security/integrity/ima/ima_policy.c | 65 +++++++++++++++-------------- 1 file changed, 33 insertions(+), 32 deletions(-)
From: Roberto Sassu <roberto.sassu@huawei.com>
Maintain a global sequence number, and set it to individual policy rules,
when they are created.
When a rule is stale, wait for the global sequence number to increase,
which happens when the LSM policy has been fully updated.
Finally, restart the ima_match_policy() loop, which should not encounter
the stale rule anymore.
On policy update, already increase the rule sequence number, so that a wait
can be implemented for the next policy update.
Signed-off-by: Roberto Sassu <roberto.sassu@huawei.com>
---
security/integrity/ima/ima_policy.c | 65 +++++++++++++++--------------
1 file changed, 33 insertions(+), 32 deletions(-)
diff --git a/security/integrity/ima/ima_policy.c b/security/integrity/ima/ima_policy.c
index c0556907c2e6..4f8275e216fe 100644
--- a/security/integrity/ima/ima_policy.c
+++ b/security/integrity/ima/ima_policy.c
@@ -51,9 +51,12 @@
#define INVALID_PCR(a) (((a) < 0) || \
(a) >= (sizeof_field(struct ima_iint_cache, measured_pcrs) * 8))
+DECLARE_WAIT_QUEUE_HEAD(wait_queue_t);
+
int ima_policy_flag;
static int temp_ima_appraise;
static int build_ima_appraise __ro_after_init;
+static int global_seqno;
atomic_t ima_setxattr_allowed_hash_algorithms;
@@ -122,6 +125,7 @@ struct ima_rule_entry {
struct ima_rule_opt_list *keyrings; /* Measure keys added to these keyrings */
struct ima_rule_opt_list *label; /* Measure data grouped under this label */
struct ima_template_desc *template;
+ int seqno;
};
/*
@@ -442,6 +446,8 @@ static int ima_lsm_update_rule(struct ima_rule_entry *entry)
if (!nentry)
return -ENOMEM;
+ nentry->seqno++;
+
list_replace_rcu(&entry->list, &nentry->list);
synchronize_rcu();
/*
@@ -497,6 +503,8 @@ int ima_lsm_policy_change(struct notifier_block *nb, unsigned long event,
return NOTIFY_DONE;
ima_lsm_update_rules();
+ global_seqno++;
+ wake_up(&wait_queue_t);
return NOTIFY_OK;
}
@@ -560,18 +568,16 @@ static bool ima_match_rule_data(struct ima_rule_entry *rule,
* @mask: requested action (MAY_READ | MAY_WRITE | MAY_APPEND | MAY_EXEC)
* @func_data: func specific data, may be NULL
*
- * Returns true on rule match, false on failure.
+ * Returns 1 on rule match, 0 on mismatch, -ESTALE on stale policy.
*/
-static bool ima_match_rules(struct ima_rule_entry *rule,
- struct mnt_idmap *idmap,
- struct inode *inode, const struct cred *cred,
- u32 secid, enum ima_hooks func, int mask,
- const char *func_data)
+static int ima_match_rules(struct ima_rule_entry *rule,
+ struct mnt_idmap *idmap,
+ struct inode *inode, const struct cred *cred,
+ u32 secid, enum ima_hooks func, int mask,
+ const char *func_data)
{
int i;
- bool result = false;
struct ima_rule_entry *lsm_rule = rule;
- bool rule_reinitialized = false;
if ((rule->flags & IMA_FUNC) &&
(rule->func != func && func != POST_SETATTR))
@@ -642,7 +648,6 @@ static bool ima_match_rules(struct ima_rule_entry *rule,
return false;
}
-retry:
switch (i) {
case LSM_OBJ_USER:
case LSM_OBJ_ROLE:
@@ -663,27 +668,13 @@ static bool ima_match_rules(struct ima_rule_entry *rule,
break;
}
- if (rc == -ESTALE && !rule_reinitialized) {
- lsm_rule = ima_lsm_copy_rule(rule);
- if (lsm_rule) {
- rule_reinitialized = true;
- goto retry;
- }
- }
- if (!rc) {
- result = false;
- goto out;
- }
+ if (!rc)
+ return false;
+ else if (rc == -ESTALE)
+ return rc;
}
- result = true;
-out:
- if (rule_reinitialized) {
- for (i = 0; i < MAX_LSM_RULES; i++)
- ima_filter_rule_free(lsm_rule->lsm[i].rule);
- kfree(lsm_rule);
- }
- return result;
+ return true;
}
/*
@@ -741,12 +732,12 @@ int ima_match_policy(struct mnt_idmap *idmap, struct inode *inode,
