do it will probably figure out how to do it manually... its REALLY not that hard.
Will add random IMEI generation as an option. The only purpose I see for this is for those who don't want to use the generic IMEI and cannot get adb working... even with the included adb in this program.
Feedback and bug reports are welcome and will help improve the program.
Thank you for this
I had to copy my AdbWinApi.dll for it to work. It did not put the new framework.jar in the zip though. It made the files, but didn't update the zip. I moved it to the root of my drive and ran it as administrator, but it still didn't update the zip. I am using Windows 7 x64. I used the IMEI.fix file and updated the zip myself. Thanks again for this nice tool.
mellopete said:
I had to copy my AdbWinApi.dll for it to work. It did not put the new framework.jar in the zip though. It made the files, but didn't update the zip. I moved it to the root of my drive and ran it as administrator, but it still didn't update the zip. I am using Windows 7 x64. I used the IMEI.fix file and updated the zip myself. Thanks again for this nice tool.
Click to expand...
Click to collapse
Did you use something prior to b3 ?
There was an issue I discovered that was preventing appending IMEI.fix to TelephoneProvider.smali that was fixed in b3.
I did my development on windows64 so that shouldn't be an issue.
As for the dll... I hadn't experience issues with that... but I can certainly add it to the program.
Both adb dll's will be included in all releases after b3.
Good job!
Can you explain more about how rom is being affected?and what to check?
Sent from my phiremod for Nook using Tapatalk
DizzyDen said:
Did you use something prior to b3 ?
There was an issue I discovered that was preventing appending IMEI.fix to TelephoneProvider.smali that was fixed in b3.
I did my development on windows64 so that shouldn't be an issue.
As for the dll... I hadn't experience issues with that... but I can certainly add it to the program.
Both adb dll's will be included in all releases after b3.
Click to expand...
Click to collapse
b3 is the first one I tried. I didn't look at the classes.dex before it was deleted. I will check.
RASTAVIPER said:
Good job!
Can you explain more about how rom is being affected?and what to check?
Sent from my phiremod for Nook using Tapatalk
Click to expand...
Click to collapse
Read here http://forum.xda-developers.com/showthread.php?t=1004102
TelephonyManager.smali did not change.
mellopete said:
TelephonyManager.smali did not change.
Click to expand...
Click to collapse
Please make sure b3 is the one you are using. When you originally posted... the thread was showing 0 downloads of that file.... or just wait a few minutes... beta 4 is on its way shortly.
To ensure TelephonyManager.smali is not changed you need to look in two places.... the easiest way is to search for getDeviceID
If it worked correctly you should find 2 instances... the first is the original function and my program renames it to getDeviceID2()... the second should be the one !IMEMe adds to the end of TelephonyManager.smali
Additionally... could you check and see if your run is actually overwriting update zip file.... see if there is a update ".zip.tmp" file left over... if it is there... the zipping is running into an issue overwriting the original file... I thought I had that issue worked out... but may need to add a check for that within my program.
I d/l b4, dropped it in a directory with just the .zip for n87 and ran it (win7 pro 64-bit). It errored out and here's the play-by-play of each of the windows which popped up one immediately after the other:
- I was warned about you being an unverified software publisher, which I OKed.
- "Windows cannot find 'java'. Make sure you typed the name correctly, and then try again." I OKed this one as well.
- window titled "DizzyDen's IMEI Generator" containing: "Return Code is:0 and Error Code is: 1"
- window titled "DizzyDen's IMEI Generator" containing: "Java is required on your system. You can download the current version from http://java.com"
I have JRE6 on my machine, though it is not in the system PATH.
Oh, and there were files for 7za, adb, .dll's and .jar files left behind.
mr_fosi said:
I d/l b4, dropped it in a directory with just the .zip for n87 and ran it (win7 pro 64-bit). It errored out and here's the play-by-play of each of the windows which popped up one immediately after the other:
- I was warned about you being an unverified software publisher, which I OKed.
- "Windows cannot find 'java'. Make sure you typed the name correctly, and then try again." I OKed this one as well.
- window titled "DizzyDen's IMEI Generator" containing: "Return Code is:0 and Error Code is: 1"
- window titled "DizzyDen's IMEI Generator" containing: "Java is required on your system. You can download the current version from http://java.com"
I have JRE6 on my machine, though it is not in the system PATH.
Oh, and there were files for 7za, adb, .dll's and .jar files left behind.
Click to expand...
