Date: Fri, 8 Mar 2002 14:55:53 -0500 From: The Anarcat <anarcat@anarcat.dyndns.org> To: stable@freebsd.org Subject: (fwd) Re: kern/35640: heavy collision rate hangs vr network interface Message-ID: <20020308195553.GA547@lenny.anarcat.dyndns.org>
next in thread | raw e-mail | index | archive | help
--AhhlLboLdkugWU4S
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
Content-Transfer-Encoding: quoted-printable
Hi.
I'm currently having curious problems with my network configuration and
I wanted to know if anyone ever stumbled upon similar problems.
The setup is simple: a client workstation networked with a ftp server
via a hub. The hub is 10baseT/UTP.
Client NIC: rl0: <D-Link DFE-530TX+ 10/100BaseTX> port 0x9400-0x94ff mem
0xd6800000-0xd68000ff irq 9 at device 9.0 on pci0
Server NIC: vr0: <VIA VT6102 Rhine II 10/100BaseTX> port 0x6600-0x66ff
mem 0xe1001000-0xe10010ff irq 10 at device 11.0 on pci0
Both are capable of:
media autoselect
media 100baseTX mediaopt full-duplex
media 100baseTX
media 10baseT/UTP mediaopt full-duplex
media 10baseT/UTP
media 100baseTX mediaopt hw-loopback
Now to the problem.
Initially, when I connected the client to the network, it wasn't
"autoselecting" properly. I have to manually set it to 10baseT/UTP. Same
with the server.
Now, when I make massive FTP uploads or downloads, I get massive
collisions counts:
client:
input (Total) output
packets errs bytes packets errs bytes colls
467 0 30822 691 0 1061314 709
server:
input (Total) output
packets errs bytes packets errs bytes colls
711 0 1073558 512 0 31396 439
Note that for some reason, the collisions count is different on the
server and the client.
After around 30 to 60 seconds, the vr0 interface on the server hangs and
all network connections hang. I have to ifconfig down/up the interface
to bring back network capabilities on the NIC.
I then tried to change wires, shuffle the ports in the hub, without any
luck. I tried to switch both NICs to 100baseTX and
100baseTX/full-duplex, which was pretty hopeless since the hub is
10baseT. NIC hangs with console message:
vr0: watchdog timeout
I tried to connect the 2 machines together with a crosslink, no
luck either.=20
After a followup on PR kern/35435 (rl driver generates large amount of
collisions), I patched my kernel to fix autonegotiation on the rl NIC.
No change. Collision rate still high.
I then tried to connect the machines with a crosslink again. I then
discovered that autonegotiation was indeed fixed on the rl, but
surprise! the media selected was 100baseTX!
Then transfers were marvelously fast and without a single collision.
input (Total) output
packets errs bytes packets errs bytes colls
2027 0 133782 2382 0 3605108 0
Great. But my hub is a 10baseT and I have a 3rd machine on this network.
So I'm stuck with the 700 colls/s and the freezes. Fortunatly, the
server usually doesn't get that much load. Such load only comes from the
internal network (10baseT) since the network external link (ADSL)
shouldn't pump more than 100K/s.=20
I suspect a bug in the vr driver *and* maybe a bug in the rl driver.=20
If anyone has seen this behavior, please speak up.
A>
--AhhlLboLdkugWU4S
Content-Type: application/pgp-signature
Content-Disposition: inline
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.0.6 (FreeBSD)
Comment: For info see http://www.gnupg.org
iEYEARECAAYFAjyJF0gACgkQttcWHAnWiGeaPACghq3AxjSncJyBxqcsXucrtwv8
u8kAoJeH9dJgZNpMh88T3Bo2H1gcuUNC
=DuiZ
-----END PGP SIGNATURE-----
--AhhlLboLdkugWU4S--
To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-stable" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20020308195553.GA547>