const char *func_data, unsigned int *allowed_algos)
{
struct ima_rule_entry *entry;
- int action = 0, actmask = flags | (flags << 1);
+ int action = 0, rc, actmask = flags | (flags << 1);
struct list_head *ima_rules_tmp;
if (template_desc && !*template_desc)
*template_desc = ima_template_desc_current();
-
+retry:
rcu_read_lock();
ima_rules_tmp = rcu_dereference(ima_rules);
list_for_each_entry_rcu(entry, ima_rules_tmp, list) {
@@ -754,9 +745,18 @@ int ima_match_policy(struct mnt_idmap *idmap, struct inode *inode,
if (!(entry->action & actmask))
continue;
- if (!ima_match_rules(entry, idmap, inode, cred, secid,
- func, mask, func_data))
+ rc = ima_match_rules(entry, idmap, inode, cred, secid,
+ func, mask, func_data);
+ if (!rc)
continue;
+ else if (rc == -ESTALE) {
+ rcu_read_unlock();
+
+ wait_event_interruptible(wait_queue_t,
+ (global_seqno == entry->seqno + 1));
+
+ goto retry;
+ }
action |= entry->flags & IMA_NONACTION_FLAGS;
@@ -1153,6 +1153,7 @@ static int ima_lsm_rule_init(struct ima_rule_entry *entry,
result = 0;
}
+ entry->seqno = global_seqno;
return result;
}
--
2.34.1On Tue, 2024-05-07 at 11:28 +0200, Roberto Sassu wrote:
> From: Roberto Sassu <roberto.sassu@huawei.com>
>
> Maintain a global sequence number, and set it to individual policy rules,
> when they are created.
Just did an attempt, to see if this path is viable.
This patch would be an alternative to:
[PATCH v3] ima: Avoid blocking in RCU read-side critical section
Roberto
> When a rule is stale, wait for the global sequence number to increase,
> which happens when the LSM policy has been fully updated.
>
> Finally, restart the ima_match_policy() loop, which should not encounter
> the stale rule anymore.
>
> On policy update, already increase the rule sequence number, so that a wait
> can be implemented for the next policy update.
>
> Signed-off-by: Roberto Sassu <roberto.sassu@huawei.com>
> ---
> security/integrity/ima/ima_policy.c | 65 +++++++++++++++--------------
> 1 file changed, 33 insertions(+), 32 deletions(-)
>
> diff --git a/security/integrity/ima/ima_policy.c b/security/integrity/ima/ima_policy.c
> index c0556907c2e6..4f8275e216fe 100644
> --- a/security/integrity/ima/ima_policy.c
> +++ b/security/integrity/ima/ima_policy.c
> @@ -51,9 +51,12 @@
> #define INVALID_PCR(a) (((a) < 0) || \
> (a) >= (sizeof_field(struct ima_iint_cache, measured_pcrs) * 8))
>
> +DECLARE_WAIT_QUEUE_HEAD(wait_queue_t);
> +
> int ima_policy_flag;
> static int temp_ima_appraise;
> static int build_ima_appraise __ro_after_init;
> +static int global_seqno;
>
> atomic_t ima_setxattr_allowed_hash_algorithms;
>
> @@ -122,6 +125,7 @@ struct ima_rule_entry {
> struct ima_rule_opt_list *keyrings; /* Measure keys added to these keyrings */
> struct ima_rule_opt_list *label; /* Measure data grouped under this label */
> struct ima_template_desc *template;
> + int seqno;
> };
>
> /*
> @@ -442,6 +446,8 @@ static int ima_lsm_update_rule(struct ima_rule_entry *entry)
> if (!nentry)
> return -ENOMEM;
>
> + nentry->seqno++;
> +
> list_replace_rcu(&entry->list, &nentry->list);
> synchronize_rcu();
> /*
> @@ -497,6 +503,8 @@ int ima_lsm_policy_change(struct notifier_block *nb, unsigned long event,
> return NOTIFY_DONE;
>
> ima_lsm_update_rules();
> + global_seqno++;
> + wake_up(&wait_queue_t);
> return NOTIFY_OK;
> }
>
> @@ -560,18 +568,16 @@ static bool ima_match_rule_data(struct ima_rule_entry *rule,
> * @mask: requested action (MAY_READ | MAY_WRITE | MAY_APPEND | MAY_EXEC)
> * @func_data: func specific data, may be NULL
> *
> - * Returns true on rule match, false on failure.
> + * Returns 1 on rule match, 0 on mismatch, -ESTALE on stale policy.
