So is there a consensus that can we get rid of the SIOCSLIFINDEX ioctl ?
Allowing someone to change the ifindex by doing SIOCSLIFINDEX when
quite a few interfaces actually use the ifindex to specify the interface to
use seems more likely to cause damage than actually nuking SIOCSLIFINDEX
and "risking" compatibility.
Given that IPMP is undergoing a rehaul, we should be able to
get rid of IPMP / DR / RCM reliance on SIOCSLIFINDEX. Except for this,
are there any other known users of SIOCSLIFINDEX ?
Thirumalai
James Carlson wrote:
Peter Memishian writes:
I'd much rather just admonish people against changing the interface index
in the first place, since there's really no reason to do it -- and indeed,
AFAIK, no one does. So, again, I think all of this is a non-issue.
I have a hard time imagining when SIOCSLIFINDEX (or SIOCSIFINDEX)
could ever be the right thing to do. The existence of it looks to me
like a bug, not a feature. To me, it's roughly equivalent to having a
setpid() interface.
_______________________________________________
networking-discuss mailing list
[email protected]
