We were able to succesfully cross compile AES Script to an embedded environment on an ARM processor. The compile was successful and the software worked well except that it made the rest of the system unstable. We could not bring up the nano text editor (received the following "can't resolve symbol '__aeabi_idiv'") and when we try to execute the script language Lua we received the following - can't resolve symbol '__aeabi_d2iz'.
This is a tough issue for us and we would very much like to continue using AES Script but why is it seemingly trashng libraries in our system? I'm sure it is something we are doing wrong but we do not know what. Any help would be greatly appreciated.
Tim McClure
Problem with cross compile on ARM processor
-
paulej
- Posts: 593
- Joined: Sun Aug 23, 2009 7:32 pm
- Location: Research Triangle Park, NC, USA
- Contact:
Re: Problem with cross compile on ARM processor
Tim,
I'm not sure what to suggest. As you know, AES Crypt does nothing more than open a file, encrypt it, and write out the encrypted file (and vice-versa). It never attempts to access other files.
AES Crypt is highly CPU intensive, though. Perhaps when you start it, it's consuming all of the available resources? On Windows, we deliberately release control to the operating system frequently so as to avoid starving other applications. Is that a similar issue in your environment?
Paul
I'm not sure what to suggest. As you know, AES Crypt does nothing more than open a file, encrypt it, and write out the encrypted file (and vice-versa). It never attempts to access other files.
AES Crypt is highly CPU intensive, though. Perhaps when you start it, it's consuming all of the available resources? On Windows, we deliberately release control to the operating system frequently so as to avoid starving other applications. Is that a similar issue in your environment?
Paul
Re: Problem with cross compile on ARM processor
Paul,
Thank you for your response. I think it is more of a library issue than a CPU intensive one. Do you of anyone who has compiled it to an ARN+M processor.
Tim
Thank you for your response. I think it is more of a library issue than a CPU intensive one. Do you of anyone who has compiled it to an ARN+M processor.
Tim
-
paulej
- Posts: 593
- Joined: Sun Aug 23, 2009 7:32 pm
- Location: Research Triangle Park, NC, USA
- Contact:
Re: Problem with cross compile on ARM processor
Tim,
I haven't, but I was thinking somebody had done that before. I just don't recall who, but I was thinking somebody contacted me about doing that once. If memory serves, they were having issues with the ASCII/Unicode conversion.
I'm really not sure what to suggest. It doesn't make sense that one would get an error like "can't resolve symbol '__aeabi_idiv'". That looks more like a linking error than a run-time error, but perhaps that library was being dynamically loaded? Still, if the symbol is there, it ought to find it. Rather perplexing.
Do you get this error merely by compiling and installing the AES Crypt code, or are you getting that only when the code is executing, or other? All I can suggest is we try to figure out when/where this is happening. There must be some logic to the madness
Paul
I haven't, but I was thinking somebody had done that before. I just don't recall who, but I was thinking somebody contacted me about doing that once. If memory serves, they were having issues with the ASCII/Unicode conversion.
I'm really not sure what to suggest. It doesn't make sense that one would get an error like "can't resolve symbol '__aeabi_idiv'". That looks more like a linking error than a run-time error, but perhaps that library was being dynamically loaded? Still, if the symbol is there, it ought to find it. Rather perplexing.
Do you get this error merely by compiling and installing the AES Crypt code, or are you getting that only when the code is executing, or other? All I can suggest is we try to figure out when/where this is happening. There must be some logic to the madness
Paul