Re: [PATCH v1] mm/gup: remove (VM_)BUG_ONs

[David Hildenbrand <david@redhat.com>]

On 06.06.25 20:31, John Hubbard wrote:
> ...
>>>>> David, what do you think?
>>>>
>>>> Well, in this patch here I deliberately want _ONCE for the unpin sanity
>>>> checks. Because if they start happening (IOW, now after 5 years observed
>>>> for the first time?) I *absolutely don't* want to get flooded and
>>>> *really* figure out what is going on by seeing what else failed.
>>>>
>>>> And crashing on VM_BUG_ON() and not observing anything else was also not
>>>> particularly helpful :)
>>>>
>>>> Because ... they shouldn't be happening ...
>>>>
>>>> (well, it goes back to my initial point about requiring individual
>>>> decisions etc ...)
>>>>
>>>> Not sure what's best now in the general case, in the end I don't care
>>>> that much.
>>>>
>>>> Roll a dice? ;)
>>>
>>> One last data point: I've often logged onto systems that were running
>>> long enough that the dmesg had long since rolled over. And this makes
>>> the WARN_ON_ONCE() items disappear.
>>
>> I think what would be *really* helpful would be quick access to the very
>> first warning that triggered. At least that's what I usually dig for ... :)
>>
> 
> These two use cases cannot be simultaneously solved, at least not
> perfectly.

I guess we'd have to store the first WARN data separately away, just 
like when tainting the kernel or so. Not sure.

Fortunately, since a VM_BUG_ON() case cannot ever happen,

I mean, also WARN and friends are for things that should never happen ... :)

-- 
Cheers,

David / dhildenb