Click to collapse
java will need to be in your path... I have no way of including all possible locations of where it could be installed... and it is way too big to include with my program.
The left over files is due to the program exiting when it did... I will fix that in next beta... should have waited until java was tested to extract them... or have it perform cleanup before exiting on any errors... sorry bout that.... you can leave them... when you have successful run (or run beta 5 or later) it will clean them up.
For now you may have to run as administrator.... I will try to add code to avoid this in the short future.
BTW. Nowhere does getDeviceID does it say that it must be a well formed IMEI.
nemith said:
BTW. Nowhere does getDeviceID does it say that it must be a well formed IMEI.
Click to expand...
Click to collapse
As much as I admire your work... I am honored that you are even checking this out.
I do understand that as of now it is not required... but I figure if I utilize standards (as much as there are anyway) we may avoid future issues if dev's start checking for well formed IMEI's.
I figure if I'm going to make this... I might as well make it right.
As far as I can determine... if a sw dev implemented IMEI checks, the only thing that could cause them to shut down someone using this would be to check that it is a "TEST" IMEI... but I don't see that happening, because hardware manufacturers do use these in testing.
DizzyDen said:
java will need to be in your path... I have no way of including all possible locations of where it could be installed... and it is way too big to include with my program.
Click to expand...
Click to collapse
Roger that. Should the instructions then note either the required change to PATH or that the file must be run in the user's jre#\bin directory?
DizzyDen said:
The left over files is due to the program exiting when it did... I will fix that in next beta...
Click to expand...
Click to collapse
I figured as much, but thought you should know.
DizzyDen said:
For now you may have to run as administrator...
Click to expand...
Click to collapse
I ran it this way and got the same behavior.
I'll keep a lookout for further versions, test them and report.
Beta 5 is up... OP updated to include Java requirements... thank you mr_fosi for pointing this out.
RASTAVIPER said:
Good job!
Can you explain more about how rom is being affected?and what to check?
Sent from my phiremod for Nook using Tapatalk
Click to expand...
Click to collapse
Did you find the information in the thread linked in response to your questions?
TY mellopete for that.
- Plugged NC into USB port.
- Copied new B5 exe and n87 zip to java\jre6\bin directory.
- Ran exe as admin.
- Prompted for .zip check ("is this correct") and it was, so I OKed it. Not OKing it gave me the option to browse for the file, which I cancelled, resulting in a termination of the prog with a few more dialogs. Any extracted files were cleaned up an prog close, except for adb.exe (which I deal with below).
- Re-ran, exe, chose the detected n87 .zip.
- Displayed correct serial.
- Displayed correct generated 17-digit IMEI.
- Dialog contents "Modifying" gave error "Unable to open file", which I OKed.
- Several more dialogs flew by in rapid succession without error, ending with "Updating ROM" overlaid by "Updated ROM file has been saved as: cm_encore_full-87-IMEI.zip".
- Not all ancillary files were cleaned up. Two files remained: 1) IMEI.fix, a plain txt file containing the correct code to insert the generated IMEI and 2)adb.exe which could not be removed because it was still running the devices server. Running "adb kill-server" in the java\jre6\bin directory allowed me to remove adb.exe.
- A check of the modified smali showed only one instance of "getDeviceId" indicating that the smali had not been modified to add the code to spoof the IMEI.
I would also not have been able to eject my NC, had I tried, until I killed the adb server. Looks like one more line of code to add before cleanup.
After some google action, I'm posting my problem here:
Scenary:
Gear: Motorola Atrix 4G AT&T - stock ROM (4.5.91, Android 2.3.4, WebTop version WT-1.2.0-110)
1. run webtop2sd (1.1.2) - OK
2. running web scripts - OK
Then when trying to install for example openoffice.org or trying to do:
# sudo dpkg-reconfigure tzdata
(setting to Europe/Ljubljana - or any other location for that matters)
Current default timezone: 'Europe/Ljubljana/
date: invalid date 'Wed Oct 5 07:06:25 UTC 2011'
This is the reason why packages that update or configure tzdata (tzdata-java) fail with apt-get or synaptic.
What I tried:
probably all locations with reconfiguring tzdata.. fail
manualy edited timezone.. fail
removing&reinstaling tzdata.. epic fail
Anybody else has this issue... solved ?
Same here! Any help would be greatly appreciated.
Not the right answer, but this works...
I was having the same problem, and after a few hours of struggling, here's what I eventually did.
Edit the file /var/lib/dpkg/info/tzdata.postinst and on the first blank line (3 rows down) just put exit 0 which makes the script stop and return a success. This isn't a good solution, but it does work.
