I upgraded my kernel (for a variety of reasons) from 20030128 sources 
up to 20030218, then
after discovering the problem, up to 20030301, in case it had been 
fixed.

Appears to be a problem with wireless and ftp, at a guess. tcpdump on 
both ends is
the same, and an scp on the same path is fine. An ftp in the other 
direction is also fine.
Client is a Apple laptop running a D-Link card with the beta 
WirelessDriver under OS X.
wi0 under NetBSD is:

wi0 at pci0 dev 8 function 0: Intersil Prism2.5 Wireless Lan
wi0: interrupting at irq 11
wi0: 802.11 address 00:05:5d:5b:c5:f5
wi0: using RF:PRISM2.5 MAC:ISL3874A(Mini-PCI)
wi0: Intersil Firmware: Primary (1.0.5), Station (1.3.4)
wi0: supported rates: 1Mbps 2Mbps 5.5Mbps 11Mbps

Problem occurs when transferring from laptop to NetBSD (ftp put with 
ftpd on NetBSD). Hosts
on the other end of a tlp0 interface on the same NetBSD system are 
fine. Even stranger, an
ftp run on the NetBSD box connecting to the laptop is also fine (in 
both directions).
tcpdump shows 0.6 - 1.0 second pauses, which looks like a timeout to 
me. 192.168.254.130 is
the laptop, 192.168.254.129 is the NetBSD box.

11:26:40.402495 192.168.254.130.53311 > 192.168.254.129.62023: . 
2897:4345(1448) ack 0 win 33304 <nop,nop,timestamp 600652400 123> (DF) 
[tos 0x8]
11:26:40.600700 192.168.254.129.62023 > 192.168.254.130.53311: . ack 
4345 win 702 <nop,nop,timestamp 124 600652400> (DF)
11:26:41.400813 192.168.254.129.62023 > 192.168.254.130.53311: . ack 
4345 win 2238 <nop,nop,timestamp 125 600652400> (DF)
11:26:41.405995 192.168.254.130.53311 > 192.168.254.129.62023: . 
4345:5793(1448) ack 0 win 33304 <nop,nop,timestamp 600652402 125> (DF) 
[tos 0x8]
11:26:41.407678 192.168.254.130.53311 > 192.168.254.129.62023: . 
5793:7241(1448) ack 0 win 33304 <nop,nop,timestamp 600652402 125> (DF) 
[tos 0x8]
11:26:41.407725 192.168.254.129.62023 > 192.168.254.130.53311: . ack 
7241 win 1814 <nop,nop,timestamp 125 600652402> (DF)
11:26:41.409438 192.168.254.130.53311 > 192.168.254.129.62023: . 
7241:8689(1448) ack 0 win 33304 <nop,nop,timestamp 600652402 125> (DF) 
[tos 0x8]
11:26:41.414399 192.168.254.130.53311 > 192.168.254.129.62023: . 
8689:10137(1448) ack 0 win 33304 <nop,nop,timestamp 600652402 125> (DF) 
[tos 0x8]
11:26:41.414447 192.168.254.129.62023 > 192.168.254.130.53311: . ack 
10137 win 366 <nop,nop,timestamp 125 600652402> (DF)
11:26:42.410917 192.168.254.129.62023 > 192.168.254.130.53311: . ack 
10137 win 1902 <nop,nop,timestamp 127 600652402> (DF)
11:26:42.416212 192.168.254.130.53311 > 192.168.254.129.62023: . 
10137:11585(1448) ack 0 win 33304 <nop,nop,timestamp 600652404 127> 
(DF) [tos 0x8]

Tell me I'm dreaming. I don't understand this. "wiconfig wi0 -o" isn't 
reporting anything
out of the ordinary.

BTW: I've also noticed that my wi0 appears to be asymmetric... I get 
approx 180 KB/s in one
direction (NetBSD sending) and 500+ KB/s in the other. ifconfig wi0 
reports:

wi0: flags=8863<UP,BROADCAST,NOTRAILERS,RUNNING,SIMPLEX,MULTICAST> mtu 
1500
         nwid xxxxxxxx
         powersave off
         bssid 00:05:5d:5b:c5:f5 chan 6
         address: 00:05:5d:5b:c5:f5
->      media: IEEE802.11 DS11 hostap (DS2 hostap)
         status: active
         inet 192.168.254.129 netmask 0xffffff80 broadcast 
192.168.254.255
         inet6 fe80::205:5dff:fe5b:c5f5%wi0 prefixlen 64 scopeid 0x1
         atalk 101.1 range 101-101 phase 2 broadcast 101.1

2 Mbit sounds about right. What happened to DS11?

Cheers,
--
Paul Ripke
Unix/OpenVMS/TSM/DBA
101 reasons why you can't find your Sysadmin:
68: It's 9AM. He/She is not working that late.
-- Koos van den Hout