Re: [PATCH] scsi: sd: Avoid sending an INQUIRY if the page is not supported

[Douglas Gilbert <dgilbert@interlog.com>]

On 2023-05-11 12:44, Benjamin Block wrote:
> On Mon, May 08, 2023 at 09:34:15AM -0700, Brian Bunker wrote:
>>> On May 8, 2023, at 3:09 AM, Benjamin Block <bblock@linux.ibm.com> wrote:
>>> On Fri, May 05, 2023 at 01:49:50PM -0700, Brian Bunker wrote:
>>>
>>>> + int ret = -EINVAL;
>>>
>>> Been wondering, whether it would make sense to have two different error
>>> levels here. One for the case where the page is not found in the loop
>>> that searches within page 0, and one for when page 0 is not present when
>>> we try to dereference the RCU protected pointer.
>>>
>>> That way we could have a safe fallback. If the page is there, we use its
>>> data, if it is not, we blindly send the INQUIRY like we do today.
>>>
>>> Not sure whether this is a bit too paranoid.. VPD page 0 is mandatory
>>> after all.
>>
>> That could be done, but the problem would still exist for the PURE target.
>> We don’t support the page 0xb9, and we don’t advertise we do in the response
>> to VPD 0. This approach would still lead to the INQUIRY being sent to devices
> 
> I wasn't meaning to send the INQUIRY regardless to what the page says,
> if it is present. I was just thinking to having fall-back for when the
> page 0 is not there at all (initially, when you call
> `rcu_dereference()`). That would support your storage, as you have page
> 0, and it would be present for the check I assume.
> 
> But anyway, it seems this is a no-go regardless. I didn't expect targets
> sending a valid page 0, but still supporting pages that are not listed in it.

SCSI standards did contain words that implied vendors would report all
VPD/log/mode pages and operation codes using the mechanisms added to
those standards over the years. I guess users pressured vendors to do
this. Then recently, in T10 approved proposal 21-063r2, the weasel
words "may or may not" were added to all those mechanisms. IMO that
was a pretty disgraceful move by the storage vendors that control T10.

Doug Gilbert


P.S. I have a new option appearing in some of my utilities: --examine
The idea is that it uses an iota function to check for unreported
VPD/log/mode pages as the "report supported" device server responses
"may or may not" be trustworthy :-)

>> who don’t support it, don’t expect it, and report an unexpected error. What I am
>> trying to avoid is the INQUIRY being sent to devices who don’t invite it.
>