Experimental Physics and
Industrial Control System

Ralph Lange <[email protected]> · Mon, 26 Jul 1999 11:58:36 +0200 (METDST)

> I changed my host computer from tcsh back to csh,

Why? Is there a good reason to do this backward step?

For performance reasons I would suggest having the IOCs boot with a user
name that has /bin/sh as login shell, anyway. Being a shared excutable,
the /bin/sh binary will very likely be already running on the host, i.e.
it will already be in memory. Plus, usually /bin/sh is the fastest shell
in terms of time needed for startup/login.

> and I found all the 162s that are booting from my host stop
> loading. Two different cases: One is if my IOC using the rsh, then it
> will say "bus error" and print out bunch of things.

We had a similar csh/tcsh shell related problem that we tracked down to
the PATH variable being limited in length (1024 chars, iirc).

> The other is if my ioc is using the ftp (user name + passwd), then it
> will hang up when it tries to load the VxWorks symbols.

This is due to a bug in the mv162 BSP: Even if you tell the IOC to boot
using ftp, the symbol table still will be loaded using rsh access. (So
you're running into the first case error again.)

> What really bothers me is that this only happens to 162. I have 167s
> as well which are using rsh. It has the same "bus error",

Which definitely makes it look like a host computer csh problem more
than an IOC problem ...

> but when I put in the passwd to use the ftp, then it loads no problem.

So the mv167 BSP doesn't seem to have the mv162 bug. Good.

> Hope someone out there could shed some light on this murky situation
> for me.

Well, I tried ...

Regards,
Ralph

Experimental Physics and Industrial Control System

Experimental Physics and
Industrial Control System