Now i can go on my merry way installing things to my hearts content, but I'd love to hear if someone has a proper solution...
conundri said:
I was having the same problem, and after a few hours of struggling, here's what I eventually did.
Edit the file /var/lib/dpkg/info/tzdata.postinst and on the first blank line (3 rows down) just put exit 0 which makes the script stop and return a success. This isn't a good solution, but it does work.
Now i can go on my merry way installing things to my hearts content, but I'd love to hear if someone has a proper solution...
Click to expand...
Click to collapse
thanks a millon man,that totally fixed my problem I was having with my webtop
I took another look at this and has anyone tried adding exit 0 to just after the second to last fi close of the if then part at the bottom, I use dthe workaround posted above and went back looked over it more after having installed a few programs and added exit 0 to line 46 and it seems to be working fine. Can some one confirm this for me if they are looking at it for the first time on their version of webtop.
Hi there, I have a new install webtop2sd, and the script still failed after adding exit 0 at line 46. I have put exit 0 at line 3 afterwards to get past the problem.
I figured out the cause of the 'invalid date' problem. Actually the bug is on the end of the file in the following lines:
Code:
UTdate=$(LC_ALL=C TZ=UTC0 date -d "$TZBase")
TZdate=$(unset TZ ; LANG=C date -d "$TZBase")
The problem is that webtop uses Busybox instead of the 'original' tools. Busybox has an incomplete implementation of 'date -d'. Replacing it with an alternative like GNU Tools is the best solution for this problem. But this could be complicated and I have not much time. Adapting this script to Busybox is easier and still better than the 'exit' workaround.
All you have to do is replacing the line
Code:
TZBase=$(LC_ALL=C TZ=UTC0 date)
with
Code:
TZBase=$(LC_ALL=C TZ=UTC0 date +%m%d%H%M%Y.%S)
Last night, I built my first Android kernel from source (opensource.samsung.com) - it was a great experience. However, I wasn't able to accomplish my goals.
I'm trying to build a few specific kernel modules, but I need them to be external modules (*.ko). The build works well, but from what I can tell by the modules.builtin file(s), the specific ones I'm trying to get out of this are built-in only. During the compile, it generates *.o files just fine. At the end of the compile, it shows it used MODPOST to create 17 .ko files. All of these are noted in the modules.order file(s).
It seems that no matter what I do, I can't change whether these are built-in or external. I tried modifying the modules.builtin and modules.order files (there's multiple copies of these that reference the specific modules I want), but these just seem to get overwritten at build time.
I've done lots of searching here and abroad, and I can't seem to find the magic I'm looking for. I've tried many things, most notably:
- make <path>/<module>.ko
- make modules
- other various commands
- cd into subdir that contains the modules I want, running make there (no targets)
Doing more research, it seems it might point to the Makefile that's in the subdirectory of the modules I want. I see various "hello world" Makefiles out there, but most of these don't seem to be completely relevant to what I'm trying to do - plus I already have a Makefile for these modules.
I'm assuming that what I probably need to do is modify the Makefile that's specific to these modules, to force it to make .ko files instead of .o files. I'm also assuming that I can't simply change the lines in the Makefile to read "<module>.ko" instead of "<module>.o". I realize that there's a difference in these files.
I also thought the answers might lie in looking at the Makefile of one of the other 17 external modules that get created. However, it appears these are no different than the Makefile of the modules I'm after.
Okay, enough rambling - I'll give some details now, in case they are needed. I've downloaded the I337 source (MF3) from opensource.samsung.com. The three modules I need (so far) are: msm_kgsl_core.o, msm_adreno.o, msm_z180.o. The makefile that includes these three came from <source-root>/drivers/gpu/msm, and it is attached. I may need more than these three files later, but this is a good start.
I don't want to hear a bajillion questions about, "What are you planning to do with these files?" or "Why do you need them anyway?" - the answer is a longer story than this thread. Trust me, I just need these.
So what do I need to do to make these three files? Thank you for your help.
EDIT/SOLUTION:
obj-y = built-in module
obj-m = external module
I still can't track down the source of the three variables in the last three lines of this Makefile, but they are obviously returning "y". I changed the last three lines manually to:
Code:
obj-m += msm_kgsl_core.o
obj-m += msm_adreno.o
obj-m += msm_z180.o
... and then ran a "make clean" and "make modules" again. Voilà: 20 modules built by MODPOST.