Re: [PATCH 11/25] scsi: move get_scsi_ml_byte() to scsi_priv.h

From: Niklas Cassel <Niklas.Cassel@wdc.com>
To: Mike Christie <michael.christie@oracle.com>
Cc: "James E.J. Bottomley" <jejb@linux.ibm.com>,
	"Martin K. Petersen" <martin.petersen@oracle.com>,
	Hannes Reinecke <hare@suse.de>,
	"linux-scsi@vger.kernel.org" <linux-scsi@vger.kernel.org>
Subject: Re: [PATCH 11/25] scsi: move get_scsi_ml_byte() to scsi_priv.h
Date: Thu, 29 Dec 2022 20:19:17 +0000	[thread overview]
Message-ID: <Y632RMabE9ansMyy@x1-carbon> (raw)
In-Reply-To: <8319d269-42b5-80ef-f103-ae9537d4f6b2@oracle.com>

On Thu, Dec 29, 2022 at 12:55:47PM -0600, Mike Christie wrote:
> On 12/28/22 2:41 PM, Niklas Cassel wrote:
> > I understand that you intentionally didn't put the ML byte getter in scsi.h,
> > as you intentionally didn't want a LLDD to be able to get the SCSI ML byte.
> > 
> > However, isn't the important thing that that a LLDD can't *set* the SCSI ML
> > byte?
> > Does it really matter if a LLDD can *get* the SCSI ML byte?
> 
> We don't want LLDDs to get the ml byte and do something with it because I think
> it's kind of just a hack to avoid parsing the result/sense multiple times, and
> the drivers have no need for accessing it.
> 
> If you put it in scsi_cmnd.h they can still do:
> 
> if (get_scsi_ml_byte(cmd))
> 	do something;
> 
> which we don't want.
> 

Ok, when I send a V2 of this patch, it will be identical to this patch,
with the exception that get_scsi_ml_byte() will be renamed to
scsi_ml_byte(). (Parameters will be the same.)

Kind regards,
Niklas