commit 901ef845fa2469c ("selinux: allow per-file labeling for cgroupfs") broke mounting of cgroup2 under older SELinux policies which lacked a genfscon rule for cgroup2. This prevents mounting of cgroup2 even when SELinux is permissive.
Change the handling when there is no genfscon rule in policy to just mark the inode unlabeled and not return an error to the caller. This permits mounting and access if allowed by policy, e.g. to unconfined domains.
I also considered changing the behavior of security_genfs_sid() to never return -ENOENT, but the current behavior is relied upon by other callers to perform caller-specific handling.
Fixes: 901ef845fa2469c ("selinux: allow per-file labeling for cgroupfs") CC: stable@vger.kernel.org Reported-by: Dmitry Vyukov dvyukov@google.com Reported-by: Waiman Long longman@redhat.com Signed-off-by: Stephen Smalley sds@tycho.nsa.gov --- security/selinux/hooks.c | 5 +++++ 1 file changed, 5 insertions(+)
diff --git a/security/selinux/hooks.c b/security/selinux/hooks.c index f78318af8254..58fee382a3bb 100644 --- a/security/selinux/hooks.c +++ b/security/selinux/hooks.c @@ -1508,6 +1508,11 @@ static int selinux_genfs_get_sid(struct dentry *dentry, } rc = security_genfs_sid(&selinux_state, sb->s_type->name, path, tclass, sid); + if (rc == -ENOENT) { + /* No match in policy, mark as unlabeled. */ + *sid = SECINITSID_UNLABELED; + rc = 0; + } } free_page((unsigned long)buffer); return rc;
On 09/04/2018 04:51 PM, Stephen Smalley wrote:
commit 901ef845fa2469c ("selinux: allow per-file labeling for cgroupfs") broke mounting of cgroup2 under older SELinux policies which lacked a genfscon rule for cgroup2. This prevents mounting of cgroup2 even when SELinux is permissive.
Change the handling when there is no genfscon rule in policy to just mark the inode unlabeled and not return an error to the caller. This permits mounting and access if allowed by policy, e.g. to unconfined domains.
I also considered changing the behavior of security_genfs_sid() to never return -ENOENT, but the current behavior is relied upon by other callers to perform caller-specific handling.
Fixes: 901ef845fa2469c ("selinux: allow per-file labeling for cgroupfs") CC: stable@vger.kernel.org Reported-by: Dmitry Vyukov dvyukov@google.com Reported-by: Waiman Long longman@redhat.com Signed-off-by: Stephen Smalley sds@tycho.nsa.gov
security/selinux/hooks.c | 5 +++++ 1 file changed, 5 insertions(+)
diff --git a/security/selinux/hooks.c b/security/selinux/hooks.c index f78318af8254..58fee382a3bb 100644 --- a/security/selinux/hooks.c +++ b/security/selinux/hooks.c @@ -1508,6 +1508,11 @@ static int selinux_genfs_get_sid(struct dentry *dentry, } rc = security_genfs_sid(&selinux_state, sb->s_type->name, path, tclass, sid);
if (rc == -ENOENT) {/* No match in policy, mark as unlabeled. */*sid = SECINITSID_UNLABELED;rc = 0;}
I have tested this patch and it works in my case.
Tested-by: Waiman Long longman@redhat.com
On Tue, Sep 4, 2018 at 4:49 PM Stephen Smalley sds@tycho.nsa.gov wrote:
commit 901ef845fa2469c ("selinux: allow per-file labeling for cgroupfs") broke mounting of cgroup2 under older SELinux policies which lacked a genfscon rule for cgroup2. This prevents mounting of cgroup2 even when SELinux is permissive.
Change the handling when there is no genfscon rule in policy to just mark the inode unlabeled and not return an error to the caller. This permits mounting and access if allowed by policy, e.g. to unconfined domains.
I also considered changing the behavior of security_genfs_sid() to never return -ENOENT, but the current behavior is relied upon by other callers to perform caller-specific handling.
