[PATCH] support for context based audit filtering
Darrel Goeddel [Fri, 24 Feb 2006 21:44:05 +0000 (15:44 -0600)]
The following patch provides selinux interfaces that will allow the audit
system to perform filtering based on the process context (user, role, type,
sensitivity, and clearance).  These interfaces will allow the selinux
module to perform efficient matches based on lower level selinux constructs,
rather than relying on context retrievals and string comparisons within
the audit module.  It also allows for dominance checks on the mls portion
of the contexts that are impossible with only string comparisons.

Signed-off-by: Darrel Goeddel <dgoeddel@trustedcs.com>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>

include/linux/audit.h
include/linux/selinux.h [new file with mode: 0644]
security/selinux/Makefile
security/selinux/avc.c
security/selinux/exports.c [new file with mode: 0644]
security/selinux/ss/mls.c
security/selinux/ss/mls.h
security/selinux/ss/services.c

index 39fef6e..740f950 100644 (file)
 #define AUDIT_PERS     10
 #define AUDIT_ARCH     11
 #define AUDIT_MSGTYPE  12
+#define AUDIT_SE_USER  13      /* security label user */
+#define AUDIT_SE_ROLE  14      /* security label role */
+#define AUDIT_SE_TYPE  15      /* security label type */
+#define AUDIT_SE_SEN   16      /* security label sensitivity label */
+#define AUDIT_SE_CLR   17      /* security label clearance label */
 
                                /* These are ONLY useful when checking
                                 * at syscall exit time (AUDIT_AT_EXIT). */
diff --git a/include/linux/selinux.h b/include/linux/selinux.h
new file mode 100644 (file)
index 0000000..9d684b1
--- /dev/null
@@ -0,0 +1,112 @@
+/*
+ * SELinux services exported to the rest of the kernel.
+ *
+ * Author: James Morris <jmorris@redhat.com>
+ *
+ * Copyright (C) 2005 Red Hat, Inc., James Morris <jmorris@redhat.com>
+ * Copyright (C) 2006 Trusted Computer Solutions, Inc. <dgoeddel@trustedcs.com>
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2,
+ * as published by the Free Software Foundation.
+ */
+#ifndef _LINUX_SELINUX_H
+#define _LINUX_SELINUX_H
+
+struct selinux_audit_rule;
+struct audit_context;
+
+#ifdef CONFIG_SECURITY_SELINUX
+
+/**
+ *     selinux_audit_rule_init - alloc/init an selinux audit rule structure.
+ *     @field: the field this rule refers to
+ *     @op: the operater the rule uses
+ *     @rulestr: the text "target" of the rule
+ *     @rule: pointer to the new rule structure returned via this
+ *
+ *     Returns 0 if successful, -errno if not.  On success, the rule structure
+ *     will be allocated internally.  The caller must free this structure with
+ *     selinux_audit_rule_free() after use.
+ */
+int selinux_audit_rule_init(u32 field, u32 op, char *rulestr,
+                            struct selinux_audit_rule **rule);
+
+/**
+ *     selinux_audit_rule_free - free an selinux audit rule structure.
+ *     @rule: pointer to the audit rule to be freed
+ *
+ *     This will free all memory associated with the given rule.
+ *     If @rule is NULL, no operation is performed.
+ */
+void selinux_audit_rule_free(struct selinux_audit_rule *rule);
+
+/**
+ *     selinux_audit_rule_match - determine if a context ID matches a rule.
+ *     @ctxid: the context ID to check
+ *     @field: the field this rule refers to
+ *     @op: the operater the rule uses
+ *     @rule: pointer to the audit rule to check against
+ *     @actx: the audit context (can be NULL) associated with the check
+ *
+ *     Returns 1 if the context id matches the rule, 0 if it does not, and
+ *     -errno on failure.
+ */
+int selinux_audit_rule_match(u32 ctxid, u32 field, u32 op,
+                             struct selinux_audit_rule *rule,
+                             struct audit_context *actx);
+
+/**
+ *     selinux_audit_set_callback - set the callback for policy reloads.
+ *     @callback: the function to call when the policy is reloaded
+ *
+ *     This sets the function callback function that will update the rules
+ *     upon policy reloads.  This callback should rebuild all existing rules
+ *     using selinux_audit_rule_init().
+ */
+void selinux_audit_set_callback(int (*callback)(void));
+
+/**
+ *     selinux_task_ctxid - determine a context ID for a process.
+ *     @tsk: the task object
+ *     @ctxid: ID value returned via this
+ *
+ *     On return, ctxid will contain an ID for the context.  This value
+ *     should only be used opaquely.
+ */
+void selinux_task_ctxid(struct task_struct *tsk, u32 *ctxid);
+
+#else
+
+static inline int selinux_audit_rule_init(u32 field, u32 op,
+                                          char *rulestr,
+                                          struct selinux_audit_rule **rule)
+{
+       return -ENOTSUPP;
+}
+
+static inline void selinux_audit_rule_free(struct selinux_audit_rule *rule)
+{
+       return;
+}
+
+static inline int selinux_audit_rule_match(u32 ctxid, u32 field, u32 op,
+                                           struct selinux_audit_rule *rule,
+                                           struct audit_context *actx)
+{
+       return 0;
+}
+
+static inline void selinux_audit_set_callback(int (*callback)(void))
+{
+       return;
+}
+
+static inline void selinux_task_ctxid(struct task_struct *tsk, u32 *ctxid)
+{
+       *ctxid = 0;
+}
+
+#endif /* CONFIG_SECURITY_SELINUX */
+
+#endif /* _LINUX_SELINUX_H */
index 688c0a2..faf2e02 100644 (file)
@@ -4,7 +4,7 @@
 
