Re: libthread fresh findings

[Marcus Brinkmann]

On Thu, Feb 07, 2002 at 02:13:41AM -0800, Justin  Langer wrote:
> hurd~# settrans -ac test hello-mt
> 
> But the program is crashing when a file system request comes (cat or
> ls -l)in. The error caught with gdb is:
> 
> Program Received Signal
> EXC_BAD_INSTRUCTION
> Illegal Instruction/Operand in cthread_fork_child () from 
> /lib/libthreads.so.0.2

uh-oh. ;)

Maybe the gcc I used produces bad code.  I don't even remember anymore which
version I used.  Actually, I am quite frightened by all the SIGILLs and
EXC_BAD_INSTRUCTIONS popping up lately.

> I think there is some problem some where? Can any _HURD_DEVELOPER_ help
> me in fixing it? I think its prob with libraries....or.....

What do you think I am?  However, all our forces are concentrated on getting
the libio/gcc-3.0 move worked out.  This might actually help you directly,
as we struggle with similar problems there, only at a much lower level.
 
> Is it because of Muli threading?

The Hurd is multi threaded throughout, so it can't be the only reason.  It
wouldn't even boot if just using threads would be the problem.

> The function ports_manage_port_operations_multi_thread()? 

Likewise, used for ext2fs etc.
 
> In both cases (hello-mt) and translator I wrote using libnetfs a bt gives:

Please compile your own Hurd with debugging symbols and use that.  Then you
can come back with a backtrace, "info reg" and "x/30i $eip", and we might be
able to say more.

Thanks,
Marcus

-- 
`Rhubarb is no Egyptian god.' Debian http://www.debian.org brinkmd@debian.org
Marcus Brinkmann              GNU    http://www.gnu.org    marcus@gnu.org
Marcus.Brinkmann@ruhr-uni-bochum.de
http://www.marcus-brinkmann.de