On Fri, Mar 14, 2003 at 06:58:58PM +0000, Jonathan Perkin wrote:
> So I'm doing my first bit of sparc64 kernel debugging, and not getting
> very far..
> 
> Narrowing down the PR, mkfifo() succeeds, but any subsequent attempts to
> open the fifo fail, open() simply hangs.
> 
> What's the best way to start debugging through sys_open() to see where
> the deadlock is occurring?  I can debug userland stuff ok, but when the
> actual syscall is hanging then I'm not sure where to go from there :-)

Wait until the system is hung, and break into DDB.
Do 'ps' to locate the process that's hung.
Dump it's PCB and find the stack pointer.
Hand the stack pointer to 'tr' or 'mach stack'
commands.

Eduardo