Thread (35 messages) 35 messages, 5 authors, 2021-09-15

Re: [RFC PATCH v2 9/9] Smack: Brutalist io_uring support with debug

From: Casey Schaufler <casey@schaufler-ca.com>
Date: 2021-08-31 15:05:13
Also in: io-uring, linux-fsdevel, selinux

On 8/31/2021 7:44 AM, Paul Moore wrote:
On Wed, Aug 11, 2021 at 4:49 PM Paul Moore [off-list ref] wrote:
quoted
From: Casey Schaufler <casey@schaufler-ca.com>

Add Smack privilege checks for io_uring. Use CAP_MAC_OVERRIDE
for the override_creds case and CAP_MAC_ADMIN for creating a
polling thread. These choices are based on conjecture regarding
the intent of the surrounding code.

Signed-off-by: Casey Schaufler <casey@schaufler-ca.com>
[PM: make the smack_uring_* funcs static]
Signed-off-by: Paul Moore <paul@paul-moore.com>

---
v2:
- made the smack_uring_* funcs static
v1:
- initial draft
---
 security/smack/smack_lsm.c |   64 ++++++++++++++++++++++++++++++++++++++++++++
 1 file changed, 64 insertions(+)
diff --git a/security/smack/smack_lsm.c b/security/smack/smack_lsm.c
index 223a6da0e6dc..7fb094098f38 100644
--- a/security/smack/smack_lsm.c
+++ b/security/smack/smack_lsm.c
@@ -4691,6 +4691,66 @@ static int smack_dentry_create_files_as(struct dentry *dentry, int mode,
        return 0;
 }

+#ifdef CONFIG_IO_URING
+/**
+ * smack_uring_override_creds - Is io_uring cred override allowed?
+ * @new: the target creds
+ *
+ * Check to see if the current task is allowed to override it's credentials
+ * to service an io_uring operation.
+ */
+static int smack_uring_override_creds(const struct cred *new)
+{
+       struct task_smack *tsp = smack_cred(current_cred());
+       struct task_smack *nsp = smack_cred(new);
+
+#if 1
+       if (tsp->smk_task == nsp->smk_task)
+               pr_info("%s: Smack matches %s\n", __func__,
+                       tsp->smk_task->smk_known);
+       else
+               pr_info("%s: Smack override check %s to %s\n", __func__,
+                       tsp->smk_task->smk_known, nsp->smk_task->smk_known);
+#endif
Casey, with the idea of posting a v3 towards the end of the merge
window next week, without the RFC tag and with the intention of
merging it into -next during the first/second week of the -rcX phase,
do you have any objections to me removing the debug code (#if 1 ...
#endif) from your patch?  Did you have any other changes?
I have no other changes. And yes, the debug code should be stripped.
Thank you.

--
paul moore
www.paul-moore.com
Keyboard shortcuts
hback out one level
jnext message in thread
kprevious message in thread
ldrill in
Escclose help / fold thread tree
?toggle this help