From: ChenXiaoSong chenxiaosong2@huawei.com
If two processes mount same superblock, memory leak occurs:
CPU0 | CPU1 do_new_mount | do_new_mount fs_set_subtype | fs_set_subtype kstrdup | | kstrdup memrory leak |
The following reproducer triggers the problem:
1. shell command: mount -t ntfs /dev/sda1 /mnt & 2. c program: mount("/dev/sda1", "/mnt", "fuseblk", 0, "...")
with kmemleak report being along the lines of
unreferenced object 0xffff888235f1a5c0 (size 8): comm "mount.ntfs", pid 2860, jiffies 4295757824 (age 43.423s) hex dump (first 8 bytes): 00 a5 f1 35 82 88 ff ff ...5.... backtrace: [<00000000656e30cc>] __kmalloc_track_caller+0x16e/0x430 [<000000008e591727>] kstrdup+0x3e/0x90 [<000000008430d12b>] do_mount.cold+0x7b/0xd9 [<0000000078d639cd>] ksys_mount+0xb2/0x150 [<000000006015988d>] __x64_sys_mount+0x29/0x40 [<00000000e0a7c118>] do_syscall_64+0xc1/0x1d0 [<00000000bcea7df5>] entry_SYSCALL_64_after_hwframe+0x44/0xa9 [<00000000803a4067>] 0xffffffffffffffff
Linus's tree already have refactoring patchset [1], one of them can fix this bug: c30da2e981a7 ("fuse: convert to use the new mount API") After refactoring, init super_block->s_subtype in fuse_fill_super.
Since we did not merge the refactoring patchset in this branch, I create this patch. This patch fix this by adding a write lock while calling fs_set_subtype.
[1] https://patchwork.kernel.org/project/linux-fsdevel/patch/20190903113640.7984...
Fixes: 79c0b2df79eb ("add filesystem subtype support") Cc: David Howells dhowells@redhat.com Cc: stable@vger.kernel.org Signed-off-by: ChenXiaoSong chenxiaosong2@huawei.com Signed-off-by: Greg Kroah-Hartman gregkh@linuxfoundation.org --- v1: Can not mount sshfs ([PATCH linux-4.19.y] VFS: Fix fuseblk memory leak caused by mount concurrency) v2: Use write lock while writing superblock ([PATCH 4.19,v2] VFS: Fix fuseblk memory leak caused by mount concurrency) v3: Update commit message
fs/namespace.c | 9 ++++++--- 1 file changed, 6 insertions(+), 3 deletions(-)
--- a/fs/namespace.c +++ b/fs/namespace.c @@ -2570,9 +2570,12 @@ static int do_new_mount(struct path *pat return -ENODEV;
mnt = vfs_kern_mount(type, sb_flags, name, data); - if (!IS_ERR(mnt) && (type->fs_flags & FS_HAS_SUBTYPE) && - !mnt->mnt_sb->s_subtype) - mnt = fs_set_subtype(mnt, fstype); + if (!IS_ERR(mnt) && (type->fs_flags & FS_HAS_SUBTYPE)) { + down_write(&mnt->mnt_sb->s_umount); + if (!mnt->mnt_sb->s_subtype) + mnt = fs_set_subtype(mnt, fstype); + up_write(&mnt->mnt_sb->s_umount); + }
put_filesystem(type); if (IS_ERR(mnt))