[openamq-dev] porting to Ubuntu

[Jonathan Schultz]

Sorry Nick I really have no immediate idea of what might be happening.

If you look at the code you should see that one thread (thread 1 in your 
output) does a wait on the condition variable s_sync_cond, one thread is 
irrelevant (only used for timing), and each of the other thread calls 
s_decrement_active_thread_count.  One of the latter should find that 
s_os_active_thread_count hits zero and therefore does a broadcast on 
s_sync_cond, allowing thread 1 to pick up and keep on going.  We see 
from your debug info that s_decrement_active_thread_count is called 4 
times from 4 different threads, which is correct.  What we don't see is 
whether the broadcast is ever done.  Could you try placing breakpoints 
on lines 2755 and 2757 of smt_os_thread.c to see what that gives?

Unfortunately my timezone here (Melbourne, Australia, UTC+10) means that 
I'm due to sleep about now.  I gather you're in England so maybe if we 
need to communicate further about this issue we could do it early in 
your working day, late in mine?  Alternatively if you're a night owl, 
late in yours, early in mine?

Regards,
Jonathan

Nick Levine wrote:
> Jonathan,
> 
> I applied your patch (and checked that the patch had applied properly
> by grepping for s_decrement_active_thread_count in libsmt.a).
> 
> But still no succeess:
> 
>   (gdb) break s_decrement_active_thread_count
>   Function "s_decrement_active_thread_count" not defined.
>   Make breakpoint pending on future shared library load? (y or [n]) y
>   Breakpoint 1 (s_decrement_active_thread_count) pending.
>   (gdb) c
>   Continuing.
> 
> <<< I atempt to connect >>>
> 
>   Breakpoint 2 at 0xb7d0febc: file smt_os_thread.c, line 2753.
>   Pending breakpoint "s_decrement_active_thread_count" resolved
>   [New Thread -1215059056 (LWP 21616)]
>   [New Thread -1223451760 (LWP 21617)]
>   [Switching to Thread -1223451760 (LWP 21617)]
> 
>   Breakpoint 2, s_decrement_active_thread_count () at smt_os_thread.c:2753
>   2753	    assert (s_os_thread_active_count > 0);
>   (gdb) c
>   Continuing.
>   [New Thread -1231844464 (LWP 21618)]
>   [Switching to Thread -1231844464 (LWP 21618)]
> 
>   Breakpoint 2, s_decrement_active_thread_count () at smt_os_thread.c:2753
>   2753	    assert (s_os_thread_active_count > 0);
>   (gdb) c
>   Continuing.
>   [New Thread -1240237168 (LWP 21619)]
>   [Switching to Thread -1240237168 (LWP 21619)]
> 
>   Breakpoint 2, s_decrement_active_thread_count () at smt_os_thread.c:2753
>   2753	    assert (s_os_thread_active_count > 0);
>   (gdb) c
>   Continuing.
>   [New Thread -1248629872 (LWP 21620)]
>   [Switching to Thread -1248629872 (LWP 21620)]
> 
>   Breakpoint 2, s_decrement_active_thread_count () at smt_os_thread.c:2753
>   2753	    assert (s_os_thread_active_count > 0);
>   (gdb) c
>   Continuing.
> 
> <<< and my code hangs >>>
> 
>   Program received signal SIGINT, Interrupt.
>   [Switching to Thread -1209776448 (LWP 21609)]
>   0xffffe410 in __kernel_vsyscall ()
>   (gdb) info threads
>     6 Thread -1248629872 (LWP 21620)  0xffffe410 in __kernel_vsyscall ()
>     5 Thread -1240237168 (LWP 21619)  0xffffe410 in __kernel_vsyscall ()
>     4 Thread -1231844464 (LWP 21618)  0xffffe410 in __kernel_vsyscall ()
>     3 Thread -1223451760 (LWP 21617)  0xffffe410 in __kernel_vsyscall ()
>     2 Thread -1215059056 (LWP 21616)  0xffffe410 in __kernel_vsyscall ()
>   * 1 Thread -1209776448 (LWP 21609)  0xffffe410 in __kernel_vsyscall ()
>   (gdb) bt
>   #0  0xffffe410 in __kernel_vsyscall ()
>   #1  0xb7f905c6 in pthread_cond_wait@@GLIBC_2.3.2 () from /lib/tls/i686/cmov/libpthread.so.0
>   #2  0xb7d11421 in smt_wait (msecs=0) at /home/nick/wiinz/openamq-src/OpenAMQ-1.2c1/base-2.2b1/_install/include/icl.h:2014
>   #3  0xb7d11ff1 in smt_initialise () at smt_os_thread.c:1760
>   #4  0xb7bdc3ce in amq_client_connection_new_ (file=0x805e060 "", line=0, host=0x805e030 "localhost", virtual_host=0x805e048 "/", auth_data=0x8075b38, 
>       instance=0x805e050 "test", trace=3, timeout=10000) at amq_client_connection.c:1663
>   #5  0x200638ed in ?? ()
>   #6  0x0805e060 in ?? ()
>   #7  0x00000000 in ?? ()
>   (gdb) 
> 
> 
> Any thoughts?
> 
> - nick
> 
> 
>