Re: [PATCH RESEND 2/2] xen: Allow EXPERT mode to be selected from the menuconfig directly

[Jan Beulich <jbeulich@xxxxxxxx>]

On 12.05.2020 13:05, George Dunlap wrote:
> 
> 
>> On May 12, 2020, at 12:03 PM, Jan Beulich <jbeulich@xxxxxxxx> wrote:
>>
>> [CAUTION - EXTERNAL EMAIL] DO NOT reply, click links, or open attachments 
>> unless you have verified the sender and know the content is safe.
>>
>> On 12.05.2020 13:00, Julien Grall wrote:
>>> Hi Jan,
>>>
>>> On 12/05/2020 11:15, Jan Beulich wrote:
>>>> On 12.05.2020 12:08, Julien Grall wrote:
>>>>> On 12/05/2020 08:18, Jan Beulich wrote:
>>>>>> On 11.05.2020 19:14, Ian Jackson wrote:
>>>>>>> Jan Beulich writes ("Re: [PATCH RESEND 2/2] xen: Allow EXPERT mode to 
>>>>>>> be selected from the menuconfig directly"):
>>>>>>>> I'm trying to make the point that your patch, to me, looks to be
>>>>>>>> trying to overcome a problem for which we have had a solution all
>>>>>>>> the time.
>>>>>>>
>>>>>>> Thanks for this clear statement of your objection.  I'm afraid I don't
>>>>>>> agree.  Even though .config exists (and is even used by osstest, so I
>>>>>>> know about it) I don't think it is as good as having it in
>>>>>>> menuconfig.
>>>>>>
>>>>>> But you realize that my objection is (was) more towards the reasoning
>>>>>> behind the change, than towards the change itself. If, as a community,
>>>>>> we decide to undo what we might now call a mistake, and if we're ready
>>>>>> to deal with the consequences, so be it.
>>>>>
>>>>> Would you mind to explain the fall out you expect from this patch? Are
>>>>> you worry more people may contact security@xxxxxxx for non-security issue?
>>>>
>>>> That's one possible thing that might happen. But even more generally
>>>> the likelihood will increase that people report issues without paying
>>>> attention that they depend on their choice of configuration.
>>> I agree that you are going to get more report because there are more 
>>> users to try new things. So inevitently, you will get more incomplete 
>>> report. This is always the downside of allowing more flexibility.
>>>
>>> But we also need to look at the upside. I can see 2 advantages:
>>>     1) It will be easier to try upcoming features (e.g Argo). The more 
>>> testing and input, the more chance a feature will be a success.
>>>     2) It will be easier to tailor Xen (such as built-in command line).
>>>
>>> In both cases, you make Xen more compelling because you allow to 
>>> experiment and make it more flexible. IHMO, this is one of the best way 
>>> to attract users and possible new contributors/reviewers to Xen community.
>>
>> I'm fully aware of the upsides.
>>
>>>> We'll
>>>> have to both take this into consideration and ask back for the
>>>> specific .config they've used.
>>> Correct me if I am wrong, but this is not very specific to EXPERT mode. 
>>> You can already select different options that will affect the behavior 
>>> of the hypervisor. For instance, on x86, you can disable PV guest 
>>> support. How do you figure that out today without asking the .config?
>>
>> I didn't say this is a new problem; I indicated this is going to
>> become more likely to be one.
> 
> I feel like there’s a misunderstanding here — Jan, are you simply
> explaining yourself and/or making sure that we all understand the
> implications of our choice?  Or are you arguing against acceptance
> in an implicitly Nack-ing manner?

The former - it would have seemed impolite if I hadn't replied to
Julien's question.

> I understood Jan to be doing the former; and that as such with
> Ian’s ack, this patch (with the modified commit message) can go in.

Indeed. Looks like I'm the only one anyway to be concerned of the
extra effort.

Jan