From: Sasha Levin <sashal@kernel.org>
To: stable@vger.kernel.org
Cc: Bryam Vargas <hexlabsecurity@proton.me>,
Stephen Smalley <stephen.smalley.work@gmail.com>,
Paul Moore <paul@paul-moore.com>, Sasha Levin <sashal@kernel.org>
Subject: [PATCH 6.6.y 9/9] selinux: require a class's permission values to cover its permission count
Date: Tue, 25 Aug 2026 07:24:40 -0400 [thread overview]
Message-ID: <20260825112440.3543785-9-sashal@kernel.org> (raw)
In-Reply-To: <20260825112440.3543785-1-sashal@kernel.org>
From: Bryam Vargas <hexlabsecurity@proton.me>
[ Upstream commit b98a8ac50775540f3804397ed08f61ef9910bcab ]
security_get_permissions() sizes an array by the class's permissions.nprim
and fills it at value - 1, from the inherited common's permission table and
then the class's own. A value no permission defines leaves a NULL that
sel_make_perm_files() passes to d_alloc_name(), an oops inside
sel_write_load() that strands selinux_state.policy_mutex and leaves every
later load in uninterruptible sleep; two permissions sharing a value
overwrite the first kstrdup(). Bounding each value by nprim catches
neither, and neither would a count: the symbol table is keyed on the
permission name, so duplicates pass.
Track the values each permission table claims and require them to cover
exactly what its count declares, rejecting a count no value can reach.
Conforming policies are unaffected.
Cc: stable@vger.kernel.org
Fixes: 55fcf09b3fe4 ("selinux: add support for querying object classes and permissions from the running policy")
Signed-off-by: Bryam Vargas <hexlabsecurity@proton.me>
Acked-by: Stephen Smalley <stephen.smalley.work@gmail.com>
Signed-off-by: Paul Moore <paul@paul-moore.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
security/selinux/ss/policydb.c | 51 ++++++++++++++++++++++++++++++----
1 file changed, 46 insertions(+), 5 deletions(-)
diff --git a/security/selinux/ss/policydb.c b/security/selinux/ss/policydb.c
index 9bf9327fae365..084ee434dfbec 100644
--- a/security/selinux/ss/policydb.c
+++ b/security/selinux/ss/policydb.c
@@ -1099,7 +1099,18 @@ int str_read(char **strp, gfp_t flags, struct policy_file *fp, u32 len)
return 0;
}
-static int perm_read(struct policydb *p, struct symtab *s, struct policy_file *fp)
+/*
+ * Bitmap of the permission values a symtab has claimed. Values are 1-based
+ * and bounded by SEL_VEC_MAX, the width of an access vector, so the whole set
+ * fits in a u32 and the callers reject an nprim past that width.
+ */
+static u32 perm_claimed_mask(u32 nprim)
+{
+ return nprim ? U32_MAX >> (SEL_VEC_MAX - nprim) : 0;
+}
+
+static int perm_read(struct policydb *p, struct symtab *s,
+ struct policy_file *fp, u32 *claimed)
{
char *key = NULL;
struct perm_datum *perdatum;
@@ -1123,6 +1134,10 @@ static int perm_read(struct policydb *p, struct symtab *s, struct policy_file *f
/* indexes an nprim-sized array in security_get_permissions() */
if (perdatum->value > s->nprim)
goto bad;
+ /* two permissions cannot share one slot of that array */
+ if (*claimed & (1U << (perdatum->value - 1)))
+ goto bad;
+ *claimed |= 1U << (perdatum->value - 1);
rc = str_read(&key, GFP_KERNEL, fp, len);
if (rc)
@@ -1143,7 +1158,7 @@ static int common_read(struct policydb *p, struct symtab *s, struct policy_file
char *key = NULL;
struct common_datum *comdatum;
__le32 buf[4];
- u32 i, len, nel;
+ u32 i, len, nel, claimed = 0;
int rc;
comdatum = kzalloc(sizeof(*comdatum), GFP_KERNEL);
@@ -1165,17 +1180,28 @@ static int common_read(struct policydb *p, struct symtab *s, struct policy_file
