mm/shmem: fix shmem_swapin() race with swapoff
authorMiaohe Lin <linmiaohe@huawei.com>
Tue, 29 Jun 2021 02:36:57 +0000 (19:36 -0700)
committerLinus Torvalds <torvalds@linux-foundation.org>
Tue, 29 Jun 2021 17:53:49 +0000 (10:53 -0700)
When I was investigating the swap code, I found the below possible race
window:

CPU 1                                         CPU 2
-----                                         -----
shmem_swapin
  swap_cluster_readahead
    if (likely(si->flags & (SWP_BLKDEV | SWP_FS_OPS))) {
                                              swapoff
                                                ..
                                                si->swap_file = NULL;
                                                ..
    struct inode *inode = si->swap_file->f_mapping->host;[oops!]

Close this race window by using get/put_swap_device() to guard against
concurrent swapoff.

Link: https://lkml.kernel.org/r/20210426123316.806267-5-linmiaohe@huawei.com
Fixes: 8fd2e0b505d1 ("mm: swap: check if swap backing device is congested or not")
Signed-off-by: Miaohe Lin <linmiaohe@huawei.com>
Reviewed-by: "Huang, Ying" <ying.huang@intel.com>
Cc: Dennis Zhou <dennis@kernel.org>
Cc: Tim Chen <tim.c.chen@linux.intel.com>
Cc: Hugh Dickins <hughd@google.com>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: Michal Hocko <mhocko@suse.com>
Cc: Joonsoo Kim <iamjoonsoo.kim@lge.com>
Cc: Alex Shi <alexs@kernel.org>
Cc: Matthew Wilcox <willy@infradead.org>
Cc: Minchan Kim <minchan@kernel.org>
Cc: Wei Yang <richard.weiyang@gmail.com>
Cc: Yang Shi <shy828301@gmail.com>
Cc: David Hildenbrand <david@redhat.com>
Cc: Yu Zhao <yuzhao@google.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
mm/shmem.c

index 5d46611..53f2101 100644 (file)
@@ -1696,7 +1696,8 @@ static int shmem_swapin_page(struct inode *inode, pgoff_t index,
        struct address_space *mapping = inode->i_mapping;
        struct shmem_inode_info *info = SHMEM_I(inode);
        struct mm_struct *charge_mm = vma ? vma->vm_mm : current->mm;
-       struct page *page;
+       struct swap_info_struct *si;
+       struct page *page = NULL;
        swp_entry_t swap;
        int error;
 
@@ -1704,6 +1705,12 @@ static int shmem_swapin_page(struct inode *inode, pgoff_t index,
        swap = radix_to_swp_entry(*pagep);
        *pagep = NULL;
 
+       /* Prevent swapoff from happening to us. */
+       si = get_swap_device(swap);
+       if (!si) {
+               error = EINVAL;
+               goto failed;
+       }
        /* Look it up and read it in.. */
        page = lookup_swap_cache(swap, NULL, 0);
        if (!page) {
@@ -1765,6 +1772,8 @@ static int shmem_swapin_page(struct inode *inode, pgoff_t index,
        swap_free(swap);
 
        *pagep = page;
+       if (si)
+               put_swap_device(si);
        return 0;
 failed:
        if (!shmem_confirm_swap(mapping, index, swap))
@@ -1775,6 +1784,9 @@ unlock:
                put_page(page);
        }
 
+       if (si)
+               put_swap_device(si);
+
        return error;
 }