Experimental Physics and
Industrial Control System

Marty Kraimer <[email protected]> · Mon, 06 Mar 2000 08:48:39 -0600

"Peregrine M. McGehee" wrote:
> 
> Dear all,
> 
> One of my mv162 based IOCs has recently started to hang whenever I issue
> a ^x or type reboot on the console. If I don't enable interrupts for a
> specific module (the GreenSprings ipOctalSerial) the reboot proceeds as
> expected.
> 
> Note that this behavior does not seem to affect the operation and data
> collecting of the IOC, only the reboot process.
> 
> Any ideas?
>         Peregrine
> --
> Peregrine M. McGehee    Coordinator, Los Alamos Astrophysics
> (505) 667-3273          MS H820, LANL, Los Alamos, NM 87545

I suspect the following:

1) You are calling rebootHookAdd
2) Your reboot hookroutine turns off interrupts but does not do anything to stop
new write requests directed to ipOctalSerial
3) Something else also is using rebootHookAdd and it's reboot hook routine does
something like taskdelay.
4) A new write request is directed to the ipOctal and started.
5) The taskdelay completes and the reboot occurs. But that ipOctal comes up with
transmit enabled.

As I recall you started with the Hideos driver for ipOctalSerial. So did MPF. In
MPF we found this problem. You may want to compare your version against the
latest MPF version, which can be obtained from:

http://www.aps.anl.gov/asd/people/mrk/epics/modules/serial/mpfSerial/

At the very least you can look at the code for handling soft reboots.

Marty Kraimer

Experimental Physics and Industrial Control System

Experimental Physics and
Industrial Control System