Date: Sat, 11 Jan 2003 02:09:52 +0700 (KRAT) From: Eugene Grosbein <eugen@grosbein.pp.ru> To: FreeBSD-gnats-submit@FreeBSD.org Subject: kern/46955: fxp(4) driver pollute logs and console with timeout messages Message-ID: <200301101909.h0AJ9qk2000621@grosbein.pp.ru>
next in thread | raw e-mail | index | archive | help
>Number: 46955
>Category: kern
>Synopsis: fxp(4) driver pollute logs and console with timeout messages
>Confidential: no
>Severity: non-critical
>Priority: low
>Responsible: freebsd-bugs
>State: open
>Quarter:
>Keywords:
>Date-Required:
>Class: sw-bug
>Submitter-Id: current-users
>Arrival-Date: Fri Jan 10 11:20:01 PST 2003
>Closed-Date:
>Last-Modified:
>Originator: Eugene Grosbein
>Release: FreeBSD 4.7-STABLE i386
>Organization:
Private person
>Environment:
System: FreeBSD grosbein.pp.ru 4.7-STABLE FreeBSD 4.7-STABLE #0: Tue Dec 31 15:11:39 KRAT 2002 eu@grosbein.pp.ru:/usr/local/obj/usr/local/src/sys/DADV i386
Intel Pro 10/100 Management NIC connected via crossover UTP cable
with unpowered notebook/PCMCIA ethernet card.
>Description:
I've got Intel Pro 10/100 Management PCI NIC for my desktop
machine to connect it to notebook that has PCMCIA NIC,
so they are connected using crossover UTP cable.
This works just fine when both machines are up and running.
However, sometimes my desktop is booted while notebook is powered off
so there is now carrier on the cable. The cable itself is in place.
This situation is pretty normal to me but not to fxp(4) driver:
it repeatedly fills console and logs with timeout messages:
Jan 11 01:41:50 <kern.crit> grosbein /kernel: fxp0: DMA timeout
Jan 11 01:41:50 <kern.crit> grosbein /kernel: fxp0: DMA timeout
Jan 11 01:41:54 <kern.crit> grosbein /kernel: fxp0: device timeout
Jan 11 01:41:54 <kern.crit> grosbein /kernel: fxp0: DMA timeout
Jan 11 01:41:54 <kern.crit> grosbein last message repeated 2 times
Jan 11 01:41:56 <kern.debug> grosbein /kernel: /dev/vmmon: Module vmmon: registe
Jan 11 01:41:56 <kern.debug> grosbein /kernel: /dev/vmmon: Module vmmon: initial
Jan 11 01:42:06 <kern.crit> grosbein /kernel: fxp0: command queue timeout
Jan 11 01:42:26 <kern.crit> grosbein /kernel: fxp0: device timeout
Jan 11 01:42:26 <kern.crit> grosbein /kernel: fxp0: DMA timeout
Jan 11 01:42:26 <kern.crit> grosbein last message repeated 2 times
Jan 11 01:42:38 <kern.crit> grosbein /kernel: fxp0: command queue timeout
Jan 11 01:42:58 <kern.crit> grosbein /kernel: fxp0: device timeout
Jan 11 01:42:58 <kern.crit> grosbein /kernel: fxp0: DMA timeout
Jan 11 01:42:58 <kern.crit> grosbein last message repeated 2 times
Jan 11 01:44:15 <kern.crit> grosbein /kernel: fxp0: command queue timeout
Jan 11 01:44:35 <kern.crit> grosbein /kernel: fxp0: device timeout
Jan 11 01:44:35 <kern.crit> grosbein /kernel: fxp0: DMA timeout
Jan 11 01:44:35 <kern.crit> grosbein last message repeated 2 times
Jan 11 01:44:47 <kern.crit> grosbein /kernel: fxp0: command queue timeout
Jan 11 01:45:07 <kern.crit> grosbein /kernel: fxp0: device timeout
Jan 11 01:45:07 <kern.crit> grosbein /kernel: fxp0: DMA timeout
Jan 11 01:45:07 <kern.crit> grosbein last message repeated 2 times
Jan 11 01:45:19 <kern.crit> grosbein /kernel: fxp0: command queue timeout
Jan 11 01:45:39 <kern.crit> grosbein /kernel: fxp0: device timeout
Jan 11 01:45:39 <kern.crit> grosbein /kernel: fxp0: DMA timeout
Jan 11 01:45:39 <kern.crit> grosbein last message repeated 2 times
Jan 11 01:45:51 <kern.crit> grosbein /kernel: fxp0: command queue timeout
Jan 11 01:46:11 <kern.crit> grosbein /kernel: fxp0: device timeout
Jan 11 01:46:11 <kern.crit> grosbein /kernel: fxp0: DMA timeout
And so on. It does not stop. Note there is no process trying
to use LAN meantime.
My kernel does not have fxp(4) so I load this driver
using if_fxp_load="YES" in /boot/loader.conf
Here is how interface is configured in /etc/rc.conf:
ifconfig_fxp0="inet 10.58.1.1 netmask 255.255.255.0 media 10baseT/UTP"
Note that media is set explicitly to disable full duplex,
notebook's NIC is 3C549D driven with ep(4) that does not
support full duplex currently. And fxp(4) wants to use
full duplex in this configuration by default.
Here is the interface state:
# ifconfig fxp0
fxp0: flags=8843<UP,BROADCAST,RUNNING,SIMPLEX,MULTICAST> mtu 1500
inet 10.58.1.1 netmask 0xffffff00 broadcast 10.58.1.255
ether 00:90:27:a7:5c:72
media: Ethernet 10baseT/UTP
status: no carrier
>How-To-Repeat:
Get Intel Pro 10/100 Management PCI NIC, connect a cable,
do not connect another end of cable, configure fxp(4)
(see description above) and reboot a computer.
power on a computer.
>Fix:
Unknown to me. The workaround is to bring interface down
using ifconfig(8). A kernel stops to complain about timeouts then.
>Release-Note:
>Audit-Trail:
>Unformatted:
To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-bugs" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?200301101909.h0AJ9qk2000621>