Hi Anshuman,

On 7/28/21 9:08 PM, Anshuman Khandual wrote:

On 7/27/21 11:43 AM, Gavin Shan wrote:

quoted

This uses struct pgtable_debug_args in the migration and thp test
functions. It's notable that the pre-allocated page is used in
swap_migration_tests() as set_pte_at() is used there.

Signed-off-by: Gavin Shan <redacted>
---
  mm/debug_vm_pgtable.c | 28 ++++++++++++++--------------
  1 file changed, 14 insertions(+), 14 deletions(-)

diff --git a/mm/debug_vm_pgtable.c b/mm/debug_vm_pgtable.c
index bc153cad9045..9136195efde3 100644
--- a/mm/debug_vm_pgtable.c
+++ b/mm/debug_vm_pgtable.c

@@ -845,7 +845,7 @@ static void __init pmd_swap_tests(struct pgtable_debug_args *args)
  static void __init pmd_swap_tests(struct pgtable_debug_args *args) { }
  #endif /* CONFIG_ARCH_ENABLE_THP_MIGRATION */
  
-static void __init swap_migration_tests(void)
+static void __init swap_migration_tests(struct pgtable_debug_args *args)
  {
  	struct page *page;
  	swp_entry_t swp;

@@ -861,9 +861,10 @@ static void __init swap_migration_tests(void)
  	 * problematic. Lets allocate a dedicated page explicitly for this
  	 * purpose that will be freed subsequently.
  	 */
-	page = alloc_page(GFP_KERNEL);
+	page = (args->pte_pfn != ULONG_MAX) ?
+	       pfn_to_page(args->pte_pfn) : NULL;
  	if (!page) {
-		pr_err("page allocation failed\n");
+		pr_err("no page available\n");
  		return;
  	}

Please check for a valid page earlier in the function and return. Otherwise
this calls out the page unavailability (after starting the test), which is
inconsistent with all other functions like pxx_advanced_tests().

[    1.051633] debug_vm_pgtable: [pte_swap_tests           ]: Validating PTE swap
[    1.052697] debug_vm_pgtable: [pmd_swap_tests           ]: Validating PMD swap
[    1.053765] debug_vm_pgtable: [swap_migration_tests     ]: Validating swap migration <=====
[    1.054900] debug_vm_pgtable: [swap_migration_tests     ]: no page available         <=====

Should do this just before pr_info("Validating swap migration\n").

......
page = (args->pte_pfn != ULONG_MAX) ? pfn_to_page(args->pte_pfn) : NULL;
if (!page)
	return;
.....

Yes. The order of error messages are sticky to original implementation, but
it'd better to be consistent with the new order we have in this series. I
will adjust in v5.

Thanks,
Gavin