Re: [PATCH] sync_smb2_mid_result

On Wed, 23 Mar 2011 13:26:19 -0500
Steve French [off-list ref] wrote:

On Wed, Mar 23, 2011 at 1:06 PM, Jeff Layton [off-list ref] wrote:>

quoted

Maybe you misunderstand me.

The struct is *exactly the same* aside from a few small differences. It
doesn't even need to be a union, IMO. That would let you use a lot of
the same code in transport.c for smb2.

I don't mind trying to condense it to one structure, but it is not that simple:
- smb2 doesn't need or sequence number, and the two transact2
multiresponse booleans
- cifs mid will grow by at least 10 bytes by expanding the mid, pid
etc. to the required minimum sizes.
- cifs mid doesn't need "async_response_received" or (that is an smb2
only feature and that or equivalent is required).  When we add chained
commands (common in smb2) then complex_mid or equivalent as well as
the command list, the number of the commands received and the last
response time of entries will be useless to cifs.

Is it better to add an entirely new set of functions simply because
you'll have some unused fields in this structure in either case?

I don't think so.

-- 
Jeff Layton [off-list ref]