Hi Turbo,
Thanks for your patch.
Note that you might want to have a To: and/or Reply-to: header in the
emails you send (apart from your own address), having all recipients in
Bcc: makes it hard to reply and determine if it is a personal email or
sent through the mailing-list (had to check all the headers to notice it
went through target-devel).
More below.
On 11/05/2014 10:01 PM, Turbo Fredriksson wrote:
RFC 3720 states:
c) iSCSI names are composed only of displayable characters. iSCSI
names allow the use of international character sets but are not
case sensitive. No whitespace characters are used in iSCSI
names.
But lio_node effectively nullifies the 'case sensitiveness' by running
'lower()' on the values.
It does not "nullify" anything, it merely does not preserve the initial
formatting. I am saying formatting, as this is what case amounts to in
an IQN, because as you noted RFC3720 states IQN is _not_ case sensitive.
How can you "nullify" something that is not in the first place ? In
effect, according to the RFC, the lower-case IQN spit out by rtslib are
strictly equivalent to their possibly mixed-case counterparts.
What remains is merely a cosmetic issue. We do not preserve the
formatting/style bit (the case). The decision to normalize to lower-case
only was taken because of non-RFC 3720 initiators that were spotted in
the wild, causing issues: these were choking up when the IQN contained
mixed case. Now, as for the specific details, I have to admit I do not
have them out of the top of my mind, and should have maybe left a
comment for posterity in the code.
Maybe Nic will remember something, I think we had a conversation about
that years ago. I think it was a windows initiator problem at the time,
but am not 100% positive.
Remove these, to conform to the RFC.
I fail to see how using an equivalent notation for those IQNs (according
to the RFC) is non conforming to the RFC. It is precisely because the
RFC states that IQNs are not case-sensitive that we can do the
case-lowering.
As the code as is should not cause any problem that I know of, is
equivalent to not lowering the IQNs according to the RFC, and solves
problems we had in the past, it is going to stay as it is. That is,
unless someone can point out to me a specific problem that it cause in
real-world usage.
Again, thanks for your report and feel free to expand your explanations
if you feel I got it wrong.
Best Regards,
--
Jerome
--
To unsubscribe from this list: send the line "unsubscribe target-devel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
