diff options
| author | Al Viro <viro@zeniv.linux.org.uk> | 2021-01-05 22:43:46 +0300 |
|---|---|---|
| committer | Al Viro <viro@zeniv.linux.org.uk> | 2021-01-16 23:11:35 +0300 |
| commit | d36a1dd9f77ae1e72da48f4123ed35627848507d (patch) | |
| tree | 7252b0d4174a46d062a76fea648fdde2fd992f9a /security | |
| parent | a959a9782fa87669feeed095ced5d78181a7c02d (diff) | |
| download | linux-d36a1dd9f77ae1e72da48f4123ed35627848507d.tar.xz | |
dump_common_audit_data(): fix racy accesses to ->d_name
We are not guaranteed the locking environment that would prevent
dentry getting renamed right under us. And it's possible for
old long name to be freed after rename, leading to UAF here.
Cc: stable@kernel.org # v2.6.2+
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Diffstat (limited to 'security')
| -rw-r--r-- | security/lsm_audit.c | 7 |
1 files changed, 5 insertions, 2 deletions
diff --git a/security/lsm_audit.c b/security/lsm_audit.c index 7d8026f3f377..a0cd28cd31a8 100644 --- a/security/lsm_audit.c +++ b/security/lsm_audit.c @@ -275,7 +275,9 @@ static void dump_common_audit_data(struct audit_buffer *ab, struct inode *inode; audit_log_format(ab, " name="); + spin_lock(&a->u.dentry->d_lock); audit_log_untrustedstring(ab, a->u.dentry->d_name.name); + spin_unlock(&a->u.dentry->d_lock); inode = d_backing_inode(a->u.dentry); if (inode) { @@ -293,8 +295,9 @@ static void dump_common_audit_data(struct audit_buffer *ab, dentry = d_find_alias(inode); if (dentry) { audit_log_format(ab, " name="); - audit_log_untrustedstring(ab, - dentry->d_name.name); + spin_lock(&dentry->d_lock); + audit_log_untrustedstring(ab, dentry->d_name.name); + spin_unlock(&dentry->d_lock); dput(dentry); } audit_log_format(ab, " dev="); |