Re: [PATCH v2] kexec: keep the next kernel off hardware-poisoned pages

On Fri, Jul 31, 2026 at 07:49:27PM +0200, Pratyush Yadav wrote:
> > + * return phys_addr_t, thus this "extra" parameter, instead of returning
> > + * the hit.
> > + */
> > +bool range_contains_hwpoison(phys_addr_t start, unsigned long size,
> > +			     phys_addr_t *poison)
> 
> Perhaps
> 
>     phys_addr_t range_last_hwpoison(phys_addr_t start, unsigned long size)  ?
> 
> And if no poisoned page lands here then a default of 0 or ~0UL?

That was my initial preference, but hypothetically phys_addr_t 0 could be
a poisoned page.

I hadn't considered ~0UL as a sentinel — that should work and is a good
suggestion. let me prototype.

Thanks for the review.
--breno