Re: [PATCH v6 6/6] selftest: add testing unsharing and counting ksm zero page

[Claudio Imbrenda]

On Fri, 10 Feb 2023 09:21:57 +0800 (CST)
<yang.yang29@xxxxxxxxxx> wrote:

> From: xu xin <xu.xin16@xxxxxxxxxx>
> 
> Add a function test_unmerge_zero_page() to test the functionality on
> unsharing and counting ksm-placed zero pages and counting of this patch
> series.
> 
> test_unmerge_zero_page() actually contains three subjct test objects:
> 	1) whether the count of ksm zero page can react correctly to cow
> 	   (copy on write);
> 	2) whether the count of ksm zero page can react correctly to unmerge;
> 	3) whether ksm zero pages are really unmerged.
> 
> Signed-off-by: xu xin <xu.xin16@xxxxxxxxxx>
> Cc: Claudio Imbrenda <imbrenda@xxxxxxxxxxxxx>
> Cc: David Hildenbrand <david@xxxxxxxxxx>
> Cc: Xuexin Jiang <jiang.xuexin@xxxxxxxxxx>
> Reviewed-by: Xiaokai Ran <ran.xiaokai@xxxxxxxxxx>
> Reviewed-by: Yang Yang <yang.yang29@xxxxxxxxxx>

[...]

> +static int unmerge_zero_page(char *start, unsigned long size)
> +{
> +	int ret;
> +
> +	ret = madvise(start, size, MADV_UNMERGEABLE);
> +	if (ret) {
> +		ksft_test_result_fail("MADV_UNMERGEABLE failed\n");
> +		return ret;
> +	}
> +
> +	/*
> +	 * Wait for two full scans such that any possible unmerging of zero
> +	 * pages happened. Why? Because the unmerge action of zero pages is not

please remove "Why? Because"

> +	 * done in the context of madvise(), but in the context of
> +	 * unshare_zero_pages() of the ksmd thread.
> +	 */
> +	return wait_two_full_scans();
> +}
> +
>  static char *mmap_and_merge_range(char val, unsigned long size)
>  {
>  	char *map;

[...]