* Fw: [Bugme-new] [Bug 3657] New: downed interfaces acting as aliases
@ 2004-10-29 6:46 Andrew Morton
2004-10-29 6:45 ` David S. Miller
0 siblings, 1 reply; 6+ messages in thread
From: Andrew Morton @ 2004-10-29 6:46 UTC (permalink / raw)
To: netdev; +Cc: mbm
Begin forwarded message:
Date: Thu, 28 Oct 2004 23:45:02 -0700
From: bugme-daemon@osdl.org
To: bugme-new@lists.osdl.org
Subject: [Bugme-new] [Bug 3657] New: downed interfaces acting as aliases
http://bugme.osdl.org/show_bug.cgi?id=3657
Summary: downed interfaces acting as aliases
Kernel Version: 2.6.8.1
Status: NEW
Severity: normal
Owner: shemminger@osdl.org
Submitter: mbm@alt.org
An interface will still reply while down, acting as an ip alias if any other
adapters have a compatible subnet.
Scenerio:
Using a laptop on a wired network, 192.168.1.0/24 subnet, eth0 set to
192.168.1.2. Moved to a different network on wireless, also a 192.168.1.0/24.
The ethernet is still configured for 192.168.1.2 but the interface is down
(ifconfig eth0 down); wireless clients on the second network are now having
issues, 192.168.1.2 requests are going to the laptop.
To confirm the issue, eth0 is changed to 192.168.1.254 and brought down again
(an ip not in use on the wireless network). All wireless clients are magically
able to ping 192.168.1.254.
Steps to reproduce:
ifconfig eth0 10.1.2.3 (or any random nonroutable ip)
ifconfig eth0 down
ping 10.1.2.3
.. wtf, how did it get a reply from a down'd interface?
------- You are receiving this mail because: -------
You are on the CC list for the bug, or are watching someone who is.
^ permalink raw reply [flat|nested] 6+ messages in thread* Re: Fw: [Bugme-new] [Bug 3657] New: downed interfaces acting as aliases
2004-10-29 6:46 Fw: [Bugme-new] [Bug 3657] New: downed interfaces acting as aliases Andrew Morton
@ 2004-10-29 6:45 ` David S. Miller
2004-10-29 7:10 ` Ben Greear
0 siblings, 1 reply; 6+ messages in thread
From: David S. Miller @ 2004-10-29 6:45 UTC (permalink / raw)
To: Andrew Morton; +Cc: netdev, mbm
People complain about this all the time, yet it is expected
and intentional behavior.
Just because you "down" the interface does not mean that
addresses assigned to that interface are no longer assosciated
with the host system.
If you wish the machine not to respond to any of the interface's
addresses, you must explicitly delete them from the interface
address list.
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: Fw: [Bugme-new] [Bug 3657] New: downed interfaces acting as aliases
2004-10-29 6:45 ` David S. Miller
@ 2004-10-29 7:10 ` Ben Greear
2004-10-29 7:06 ` David S. Miller
0 siblings, 1 reply; 6+ messages in thread
From: Ben Greear @ 2004-10-29 7:10 UTC (permalink / raw)
To: David S. Miller; +Cc: Andrew Morton, netdev, mbm
David S. Miller wrote:
> People complain about this all the time, yet it is expected
> and intentional behavior.
>
> Just because you "down" the interface does not mean that
> addresses assigned to that interface are no longer assosciated
> with the host system.
>
> If you wish the machine not to respond to any of the interface's
> addresses, you must explicitly delete them from the interface
> address list.
Why would you want this behaviour? If it's configured down, it would
seem that the user is trying to tell the system not to use it :)
Ben
--
Ben Greear <greearb@candelatech.com>
Candela Technologies Inc http://www.candelatech.com
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: Fw: [Bugme-new] [Bug 3657] New: downed interfaces acting as aliases
2004-10-29 7:10 ` Ben Greear
@ 2004-10-29 7:06 ` David S. Miller
2004-10-29 7:28 ` Ben Greear
2004-10-29 7:39 ` Andi Kleen
0 siblings, 2 replies; 6+ messages in thread
From: David S. Miller @ 2004-10-29 7:06 UTC (permalink / raw)
To: Ben Greear; +Cc: akpm, netdev, mbm
On Fri, 29 Oct 2004 00:10:58 -0700
Ben Greear <greearb@candelatech.com> wrote:
> Why would you want this behaviour? If it's configured down, it would
> seem that the user is trying to tell the system not to use it :)
Because IP addresses are assosciated with the host, not
a specific interface or link. That is the model that
we've implemented since day one.
According to the RFCs, this is one of several valid models.
People hate it that when there's a decision of whether to
reply to something or not, we do whatever we can to reply
to packets we receive if we can find a way to do so. This
approache increases the likelyhood that two hosts can
communicate successfully. I need not remind people about
how much people dislike our default ARP behavior :-) but
it is done that way for the same reason.
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: Fw: [Bugme-new] [Bug 3657] New: downed interfaces acting as aliases
2004-10-29 7:06 ` David S. Miller
@ 2004-10-29 7:28 ` Ben Greear
2004-10-29 7:39 ` Andi Kleen
1 sibling, 0 replies; 6+ messages in thread
From: Ben Greear @ 2004-10-29 7:28 UTC (permalink / raw)
To: David S. Miller; +Cc: akpm, netdev, mbm
David S. Miller wrote:
> On Fri, 29 Oct 2004 00:10:58 -0700
> People hate it that when there's a decision of whether to
> reply to something or not, we do whatever we can to reply
> to packets we receive if we can find a way to do so. This
> approache increases the likelyhood that two hosts can
> communicate successfully. I need not remind people about
> how much people dislike our default ARP behavior :-) but
> it is done that way for the same reason.
Note I was distictly kind enough not to bring it up either :)
(I have already patched & hacked around that problem some years ago!)
Since you finally accepted the arp fixup, maybe you'd at least let
in a flag to change the default behaviour of downed interfaces too?
Ben
--
Ben Greear <greearb@candelatech.com>
Candela Technologies Inc http://www.candelatech.com
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: Fw: [Bugme-new] [Bug 3657] New: downed interfaces acting as aliases
2004-10-29 7:06 ` David S. Miller
2004-10-29 7:28 ` Ben Greear
@ 2004-10-29 7:39 ` Andi Kleen
1 sibling, 0 replies; 6+ messages in thread
From: Andi Kleen @ 2004-10-29 7:39 UTC (permalink / raw)
To: David S. Miller; +Cc: Ben Greear, akpm, netdev, mbm
> People hate it that when there's a decision of whether to
> reply to something or not, we do whatever we can to reply
> to packets we receive if we can find a way to do so. This
> approache increases the likelyhood that two hosts can
> communicate successfully.
I bet it's actually only a very small minority that is complaining
about this. The silent majority just sees a working network.
-Andi
^ permalink raw reply [flat|nested] 6+ messages in thread
end of thread, other threads:[~2004-10-29 7:39 UTC | newest]
Thread overview: 6+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2004-10-29 6:46 Fw: [Bugme-new] [Bug 3657] New: downed interfaces acting as aliases Andrew Morton
2004-10-29 6:45 ` David S. Miller
2004-10-29 7:10 ` Ben Greear
2004-10-29 7:06 ` David S. Miller
2004-10-29 7:28 ` Ben Greear
2004-10-29 7:39 ` Andi Kleen
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).