Re: [PATCH] maple_tree: add mas_node_count() before going to slow_path in mas_wr_modify()

[JaeJoon Jung]

Hello, Liam.
As you pointed out, I did enough testing in the userspace and in
previous versions of the kernel, but I was surprised that a syntax
error occurred in making newest patch because of my careless mistake.
I will be careful not to make this mistake again.
And I am going to study the low memory situation in kernel space in more detail.
Thank you very much for your detailed answer and explanation.

Best regards,
JaeJoon Jung

On Mon, 3 Jun 2024 at 21:29, Liam R. Howlett <Liam.Howlett@xxxxxxxxxx> wrote:
>
> * JaeJoon Jung <rgbi3307@xxxxxxxxx> [240602 05:06]:
> > Hello, Liam.
> > Thank you very much for the detailed answer and explanation.
> >
> > I tested this patch in userspace.
> > In user space, this phenomenon always occurs when kmem_cache_alloc()
> > is executed to allocate a new node.
>
> This is expected in the userspace test program.  We test the error path
> most frequently, with only a bypass for those who wish to test an
> initial success - such as preallocations.  I was concerned about the
> testing since your first patch had a syntax error with a correction
> quickly after.  It is good news that you were able to find and use the
> maple tree testing framework, though.
>
> > I will try to test it in more detail in kernel space.
>
> If you test in kernel space, you will have to hit a low memory scenario
> to see a difference.  stress-ng would probably help.
>
> > I will also refer to the notes from the email list you shared
> > and send results once a more clear analysis has been made.
>
> I don't think you need to continue with this work as you will find that
> the low memory situation is going to be rare and in a very slow path
> already.
>
> Thanks,
> Liam