Re: Linux MM Alignment Session on hwpoison

[David Hildenbrand (Arm)]

On 8/29/26 01:36, David Rientjes wrote:
> Following up on this thread and the suggestion from Matthew (thank you!), 
> I think it would be useful to have a hwpoison requirements and roadmap 
> discussion on how its support is currently integrated into the core MM, 
> get an understanding of what the areas to improve are, and get a sense of 
> future work in this area that people are planning.
> 
> This particular patch series is addressing an area of interest to Google, 
> but it's also quite clear from the thread that there are multiple areas to 
> improve for hwpoison in the core MM and it wouldn't hurt to get more 
> insight into what dependencies users have on it.
> 
> I'd propose this for discussion in one of our upcoming Linux MM Alignment 
> Sessions: Wednesday, September 16 9-10am PDT (UTC-7).
> 
> No specific agenda other than to gather core MM developers with hwpoison 
> developers and discuss:
>  - where validation should live, esp with regard to the page allocator, or
>    strict invariants in the page allocator
>    + implications for the fastpath
>  - handling for folios of arbitary orders
>  - handling partial folio poisoning and split failures
>  - race conditions between memory_failure() and concurrent
>    alloc/free/compaction
>  - soft offline for pages in the pcp lists

There is also the hugetlb hwpoison oddity.

-- 
Cheers,

David