On Sun, 3 Jan 1999 [EMAIL PROTECTED] wrote:
> >I'll be able to test it on Monday. My problem is that execution never
> >gets out of pthread_create() nor a thread is started. I started to trace
> >glibc v.106 and something hangs in thread_manager() or similar. I'll get
> >more details on Monday too.
>
> OK. If you can strace the failing process that would be very helpful. Also
> if you mail me your binary I can see if it works on another machine. (The
> gnu.org mail server seems to be experiencing some downtime, so use this
> address rather than my usual one.)
OK. Finally found it. It seems that threads can't work if they are
started from a process which PID is 1. At some point in pthread_manager()
there is a test to see if getppid() returns 1 and if so, all threads are
killed because the main thread would have died. In my case, the main
thread isn't dead but rather init itself. Kind of subtile...
Don't know if it should be considered as a bug, but at least noted
somewhere as a limitation.
Thanks for your help!
P.S.: Don't know if I missed it, but what's needed to make floating point
work correctly with statically linked binaries?
Nicolas Pitre, B. ing.
[EMAIL PROTECTED]
unsubscribe: body of `unsubscribe linux-arm' to [EMAIL PROTECTED]
