Re: [PATCH] [14/21] HWPOISON: shmem: call set_page_dirty() withlocked page

[Andi Kleen]

> Will it, from now on, be mandatory to dirty pages under PG_lock, and

Only data pages with mapping.

It's not a strict requirement, but if it's not done then in theory
you could not propagate an error if the page gets an memory error
at exactly the wrong time. It's probably only worth caring if you
expect to have a lot of pages like this.

> does this patch correct the last few such cases that didn't yet do that,
> or does it still leave holes?

I believe in metadata it's still widely there, but hwpoison ignores
those pages anyways.

> 
> Can we document the requirements and or holes to it someplace near
> set_page_dirty()?

Ok I'll add a comment there.

-Andi

-- 
ak@xxxxxxxxxxxxxxx -- Speaking for myself only.
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/