On 7 Aug 2021, at 6:32, Mike Rapoport wrote:

On Thu, Aug 05, 2021 at 03:02:46PM -0400, Zi Yan wrote:

quoted

From: Zi Yan <ziy@nvidia.com>

It keeps the existing behavior after MAX_ORDER is increased beyond
a section size.

Signed-off-by: Zi Yan <ziy@nvidia.com>
Cc: Mike Rapoport <rppt@kernel.org>
Cc: David Hildenbrand <redacted>
Cc: Oscar Salvador <osalvador@suse.de>
Cc: Ying Chen <redacted>
Cc: Feng Zhou <redacted>
Cc: linux-fsdevel@vger.kernel.org
Cc: linux-mm@kvack.org
Cc: linux-kernel@vger.kernel.org
---
 fs/proc/kcore.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/fs/proc/kcore.c b/fs/proc/kcore.c
index 3f148759a5fd..77b7ba48fb44 100644
--- a/fs/proc/kcore.c
+++ b/fs/proc/kcore.c

@@ -486,7 +486,7 @@ read_kcore(struct file *file, char __user *buffer, size_t buflen, loff_t *fpos)
 			}
 		}

-		if (page_offline_frozen++ % MAX_ORDER_NR_PAGES == 0) {
+		if (page_offline_frozen++ % PAGES_PER_SECTION == 0) {

The behavior changes here. E.g. with default configuration on x86 instead
of cond_resched() every 2M we get cond_resched() every 128M.

I'm not saying it's wrong but at least it deserves an explanation why.

Sure. I will also think about whether I should use PAGES_PER_SECTION or pageblock_nr_pages
to replace MAX_ORDER in this and other patches. pageblock_nr_pages will be unchanged,
so at least in x86_64, using pageblock_nr_pages would not change code behaviors.

—
Best Regards,
Yan, Zi