 obj-$(CONFIG_SECURITY_SELINUX) := selinux.o ss/
 
-selinux-y := avc.o hooks.o selinuxfs.o netlink.o nlmsgtab.o netif.o
+selinux-y := avc.o hooks.o selinuxfs.o netlink.o nlmsgtab.o netif.o exports.o
 
 selinux-$(CONFIG_SECURITY_NETWORK_XFRM) += xfrm.o
 
index ac5d69b..a300702 100644 (file)
@@ -800,7 +800,7 @@ out:
 int avc_ss_reset(u32 seqno)
 {
        struct avc_callback_node *c;
-       int i, rc = 0;
+       int i, rc = 0, tmprc;
        unsigned long flag;
        struct avc_node *node;
 
@@ -813,15 +813,16 @@ int avc_ss_reset(u32 seqno)
 
        for (c = avc_callbacks; c; c = c->next) {
                if (c->events & AVC_CALLBACK_RESET) {
-                       rc = c->callback(AVC_CALLBACK_RESET,
-                                        0, 0, 0, 0, NULL);
-                       if (rc)
-                               goto out;
+                       tmprc = c->callback(AVC_CALLBACK_RESET,
+                                           0, 0, 0, 0, NULL);
+                       /* save the first error encountered for the return
+                          value and continue processing the callbacks */
+                       if (!rc)
+                               rc = tmprc;
                }
        }
 
        avc_latest_notif_update(seqno, 0);
-out:
        return rc;
 }
 
diff --git a/security/selinux/exports.c b/security/selinux/exports.c
new file mode 100644 (file)
index 0000000..333c4c7
--- /dev/null
@@ -0,0 +1,28 @@
+/*
+ * SELinux services exported to the rest of the kernel.
+ *
+ * Author: James Morris <jmorris@redhat.com>
+ *
+ * Copyright (C) 2005 Red Hat, Inc., James Morris <jmorris@redhat.com>
+ * Copyright (C) 2006 Trusted Computer Solutions, Inc. <dgoeddel@trustedcs.com>
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2,
+ * as published by the Free Software Foundation.
+ */
+#include <linux/types.h>
+#include <linux/kernel.h>
+#include <linux/module.h>
+#include <linux/selinux.h>
+
+#include "security.h"
+#include "objsec.h"
+
+void selinux_task_ctxid(struct task_struct *tsk, u32 *ctxid)
+{
+       struct task_security_struct *tsec = tsk->security;
+       if (selinux_enabled)
+               *ctxid = tsec->sid;
+       else
+               *ctxid = 0;
+}
index 84047f6..7bc5b64 100644 (file)
@@ -8,7 +8,7 @@
  *
  *     Support for enhanced MLS infrastructure.
  *
- * Copyright (C) 2004-2005 Trusted Computer Solutions, Inc.
+ * Copyright (C) 2004-2006 Trusted Computer Solutions, Inc.
  */
 
 #include <linux/kernel.h>
@@ -385,6 +385,34 @@ out:
 }
 
 /*
+ * Set the MLS fields in the security context structure
+ * `context' based on the string representation in
+ * the string `str'.  This function will allocate temporary memory with the
+ * given constraints of gfp_mask.
+ */
+int mls_from_string(char *str, struct context *context, gfp_t gfp_mask)
+{
+       char *tmpstr, *freestr;
+       int rc;
+
+       if (!selinux_mls_enabled)
+               return -EINVAL;
+
+       /* we need freestr because mls_context_to_sid will change
+          the value of tmpstr */
+       tmpstr = freestr = kstrdup(str, gfp_mask);
+       if (!tmpstr) {
+               rc = -ENOMEM;
+       } else {
+               rc = mls_context_to_sid(':', &tmpstr, context,
+                                       NULL, SECSID_NULL);
+               kfree(freestr);
+       }
+
+       return rc;
+}
+
+/*
  * Copies the effective MLS range from `src' into `dst'.
  */
 static inline int mls_scopy_context(struct context *dst,
index 03de697..fbb42f0 100644 (file)
@@ -8,7 +8,7 @@
  *
  *     Support for enhanced MLS infrastructure.
  *
- * Copyright (C) 2004-2005 Trusted Computer Solutions, Inc.
+ * Copyright (C) 2004-2006 Trusted Computer Solutions, Inc.
  */
 
