[PATCH v4 3/5] mm/memory-failure: efi: record hardware-poisoned frames into the poisoned-memory table

[Breno Leitao]

On Wed, Sep 09, 2026 at 01:21:28PM +0000, sashiko-bot at kernel.org wrote:
> > --- /dev/null
> > +++ b/drivers/firmware/efi/poison.c
> >
> > +	/* Whole words, and a bit count that can be taken without wrapping. */
> > +	if (!pm->size || !IS_ALIGNED(pm->size, sizeof(unsigned long)) ||
> > +	    check_mul_overflow(pm->size, (u64)BITS_PER_BYTE, &nbits))
> > +		return false;
> 
> [Severity: High]
> Does this validation fully protect against an excessively large pm->size?

I was clamping the table before, but we decided to drop it in the last
revision. See the discusion in here:

https://lore.kernel.org/all/apGWUWi5-RbhFHpe@thinkstation/

> > +	start = PAGE_ALIGN_DOWN(efi.poisoned_memory);
> > +	end = PAGE_ALIGN(efi.poisoned_memory + sizeof(*pm) + pm->size);
> 
> [Severity: High]
> Can this computation wrap around if pm->size is maliciously large?

Yes, but that means that someone was able to write to EFI memory to
update size?

I can definitely BUG_ON() on "large"  pm->size, but, I don't think this
is what we want here.