Re: [PATCH mmotm v1] mm/hwpoison: disable pcp for page_handle_poison()

From
Mike Kravetz <hidden>
Date
2021-06-18 04:49:12
Also in
lkml
Reply to
David Hildenbrand, 2021-06-17
On 6/17/21 5:21 AM, David Hildenbrand wrote:
On 17.06.21 11:26, Naoya Horiguchi wrote:
quoted
From: Naoya Horiguchi <redacted>

Recent changes by patch "mm/page_alloc: allow high-order pages to be
stored on the per-cpu lists" makes kernels determine whether to use pcp
by pcp_allowed_order(), which breaks soft-offline for hugetlb pages.

Soft-offline dissolves a migration source page, then removes it from
buddy free list, so it's assumed that any subpage of the soft-offlined
hugepage are recognized as a buddy page just after returning from
dissolve_free_huge_page().  pcp_allowed_order() returns true for
hugetlb, so this assumption is no longer true.

So disable pcp during dissolve_free_huge_page() and
take_page_off_buddy() to prevent soft-offlined hugepages from linking to
pcp lists.  Soft-offline should not be common events so the impact on
performance should be minimal.  And I think that the optimization of
Mel's patch could benefit to hugetlb so zone_pcp_disable() is called
only in hwpoison context.

Signed-off-by: Naoya Horiguchi <redacted>
---
  mm/memory-failure.c | 19 ++++++++++++++++---
  1 file changed, 16 insertions(+), 3 deletions(-)
diff --git v5.13-rc6-mmotm-2021-06-15-20-24/mm/memory-failure.c v5.13-rc6-mmotm-2021-06-15-20-24_patched/mm/memory-failure.c
index 1842822a10da..593079766655 100644
--- v5.13-rc6-mmotm-2021-06-15-20-24/mm/memory-failure.c
+++ v5.13-rc6-mmotm-2021-06-15-20-24_patched/mm/memory-failure.c
@@ -66,6 +66,19 @@ int sysctl_memory_failure_recovery __read_mostly = 1;
    atomic_long_t num_poisoned_pages __read_mostly = ATOMIC_LONG_INIT(0);
  +static bool __page_handle_poison(struct page *page)
+{
+    bool ret;
+
+    zone_pcp_disable(page_zone(page));
+    ret = dissolve_free_huge_page(page);
+    if (!ret)
+        ret = take_page_off_buddy(page);
+    zone_pcp_enable(page_zone(page));
+
+    return ret;
+}
+
  static bool page_handle_poison(struct page *page, bool hugepage_or_freepage, bool release)
  {
      if (hugepage_or_freepage) {
@@ -73,7 +86,7 @@ static bool page_handle_poison(struct page *page, bool hugepage_or_freepage, boo
           * Doing this check for free pages is also fine since dissolve_free_huge_page
           * returns 0 for non-hugetlb pages as well.
           */
-        if (dissolve_free_huge_page(page) || !take_page_off_buddy(page))
+        if (!__page_handle_poison(page))
              /*
               * We could fail to take off the target page from buddy
               * for example due to racy page allocation, but that's
@@ -986,7 +999,7 @@ static int me_huge_page(struct page *p, unsigned long pfn)
           */
          if (PageAnon(hpage))
              put_page(hpage);
-        if (!dissolve_free_huge_page(p) && take_page_off_buddy(p)) {
+        if (__page_handle_poison(p)) {
              page_ref_inc(p);
              res = MF_RECOVERED;
          }
@@ -1441,7 +1454,7 @@ static int memory_failure_hugetlb(unsigned long pfn, int flags)
          res = get_hwpoison_page(p, flags);
          if (!res) {
              res = MF_FAILED;
-            if (!dissolve_free_huge_page(p) && take_page_off_buddy(p)) {
+            if (__page_handle_poison(p)) {
                  page_ref_inc(p);
                  res = MF_RECOVERED;
              }
Just to make sure: all call paths are fine that we are taking a mutex, right?
That should be the case.  dissolve_free_huge_page can sleep, so if not
we are already broken.

-- 
Mike Kravetz
Thread (6 messages) 6 messages, 4 authors, last reply 2021-06-18
Keyboard shortcuts
hback out one level
jnext message in thread / next series-diff position
kprevious message in thread / previous series-diff position
ldrill in
Escclose help / fold thread tree
?toggle this help