c/mirror_zfs
1
0
Fork 0
mirror of https://git.proxmox.com/git/mirror_zfs.git synced 2024-12-26 19:19:32 +03:00
Code Issues Packages Projects Releases Wiki Activity

Prevent incorrect datasets being mounted

Browse Source 
During a mount, zpl_mount_impl(), uses sget() with the callback
zpl_test_super() to find a super_block with a matching objset,
stored in z_os. It does so without taking the teardown lock on
the zfsvfs.

The problem is that operations like rollback will replace the
z_os. And, there is a window where the objset in the rollback
is freed, but z_os still points to it. Then, a mount like
operation, for instance a clone, can reallocate that exact same
pointer and zpl_test_super() will then match the super_block
associated with the rollback as opposed to the clone.

This fix tests for a match and if so, takes the teardown lock
before doing the final match test.

Reviewed-by: Richard Yao <richard.yao@alumni.stonybrook.edu>
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: John Poduska <jpoduska@datto.com>
Closes #14518
This commit is contained in:
John Poduska 2023-02-27 19:49:34 -05:00 committed by GitHub
parent 4b9bc6345e
commit 73c383f541
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
1 changed files with 20 additions and 2 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

22
module/os/linux/zfs/zpl_super.c
View File

@ -20,6 +20,7 @@
*/ */
/* /*
* Copyright (c) 2011, Lawrence Livermore National Security, LLC. * Copyright (c) 2011, Lawrence Livermore National Security, LLC.
* Copyright (c) 2023, Datto Inc. All rights reserved.
*/ */
@ -276,11 +277,28 @@ zpl_test_super(struct super_block *s, void *data)
{ {
zfsvfs_t *zfsvfs = s->s_fs_info; zfsvfs_t *zfsvfs = s->s_fs_info;
objset_t *os = data; objset_t *os = data;
int match;
if (zfsvfs == NULL) /*
* If the os doesn't match the z_os in the super_block, assume it is
* not a match. Matching would imply a multimount of a dataset. It is
* possible that during a multimount, there is a simultaneous operation
* that changes the z_os, e.g., rollback, where the match will be
* missed, but in that case the user will get an EBUSY.
*/
if (zfsvfs == NULL || os != zfsvfs->z_os)
return (0); return (0);
return (os == zfsvfs->z_os); /*
* If they do match, recheck with the lock held to prevent mounting the
* wrong dataset since z_os can be stale when the teardown lock is held.
*/
if (zpl_enter(zfsvfs, FTAG) != 0)
return (0);
match = (os == zfsvfs->z_os);
zpl_exit(zfsvfs, FTAG);
return (match);
} }
static struct super_block * static struct super_block *