On Mon, 22 May 2000, Petko Manolov wrote:
> Jason Sharpee wrote:
> >
> > On Mon, 22 May 2000, Petko Manolov wrote:
> >
> > > > I bet you it needs something like this line in pegasus_start_xmit:
> > > >
> > > > warn("%s: Tx timed out. Reseting...", net->name);
> > > > + usb_unlink_urb(&pegasus->tx_urb);
> > > >
> > > > Of course, I know nothing about network drivers under Linux so this is
> > > > just an educated guess.
> > >
> > >
> > > This looks reasonable. I'll try and will post comments ;-)
> >
> > I added this code to start_xmit and tried it out with both the usb-uchi
> > and uhci drivers and my transfer of 750Megs finished. It still
> > timed out a couple of times during transfer, but it completed with an
> > average speed of 370KBytes/sec for usb-uchi and 320KBytes/sec for uhci.
> > For reason the uhci driver caused more timeout messages.
>
>
> This is good. I am ready with patch which i'll send to Randy.
Here is the patch that I used.
*** pegasus.orig Sun May 21 11:39:54 2000
--- pegasus.c Mon May 22 10:59:23 2000
***************
*** 303,308 ****
--- 303,309 ----
if ((res = usb_submit_urb(&pegasus->tx_urb))) {
warn("failed tx_urb %d", res);
pegasus->stats.tx_errors++;
+ usb_unlink_urb(&pegasus->tx_urb);
netif_start_queue(net);
} else {
pegasus->stats.tx_packets++;
> Can you check the transfer with ohci HW. I've got ~600KB/sec
> with ALI chipset (ohci HC) and K6-350. Sounds weird!
Ill check the transfer rate on my OHCI system tonight if I can and let
everyone know. I was having a problem with that RX status of -110 that
prevented a large transfer on my OHCI system as it would just loop
with that message and FTP would timeout. I had put my test machines on
their own 100BT switch, so I dont think its traffic related.
Thanks,
-Jason
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
