hi, 
      
this is regarding debugging multi-threaded applications using gdbserver.
I am facing a weird situation with respect to the same. When i try and
run gdbserver on a multi-threaded application and try to debug the same,
gdb is unable to stop all the threads on hitting a breakpoint in one of
the threads. 
	
the steps i followed are the following. 
$>gdbserver localhost:1234 thr
$>gdb thr

(gdb) target remote localhost:1234
Remote debugging using localhost:1234
0x40000be0 in ?? ()

(gdb) set remotelogfile logfile
(gdb) b main
Breakpoint 1 at 0x80484b0: file thr.c, line 11.
(gdb) c
Continuing.
Ignoring packet error, continuing...
Ignoring packet error, continuing...
Ignoring packet error, continuing...

Breakpoint 1, main (argc=1, argv=0xbffff144) at thr.c:11
11           printf("Creating...\n");
(gdb) n
12           pthread_create( &id, NULL, work2, ((void*) 0) );
(gdb) n
13           printf("Created thread.\n"); fflush(stdout);
(gdb) n
15           printf("Waiting...\n");


At this point in time if i do an info threads the only thread it shows
me is the main thread though in gdb it shows me both the threads. 
At the same point of time when the main thread is stopped at this
position the thread which has just been created runs away. ! 
and on doing an info threads i do not get the information about the 
newly created thread. 

Is this a limitation of gdbserver with respect to kernel space threads.
 I saw the comment in thread-db.c in the function
thread_db_find_new_threads. which says "iterate" over user space threads
??
The foll. is the information regarding my machine / environment. 

RedHat linux 9.0 .  gdb v 5.3 gcc version 3.2.2 20030222
GNU gdb Red Hat Linux (5.3post-0.20021129.18rh)


I wonder if anyone else is getting a similar problem ..I am attaching
the 2 files that were used for this purpose. 

build the 2 files as
 gcc -g new.c thr.c -lpthread (obviously !)


regards
Ramana

P.S. the same works fine on a native version of gdb.