linux-stable.git/security/apparmor, branch v7.1.3

apparmor: advertise the tcp fast open fix is applied

2026-07-04T11:45:09+00:00

commit 2f6701a5ce6257ae7a64ddc6d89d0a08d2a034f8 upstream.

The fix for tcp-fast-open ensures that the connect permission is being
mediated correctly but it didn't add an artifact to the feature set to
advertise the fix is available. Add an artifact so that the test suite
can identify if the fix has not been properly applied or a new
unexpected regression has occurred.

Fixes: 4d587cd8a7215 ("apparmor: mediate the implicit connect of TCP fast open sendmsg")
Signed-off-by: John Johansen 
Signed-off-by: Greg Kroah-Hartman

apparmor: fix use-after-free in rawdata dedup loop

2026-07-04T11:45:03+00:00

commit 6f060496d03e4dc560a40f73770bd08335cb7a27 upstream.

aa_replace_profiles() walks ns->rawdata_list to dedup the incoming
policy blob against entries already attached to existing profiles.
Per the kernel-doc on struct aa_loaddata, list membership does not
hold a reference: profiles hold pcount, and when the last pcount
drops, do_ploaddata_rmfs() is queued on a workqueue that takes
ns->lock and removes the entry. Between dropping the last pcount
and the workqueue running, an entry remains on the list with
pcount == 0.

aa_get_profile_loaddata() is an unconditional kref_get() on
pcount, so when the dedup loop hits such an entry, refcount
hardening reports

  refcount_t: addition on 0; use-after-free.

inside aa_replace_profiles(), and the poisoned counter then
trips "saturated" and "underflow" warnings on the subsequent
uses of the same loaddata.

Before commit a0b7091c4de4 ("apparmor: fix race on rawdata
dereference") the dedup path used a get_unless_zero-style helper
on a single counter, so the existing "if (tmp)" guard was
meaningful. The split-refcount refactor introduced
aa_get_profile_loaddata(), which has plain kref_get() semantics,
and the guard quietly became a no-op.

Introduce aa_get_profile_loaddata_not0(), matching the existing
_not0 convention used by aa_get_profile_not0(), and use it for
the rawdata_list dedup lookup so dying entries are skipped.

Reproduced on x86_64 with v7.1-rc5 in QEMU+KVM running Ubuntu
24.04 + stress-ng 0.17.06:

  stress-ng --apparmor 1 --klog-check --timeout 60s

Without this patch the three refcount_t warnings fire within a
few seconds. With it the same 60 s run is clean. Coverage is a
smoke-test only; a longer soak with CONFIG_KASAN, CONFIG_KCSAN
and CONFIG_PROVE_LOCKING would be welcome from anyone with the
cycles.

Fixes: a0b7091c4de4 ("apparmor: fix race on rawdata dereference")
Reported-by: Colin Ian King 
Closes: https://bugzilla.kernel.org/show_bug.cgi?id=221513
Cc: stable@vger.kernel.org
Signed-off-by: Ruslan Valiyev 
Signed-off-by: John Johansen 
Signed-off-by: Greg Kroah-Hartman

apparmor: mediate the implicit connect of TCP fast open sendmsg

2026-07-04T11:45:03+00:00

commit 4d587cd8a72155089a627130bbd4716ec0856e21 upstream.

sendmsg()/sendto() with MSG_FASTOPEN is a combination of connect(2) and
write(2): it opens the connection in the SYN. apparmor_socket_sendmsg()
only checks AA_MAY_SEND, so a profile that grants send but denies connect
lets a confined task open an outbound TCP/MPTCP connection that connect(2)
would have refused, bypassing connect mediation.

Mediate the implicit connect when MSG_FASTOPEN is set and a destination
is supplied. Add it to apparmor_socket_sendmsg() (not the shared
aa_sock_msg_perm() helper, which recvmsg also uses) and call aa_sk_perm()
directly, mirroring the selinux and tomoyo fixes. sk_is_tcp() does not
cover MPTCP fast open, so the SOCK_STREAM/IPPROTO_MPTCP arm is explicit.

