Re: [PATCH 06/14] mm: handle_pte_fault() use pte_offset_map_maywrite_nolock()

[Qi Zheng <zhengqi.arch@bytedance.com>]

Hi David,

On 2024/8/22 17:29, David Hildenbrand wrote:
> On 21.08.24 12:03, Qi Zheng wrote:
>>
>>

[...]

>>>>>>>> -        vmf->pte = pte_offset_map_nolock(vmf->vma->vm_mm, 
>>>>>>>> vmf->pmd,
>>>>>>>> -                         vmf->address, &vmf->ptl);
>>>>>>>> +        vmf->pte = pte_offset_map_maywrite_nolock(vmf->vma->vm_mm,
>>>>>>>> +                              vmf->pmd, vmf->address,
>>>>>>>> +                              NULL, &vmf->ptl);
>>>>>
>>>>> I think we discussed that passing NULL should be forbidden for that
>>>>> function.
>>>>
>>>> Yes, but for some maywrite case, there is no need to get pmdval to
>>>> do pmd_same() check. So I passed NULL and added a comment to
>>>> explain this.
>>>
>>> I wonder if it's better to pass a dummy variable instead. One has to
>>> think harder why that is required compared to blindly passing "NULL" :)
>>
>> You are afraid that subsequent caller will abuse this function, right?
> 
> Yes! "oh, I don't need a pmdval, why would I? let's just pass NULL, 
> easy" :)
> 
>> My initial concern was that this would add a useless local vaiable, but
>> perhaps that is not a big deal.
> 
> How many of these "special" instances do we have?

We have 5 such special instances.

> 
>>
>> Both are fine for me. ;)
> 
> Also no strong opinion, but having to pass a variable makes you think 
> what you are supposed to do with it and why it is not optional.

Yeah, I added 'BUG_ON(!pmdvalp);' in pte_offset_map_ro_nolock(), and
have updated the v2 version [1].

[1]. 
https://lore.kernel.org/lkml/cover.1724310149.git.zhengqi.arch@bytedance.com/

Thanks,
Qi

>