From: Henry Martin <bsdhenrymartin@gmail.com>
To: john.johansen@canonical.com, john@apparmor.net,
georgia.garcia@canonical.com
Cc: apparmor@lists.ubuntu.com, linux-kernel@vger.kernel.org,
Henry Martin <bsdhenrymartin@gmail.com>,
stable@vger.kernel.org
Subject: [PATCH] apparmor: fix out-of-bounds read of xmatch perms during exec attachment
Date: Sat, 7 Nov 2026 02:54:06 +0800 [thread overview]
Message-ID: <20261106185406.3332779-1-bsdhenrymartin@gmail.com> (raw)
verify_profile() validates the accept table indexes of the file
and policy dfas, but never of the attach xmatch dfa. When a policy
carries an explicit xmatch perms table, the blob accept values are
used as indexes into that table unchecked. At exec time,
aa_lookup_perms() uses one as an index into the xmatch perms array
with no bounds check.
The index is a full 32-bit value from the blob, so the read can be
placed gigabytes past the table, across the entire direct map. The
bytes read are interpreted as the allow/deny bits of the exec
request, and the transition outcome is observable by the task,
turning the bug into a bit-per-query read oracle over kernel
memory. An index pointing at unmapped memory oopses in
find_attach() at exec time. E.g. on mainline v7.4-rc1 (KASAN
build):
BUG: KASAN: slab-out-of-bounds in find_attach+0xc50/0xfe0
Read of size 4 at addr ffff88810fb2f934 by task su/495
CPU: 5 UID: 1000 PID: 495 Comm: su
Call Trace:
kasan_report
find_attach
profile_transition
apparmor_bprm_creds_for_exec
bprm_execve
__x64_sys_execve
Run verify_dfa_accept_index() on the attach xmatch dfa in
verify_profile(), the same check the file and policy dfas already
get.
This issue was discovered by Tencent CodeBuddy Security.
Cc: stable@vger.kernel.org
Fixes: 6f442d42c0d89 ("apparmor: fix profile verification and enable it")
Signed-off-by: Henry Martin <bsdhenrymartin@gmail.com>
---
security/apparmor/policy_unpack.c | 9 +++++++++
1 file changed, 9 insertions(+)
diff --git a/security/apparmor/policy_unpack.c b/security/apparmor/policy_unpack.c
index 35dfa3b451f39..054aad13f5e48 100644
--- a/security/apparmor/policy_unpack.c
+++ b/security/apparmor/policy_unpack.c
@@ -1593,6 +1593,15 @@ static int verify_profile(struct aa_profile *profile)
return -EPROTO;
}
+ if (profile->attach.xmatch->dfa &&
+ !verify_dfa_accept_index(profile->attach.xmatch->dfa,
+ profile->attach.xmatch->size)) {
+ audit_iface(profile, NULL, NULL,
+ "Unpack: xmatch Invalid named transition", NULL,
+ -EPROTO);
+ return -EPROTO;
+ }
+
if (!verify_perms(rules->file)) {
audit_iface(profile, NULL, NULL,
"Unpack: Invalid perm index", NULL, -EPROTO);
--
2.43.7
reply other threads:[~2026-11-06 18:54 UTC|newest]
Thread overview: [no followups] expand[flat|nested] mbox.gz Atom feed
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=20261106185406.3332779-1-bsdhenrymartin@gmail.com \
--to=bsdhenrymartin@gmail.com \
--cc=apparmor@lists.ubuntu.com \
--cc=georgia.garcia@canonical.com \
--cc=john.johansen@canonical.com \
--cc=john@apparmor.net \
--cc=linux-kernel@vger.kernel.org \
--cc=stable@vger.kernel.org \
/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
all inboxes | Powered by JetHome®