Re: [PATCH v10 16/50] x86/sev: Introduce snp leaked pages list

[Borislav Petkov]

On Mon, Oct 16, 2023 at 08:27:45AM -0500, Michael Roth wrote:
> +	spin_lock(&snp_leaked_pages_list_lock);
> +	while (npages--) {
> +		/*
> +		 * Reuse the page's buddy list for chaining into the leaked
> +		 * pages list. This page should not be on a free list currently
> +		 * and is also unsafe to be added to a free list.
> +		 */
> +		list_add_tail(&page->buddy_list, &snp_leaked_pages_list);
> +		sev_dump_rmpentry(pfn);
> +		pfn++;
> +	}
> +	spin_unlock(&snp_leaked_pages_list_lock);
> +	atomic_long_inc(&snp_nr_leaked_pages);

How is this supposed to count?

You're leaking @npages as the function's parameter but are incrementing
snp_nr_leaked_pages only once?

Just make it a bog-normal unsigned long and increment it inside the
locked section.

Or do at the beginning of the function:

	atomic_long_add(npages, &snp_nr_leaked_pages);

-- 
Regards/Gruss,
    Boris.

https://people.kernel.org/tglx/notes-about-netiquette