[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
another padding question ...
Folks,
I've reworded the padding field text as per my earlier message, to
make padding content algorithm/mode specific. However, Bill also has
suggested that padding be used to ensure that the Auth Data field, if
present, be aligned on an 8-byte boundary. Previously published transform
I-Ds for ESP do not call for 64 bit alignment, but rather refer to
"approrpiate alignment" when discussing padding. The current version of
the ESP I-D calls fo the Next Header field to end on 4-byte boundary,
matching the diagram in the I-D, irrespective of the presence of the Auth
Data field. I can see a good motivation for 4-byte alignment here, in any
case, but 8-byte alignment seems less of a critical issue. It's just
another field within the ESP format and we don't generally require such
alignment for each of our protocol fields. In fact, the suggestion of
requiring 8-byte alignment for the start of the Payload was not adopted.
So, shall we sitck with the 4-byte alignment that has been called for
previously, or is there a strong sentiment to up it to 8-bytes?
Steve