On 04/07/2014 07:34 PM, Mike Christie wrote:
<>
> 
> Ccing Boaz to see if there is some easy osd test script to test it out
> and fix.
> 


Hi

up until now I was not aware at all that iscsi responds have AHSs as well.
At the time what I was coding was the addition of two AHSs needed for
OSD that are needed in the *command send*.

the one - is extended_cdb when the CDB spills over the 16 bytes. (struct 
iscsi_ecdb_ahdr)
the 2nd - is the bidi_length when command is bidi (struct iscsi_rlength_ahdr)
          (write_length is in the usual main header, read_length spills into an 
AHS)

but these are outgoing.
(BTW do you support those two in iser I have a very old patch for that, though
 iser code totally changed since then. Not that it would take more than
 couple minutes + testing)

With OSD, all its signatures and extra data segments are encoded into the
return command's payload and initiator extracts those there.

So I'm not sure at all what SCSI command set utilizes AHSs on it's response.
Does anybody know? (except iscsi with the scsi_sense which is kind of an
AHS of sorts). If there was such an AHS I would imagine that some code
at some scsi_initiator would need to see it, no? which means we need
a way to return it to a ULD. Are we just trying for a way to properly
drop it on the floor?

Yes us too in OSD we kind of CAP our sense_len to Linux max it is an
compilation constant. That said it is kind of hard to reach that maximum
it might be produced with erroring OSD PDUs that are not currently used
at all. So in practice this maximum is never reached.

Thanks
Boaz

-- 
You received this message because you are subscribed to the Google Groups 
"open-iscsi" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to open-iscsi+unsubscr...@googlegroups.com.
To post to this group, send email to open-iscsi@googlegroups.com.
Visit this group at http://groups.google.com/group/open-iscsi.
For more options, visit https://groups.google.com/d/optout.

Reply via email to