 #ifndef _SS_MLS_H_
@@ -27,6 +27,8 @@ int mls_context_to_sid(char oldc,
                       struct sidtab *s,
                       u32 def_sid);
 
+int mls_from_string(char *str, struct context *context, gfp_t gfp_mask);
+
 int mls_convert_context(struct policydb *oldp,
                        struct policydb *newp,
                        struct context *context);
index 6149248..7177e98 100644 (file)
@@ -7,12 +7,13 @@
  * Updated: Trusted Computer Solutions, Inc. <dgoeddel@trustedcs.com>
  *
  *     Support for enhanced MLS infrastructure.
+ *     Support for context based audit filters.
  *
  * Updated: Frank Mayer <mayerf@tresys.com> and Karl MacMillan <kmacmillan@tresys.com>
  *
  *     Added conditional policy language extensions
  *
- * Copyright (C) 2004-2005 Trusted Computer Solutions, Inc.
+ * Copyright (C) 2004-2006 Trusted Computer Solutions, Inc.
  * Copyright (C) 2003 - 2004 Tresys Technology, LLC
  * Copyright (C) 2003 Red Hat, Inc., James Morris <jmorris@redhat.com>
  *     This program is free software; you can redistribute it and/or modify
@@ -1811,3 +1812,235 @@ out:
        POLICY_RDUNLOCK;
        return rc;
 }
+
+struct selinux_audit_rule {
+       u32 au_seqno;
+       struct context au_ctxt;
+};
+
+void selinux_audit_rule_free(struct selinux_audit_rule *rule)
+{
+       if (rule) {
+               context_destroy(&rule->au_ctxt);
+               kfree(rule);
+       }
+}
+
+int selinux_audit_rule_init(u32 field, u32 op, char *rulestr,
+                            struct selinux_audit_rule **rule)
+{
+       struct selinux_audit_rule *tmprule;
+       struct role_datum *roledatum;
+       struct type_datum *typedatum;
+       struct user_datum *userdatum;
+       int rc = 0;
+
+       *rule = NULL;
+
+       if (!ss_initialized)
+               return -ENOTSUPP;
+
+       switch (field) {
+       case AUDIT_SE_USER:
+       case AUDIT_SE_ROLE:
+       case AUDIT_SE_TYPE:
+               /* only 'equals' and 'not equals' fit user, role, and type */
+               if (op != AUDIT_EQUAL && op != AUDIT_NOT_EQUAL)
+                       return -EINVAL;
+               break;
+       case AUDIT_SE_SEN:
+       case AUDIT_SE_CLR:
+               /* we do not allow a range, indicated by the presense of '-' */
+               if (strchr(rulestr, '-'))
+                       return -EINVAL;
+               break;
+       default:
+               /* only the above fields are valid */
+               return -EINVAL;
+       }
+
+       tmprule = kzalloc(sizeof(struct selinux_audit_rule), GFP_KERNEL);
+       if (!tmprule)
+               return -ENOMEM;
+
+       context_init(&tmprule->au_ctxt);
+
+       POLICY_RDLOCK;
+
+       tmprule->au_seqno = latest_granting;
+
+       switch (field) {
+       case AUDIT_SE_USER:
+               userdatum = hashtab_search(policydb.p_users.table, rulestr);
+               if (!userdatum)
+                       rc = -EINVAL;
+               else
+                       tmprule->au_ctxt.user = userdatum->value;
+               break;
+       case AUDIT_SE_ROLE:
+               roledatum = hashtab_search(policydb.p_roles.table, rulestr);
+               if (!roledatum)
+                       rc = -EINVAL;
+               else
+                       tmprule->au_ctxt.role = roledatum->value;
+               break;
+       case AUDIT_SE_TYPE:
+               typedatum = hashtab_search(policydb.p_types.table, rulestr);
+               if (!typedatum)
+                       rc = -EINVAL;
+               else
+                       tmprule->au_ctxt.type = typedatum->value;
+               break;
+       case AUDIT_SE_SEN:
+       case AUDIT_SE_CLR:
+               rc = mls_from_string(rulestr, &tmprule->au_ctxt, GFP_ATOMIC);
+               break;
+       }
+
+       POLICY_RDUNLOCK;
+
+       if (rc) {
+               selinux_audit_rule_free(tmprule);
+               tmprule = NULL;
+       }
+
+       *rule = tmprule;
+
+       return rc;
+}
+
+int selinux_audit_rule_match(u32 ctxid, u32 field, u32 op,
+                             struct selinux_audit_rule *rule,
+                             struct audit_context *actx)
+{
+       struct context *ctxt;
+       struct mls_level *level;
+       int match = 0;
+
+       if (!rule) {
+               audit_log(actx, GFP_ATOMIC, AUDIT_SELINUX_ERR,
+                         "selinux_audit_rule_match: missing rule\n");
