Erik Nordmark wrote:
Rao Shoaib wrote:
Attached is a high level description of proposed changes to Solaris
to make it compliant with the latest MIB RFC's.
Thanks to Jim Carlson for his comments on the initial versions of
this draft which helped in refining the design.
Please send in your comments by COB 9/1/2006.
Perhaps I'm misunderstanding this part:
The current per ill ipv6 statistics structure
(mib2_ipv6IfStatsEntry_t) pointed
by ill_ip6_mib will be renamed and will be used to gather ipv4 and ipv6
statistics. New fields will be added to the end of this structure. For
IPv4,
mib2_ip_t (ip_mib) will continue to exist to provide backward
compatibility.
macros will be provided to map new structure and fields to previous
mib2_ipv6IfStatsEntry_t fields.
Shouldn't we have a single structure definition for the statistics,
and then use that both hanging off the ill_t, and also as a single
global one? (The latter is needed for IPv4 for the same reason as it
is needed for IPv6; some errors occur before we know the ill.)
Erik,
Sorry that the text is not clear but we are doing exactly what you have
stated above.
However
A) We will name rename the mib2_ipv6IfStatsEntry_t structure and it's
fields provide macros for older programs to keep accessing the fields.
B) For backward compatibility we can not EOL the current ipv4 stats
structure hence it will exist.
Rao.
Thus I don't see why we'd need to use a mib2_ip_t or macros.
Erik
_______________________________________________
networking-discuss mailing list
[email protected]
