Re: [PATCH v6 1/2] mm,hwpoison: fix race with hugetlb page allocation

[Luck, Tony]

On Tue, Aug 17, 2021 at 02:12:07AM +0900, Naoya Horiguchi wrote:
> This dump indicates that HWPoisonHandlable() returned false due to
> the lack of PG_lru flag.  In older code before 5.13, get_any_page() does
> retry with shake_page(), but does not since 5.13, which seems to me
> the root cause of the issue. So my suggestion is to call shake_page()
> when HWPoisonHandlable() is false.
> 
> Could you try checking that the following diff fixes the issue?
> I could still have better fix (like inserting shake_page() to other
> retry paths in get_any_page()), but the below is the minimum one.

Tried it ... and it works! Injected and recovered from a thousand
errors without seeing any problems.

-Tony

P.S. Somewhere in the mail system your patch arrived with <TAB>s
changed to spaces. Here's what I applied to v5.14-rc6 (hopefully with
TABS preserved) ... just in case anyone else is following along with
this thread and wants to try some tests.

diff --git a/mm/memory-failure.c b/mm/memory-failure.c
index eefd823deb67..aa6592540f17 100644
--- a/mm/memory-failure.c
+++ b/mm/memory-failure.c
@@ -1146,7 +1146,7 @@ static int __get_hwpoison_page(struct page *page)
 	 * unexpected races caused by taking a page refcount.
 	 */
 	if (!HWPoisonHandlable(head))
-		return 0;
+		return -EBUSY;
 
 	if (PageTransHuge(head)) {
 		/*
@@ -1199,9 +1199,14 @@ static int get_any_page(struct page *p, unsigned long flags)
 			}
 			goto out;
 		} else if (ret == -EBUSY) {
-			/* We raced with freeing huge page to buddy, retry. */
-			if (pass++ < 3)
+			/*
+			 * We raced with (possibly temporary) unhandlable
+			 * page, retry.
+			 */
+			if (pass++ < 3) {
+				shake_page(p, 1);
 				goto try_again;
+			}
 			goto out;
 		}
 	}