Some problems with 2.3.29
vasil@eos.dobrich.net
Fri, 3 Dec 1999 13:27:37 +0200 (EET)
Hello,
I've been running a P III Xeon machine with 51MB ramfor some time with
2.2.13 kernel,and eepro100 NIC.When i tried to switch to 2.3.29 kernel(I
wanted to have more processes avialable,and I 've heard that 2.3.29's
networking is faster (and of course ,the filesystem is faster than in
2.2.x)) the card started giving me overruns and error.I wrote Don Becker
about it,and he said sth like 'Ask Linus,the driver in 2.3.xx is
modified'.
I changed the NIC with 3COM 905B-TX and the problem reappeared in even
stranger form.There were some messages about TCP/IPv4 checksums like
this:
Dec 2 20:54:01 horny kernel: TCPv4 bad checksum from 194.217.194.139:9bc5
to 216.65.123.66:0077, len=24/44
Dec 2 20:56:02 horny kernel: TCPv4 bad checksum from 194.217.194.139:b695
to 216.65.123.66:0050, len=24/44
And because this machine is 15K km from me,the only thing I see when the
machine goes down (which happens when I start apache) is the followiung (i
use debian 2.2.potato):
Starting apache
and even the dot doesn't appear...When i check ,the TCP connection is
running (e.g. the server ACKs my packets),but she sshd at the outer side
doesn't echo my characters..it just dies...The machine still responds to
pings,etc,but nothing else.The machine is under constant load of about 3-4
Mbit/s..
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/