linux-kernel.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Oscar Salvador <osalvador@suse.de>
To: n-horiguchi@ah.jp.nec.com
Cc: mhocko@kernel.org, mike.kravetz@oracle.com, linux-mm@kvack.org,
	linux-kernel@vger.kernel.org, Oscar Salvador <osalvador@suse.de>
Subject: [PATCH 10/10] mm,hwpoison: Use hugetlb_replace_page to replace free hugetlb pages
Date: Tue, 10 Sep 2019 12:30:16 +0200	[thread overview]
Message-ID: <20190910103016.14290-11-osalvador@suse.de> (raw)
In-Reply-To: <20190910103016.14290-1-osalvador@suse.de>

When soft offlining a free hugtlb, try first to allocate a new hugetlb
to the pool and pass the old state to the new one by move_hugetlb_state().
Either we succeed or not, we dissolve the poisoned hugetlb page.

Worst-scenario case is that we cannot allocate a new fresh hugetlb page
as a replacement.

Signed-off-by: Oscar Salvador <osalvador@suse.de>
---
 mm/hugetlb.c        | 16 ++++++++++++++++
 mm/memory-failure.c | 34 ++++++++++++++++++++++++++++------
 2 files changed, 44 insertions(+), 6 deletions(-)

diff --git a/mm/hugetlb.c b/mm/hugetlb.c
index 139e1c05c9a1..d0844aec7531 100644
--- a/mm/hugetlb.c
+++ b/mm/hugetlb.c
@@ -5154,3 +5154,19 @@ void move_hugetlb_state(struct page *oldpage, struct page *newpage, int reason)
 		spin_unlock(&hugetlb_lock);
 	}
 }
+
+#ifdef CONFIG_MEMORY_FAILURE
+int hugetlb_replace_page(struct page *page, int reason)
+{
+	int nid = page_to_nid(page);
+	struct hstate *h = page_hstate(page);
+	struct page *new_page;
+
+	new_page = alloc_huge_page_nodemask(h, nid, &node_states[N_MEMORY]);
+	if (!new_page)
+		return -ENOMEM;
+
+	move_hugetlb_state(page, new_page, reason);
+	return 0;
+}
+#endif
diff --git a/mm/memory-failure.c b/mm/memory-failure.c
index 03f07015a106..fe73fe19c6e9 100644
--- a/mm/memory-failure.c
+++ b/mm/memory-failure.c
@@ -79,6 +79,7 @@ EXPORT_SYMBOL_GPL(hwpoison_filter_flags_mask);
 EXPORT_SYMBOL_GPL(hwpoison_filter_flags_value);
 
 extern bool take_page_off_buddy(struct page *page);
+extern int hugetlb_replace_page(struct page *page, int reason);
 
 static bool page_set_poison(struct page *page)
 {
@@ -1804,16 +1805,37 @@ static int soft_offline_in_use_page(struct page *page)
 	return __soft_offline_page(page);
 }
 
+static int soft_offline_free_huge_page(struct page *page)
+{
+	struct page *hpage = compound_head(page);
+
+	/*
+	 * Try to add a new hugetlb page to the pool
+	 */
+	if (hugetlb_replace_page(hpage, MR_MEMORY_FAILURE))
+		return -EBUSY;
+
+	/*
+	 * Remove old hugetlb from the pool
+	 */
+	if (!page_set_poison(hpage))
+		return -EBUSY;
+
+	return 0;
+}
+
 static int soft_offline_free_page(struct page *page)
 {
-	int rc = dissolve_free_huge_page(page);
+	int rc = -EBUSY;
 
-	if (!rc) {
-		if (take_page_off_buddy(page))
+	if (PageHuge(page))
+		rc = soft_offline_free_huge_page(page);
+	else
+		if (take_page_off_buddy(page)) {
 			page_set_poison(page);
-		else
-			rc = -EBUSY;
-	}
+			rc = 0;
+		}
+
 	return rc;
 }
 
-- 
2.12.3


  parent reply	other threads:[~2019-09-10 10:31 UTC|newest]

Thread overview: 26+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2019-09-10 10:30 [PATCH 00/10] Hwpoison soft-offline rework Oscar Salvador
2019-09-10 10:30 ` [PATCH 01/10] mm,hwpoison: cleanup unused PageHuge() check Oscar Salvador
2019-09-11 10:17   ` David Hildenbrand
2019-09-10 10:30 ` [PATCH 02/10] mm,madvise: call soft_offline_page() without MF_COUNT_INCREASED Oscar Salvador
2019-09-11 10:23   ` David Hildenbrand
2019-09-12  1:28     ` Naoya Horiguchi
2019-09-11 10:27   ` David Hildenbrand
2019-09-12  1:37     ` Naoya Horiguchi
2019-09-13  8:37       ` David Hildenbrand
2019-09-10 10:30 ` [PATCH 03/10] mm,hwpoison-inject: don't pin for hwpoison_filter Oscar Salvador
2019-09-16  7:41   ` David Hildenbrand
2019-09-10 10:30 ` [PATCH 04/10] mm,hwpoison: remove MF_COUNT_INCREASED Oscar Salvador
2019-09-11 10:24   ` David Hildenbrand
2019-09-10 10:30 ` [PATCH 05/10] mm: remove flag argument from soft offline functions Oscar Salvador
2019-09-11 10:24   ` David Hildenbrand
2019-09-10 10:30 ` [PATCH 06/10] mm,hwpoison: Unify THP handling for hard and soft offline Oscar Salvador
2019-09-10 10:30 ` [PATCH 07/10] mm,hwpoison: Rework soft offline for in-use pages Oscar Salvador
2019-09-10 10:30 ` [PATCH 08/10] mm,hwpoison: Refactor soft_offline_huge_page and __soft_offline_page Oscar Salvador
2019-09-10 10:30 ` [PATCH 09/10] mm,hwpoison: Rework soft offline for free pages Oscar Salvador
2019-09-10 10:30 ` Oscar Salvador [this message]
2019-09-10 10:32 ` [PATCH 00/10] Hwpoison soft-offline rework Oscar Salvador
2019-09-11  5:29 ` Naoya Horiguchi
2019-09-11  6:22   ` Naoya Horiguchi
2019-09-11  6:35     ` osalvador
2019-09-11  7:21       ` Naoya Horiguchi
2019-09-12 13:16         ` Oscar Salvador

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20190910103016.14290-11-osalvador@suse.de \
    --to=osalvador@suse.de \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-mm@kvack.org \
    --cc=mhocko@kernel.org \
    --cc=mike.kravetz@oracle.com \
    --cc=n-horiguchi@ah.jp.nec.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).