> The first suspect is the stacksize of that thread in which you crash. I'd
> do something like:
> 
> p *lwp_cpptr
> > #0  0x4011518c in __libc_internal_tsd_get () at specific.c:157
> > #1  0x400aa5b4 in malloc () at malloc.c:2349

You guys are doing user-space threads on Linux ??  I had a quick look at
the LWP library in Coda.

Have you considered using the native Pthreads support in Redhat 5.0 ??
I dont think this would be hard, just map LWPs onto Pthreads.

I've been having great luck with Pthreads... I have DCE RPC working on
RH Linux 5.0 (Intel and Alpha platform). DCE RPC uses a bunch of threads
underneath - for packet-retransmission timers, thread dispatch per RPC,
etc. 

GNU Libc uses Pthreads internally... Thats whats going on at the last
activation record in the gdb dump. I'm wondering if somehow the user-space
threads library is messing with the internal pthreads stuff in Glibc?

I wish I had time to contribute to Coda - its truly an awesome and
worthy project.. But my time is already maxxed on other committments,
mostly DCE for Linux. I like to follow whats going on here though.

-- Jim

+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
Jim Doyle                         Boston University   Information Technology
Systems Analyst/Programmer        email: jrd_at_bu.edu   Distributed Systems
						      tel. (617)-353-8248
-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-++--+-+-+-+-+-+-