> */
> -static bool ima_match_rules(struct ima_rule_entry *rule,
> - struct mnt_idmap *idmap,
> - struct inode *inode, const struct cred *cred,
> - u32 secid, enum ima_hooks func, int mask,
> - const char *func_data)
> +static int ima_match_rules(struct ima_rule_entry *rule,
> + struct mnt_idmap *idmap,
> + struct inode *inode, const struct cred *cred,
> + u32 secid, enum ima_hooks func, int mask,
> + const char *func_data)
> {
> int i;
> - bool result = false;
> struct ima_rule_entry *lsm_rule = rule;
> - bool rule_reinitialized = false;
>
> if ((rule->flags & IMA_FUNC) &&
> (rule->func != func && func != POST_SETATTR))
> @@ -642,7 +648,6 @@ static bool ima_match_rules(struct ima_rule_entry *rule,
> return false;
> }
>
> -retry:
> switch (i) {
> case LSM_OBJ_USER:
> case LSM_OBJ_ROLE:
> @@ -663,27 +668,13 @@ static bool ima_match_rules(struct ima_rule_entry *rule,
> break;
> }
>
> - if (rc == -ESTALE && !rule_reinitialized) {
> - lsm_rule = ima_lsm_copy_rule(rule);
> - if (lsm_rule) {
> - rule_reinitialized = true;
> - goto retry;
> - }
> - }
> - if (!rc) {
> - result = false;
> - goto out;
> - }
> + if (!rc)
> + return false;
> + else if (rc == -ESTALE)
> + return rc;
> }
> - result = true;
>
> -out:
> - if (rule_reinitialized) {
> - for (i = 0; i < MAX_LSM_RULES; i++)
> - ima_filter_rule_free(lsm_rule->lsm[i].rule);
> - kfree(lsm_rule);
> - }
> - return result;
> + return true;
> }
>
> /*
> @@ -741,12 +732,12 @@ int ima_match_policy(struct mnt_idmap *idmap, struct inode *inode,
> const char *func_data, unsigned int *allowed_algos)
> {
> struct ima_rule_entry *entry;
> - int action = 0, actmask = flags | (flags << 1);
> + int action = 0, rc, actmask = flags | (flags << 1);
> struct list_head *ima_rules_tmp;
>
> if (template_desc && !*template_desc)
> *template_desc = ima_template_desc_current();
> -
> +retry:
> rcu_read_lock();
> ima_rules_tmp = rcu_dereference(ima_rules);
> list_for_each_entry_rcu(entry, ima_rules_tmp, list) {
> @@ -754,9 +745,18 @@ int ima_match_policy(struct mnt_idmap *idmap, struct inode *inode,
> if (!(entry->action & actmask))
> continue;
>
> - if (!ima_match_rules(entry, idmap, inode, cred, secid,
> - func, mask, func_data))
> + rc = ima_match_rules(entry, idmap, inode, cred, secid,
> + func, mask, func_data);
> + if (!rc)
> continue;
> + else if (rc == -ESTALE) {
> + rcu_read_unlock();
> +
> + wait_event_interruptible(wait_queue_t,
> + (global_seqno == entry->seqno + 1));
> +
> + goto retry;
> + }
>
> action |= entry->flags & IMA_NONACTION_FLAGS;
>
> @@ -1153,6 +1153,7 @@ static int ima_lsm_rule_init(struct ima_rule_entry *entry,
> result = 0;
> }
>
> + entry->seqno = global_seqno;
> return result;
> }
>
Hi Roberto,
On Tue, 2024-05-07 at 11:32 +0200, Roberto Sassu wrote:
> On Tue, 2024-05-07 at 11:28 +0200, Roberto Sassu wrote:
> > From: Roberto Sassu <roberto.sassu@huawei.com>
> >
> > Maintain a global sequence number, and set it to individual policy rules,
> > when they are created.
>
> Just did an attempt, to see if this path is viable.
>
> This patch would be an alternative to:
>
> [PATCH v3] ima: Avoid blocking in RCU read-side critical section
Stephen had said,
"Sidebar: the refactoring of the SELinux policy loading logic may have
made it possible to revisit the approaches here to permit holding a
reference to the policy from which the rule was derived so that we
don't have to return -ESTALE in this scenario."
Removing -ESTALE would be the best solution. We could then remove the -ESTALE
detection.
I assume the change would be in selinux_policy_commit(). Instead of freeing the
old policy, define and increment a per policy reference count for each
registered notifier callback.
/* Free the old policy */
synchronize_rcu();
selinux_policy_free(oldpolicy);
kfree(load_state->convert_data);
/* Notify others of the policy change */
selinux_notify_policy_change(seqno);
Mimi
© 2016 - 2024 Red Hat, Inc.