Christoph,

Qlogic was asked to support other Target Stack (SC ST) using the same 
qla2xxx.ko upstream driver.  I had mentioned this during the last 2015 LSF get 
together.  Hopefully, that gives a better context.  The understanding that came 
away was “Qlogic do what is needed to provide this support without creating any 
dead code”.

As for the cleanup patch(es), we fail to mention that conversation.  The patch 
attempts to remove any knowledge of TCM’s specific symbol and try to formalize 
the interface.  Qla2xxx driver will behave in a more generic manner when it 
comes to Target Mode.  All TCM knowledge will be confined to the tcm_qla2xxx 
driver before communicating with QLA driver.   The same goes for other target 
stack.

The patches look like its copying a lot of fields without doing a whole lot, 
instead the patches are translating the request from tcm_qla2xxx to qla2xxx.

As for T10-PI change in this series, the goal were the same : 1) remove TCM 
knowledge/cleanup, 2) additional bug fixes were needed while testing with other 
stacks. 

Regards,
Quinn Tran

-----Original Message-----
From: <[email protected]> on behalf of "Madhani, Himanshu" 
<[email protected]>
Date: Thursday, December 15, 2016 at 11:29 AM
To: Christoph Hellwig <[email protected]>
Cc: target-devel <[email protected]>, Nicholas Bellinger 
<[email protected]>, "Malavali, Giridhar" <[email protected]>, 
"[email protected]" <[email protected]>
Subject: Re: [PATCH 08/22] qla2xxx: Refactore target code to remove symbol 
dependency
Resent-From: <[email protected]>

    Hi Christoph, 
    
    
    
    On 12/14/16, 1:18 PM, "Christoph Hellwig" <[email protected]> wrote:
    
    > - the new qla2x00_free_fcport is entirely pointless, please drop
    >   that part of the patch (and even if it wasn't pointless it should
    >   have been a patch on it's own)
    > - please use struct names and not typedefs for your new structures
    > - pretty much avery single items in your list should be a separate
    >   patch.  And some of them are actively counterproductive and should
    >   be dropped:
    >
    >        - Remove direct access of scsi_status field in se_cmd
    >   - Remove se_cmd from qlt_do_ctio_completion
    >   - Remove se_cmd access in srr code section
    >   - Move se_cmd struct outside of qla_tgt_cmd/qla_tgt_mgmt_cmd.
    
    We combined the patches which includes fixes, enhancements and cleanups to 
    Support multiple target stacks, reviews from customer, and to improve code
    maintainability. This patch tries to organize code logically which is spread
    across qla_target and tcm_qla2xxx. We will split this patch into multiple 
    subset and resubmit the series. 
    
    Thanks, 
    - Himanshu
    >
    
?�{.n�+�������+%��lzwm��b�맲��r��zX��??�(��?��ܨ}���Ơz�&j:+v���?����zZ+��+zf���h���~����i���z�?�w���?����&�)ߢ?f

Reply via email to