Re: [PATCH v4 1/3] kselftest: mm: prevent random failure of huge page split for khugepaged

[Lorenzo Stoakes (ARM)]

On Wed, Sep 02, 2026 at 01:09:35PM +0100, Yeoreum Yun wrote:
> There're some random failure for split_huge_page_test when khugepaged
> collapses pages into pmd again which had split by the test.
>
> Prevent the khugepaged's collapses for split page by setting the
> mapped pmd-huge-page with MADV_NOHUGEPAGE before split.

Maybe worth saying that MADV_NOHUGEPAGE doesn't immediately split anything, just
prevents future khugepaged collapses?

>
> Suggested-by: Kevin Brodsky <kevin.brodsky@xxxxxxx>
> Suggested-by: Lorenzo Stoakes (ARM) <ljs@xxxxxxxxxx>
> Signed-off-by: Yeoreum Yun <yeoreum.yun@xxxxxxx>

LGTM, thanks, with a few nits above/below (sorry! :P)

Reviewed-by: Lorenzo Stoakes (ARM) <ljs@xxxxxxxxxx>

> ---
>  tools/testing/selftests/mm/split_huge_page_test.c | 19 +++++++++++++++++++
>  1 file changed, 19 insertions(+)
>
> diff --git a/tools/testing/selftests/mm/split_huge_page_test.c b/tools/testing/selftests/mm/split_huge_page_test.c
> index 86a603692826..4bbad892bc57 100644
> --- a/tools/testing/selftests/mm/split_huge_page_test.c
> +++ b/tools/testing/selftests/mm/split_huge_page_test.c
> @@ -168,6 +168,14 @@ static char *allocate_zero_filled_hugepage(size_t len)
>  	return result;
>  }
>
> +static void madv_nohuge(void *addr, size_t len)

Maybe call it disable_khugepaged() to make it explicit that's what this is doing
rather than eliminating any THP pages in the range?

> +{

Maybe a comment here like:

	/*
	 * Disables khugepaged from collapsing THPs in range, existing THP
	 * pages remain.
	 */

> +	if (!madvise(addr, len, MADV_NOHUGEPAGE))
> +		return;
> +
> +	ksft_exit_fail_msg("MADV_NOHUGEPAGE failed, err=%d\n", errno);
> +}
> +
>  static void verify_rss_anon_split_huge_page_all_zeroes(char *one_page, int nr_hpages, size_t len)
>  {
>  	unsigned long rss_anon_before, rss_anon_after;
> @@ -180,6 +188,9 @@ static void verify_rss_anon_split_huge_page_all_zeroes(char *one_page, int nr_hp
>  	if (!rss_anon_before)
>  		ksft_exit_fail_msg("No RssAnon is allocated before split\n");
>
> +	/* Prevent khugepaged from collapsing the pages. */
> +	madv_nohuge(one_page, len);
> +
>  	/* split all THPs */
>  	write_debugfs(PID_FMT, getpid(), (uint64_t)one_page,
>  		      (uint64_t)one_page + len, 0);
> @@ -227,6 +238,9 @@ static void split_pmd_thp_to_order(int order)
>  	if (!check_huge_anon(one_page, 4 * pmd_pagesize, 4, pmd_pagesize))
>  		ksft_exit_fail_msg("No THP is allocated\n");
>
> +	/* Prevent khugepaged from collapsing the pages. */
> +	madv_nohuge(one_page, len);
> +
>  	/* split all THPs */
>  	write_debugfs(PID_FMT, getpid(), (uint64_t)one_page,
>  		(uint64_t)one_page + len, order);
> @@ -275,6 +289,9 @@ static void split_pte_mapped_thp(void)
>  		goto out;
>  	}
>
> +	/* Prevent khugepaged from collapsing the pages. */
> +	madv_nohuge(thp_area, thp_area_size);
> +
>  	/*
>  	 * To challenge spitting code, we will mremap a single page of each
>  	 * THP (page[i] of thp[i]) in the thp_area into page_area. This will
> @@ -542,6 +559,8 @@ static int create_pagecache_thp_and_fd(const char *testfile, size_t fd_size,
>  		ksft_test_result_skip("Pagecache folio split skipped\n");
>  		return -2;
>  	}
> +	/* Prevent khugepaged from collapsing the pages. */
> +	madv_nohuge(*addr, fd_size);
>  	return 0;
>  err_out_close:
>  	close(*fd);
>
> --
> 2.43.0
>

--
Cheers, Lorenzo