RE: [hpoj-devel] SMP patches?

[Robert G. B. <rg...@ph...>]

On Wed, 29 Nov 2000, PASCHAL,DAVID (HP-Roseville,ex1) wrote:

> > I have a bit of kernel module and smp expertise myself (at
> > least enough
> > to know what a "bottom half handler" is:-) and can possibly even help
> > intelligently in the event that the patch doesn't fix my problem(s).
> All input is welcome.  :-)  If nothing else perhaps you can review the
> changes to ieee12844pp.c and ieee12844.c (particularly the functions
> mlc_sendmsg and mlc_recvmsg in the latter file) and let us know if you see
> any problems.

When I looked I realized that I had actually downloaded and installed
the SMP patch when I first grabbed the software; I just didn't recognize
it from the description.  So it was installed all along.

So I was (and still am) getting the following message from the
mlcpp_transmit() routine even with the patched driver:

Dec  1 18:23:30 lucifer kernel: mlcpp0: mlcpp_transmit while TX busy

This invariably causes the image I'm printing to scramble at that point
and the printer starts trying to print the incoming message as if it is
a binary/ascii data stream (that is, it rips through paper putting all
those funny little ascii block characters).

I still suspect an SMP problem but I'm not certain.  It could either be
that having two processors allows a second transmit interrupt to occur
before the line is fully clean from the first -- not necessarily a
proper spinlock failure but a mix of spinlock failure and handshaking
failure.  Or something.  It could also just be some sort of flow control
failure all by itself -- with megabyte sized images, the printer may be
trying to say "enough" while it processes what it's got but the flow
stop may not be getting through.

Over the weekend I'll try a couple of things:

  a) To make sure that the problem is really likely the SMP box, I'll
run the printer from an almost identical (but UP) box next door.  If
this works I could, of course, just leave things that way but I'd rather
help debug the SMP problem.

  b) If the problem occurs on the SMP box but not the UP box, I'll try
to run the modules with a higher debug mode.  I assume that this traces
the operation fairly tightly.  I have lots of cheap paper and although
the cartridges aren't cheap, I'm willing to burn at least one set
getting the thing to work.

Any helpful suggestions beyond that would be appreciated, especially
from folks who already know why the transmit while TX busy message might
be occurring.  I did already note the new requirement that the printer
port be configured EPC (mine was SPP in the bios) and reset it but it
made no difference.  I'm guessing that was more for scanner function
anyway.  This seems like an overrun failure of the plain old printer
device.

   rgb

-- 
Robert G. Brown	                       http://www.phy.duke.edu/~rgb/
Duke University Dept. of Physics, Box 90305
Durham, N.C. 27708-0305
Phone: 1-919-660-2567  Fax: 919-660-2525     email:rg...@ph...