[LTP] [PATCH 5/5] syscalls/move_pages12: Add one more kernel git hash
Cyril Hrubis
chrubis@suse.cz
Wed Dec 16 11:10:12 CET 2020
+ Reformat the comment to asciidoc
Signed-off-by: Cyril Hrubis <chrubis@suse.cz>
---
.../kernel/syscalls/move_pages/move_pages12.c | 52 ++++++++++++-------
1 file changed, 32 insertions(+), 20 deletions(-)
diff --git a/testcases/kernel/syscalls/move_pages/move_pages12.c b/testcases/kernel/syscalls/move_pages/move_pages12.c
index c906acbe3..d209426c7 100644
--- a/testcases/kernel/syscalls/move_pages/move_pages12.c
+++ b/testcases/kernel/syscalls/move_pages/move_pages12.c
@@ -7,16 +7,17 @@
* Ported: Yang Xu <xuyang2018.jy@cn.jujitsu.com>
*/
-/*
- * Description:
+/*\
+ * [DESCRIPTION]
+ *
+ * *Test 1*
*
- * Test #1:
- * This is a regression test for the race condition between move_pages()
- * and freeing hugepages, where move_pages() calls follow_page(FOLL_GET)
- * for hugepages internally and tries to get its refcount without
- * preventing concurrent freeing.
+ * This is a regression test for the race condition between move_pages()
+ * and freeing hugepages, where move_pages() calls follow_page(FOLL_GET)
+ * for hugepages internally and tries to get its refcount without
+ * preventing concurrent freeing.
*
- * This test can crash the buggy kernel, and the bug was fixed in:
+ * This test can crash the buggy kernel, and the bug was fixed in:
*
* commit e66f17ff71772b209eed39de35aaa99ba819c93d
* Author: Naoya Horiguchi <n-horiguchi@ah.jp.nec.com>
@@ -24,25 +25,25 @@
*
* mm/hugetlb: take page table lock in follow_huge_pmd()
*
- * Test #2:
- * #2.1:
- * This is a regression test for the race condition, where move_pages()
- * and soft offline are called on a single hugetlb page concurrently.
+ * *Test 2.1*
*
- * This bug can crash the buggy kernel, and was fixed by:
+ * This is a regression test for the race condition, where move_pages()
+ * and soft offline are called on a single hugetlb page concurrently.
+ *
+ * This test can crash the buggy kernel, and was fixed by:
*
* commit c9d398fa237882ea07167e23bcfc5e6847066518
* Author: Naoya Horiguchi <n-horiguchi@ah.jp.nec.com>
* Date: Fri Mar 31 15:11:55 2017 -0700
*
- * mm, hugetlb: use pte_present() instead of pmd_present() in
- * follow_huge_pmd()
+ * mm, hugetlb: use pte_present() instead of pmd_present() in follow_huge_pmd()
+ *
+ * *Test 2.2*
*
- * #2.2:
- * This is also a regression test for an race condition causing SIGBUS
- * in hugepage migration/fault.
+ * This is also a regression test for an race condition causing SIGBUS
+ * in hugepage migration/fault.
*
- * This bug was fixed by:
+ * This bug was fixed by:
*
* commit 4643d67e8cb0b3536ef0ab5cddd1cedc73fa14ad
* Author: Mike Kravetz <mike.kravetz@oracle.com>
@@ -50,7 +51,17 @@
*
* hugetlbfs: fix hugetlb page migration/fault race causing SIGBUS
*
- */
+ * *Test 2.3*
+ *
+ * The madvise() in the do_soft_online() was also triggering cases where soft
+ * online returned EIO when page migration failed, which was fixed in:
+ *
+ * commit 3f4b815a439adfb8f238335612c4b28bc10084d8
+ * Author: Oscar Salvador <osalvador@suse.de>
+ * Date: Mon Dec 14 19:11:51 2020 -0800
+ *
+ * mm,hwpoison: return -EBUSY when migration fails
+\*/
#include <errno.h>
#include <unistd.h>
@@ -334,6 +345,7 @@ static struct tst_test test = {
{"linux-git", "e66f17ff7177"},
{"linux-git", "c9d398fa2378"},
{"linux-git", "4643d67e8cb0"},
+ {"linux-git", "3f4b815a439a"},
{}
}
};
--
2.26.2
More information about the ltp
mailing list