llvm dev - Mar 2013 - [LLVMdev] Running cross compiled binaries for ARM on gem5

Hi,

I am trying to optimize some benchmarks using LLVM and run them on gem5
simulator (build for ARM). I am using Sourcery Codebench cross-compiler for ARM
on my x-86 machine.
My steps up till now have been using the following commands.

1. clang -static -emit-llvm -march=armv7-a -mfloat-abi=soft -target arm-elf
a.cpp -c  -integrated-as  \
 --sysroot=/home/silky/CodeSourcery/Sourcery_CodeBench_Lite_for_ARM_GNU_Linux/arm-none-linux-gnueabi/libc
\
 -I/home/silky/CodeSourcery/Sourcery_CodeBench_Lite_for_ARM_GNU_Linux/arm-none-linux-gnueabi/include/c++/4.7.2/
\
-I/home/silky/CodeSourcery/Sourcery_CodeBench_Lite_for_ARM_GNU_Linux/arm-none-linux-gnueabi/include/c++/4.7.2/arm-none-linux-gnueabi/armv4t/
-o a.bc

2. llc -O3 -o a.s -march=arm a.bc

3. arm-none-linux-gnueabi-g++ -static a.s -o a_st

To run the final executable on gem5, I use

4. /gem5$   build/ARM/gem5.debug configs/example/se.py -c
/home/silky/mbench/vect/a_st

This results in a page fault -

Global frequency set at 1000000000000 ticks per second
0: system.remote_gdb.listener: listening for remote gdb #0 on port 7000
**** REAL SIMULATION ****
info: Entering event queue @ 0.  Starting simulation...
panic: Page table fault when accessing virtual address 0xfffffff4
 @ cycle 2509500
[invoke:build/ARM/sim/faults.cc, line 70]

Trace :
9405000: system.cpu T0 : 0x8118    :   bl                       : IntAlu : 
D=0x000000000000811c
gem5.debug: build/ARM/mem/request.hh:514: Addr Request::getPC() const: Assertion
`privateFlags.isSet(VALID_PC)' failed.

Most of the search results talk about the executable not being statically
compiled. Could anyone please tell me if any of the compilation commands is
incorrect.
Is there any other way I can check the executable generated is not buggy?

Thanks and Regards,
Silky

-------------- next part --------------
An HTML attachment was scrubbed...
URL:
<http://lists.llvm.org/pipermail/llvm-dev/attachments/20130309/8a722bf1/attachment.html>

Hi Silky,

If I got correctly, you seem to be trying to run a bare-metal image on your
model, but you compile with linux-gnueabi GCC. I don't know if that will
make a difference, but I'd try to use none-eabi GCC toolchain and set the
-target armv7a-none-eabi just in case.


On 10 March 2013 00:26, Silky Arora <silkyar at umich.edu> wrote:
> Most of the search results talk about the executable not being statically
> compiled. Could anyone please tell me if any of the compilation commands is
> incorrect.
> Is there any other way I can check the executable generated is not buggy?
>
The options -target may not be correct. LLVM/Clang will silently ignore
unknown triples and will choose the default for whatever else it thinks is
appropriate. I'd use the one I mentioned above.

To make sure it's not a bug in the model, I'd run on other models (like
QEMU) or run in real hardware.

cheers,
--renato
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
<http://lists.llvm.org/pipermail/llvm-dev/attachments/20130311/9d0a84db/attachment.html>

Thanks Renato and Giang.

I was able to run a simple HelloWorld program on gem5 (in SE mode) by using 
-target -arm-none-linux-gnueabi as the option during cross compilation.

I am trying to cross compile some Computer Vision algorithms, which use
OpenCV libraries, and I tried using the same command line options - 

clang++ *-emit-llvm -static -c -Wall -g  -O0 -mfpu=vfp -mfloat-abi=soft
-march=armv7-a -target arm-none-linux-gnueabi* -DCLOCK_GETTIME_TIMING
-DOPENCV_2_4 
-c ../MEVBench/Benchmarks/FeatureExtraction/FeatureExtraction.cpp 
-I/home/silky/VecProject/opencv/OpenCVInstall/arm/include 
-I/home/silky/CodeSourcery/Sourcery_CodeBench_Lite_for_ARM_GNU_Linux/arm-none-linux-gnueabi/include/c++/4.7.2/
-I
/home/silky/CodeSourcery/Sourcery_CodeBench_Lite_for_ARM_GNU_Linux/arm-none-linux-gnueabi/include/c++/4.7.2/arm-none-linux-gnueabi/armv4t/
-o FE.bc

However, this errors out saying 

/home/silky/VecProject/opencv/OpenCVInstall/arm/include/opencv2/core/mat.hpp:117:9:
error: cannot compile this atomic library call yet
        CV_XADD(refcount, 1);
        ^~~~~~~~~~~~~~~~~~~~
/home/silky/VecProject/opencv/OpenCVInstall/arm/include/opencv2/core/operations.hpp:61:38:
note: expanded from macro 'CV_XADD'
        #define CV_XADD(addr, delta)
__c11_atomic_fetch_add((_Atomic(int)*)(addr), (delta), __ATOMIC_SEQ_CST)

Could someone please suggest what I am missing here, or what the error
indicates?

Thank you all.




--
View this message in context:
http://llvm.1065342.n5.nabble.com/Running-cross-compiled-binaries-for-ARM-on-gem5-tp55767p56023.html
Sent from the LLVM - Dev mailing list archive at Nabble.com.