Hello again,

here is a small HOWTO for debugging Genode apps on Linux:

* The first thing to do is to interrupt Genode at a certain point as you cannot start core with gdb. So, if you want to debug an application you have to call 'wait_for_continue' from within your code. Somewhere in your code write something like:

extern "C" void wait_for_continue();

void some_func_in_your_program() { wait_for_continue(); }

This will cause your program to stall in the function above.

* Compile your app * Start './core' from the command line * As mentioned before your program will stop in the 'wait_for_continue' function * Now find out the process id of your program, ('ps -efL | grep Genode') * Start gdb with 'gdb -p <Process ID from ps>' , this will attach gdb to the process * Switch to the Genode output console (where you started core) and hit 'enter', this will cause your program to leave the 'wait_for_continue' function * Switch back to GDB, you now may step and set breakpoints (note: Does not work with all gdb versions), inspect variables, ... .

I hope this helps,

Sebastian

On 09/10/2010 05:34 PM, stephane frenot wrote:

Hello, I have a problem while debugging genode with gdb.

When starting gdb ./core it freeezes on 3 LWP.

Is there a specific gdb configuration in order to reach sub services ?

/stephane

http://perso.citi.insa-lyon.fr/sfrenot work here : http://tinyurl.com/2ts4xf sleep here : http://tinyurl.com/33p2gt play here : http://tinyurl.com/3yarjz

---

Stephane Frenot Université de Lyon, INRIA INSA-Lyon, CITI, F-69621, France

---

Automate Storage Tiering Simply Optimize IT performance and efficiency through flexible, powerful, automated storage tiering capabilities. View this brief to learn how you can reduce costs and improve performance. http://p.sf.net/sfu/dell-sfdev2dev

---

Genode-main mailing list Genode-main@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/genode-main

Hello,

On Tue, Sep 14, 2010 at 09:45:10AM +0200, stephane frenot wrote:

(gdb) info thread

• 1 process 2370 pseudo_end () at

/opt/genode/genode-r121/base-linux/../base-linux/src/platform/x86_64/lx_syscall.S:29

(gdb) info stack #0 pseudo_end () at /opt/genode/genode-r121/base-linux/../base-linux/src/platform/x86_64/lx_syscall.S:29 #1 0x0000000001021442 in lx_nanosleep (this=0x103c400) at /opt/genode/genode-r121/base-linux/src/platform/linux_syscalls.h:371 #2 thread_stop_myself (this=0x103c400) at /opt/genode/genode-r121/base-linux/src/base/lock/lock_helper.h:70 #3 Genode::Cancelable_lock::lock (this=0x103c400) at /opt/genode/genode-r121/base/src/base/lock/lock.cc:137 #4 0x000000000101022e in Genode::Lock::lock (this=<value optimized out>) at /opt/genode/genode-r121/base/include/base/lock.h:35 #5 Genode::Platform::wait_for_exit (this=<value optimized out>) at /opt/genode/genode-r121/base-linux/src/core/platform.cc:50 #6 0x000000000100dfde in main () at /opt/genode/genode-r121/base/src/core/main.cc:229

Looking at your backtrace, I assume 'core' exits for any reason. Typically, this is caused by pressing CTRL-c on the command line. Technically, core's main thread blocks on a lock (that is initialized locked), which is only unlocked by the signal handler. So, there may be the very unlikely case that the lock implementation does not work on your platform... To debug the situation in core you may run it from the very beginning in GDB.

Greets

Thanks, it works now. The only remaining thing is to think to run gdb in the child process directory.

/stephane

On Tue, Sep 14, 2010 at 10:28 AM, Norman Feske <norman.feske@...1...

wrote:

Hi Stephane,

I suspect that have you placed the 'wait_for_continue' call in a program that you started via init's 'config' file but attached gdb to core's PID? You'll need to attach gdb to the thread that executes 'wait_for_continue'. Each Genode process is executed as a separate Linux process. If you want to debug a non-core process, attach gdb to the corresponding Linux thread ID. To find out about which threads there are, use 'ps -eLf' and look for thread names starting with "[Genode]".

Best regards Norman

On 09/14/2010 09:45 AM, stephane frenot wrote:

...

Actually, it is not really working.

Everything works find, until I attach gdb to the process.

I am stucked to this:

(gdb) info thread

• 1 process 2370 pseudo_end () at

/opt/genode/genode-r121/base-linux/../base-linux/src/platform/x86_64/lx_syscall.S:29

...

(gdb) info stack #0 pseudo_end () at

/opt/genode/genode-r121/base-linux/../base-linux/src/platform/x86_64/lx_syscall.S:29

...

#1 0x0000000001021442 in lx_nanosleep (this=0x103c400) at

/opt/genode/genode-r121/base-linux/src/platform/linux_syscalls.h:371

...

#2 thread_stop_myself (this=0x103c400) at /opt/genode/genode-r121/base-linux/src/base/lock/lock_helper.h:70 #3 Genode::Cancelable_lock::lock (this=0x103c400) at /opt/genode/genode-r121/base/src/base/lock/lock.cc:137 #4 0x000000000101022e in Genode::Lock::lock (this=<value optimized out>) at /opt/genode/genode-r121/base/include/base/lock.h:35 #5 Genode::Platform::wait_for_exit (this=<value optimized out>) at /opt/genode/genode-r121/base-linux/src/core/platform.cc:50 #6 0x000000000100dfde in main () at /opt/genode/genode-r121/base/src/core/main.cc:229

And I am not able to reach the line where I put "the wait_for_continue" instruction.

/stephane

---

Start uncovering the many advantages of virtual appliances and start using them to simplify application deployment and accelerate your shift to cloud computing. http://p.sf.net/sfu/novell-sfdev2dev _______________________________________________ Genode-main mailing list Genode-main@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/genode-main