virtualbox.org

Situation:
VirtualBox 2.2.0
Linux Host (Slackware-based with self-compiled stock kernel)
Windows 2000 Guest
Bridged networking
Guest gets IP-Nr via DHCP

Everything works fine with Linux kernel 2.6.28.

But with a 2.6.29 kernel the direct network connections between guest and host are extremely slow and not usable.
Examples:
A longer output in a terminal window comes with great pauses and eventually stalls.
Filesystem connections to a samba server can't display large directory listings, in the samba log files I see timeout errors and frequent server restarts.

Astonishingly at the same time connections from the guest to other remote hosts work without problems.

Any ideas where to look and whom to blame?

JRO

burghj wrote:Situation:
VirtualBox 2.2.0
Linux Host (Slackware-based with self-compiled stock kernel)
Windows 2000 Guest
Bridged networking
Guest gets IP-Nr via DHCP

But with a 2.6.29 kernel the direct network connections between guest and host are extremely slow and not usable.

JRO

Additional remarks:

I think there must be at least one more factor involved, because with the same configuration on my home system the problem does not occur.

Main difference that I see are the real network chips on the hosts.
On the working system I have some Realtek gigabit chip, on the non-working an Intel chip, namely "Intel Corporation 82566DM-2 Gigabit Network Connection".

So it could well be possible that some "unexpected feature" in the Linux network driver (e1000e) is the cause for the problem.

JRO

burghj wrote:

burghj wrote:Situation:
VirtualBox 2.2.0
Linux Host (Slackware-based with self-compiled stock kernel)
Windows 2000 Guest
Bridged networking
Guest gets IP-Nr via DHCP

But with a 2.6.29 kernel the direct network connections between guest and host are extremely slow and not usable.

JRO

Additional remarks:

I think there must be at least one more factor involved, because with the same configuration on my home system the problem does not occur.

Main difference that I see are the real network chips on the hosts.
On the working system I have some Realtek gigabit chip, on the non-working an Intel chip, namely "Intel Corporation 82566DM-2 Gigabit Network Connection".

JRO

I've searched a little bit more in the direction "Intel 82566DM-2 chip" and found this from last year:

http://bugzilla.kernel.org/show_bug.cgi?id=9940

Main point there was:

The e1000e driver doesn't strip the ethernet frame crc is used with the
82566DM-2 chip.
This will result, among other things, that bridging doesn't work together with
this chip.

This sounds similar to my current problem. Certainly it is not the same, this was more than a year ago with kernel 2.6.24 and was obviously corrected then. But I think my problem could be in the same class, something with ethernet packets not surviving the path through vbox's netfilter. Any idea how to diagnose such a possible packet loss between host and guest?

JRO

Why not open the bug report and state that it's a regression in 2.6.29? It wouldn't be the first time they fix something in a certain release, only to break it again a few releases later.

This is now outside the scope of this forum, so I'm locking it.