+               return -ENOENT;
+       }
+
+       POLICY_RDLOCK;
+
+       if (rule->au_seqno < latest_granting) {
+               audit_log(actx, GFP_ATOMIC, AUDIT_SELINUX_ERR,
+                         "selinux_audit_rule_match: stale rule\n");
+               match = -ESTALE;
+               goto out;
+       }
+
+       ctxt = sidtab_search(&sidtab, ctxid);
+       if (!ctxt) {
+               audit_log(actx, GFP_ATOMIC, AUDIT_SELINUX_ERR,
+                         "selinux_audit_rule_match: unrecognized SID %d\n",
+                         ctxid);
+               match = -ENOENT;
+               goto out;
+       }
+
+       /* a field/op pair that is not caught here will simply fall through
+          without a match */
+       switch (field) {
+       case AUDIT_SE_USER:
+               switch (op) {
+               case AUDIT_EQUAL:
+                       match = (ctxt->user == rule->au_ctxt.user);
+                       break;
+               case AUDIT_NOT_EQUAL:
+                       match = (ctxt->user != rule->au_ctxt.user);
+                       break;
+               }
+               break;
+       case AUDIT_SE_ROLE:
+               switch (op) {
+               case AUDIT_EQUAL:
+                       match = (ctxt->role == rule->au_ctxt.role);
+                       break;
+               case AUDIT_NOT_EQUAL:
+                       match = (ctxt->role != rule->au_ctxt.role);
+                       break;
+               }
+               break;
+       case AUDIT_SE_TYPE:
+               switch (op) {
+               case AUDIT_EQUAL:
+                       match = (ctxt->type == rule->au_ctxt.type);
+                       break;
+               case AUDIT_NOT_EQUAL:
+                       match = (ctxt->type != rule->au_ctxt.type);
+                       break;
+               }
+               break;
+       case AUDIT_SE_SEN:
+       case AUDIT_SE_CLR:
+               level = (op == AUDIT_SE_SEN ?
+                        &ctxt->range.level[0] : &ctxt->range.level[1]);
+               switch (op) {
+               case AUDIT_EQUAL:
+                       match = mls_level_eq(&rule->au_ctxt.range.level[0],
+                                            level);
+                       break;
+               case AUDIT_NOT_EQUAL:
+                       match = !mls_level_eq(&rule->au_ctxt.range.level[0],
+                                             level);
+                       break;
+               case AUDIT_LESS_THAN:
+                       match = (mls_level_dom(&rule->au_ctxt.range.level[0],
+                                              level) &&
+                                !mls_level_eq(&rule->au_ctxt.range.level[0],
+                                              level));
+                       break;
+               case AUDIT_LESS_THAN_OR_EQUAL:
+                       match = mls_level_dom(&rule->au_ctxt.range.level[0],
+                                             level);
+                       break;
+               case AUDIT_GREATER_THAN:
+                       match = (mls_level_dom(level,
+                                             &rule->au_ctxt.range.level[0]) &&
+                                !mls_level_eq(level,
+                                              &rule->au_ctxt.range.level[0]));
+                       break;
+               case AUDIT_GREATER_THAN_OR_EQUAL:
+                       match = mls_level_dom(level,
+                                             &rule->au_ctxt.range.level[0]);
+                       break;
+               }
+       }
+
+out:
+       POLICY_RDUNLOCK;
+       return match;
+}
+
+static int (*aurule_callback)(void) = NULL;
+
+static int aurule_avc_callback(u32 event, u32 ssid, u32 tsid,
+                               u16 class, u32 perms, u32 *retained)
+{
+       int err = 0;
+
+       if (event == AVC_CALLBACK_RESET && aurule_callback)
+               err = aurule_callback();
+       return err;
+}
+
+static int __init aurule_init(void)
+{
+       int err;
+
+       err = avc_add_callback(aurule_avc_callback, AVC_CALLBACK_RESET,
+                              SECSID_NULL, SECSID_NULL, SECCLASS_NULL, 0);
+       if (err)
+               panic("avc_add_callback() failed, error %d\n", err);
+
+       return err;
+}
+__initcall(aurule_init);
+
+void selinux_audit_set_callback(int (*callback)(void))
+{
+       aurule_callback = callback;
+}