[PATCH v15 07/11] secretmem: use PMD-size pages to amortize direct map fragmentation

[Matthew Wilcox]

On Wed, Jan 20, 2021 at 08:06:08PM +0200, Mike Rapoport wrote:
> +static int secretmem_pool_increase(struct secretmem_ctx *ctx, gfp_t gfp)
>  {
> +	unsigned long nr_pages = (1 << PMD_PAGE_ORDER);
> +	struct gen_pool *pool = ctx->pool;
> +	unsigned long addr;
> +	struct page *page;
> +	int err;
> +
> +	page = cma_alloc(secretmem_cma, nr_pages, PMD_SIZE, gfp & __GFP_NOWARN);
> +	if (!page)
> +		return -ENOMEM;

Does cma_alloc() zero the pages it allocates?  If not, where do we avoid
leaking kernel memory to userspace?