if (rc)
goto bad;
comdatum->permissions.nprim = le32_to_cpu(buf[2]);
+ /* no permission value can reach a slot past SEL_VEC_MAX */
+ rc = -EINVAL;
+ if (comdatum->permissions.nprim > SEL_VEC_MAX)
+ goto bad;
rc = str_read(&key, GFP_KERNEL, fp, len);
if (rc)
goto bad;
for (i = 0; i < nel; i++) {
- rc = perm_read(p, &comdatum->permissions, fp);
+ rc = perm_read(p, &comdatum->permissions, fp, &claimed);
if (rc)
goto bad;
}
+ rc = -EINVAL;
+ if (claimed != perm_claimed_mask(comdatum->permissions.nprim)) {
+ pr_err("SELinux: common %s does not define every permission it declares\n",
+ key);
+ goto bad;
+ }
+
rc = symtab_insert(s, key, comdatum);
if (rc)
goto bad;
@@ -1309,7 +1335,7 @@ static int class_read(struct policydb *p, struct symtab *s, struct policy_file *
char *key = NULL;
struct class_datum *cladatum;
__le32 buf[6];
- u32 i, len, len2, ncons, nel, val;
+ u32 i, len, len2, ncons, nel, val, claimed = 0, inherited = 0;
int rc;
cladatum = kzalloc(sizeof(*cladatum), GFP_KERNEL);
@@ -1337,6 +1363,10 @@ static int class_read(struct policydb *p, struct symtab *s, struct policy_file *
if (rc)
goto bad;
cladatum->permissions.nprim = le32_to_cpu(buf[3]);
+ /* no permission value can reach a slot past SEL_VEC_MAX */
+ rc = -EINVAL;
+ if (cladatum->permissions.nprim > SEL_VEC_MAX)
+ goto bad;
ncons = le32_to_cpu(buf[5]);
@@ -1371,11 +1401,22 @@ static int class_read(struct policydb *p, struct symtab *s, struct policy_file *
}
}
for (i = 0; i < nel; i++) {
- rc = perm_read(p, &cladatum->permissions, fp);
+ rc = perm_read(p, &cladatum->permissions, fp, &claimed);
if (rc)
goto bad;
}
+ /* the class's own permissions must claim the slots the common leaves */
+ if (cladatum->comdatum)
+ inherited = cladatum->comdatum->permissions.nprim;
+ rc = -EINVAL;
+ if (claimed != (perm_claimed_mask(cladatum->permissions.nprim) &
+ ~perm_claimed_mask(inherited))) {
+ pr_err("SELinux: class %s does not define every permission it declares\n",
+ key);
+ goto bad;
+ }
+
rc = read_cons_helper(p, &cladatum->constraints, ncons, 0, fp);
if (rc)
goto bad;
--
2.53.0
prev parent reply other threads:[~2026-08-25 11:24 UTC|newest]
Thread overview: 10+ messages / expand[flat|nested] mbox.gz Atom feed top
2026-08-20 10:41 FAILED: patch "[PATCH] selinux: require a class's permission values to cover its" failed to apply to 6.6-stable tree gregkh
2026-08-25 11:24 ` [PATCH 6.6.y 1/9] selinux: fix style issues in security/selinux/ss/policydb.h Sasha Levin
2026-08-25 11:24 ` [PATCH 6.6.y 2/9] selinux: fix style issues in security/selinux/ss/policydb.c Sasha Levin
2026-08-25 11:24 ` [PATCH 6.6.y 3/9] selinux: use known type instead of void pointer Sasha Levin
2026-08-25 11:24 ` [PATCH 6.6.y 4/9] selinux: avoid unnecessary indirection in struct level_datum Sasha Levin
2026-08-25 11:24 ` [PATCH 6.6.y 5/9] selinux: make more use of str_read() when loading the policy Sasha Levin
2026-08-25 11:24 ` [PATCH 6.6.y 6/9] selinux: use u16 for security classes Sasha Levin
2026-08-25 11:24 ` [PATCH 6.6.y 7/9] selinux: more strict policy parsing Sasha Levin
2026-08-25 11:24 ` [PATCH 6.6.y 8/9] selinux: reject a permission value exceeding the class permission count Sasha Levin
2026-08-25 11:24 ` Sasha Levin [this message]
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20260825112440.3543785-9-sashal@kernel.org \
--to=sashal@kernel.org \
--cc=hexlabsecurity@proton.me \
--cc=paul@paul-moore.com \
--cc=stable@vger.kernel.org \
--cc=stephen.smalley.work@gmail.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox