[Linux-parport] OxPCIe952 EPP

[Matwey V. Kornilov]

Hi,

Many thanks for your comment. It seems that parport_pc needs some quirk
detection for such broken Oxford chips. I have no idea how it could be
detected. Fortunately, I don't need to use LPT anymore and this is not
an issue for me anymore. :)


13.04.2016 04:06, Jon Elson пишет:
> I saw Matwey Kornilov's post on linux-parport from Dec 2014 about the
> Oxford 952 chip.
> 
> I have been selling motion control boards using the parallel port in EPP
> mode since 2001.
> I have been using roughly the same code to drive the parallel port all
> that time, from
> ISA multi-I/O chips through PCI to PCIe.  I have only tried a couple
> PCIe parports.
> I just discovered the same problem he had with reading at the PC end.  I
> did use a logic analyzer to observe the problem.  To read, bit 5 of the
> ctrl reg is set to 1, and then the parport chip will not produce a pulse
> on the DATASTB/ pin when the EPP data port register is read with an INB
> instruction.
> 
> I don't know how Oxford Semi. could have screwed up this badly! They
> have been making PCI multi-I/O chips for at least a decade and they all
> worked.  I used to recommend SIIG PCI parallel ports as the most
> reliable, and they all used Oxford chips.  As far as I can tell, these
> PCIe chips will not work with any printer or scanner that actually uses
> EPP mode, and needs to ID the printer.
> 
> I'm amazed that there aren't more people having problems with this.
> 
> I found the Syba boards with the NetMos 9900 chip work fine with my code
> and my boards.
> 
> Jon