Fixes: cf60af03ca4e ("net-tcp: Fast Open client - sendmsg(MSG_FASTOPEN)")
Cc: stable@vger.kernel.org
Signed-off-by: Bryam Vargas 
Signed-off-by: John Johansen 
Signed-off-by: Greg Kroah-Hartman

Merge tag 'apparmor-pr-2026-04-23' of git://git.kernel.org/pub/scm/linux/kernel/git/jj/linux-apparmor

2026-04-24T16:22:21+00:00

Pull apparmor updates from John Johansen:
 "Cleanups
   - Use sysfs_emit in param_get_{audit,mode}
   - Remove redundant if check in sk_peer_get_label
   - Replace memcpy + NUL termination with kmemdup_nul in do_setattr

  Bug Fixes:
   - Fix aa_dfa_unpack's error handling in aa_setup_dfa_engine
   - Fix string overrun due to missing termination
   - Fix wrong dentry in RENAME_EXCHANGE uid check
   - fix unpack_tags to properly return error in failure cases
   - fix dfa size check
   - return error on namespace mismatch in verify_header
   - use target task's context in apparmor_getprocattr()"

* tag 'apparmor-pr-2026-04-23' of git://git.kernel.org/pub/scm/linux/kernel/git/jj/linux-apparmor:
  apparmor/lsm: Fix aa_dfa_unpack's error handling in aa_setup_dfa_engine
  apparmor: Fix string overrun due to missing termination
  apparmor: Fix wrong dentry in RENAME_EXCHANGE uid check
  apparmor: fix unpack_tags to properly return error in failure cases
  apparmor: fix dfa size check
  apparmor: Use sysfs_emit in param_get_{audit,mode}
  apparmor: Remove redundant if check in sk_peer_get_label
  apparmor: Replace memcpy + NUL termination with kmemdup_nul in do_setattr
  apparmor: return error on namespace mismatch in verify_header
  apparmor: use target task's context in apparmor_getprocattr()

apparmor/lsm: Fix aa_dfa_unpack's error handling in aa_setup_dfa_engine

2026-04-23T03:11:08+00:00

aa_dfa_unpack returns ERR_PTR not NULL when it fails, but aa_put_dfa
only checks NULL for its input, which would cause invalid memory access
in aa_put_dfa. Set nulldfa to NULL explicitly to fix that.

Fixes: 98b824ff8984 ("apparmor: refcount the pdb")
Signed-off-by: GONG Ruiqi 
Signed-off-by: John Johansen

apparmor: Fix string overrun due to missing termination

2026-04-23T03:09:05+00:00

When booting Ubuntu 26.04 with Linux 7.0-rc4 on an ARM64 Qualcomm
Snapdragon X1 we see a string buffer overrun:

BUG: KASAN: slab-out-of-bounds in aa_dfa_match (security/apparmor/match.c:535)
Read of size 1 at addr ffff0008901cc000 by task snap-update-ns/2120

CPU: 5 UID: 60578 PID: 2120 Comm: snap-update-ns Not tainted 7.0.0-rc4+ #22 PREEMPTLAZY
Hardware name: LENOVO 83ED/LNVNB161216, BIOS NHCN60WW 09/11/2025
Call trace:
show_stack (arch/arm64/kernel/stacktrace.c:501) (C)
dump_stack_lvl (lib/dump_stack.c:122)
print_report (mm/kasan/report.c:379 mm/kasan/report.c:482)
kasan_report (mm/kasan/report.c:597)
__asan_report_load1_noabort (mm/kasan/report_generic.c:378)
aa_dfa_match (security/apparmor/match.c:535)
match_mnt_path_str (security/apparmor/mount.c:244 security/apparmor/mount.c:336)
match_mnt (security/apparmor/mount.c:371)
aa_bind_mount (security/apparmor/mount.c:447 (discriminator 4))
apparmor_sb_mount (security/apparmor/lsm.c:719 (discriminator 1))
security_sb_mount (security/security.c:1062 (discriminator 31))
path_mount (fs/namespace.c:4101)
__arm64_sys_mount (fs/namespace.c:4172 fs/namespace.c:4361 fs/namespace.c:4338 fs/namespace.c:4338)
invoke_syscall.constprop.0 (arch/arm64/kernel/syscall.c:35 arch/arm64/kernel/syscall.c:49)
el0_svc_common.constprop.0 (./include/linux/thread_info.h:142 (discriminator 2) arch/arm64/kernel/syscall.c:140 (discriminator 2))
do_el0_svc (arch/arm64/kernel/syscall.c:152)
el0_svc (arch/arm64/kernel/entry-common.c:80 arch/arm64/kernel/entry-common.c:725)
el0t_64_sync_handler (arch/arm64/kernel/entry-common.c:744)
el0t_64_sync (arch/arm64/kernel/entry.S:596)

