Re: [opencbm-user] PCI parallel port (LPT) card reported to work

[Wolfgang M. <wo...@ne...>]

Hello Antti,


Antti Kultanen schrieb:
> 
> Sorry for bringing up this very old topic, but I wanted to confirm the 
> findings made by Olle Svensson back in February 2011 regarding this 
> parallel port PCI card:

thaaaaank you very much for your reporting. Although the LPT driver
thing is fading out and even if personally thinking (this is not a plan,
but just an option I'm discussing to myself) about stripping down
OpenCBM to an USB-only architecture, I'm still very interested in LPT
support and card compatibility.

> http://www.vscom.de/4409.htm
> 07:02.0 Parallel controller: Oxford Semiconductor Ltd OX9162 Mode 0 (parallel port)
> 
> I am able to transfer d64 images flawlessly from/to 1541 with d64copy 
> using the PCI card.
> 
> For cbmctrl reset/detect/status/dir/download/upload/change I get the 
> same results as Olle did: upload fails, everything else works.

I don't have a proof for this, but I believe that any bus transfer that
does not rely on the Listener hold-off bus state will work, at least
with limited reliability.
In older versions of cbm4win Spiro had support for IRQ-less transaction
mode within cbm4win. Since reliability was always and issue with this
mode I strongly recommended to not support this mode any further and so
it went out. Don't know, if Spiro did disable it somehow or if it was
left as it was at that day.

You may provoke the Listener hold-off with the "detect" command. If the
driver doesn't detect, when you insert a disk and close the drive lever
or if the change is detected only very rarely then this would mean that
Listener hold-off is tried to be supported without using IRQs.

That's the reason why some testers may believie that an arbitrary LPT
card may work.


In 2010 Frederic Briere made a massive investigation about PCI LPT port
card support in Linux. He also found a hardware solution to the problem,
but the end of it would have been the invention of another additional
X-cable tape user would have to choose from and furthermore incompatible
to other LPT ports.
Since Frederic found that it is a pure hardware issue of migrating the
ISA bus based LPT port to the PCI bus that OpenCBM would not work with
these, it makes no sense to invent a special PCI-LPT-X-cable since for
the price of such a card and the additional cable you can also buy a
xum1541 cable, if not the ZoomFloppy.


Womo

-- 
  ------ to obtain more infos about me, look up the page ------
    ------ http://www.wmsr.de | wm (at) wmsr (dot) de  ------