Re: [PATCH 0/7] arm64: hibernate: dust off the suspend/cleanup rework

[Mark Rutland]

Hi Will,

On Thu, Sep 03, 2026 at 11:53:59AM +0100, Will Deacon wrote:
> it's worth noting that Bradley's patch _does_ fix one of the
> hibernation problems and has been tested by Vladimir:
> 
> https://lore.kernel.org/all/20260809213615.11646-1-include@xxxxxxxxx/
> 
> I would prefer to take your complete set of patches, but if you don't
> want that then should I just take the above for now? Prior to this
> series, there has been no activity for nearly two years on your branch,
> so I'd prefer to get _something_ in, even if it only fixes some of the
> problems.

Sorry for the delay.

I'd prefer if you picked up Bradley's single patch for now. I'll follow
up with a respin of the rest of the cleanup and fixes shortly.

Could you add a note to the commit message as to how Vladimir was able
to trigger a hang with CONFIG_PAGE_POISONING=y and page_poison=on? It's
helpful for anyone testing a backport, etc.

With that:

Acked-by: Mark Rutland <mark.rutland@xxxxxxx>

For context, when I originally wrote my patches, I hadn't managed to
trigger failures at top level without additional hacks, and so this
hadn't been a high priority. Given that Vladimir can trigger this
specific issue with existing options, it makes sense to pick that up.

I had a vague recollection that this specific fix lead to a secondary
issue, but digging through my notes I can't see what that would have
been, and I suspect I had just gotten confused w.r.t. what KVM was
doing. If I find there was a real secondary issue I'll follow up with a
fix .

Thanks,
Mark.