[222] in linux-net channel archive

home help back first fref pref prev next nref lref last post

1.2.6 and ne2000 driver

daemon@ATHENA.MIT.EDU (Dan Root)
Tue Apr 25 03:28:39 1995

Date: Tue, 25 Apr 1995 01:37:30 -0400
From: Dan Root <refugee@ghost.pc.cc.cmu.edu>
To: linux-net@vger.rutgers.edu

Hello all.

	I just installed the 1.2.6 kernel last night and I'm getting
some real weirdness from my ne2000 clone.  I'm getting quite a number
of messages that look like this:

Apr 25 00:36:48 ghost linux: eth0: Too much work at interrupt, status 0x01
Apr 25 00:37:13 ghost linux: eth0: Too much work at interrupt, status 0x40

Even one or two net active connects can cause a message like this (a
hit on my httpd while I'm reading news almost always seems to do it)
The status 0x40 occur more frequently, on about a 5:2 ratio.  The
other thing is that netstat -i shows:

Kernel Interface table
Iface   MTU Met  RX-OK RX-ERR RX-DRP RX-OVR  TX-OK TX-ERR TX-DRP TX-OVR Flags
lo     2000   0      0      0      0      0    768      0      0      0 BLRU
eth0   1500   0 136856      0      0      0  13043      0      0      0 BRU

which seems to be a rather high number of received packets for a
machine that's only been up for an hour.  I'm fairly certain it's just
my machine and not the local net since my roommate (running 1.2.3 on a
different card) shows about ~14 Million RX-OK and ~6 Million TX-OK
after 26 DAYS of uptime with about 5-6 active users.  At this rate
I'll reach that level of activity in well under a week on machine with
only 1 or 2 active users at a time.

Is it possible that the card is somehow being set into
promiscuous/multicast mode for some reason?  Anyone have any ideas
what I can do to fix/work around this?

Thanks for any help

	Dan


Some other possibly relavant info:

'cat /proc/interrupts' right after the above netstat shows

 0:   633435   timer
 1:    17335   keyboard
 2:        0 + cascade
 4:     2944 + serial
11:   103226   ne
13:        0   math error
14:    12922 + ide0

parts of dmesg:

Console: colour EGA+ 100x40, 1 virtual console (max 63)
Calibrating delay loop.. ok - 33.22 BogoMips
Serial driver version 4.11 with no serial options enabled
tty00 at 0x03f8 (irq = 4) is a 16450
snd3 <Pro AudioSpectrum 16 rev 0> at 0x388 irq 10 drq 3
hda: WDC AC2420H, 405MB w/128KB Cache, CHS=989/15/56, MaxMult=16
hdb: WDC AC2250, 243MB w/63KB Cache, CHS=989/9/56, MaxMult=16
ide0: primary interface on irq 14
...
Memory: 19252k/20480k available (556k kernel code, 384k reserved, 288k data)
This processor honours the WP bit even when in supervisor mode. Good.
...
NE*000 ethercard probe at 0x340: 00 40 95 26 bc d8
eth0: NE2000 found at 0x340, using IRQ 11.
ne.c:v1.10 9/23/94 Donald Becker (becker@cesdis.gsfc.nasa.gov)
...
Linux version 1.2.6 (root@ghost) (gcc version 2.5.8) #1 Sun Apr 23 22:44:18 EDT 1995

-- 
Dan Root - refugee@cmu.edu - Linux Hacker and Anime Junkie

PGPmail prefered - "It's like the envelope for your email"
Finger refugee@cmu.edu for my key.  Also on the keyservers

home help back first fref pref prev next nref lref last post