Problem w/0.89H [long]

B. James Phillippe
Wed Jul 1 18:02:16 1998

On Wed, 1 Jul 1998, Robert G. Brown wrote:

> Did you try the ioport patch that I posted yesterday?  It should do a
> scan of each possible ioport in the list returned by the PCI bus to
> find out which one is writable, in the event that the first one (the
> only one checked in 0.89H) is not.  It "might" be that this is your
> problem here, since the register contents are vaguely reminiscent of
> what I see when the card initializes on an unwriteable ioport.

Hi Robert,

	I found your patch on a mailing-list archive, and although badly
mangled by a conversion to HTML, I was able to hand-apply it.  It did not
solve the problem as I had hoped.  It did cause the driver act differently
though, which could be a mixed blessing.  The module loads, finds the same
ethernet card it found before, but prints no diagnostic messages of any
kind (not even the usual "tulip.c:v0.89G ...".  It is quite conceivable
that I botched the patch.  Any other ideas?  Again, the de4x5.c:0.536
driver (included with 2.1.106) works flawlessly.  However, I can't use that
driver in 2.0.x without backporting it (the 2.0.x driver doesn't find any
devices).  Oh well, perhaps I have my weekend cut out for me...

Thanks for the suggestion,
B. James Phillippe <>
Linux Software Engineer, WGT Inc.