On 09/23/08 13:13, James Carlson wrote:
Jim Moore writes:
I'm going through the feedback you provided in idm_so.c. I've
played with the code that skips loopback address while determining
local interfaces and I see an entry named "lo0" that does not have
the IFF_LOOPBACK flag set in the control structure.
Are you _sure_? I don't see how that could happen if you're getting
the flags correctly.
I was wondering if
we could keep our same algorithm for now of matching the name "lo0"
while filing an RFE to investigate this further. The current code was taken
from the user mode target driver.
I don't think the current code is right, and at least for this review,
I don't believe it matters where the code came from.
Filtering out just lo0 doesn't make sense. What will this code do if
it runs into "lo0:1" configured in a zone? Is that one ok where "lo0"
outside of the zone with exactly the same address is somehow not ok?
What will it do if it runs into an interface with the address set to
all zeros and the IFF_UP flag is set (as happens with DHCP)?
Excluding by interface name is questionable at best when what you're
really trying to get is a list of addresses, and I doubt it really
does what's desired. I suspect that if you're trying to filter out
'unusable' source addresses from the list, you'll need to check for at
least INADDR_ANY and INADDR_LOOPBACK, and may want to exclude all of
INADDR_LOOPBACKNET.
Better still, just lose that code. I see no obvious reason to be
trying to skip over the loopback interface.
Okay, I'll look at the other attributes. We need to filter loopback
addresses because
we are going to supply them to the iSCSI initiator via an iSCSI inquiry
or iSNS registration.
thanks,
-jim
_______________________________________________
networking-discuss mailing list
[email protected]
