Re: [lwip-users] patch etharp

[Leon Woestenberg]

Hello Andreas,

I have read both this and your subsequent email.

On Mon, 24 Jan 2005 11:04:21 +0100, "Andreas Becker"
<address@hidden> said:

> I noticed an wrong behavior, if we do an DHCP_DOES_ARP_CHECK with an 
> unconfigured interface.

We only want to pass ARP replies to DHCP when needed (to avoid
unnecessary overhead).

In a strict sense, this is when the DHCP client attached to an interface
is in 'DHCP_CHECKING' state.

For performance reasons, we want to evaluate the strongest condition
first. The strongest
condition, without doubt, is whether an interface is DHCP_CHECKING, not
if the IP address or
Ethernet address of the ARP reply matches (as there is constant ARP
traffic).

I therefore simply removed the redundant condition in etharp.c.

It was wrong for a interface that were reconfigured (as it only passed
when the netif IP address
was zero).

Committed in CVS HEAD.

Leon.