pc386/i386ex ne2000 drivers

[erik.ivanenko at utoronto.ca (erik.ivanenko)]

Ian Lance Taylor wrote:

>    Date:        Fri, 06 Aug 1999 12:09:33 -0700
>    From: "erik.ivanenko" <erik.ivanenko at utoronto.ca>
>
>    The ne2k driver I use is quite old.  It did originally appear with the pc386
>    BSP.  I altered it slightly so that bootp would work properly.  The latest
>    incarnation of the driver does not work at all on my card:
>
>    PCM 3660 PC104 version.  Chipset = RTL8010AS
>
>    It does not even get the ethernet address from the board correctly... bootpc
>    reports that it is using a hw address of FF:FF:FF:FF:FF:FF.
>
>    There were quite a few changes to the code.  Does anyone recall what bugs were
>    being addressed by the changes?
>
>    I wish I could tell you what version I use, but it was back when there was
>    only a "network" subdirectory.  I submitted a patch to make ne2k work with
>    bootp.  That is that last change I was aware of,  since my i386ex BSP sucked
>    up and froze that particular incarnation of the ne2k.c code.
>
> Here are the patches I am using against the file in the 19990528 RTEMS
> snapshot.  This driver does work for me.  However, I do not use bootp.
> I don't think any of these differences would affect the reported
> hardware address.
>
> I do vaguely remember your bootp patches, and I thought they were
> incorporated into the official distribution.
>
> What sorts of changes are you seeing in the code?  Most of the changes
> I'm aware of are the ones Joel mentioned, which added support for byte
> transfers as opposed to word transfers.


Neither the patch sent by Joel, which had the 8 bit stuff, nor this one could be
applied cleanly to my driver source.  In fact, in the patch that you gave me, all
hunks fail, unless I use -R.  Then only half of them fail.   In the patch that Joel
gave me, only 3 fail.

I am certain that the code I use is different from the distribution, in at least one
way.... I do not allow for multiple ne2000 cards.  ie. I removed the typedef of the
ne_softc and also removed the array definition.  This is one major difference between
my code and the rest.

The other huge difference is in the order of initialization.  (May or may not be
important ):

my code:
start daemons
ne_init_hardware
install irq handler

dist:
ne_init_hardware
install irq handler
start daemons.

When I run the distibution, and trigger debug messages, I get an infinite loop
detailing that the: NE200 status is 0xFF (8259 enabled: yes;  mask : FFD2 )


-------------- next part --------------
A non-text attachment was scrubbed...
Name: vcard.vcf
Type: text/x-vcard
Size: 291 bytes
Desc: Card for Erik Ivanenko
Url : http://rtems.rtems.org/pipermail/rtems-users/attachments/19990806/48f24ecc/attachment.vcf