urtwn on VAX not working properly

[John Klos]

Hi,

I have a VAXstation 4000/90 which is running NetBSD 8 from 15-August-2017.
It has a TURBOchannel adapter and a TC-USB card from Felix Deichmann
(thanks, Felix!), and USB works well.

At first I used a Ralink run* USB wireless device and had no problems
other than the fact that it stuck out very far and made me worry something
might get knocked and damaged.

I switched that for a Realtek urtwn* USB wireless which barely sticks out
of the USB port. It came up, wpa_supplicant negotiated to get on the
network, then dhcpcd was able to get an address. I could ping on the
network.

Any sort of transfer of data larger than a certain size, though, fails. I
could see the packets leave the VAX (well, at least the VAX thought the
data was going out), but nothing appeared at the other end.

dmesg:
https://sage.klos.com/~john/vax_urtwn/dmesg_urtwn
VAX end:
https://sage.klos.com/~john/vax_urtwn/vax0.pcap.gz
Laptop end:
https://sage.klos.com/~john/vax_urtwn/vax1.pcap.gz
ssh log:
https://sage.klos.com/~john/vax_urtwn/ssh_output

Thoughts?

John

--
Posted automagically by a mail2news gateway at muc.de e.V.
Please direct questions, flames, donations, etc. to news-***@muc.de

[Felix Deichmann]

Any sort of transfer of data larger than a certain size [on urtwn(4)], though, fails. I
could see the packets leave the VAX (well, at least the VAX thought the data
was going out), but nothing appeared at the other end.

As you mention other hardware/drivers work better, it seems to be a
problem with urtwn(4), probably only in connection with slhci(4).
Can you test your urtwn(4) with another machine (with a "proper"
OHCI/UHCI/EHCI) to prove the latter?

Before Nick fixed it, umass(4) with slhci(4) had a very similar
problem that small transfers (block sizes) worked, while large ones
broke it after ~1 sec.

slhci(4) is a strange beast, I also have PR kern/51664 open for a USB
10/100 ethernet adapter (axe(4) at slhci(4))...
Nick is probably your first address here, he also has ISA hardware. :)
http://www.flxd.de/isa-ws-sl811/

Felix

--
Posted automagically by a mail2news gateway at muc.de e.V.
Please direct questions, flames, donations, etc. to news-***@muc.de

[John Klos]

Post by Felix Deichmann
As you mention other hardware/drivers work better, it seems to be a
problem with urtwn(4), probably only in connection with slhci(4).
Can you test your urtwn(4) with another machine (with a "proper"
OHCI/UHCI/EHCI) to prove the latter?

Apologies - I postponed this when I thought I sent it.

I've tested that very urtwn and other urtwn on other machines (amd64,
evbarm) with occasional issues (sometimes I get messages about firmware
not loading), but usually they work fine.

Post by Felix Deichmann
Before Nick fixed it, umass(4) with slhci(4) had a very similar
problem that small transfers (block sizes) worked, while large ones
broke it after ~1 sec.
slhci(4) is a strange beast, I also have PR kern/51664 open for a USB
10/100 ethernet adapter (axe(4) at slhci(4))...
Nick is probably your first address here, he also has ISA hardware. :)
http://www.flxd.de/isa-ws-sl811/

Thanks!

John

--
Posted automagically by a mail2news gateway at muc.de e.V.
Please direct questions, flames, donations, etc. to news-***@muc.de