Fixes: 901ef845fa2469c ("selinux: allow per-file labeling for cgroupfs") CC: stable@vger.kernel.org Reported-by: Dmitry Vyukov dvyukov@google.com Reported-by: Waiman Long longman@redhat.com Signed-off-by: Stephen Smalley sds@tycho.nsa.gov
security/selinux/hooks.c | 5 +++++ 1 file changed, 5 insertions(+)
Looks like a reasonable approach to me, merged into selinux/next, thanks.
As a FYI, since the US holiday and LSS-NA delayed the start of merging things into selinux/next I've updated selinux/next on top of v4.19-rc2 instead of -rc1 this time around.
diff --git a/security/selinux/hooks.c b/security/selinux/hooks.c index f78318af8254..58fee382a3bb 100644 --- a/security/selinux/hooks.c +++ b/security/selinux/hooks.c @@ -1508,6 +1508,11 @@ static int selinux_genfs_get_sid(struct dentry *dentry, } rc = security_genfs_sid(&selinux_state, sb->s_type->name, path, tclass, sid);
if (rc == -ENOENT) {/* No match in policy, mark as unlabeled. */*sid = SECINITSID_UNLABELED;rc = 0;} } free_page((unsigned long)buffer); return rc;-- 2.14.4
On Tue, Sep 4, 2018 at 6:18 PM Paul Moore paul@paul-moore.com wrote:
On Tue, Sep 4, 2018 at 4:49 PM Stephen Smalley sds@tycho.nsa.gov wrote:
commit 901ef845fa2469c ("selinux: allow per-file labeling for cgroupfs") broke mounting of cgroup2 under older SELinux policies which lacked a genfscon rule for cgroup2. This prevents mounting of cgroup2 even when SELinux is permissive.
Change the handling when there is no genfscon rule in policy to just mark the inode unlabeled and not return an error to the caller. This permits mounting and access if allowed by policy, e.g. to unconfined domains.
I also considered changing the behavior of security_genfs_sid() to never return -ENOENT, but the current behavior is relied upon by other callers to perform caller-specific handling.
Fixes: 901ef845fa2469c ("selinux: allow per-file labeling for cgroupfs") CC: stable@vger.kernel.org Reported-by: Dmitry Vyukov dvyukov@google.com Reported-by: Waiman Long longman@redhat.com Signed-off-by: Stephen Smalley sds@tycho.nsa.gov
security/selinux/hooks.c | 5 +++++ 1 file changed, 5 insertions(+)
Looks like a reasonable approach to me, merged into selinux/next, thanks.
I probably should expand a bit on this as Stephen's stable CC marking and the patch's inclusion in selinux/next (as opposed to selinux/stable-4.19) don't quite match. I merged this into the next branch because I didn't feel that this was a severe enough problem to warrant immediate inclusion into the stable-4.19 branch and since this is a user visible change I felt some additional time in the next branch would be valuable. However, I did leave the CC stable marking intact so that when this patch does hit Linus tree (expected during the v4.20 merge window) it should get backported to the various stable trees.
As a FYI, since the US holiday and LSS-NA delayed the start of merging things into selinux/next I've updated selinux/next on top of v4.19-rc2 instead of -rc1 this time around.
diff --git a/security/selinux/hooks.c b/security/selinux/hooks.c index f78318af8254..58fee382a3bb 100644 --- a/security/selinux/hooks.c +++ b/security/selinux/hooks.c @@ -1508,6 +1508,11 @@ static int selinux_genfs_get_sid(struct dentry *dentry, } rc = security_genfs_sid(&selinux_state, sb->s_type->name, path, tclass, sid);
if (rc == -ENOENT) {/* No match in policy, mark as unlabeled. */*sid = SECINITSID_UNLABELED;rc = 0;} } free_page((unsigned long)buffer); return rc;-- 2.14.4
-- paul moore www.paul-moore.com
linux-stable-mirror@lists.linaro.org
-
Paul Moore -
Stephen Smalley -
Waiman Long