Re: [PATCH v3 3/3] virtio: introduce SUSPEND and RESUME feature

[Cornelia Huck <cohuck@redhat.com>]

On Thu, Jun 26 2025, "Zhu, Lingshan" <lingshan.zhu@amd.com> wrote:

> On 6/26/2025 7:59 PM, Parav Pandit wrote:
>
>>> From: Zhu Lingshan<lingshan.zhu@amd.com>
>>> Sent: 23 June 2025 02:07 PM
>>> @@ -629,6 +632,54 @@ \section{Device Cleanup}\label{sec:General
>>> Initialization And Device Operation /
>>>
>>>   Thus a driver MUST ensure a virtqueue isn't live (by device reset) before
>>> removing exposed buffers.
>>>
>>> +\section{Device Suspend}\label{sec:General Initialization And Device
>>> +Operation / Device Suspend}
>>> +
>>> +If VIRTIO_F_SUSPEND is negotiated, the driver is eligible to suspend the
>>> device by setting the SUSPEND bit in \field{device status} to 1, and the device
>>> SHOULD set the DRIVER_OK bit to 0 once it has been suspended.
>>> +
>> You ignored the inputs.
>> I do not agree to add the "SHOULD" normative wording in the general description area.
>> The reasoning is explained already.
>> Please adapt to the existing style of this spec to keep the normative in requirements section.
>>
>> If VIRTIO_F_SUSPEND is negotiated, the driver is eligible to suspend the
>> device by setting the SUSPEND bit in \field{device status} to 1, and the device
>> sets the DRIVER_OK bit to 0 once it has been suspended.
>>
>> Is this really that hard to write above way?
>
> I believe you totally ignored my replies in the last thread.
> There are no rules forbid using "SHOULD" in any non-normative sections.
> Now here I copy the reply here again:
>
> In the spec section 1.3 Terminology, it says:
>
> The key words “MUST”, “MUST NOT”, “REQUIRED”, “SHALL”, “SHALL NOT”, “SHOULD”, “SHOULD NOT”, “RECOMMENDED”, “NOT RECOMMENDED”, “MAY”, and “OPTIONAL” in this document are to be interpreted as described in [RFC2119] and [RFC8174] when, and only when, they appear in all capitals, as shown here.
>
> RFC 2119 says:
>
> SHOULD This word, or the adjective "RECOMMENDED", mean that there
>     may exist valid reasons in particular circumstances to ignore a
>     particular item, but the full implications must be understood and
>     carefully weighed before choosing a different course.
>
>
> Here this "SHOULD" exactly conforms to the definition.
>
> "SHOULD" has already been used in non-normative sections, for example:
> 2.5.4 Legacy Interfaces: when using the legacy interface, drivers SHOULD read these fields multiple times until two reads generate a consistent result.
>
> The spec does not say: “SHOULD” can only be used in driver or device requirements section.
>
> @MST, we need your input
>

Not MST, but you can have my input.

We refer to the RFC for the definitions of SHOULD and friends, it does
not say where to use them.

The intention is to use them in normative statements only, so that
implementers have clear guidelines on the requirements. The legacy
sections are arguably a special case (for implementers who want to
support legacy implementations.) I don't think we want the all-caps key
words leaking into other sections (any more than what might have
happened already.)