If sqlite is the evil, you should be able to reproduce it by writing a
pure C program that calls the same function and dynamically linked to
/system/lib/libsqlite.so when loading. Just several lines can confirm
this.
And, well, here is my question: after inquired Google and Baidu and
several other ones I still have no idea how to clone an Android source
code repository on dev.lemote.com via the Git interface. Could someone
please point me a tutorial/example in detail or just do this for me?
Thanks...
在 2012年1月12日 上午12:38，李晖 写道：
> hi, all
>by further debuging, I found this problem was cased by calling
> sqlite3_open_v2(path8, &handle, sqliteFlags, NULL) function in dbopen
> of frameworks\base\core\jni\android_database_SQLiteDatabase.cpp file,
> it not crash directly in the sqlite3_open_v2 function, but died some
> where
> after calling sqlite3_open_v2, if I replaced the sqlite3_open_v2
> function with the following code:
> for(int i=0;i<1000;i++)
>LOGI("useless code");
>
> process will not die in the for loop at least, so I think the
> sqlite3_open_v2 function is the evil, but I do not know what to do
> next!
>
> BTW, the linux kernel is 64bit and android filesystem is 32bit, did it
> affect?
>
> On 1月11日, 上午12时33分, 李晖 wrote:
>> hi, all
>> I am porting android to loongson2f, when the System Server bring up the
>> Account Manager
>>
>> Service, the System Service process was terminated by signal 11, by adding
>> log to the source code,
>>
>> I found it died in the JNI C code, following is the calling stack:
>>
>> > ServerThread:run
>> > PowerManagerService()
>> > validateAccounts()
>> > DatabaseHelper::getWritableDatabase()
>> > SQLiteOpenHelper::getWritableDatabase()
>> > ContextImpl::openOrCreateDatabase()
>> > SQLiteDatabase.openOrCreateDatabase()
>> > SQLiteDatabase.openDatabase()
>> > SQLiteDatabase() //Constructor
>> > dbopen()//JNI C code implemented
>> > in "frameworks\base\core\jni\android_database_SQLiteDatabase.cpp
>>
>> I am sure the System Service died in the dbopen code as I called the Log.i
>> function millions time
>>
>> before the dbopen, log are all printed correctly, and I also added the
>> following the code at the
>>
>> end of dbopen function:
>>
>> > for(int i=0; i<1000; i++)
>> > {
>> > fprintf(fp, "dbopen %d\n", i);
>> > fflush(fp);
>> > }
>>
>> not all logs are recored, if I removes the for loop at the end of dbopen
>> function, the system will die some statement after the dbopen call, so, I
>> think the system died in this function, but I don't known how to fix it,
>> did it have to do with the JNI system?
>>
>> any suggestion or tips will be appreciated!
>> thanks!
>>
>> Best Regards
>> ---
>> ---
>> 北京航空航天大学大学嵌入式系统实验室
>> 新主楼G座1028室
>> 李晖
>> MSN:lihui205...@163.com
>
> --
> You received this message because you are subscribed to the Google Groups
> "loongson-dev" group.
> To post to this group, send email to loongson-dev@googlegroups.com.
> To unsubscribe from this group, send email to
> loongson-dev+unsubscr...@googlegroups.com.
> For more options, visit this group at
> http://groups.google.com/group/loongson-dev?hl=en.
>
--
https://fedoraproject.org/wiki/User:Lonestar
--
You received this message because you are subscribed to the Google Groups
"loongson-dev" group.
To post to this group, send email to loongson-dev@googlegroups.com.
To unsubscribe from this group, send email to
loongson-dev+unsubscr...@googlegroups.com.
For more options, visit this group at
http://groups.google.com/group/loongson-dev?hl=en.
