I was having a similar problem myself, but it took removing the SPD_DISABLE flag!
Which kernel version/version of your driver are you using? The fact that the
spin_locks aren't there also makes me suspicious. I was using 2.3.99-pre5 at the
time. It also compiles fine under 2.3.99-pre6, but i can't test it since a freind
took my work dongle home to try and get it working on his iOpener. So i'll have to
try it from home tonight...
Mike
On Thu, Apr 20, 2000 at 03:46:34PM -0700, Johannes Erdfelt wrote:
> On Mon, Apr 10, 2000, Michael W Zappe <[EMAIL PROTECTED]> wrote:
> > A new version of the Kawasaki KL5USB101 ethernet driver is out. It's mostly fixes
>to 0.1 to make it work. These include:
> >
> > - No longer hoses your network interfaces! (the autoincrementing ethXXX bug has
>gone away) It was a pretty stupid bug I should have seen off the bat anyway...
> > - Downloads the firmware correctly, and only when needed.
> > - The packet loss has been greatly reduced. (From ~10% to >=2%)
> >
> > I've gotten it working with both uhci.o and usb-uhci.o. I ran it on uhci.o over
>lunch and it didn't drop a single packet. With usb-uhci.o i saw about 1 in 50 total
>get dropped. Also the differences in latency may be of interest to the driver
>writers. I was seeing a 6-11ms latency with uhci.o, and a 1.8-5ms latency with
>usb-uhci.o.
> >
> > Multicast filtering still doesn't work, but i should get to that later this week.
>
> So, I've been working on getting this driver working with my Netgear
> EA-101 but I've run into some problems.
>
> Before I start, I'm curious what tab spacing the driver was coded with? Looks
> like it's twice the size of my tabs :)
>
> The first problem was the superfluous spin_lock_init calls to nonexistant
> variables. Removing them fixed that. I'm kinda surprised this one wasn't
> caught since the driver won't even compile.
>
> The second problem I've run into was a small bug in my UHCI driver
> regarding SPD being turned off on control transfers. Fixed that.
>
> The first kaweth_read_configuration gets called, and we get a short packet
> back. Since SPD is not set, it retries the IN token a couple of times. A
> couple of them receive no reply. Eventually the transfer times out.
>
> Then the control pipe gets unhalted, those transactions work fine, but then
> when the code tries to download the firmware, it doesn't respond to the
> SETUP packet.
>
> If I completely remove the first kaweth_read_configuration, everything works
> fine.
>
> I get the same exact problem with usb-uhci, so I don't think it's something
> with my driver.
>
> Has anyone gotten this working with the 0.2 driver?
>
> JE
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
