Re: [PATCH] erofs: introduce nolargefolio mount option

[Chao Yu]

On 3/10/26 15:02, Gao Xiang wrote:
> 
> 
> On 2026/3/10 14:43, Chao Yu wrote:
>> Xiang,
>>
>> On 3/9/26 11:03, Gao Xiang wrote:
>>> Hi Chao,
>>>
>>> (+cc -fsdevel, willy, Jan kara)
>>>
>>> On 2026/3/9 10:30, Chao Yu wrote:
>>>> This patch introduces a new mount option 'nolargefolio' for EROFS.
>>>> When this option is specified, large folio will be disabled by
>>>> default for all inodes, this option can be used for environments
>>>> where large folio resources are limited, it's necessary to only
>>>> let specified user to allocate large folios on demand.
>>>
>>> For this kind of options, I think more real backgrounds
>>> about avoiding high-order allocations are needed in the
>>> commit message (at least for later reference) also like
>>> what I observed in:
>>> https://android-review.googlesource.com/c/kernel/common/+/3877981
>>
>> Basically, the background is about contention scenario on large folio allocation,
>> it's among multiple users including EROFS in Android-system, as it's related to
>> internal scene of product, so I can not provide more details now, I'm sorry
>> about that, but I'm glad to discuss based on the background and pain point once
>> if I can share more, let's see. :)
> 
> Understood, but I think it's hard to justify an upstream
> solution without a public load for discussion.  Anyway,
> I can imagine some real workloads which large folios could
> cause unnecessary pressure since I once worked for Android,
> but I think others need an explicit one anyway to justify
> this.

Yes,

> 
> As Matthew and Jan mentioned, it's hard to add a per-fs
> knob like this.  If it's Android-specific and no possible
> public infos, I suggest leaving the changes Android
> downstream for now, until the workloads can be made public.

Sure, I can understand that we're not going to accept per-fs change
on large folio policy, as if there is conclusion or agreement about
this in previous discussion from community.

Thanks for the suggestion, I can take a look from downstream side.

Thanks,

> 
> Thanks,
> Gao Xiang