Re: [PATCH 2/2] kselftest: mm: fix intermittent failure khugepaged test

[Baolin Wang <baolin.wang@linux.alibaba.com>]

On 9/15/26 5:21 PM, Yeoreum Yun wrote:
> There are intermittent failures in collapse_max_ptes_swap() and
> collapse_max_ptes_shared() when using the khugepaged_context:
> 
>    // while running ./khugepaged -s 2
> 
>    # Run test: collapse_max_ptes_shared (khugepaged:anon)
>    # Allocate huge page... OK
>    # Share huge page over fork()... OK
>    # Trigger CoW on page 1023 of 2048... OK
>    # Maybe collapse with max_ptes_shared exceeded.... OK
>    # Trigger CoW on page 1024 of 2048... Fail
>    Bail out! Unexpected huge page
>    # Planned tests != run tests (26 != 23)
>    # Totals: pass:23 fail:0 xfail:0 xpass:0 skip:0 error:0
> 
>    # Run test: collapse_max_ptes_swap (khugepaged:anon)
>    # Swapout 257 of 2048 pages... OK
>    # Maybe collapse with max_ptes_swap exceeded.... OK
>    # Swapout 256 of 2048 pages... OK
>    Bail out! Unexpected huge page
>    # Planned tests != run tests (26 != 17)
>    # Totals: pass:17 fail:0 xfail:0 xpass:0 skip:0 error:0
> 
> This happens because khugepaged may collapse the pages before wait_for_scan()
> is called, causing a sanity check that expects uncollapsed pages to fail.
> 
> For example, in collapse_max_ptes_swap(), after faulting the pages back in
> and paging out up to max_ptes_swap pages, khugepaged may collapse them again
> before c->collapse() is called.
> 
> To prevent this, change the khugepaged setting from ALWAYS to MADVICE for
> the affected tests, and mark the VMA with MADV_NOHUGEPAGE after it has been
> collapsed by wait_for_scan(). This prevents khugepaged from collapsing it
> again before c->collapse() is called.
> 
> This failure was observed on NVIDIA Spark with 16KB page.
> 
> Signed-off-by: Yeoreum Yun <yeoreum.yun@arm.com>
> ---
>   tools/testing/selftests/mm/khugepaged.c | 15 +++++++++++++++
>   1 file changed, 15 insertions(+)
> 
> diff --git a/tools/testing/selftests/mm/khugepaged.c b/tools/testing/selftests/mm/khugepaged.c
> index c32244b565658..83e9386bbc842 100644
> --- a/tools/testing/selftests/mm/khugepaged.c
> +++ b/tools/testing/selftests/mm/khugepaged.c
> @@ -578,6 +578,8 @@ static bool wait_for_scan(const char *msg, char *p, size_t len,
>   		usleep(TICK);
>   	}
>   
> +	madvise(p, len, MADV_NOHUGEPAGE);

This looks incorrect to me and would reintroduce the previous problem. 
Please see commit 7962e05a835f ("selftests: khugepaged: fix the shmem 
collapse failure").

> +
>   	return timeout == -1;
>   }
>   
> @@ -839,6 +841,7 @@ static void collapse_swapin_single_pte(struct collapse_context *c, struct mem_op
>   
>   static void collapse_max_ptes_swap(struct collapse_context *c, struct mem_ops *ops)
>   {
> +	struct thp_settings settings = *thp_current_settings();
>   	int max_ptes_swap = thp_read_num("khugepaged/max_ptes_swap");
>   	void *p;
>   
> @@ -860,6 +863,9 @@ static void collapse_max_ptes_swap(struct collapse_context *c, struct mem_ops *o
>   	validate_memory(p, 0, hpage_pmd_size);
>   
>   	if (c->enforce_pte_scan_limits) {
> +		settings.hugepages[collapse_order].enabled = THP_MADVISE;
> +		thp_push_settings(&settings);

I'm not sure why the collapse_order setting needs to be changed here. In 
your test case, you did not use the '-c' parameter to specify the 
collapse order.

> +
>   		ops->fault(p, 0, hpage_pmd_size);
>   		ksft_print_msg("Swapout %d of %d pages...", max_ptes_swap,
>   		       hpage_pmd_nr);
> @@ -869,12 +875,15 @@ static void collapse_max_ptes_swap(struct collapse_context *c, struct mem_ops *o
>   			success("OK");
>   		} else {
>   			fail("Fail");
> +			thp_pop_settings();
>   			goto out;
>   		}
>   
>   		c->collapse("Collapse with max_ptes_swap pages swapped out", p,
>   			    1, ops, true);
>   		validate_memory(p, 0, hpage_pmd_size);
> +
> +		thp_pop_settings();
>   	}
>   out:
>   	ops->cleanup_area(p, hpage_pmd_size);
> @@ -1075,6 +1084,7 @@ static void collapse_fork_compound(struct collapse_context *c, struct mem_ops *o
>   
>   static void collapse_max_ptes_shared(struct collapse_context *c, struct mem_ops *ops)
>   {
> +	struct thp_settings settings = *thp_current_settings();
>   	int max_ptes_shared = thp_read_num("khugepaged/max_ptes_shared");
>   	int wstatus;
>   	void *p;
> @@ -1100,6 +1110,9 @@ static void collapse_max_ptes_shared(struct collapse_context *c, struct mem_ops
>   			    1, ops, !c->enforce_pte_scan_limits);
>   
>   		if (c->enforce_pte_scan_limits) {
> +			settings.hugepages[collapse_order].enabled = THP_MADVISE;
> +			thp_push_settings(&settings);

Ditto.

> +
>   			ksft_print_msg("Trigger CoW on page %d of %d...",
>   			       hpage_pmd_nr - max_ptes_shared, hpage_pmd_nr);
>   			ops->fault(p, 0, (hpage_pmd_nr - max_ptes_shared) *
> @@ -1111,6 +1124,8 @@ static void collapse_max_ptes_shared(struct collapse_context *c, struct mem_ops
>   
>   			c->collapse("Collapse with max_ptes_shared PTEs shared",
>   				    p, 1, ops, true);
> +
> +			thp_pop_settings();
>   		}
>   
>   		validate_memory(p, 0, hpage_pmd_size);
>