Allocated by task 2120:
kasan_save_stack (mm/kasan/common.c:58)
kasan_save_track (./arch/arm64/include/asm/current.h:19 mm/kasan/common.c:70 mm/kasan/common.c:79)
kasan_save_alloc_info (mm/kasan/generic.c:571)
__kasan_kmalloc (mm/kasan/common.c:419)
__kmalloc_noprof (./include/linux/kasan.h:263 mm/slub.c:5260 mm/slub.c:5272)
aa_get_buffer (security/apparmor/lsm.c:2201)
aa_bind_mount (security/apparmor/mount.c:442)
apparmor_sb_mount (security/apparmor/lsm.c:719 (discriminator 1))
security_sb_mount (security/security.c:1062 (discriminator 31))
path_mount (fs/namespace.c:4101)
__arm64_sys_mount (fs/namespace.c:4172 fs/namespace.c:4361 fs/namespace.c:4338 fs/namespace.c:4338)
invoke_syscall.constprop.0 (arch/arm64/kernel/syscall.c:35 arch/arm64/kernel/syscall.c:49)
el0_svc_common.constprop.0 (./include/linux/thread_info.h:142 (discriminator 2) arch/arm64/kernel/syscall.c:140 (discriminator 2))
do_el0_svc (arch/arm64/kernel/syscall.c:152)
el0_svc (arch/arm64/kernel/entry-common.c:80 arch/arm64/kernel/entry-common.c:725)
el0t_64_sync_handler (arch/arm64/kernel/entry-common.c:744)
el0t_64_sync (arch/arm64/kernel/entry.S:596)

The buggy address belongs to the object at ffff0008901ca000
which belongs to the cache kmalloc-rnd-06-8k of size 8192
The buggy address is located 0 bytes to the right of
allocated 8192-byte region [ffff0008901ca000, ffff0008901cc000)

The buggy address belongs to the physical page:
page: refcount:0 mapcount:0 mapping:0000000000000000 index:0x0 pfn:0x9101c8
head: order:3 mapcount:0 entire_mapcount:0 nr_pages_mapped:-1 pincount:0
flags: 0x8000000000000040(head|zone=2)
page_type: f5(slab)
raw: 8000000000000040 ffff000800016c40 fffffdffe2d14e10 ffff000800015c70
raw: 0000000000000000 0000000800010001 00000000f5000000 0000000000000000
head: 8000000000000040 ffff000800016c40 fffffdffe2d14e10 ffff000800015c70
head: 0000000000000000 0000000800010001 00000000f5000000 0000000000000000
head: 8000000000000003 fffffdffe2407201 fffffdffffffffff 00000000ffffffff
head: ffffffffffffffff 0000000000000000 00000000ffffffff 0000000000000008
page dumped because: kasan: bad access detected

Memory state around the buggy address:
ffff0008901cbf00: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
ffff0008901cbf80: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
>ffff0008901cc000: fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc
^
ffff0008901cc080: fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc
ffff0008901cc100: fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc

This was introduced by previous incorrect conversion from strcpy(). Fix it
by adding the missing terminator.

Cc: stable@vger.kernel.org
Reviewed-by: Georgia Garcia 
Signed-off-by: Daniel J Blueman 
Fixes: 93d4dbdc8da0 ("apparmor: Replace deprecated strcpy in d_namespace_path")
Signed-off-by: John Johansen

apparmor: Fix wrong dentry in RENAME_EXCHANGE uid check

2026-04-23T03:08:09+00:00

In apparmor_path_rename(), when handling RENAME_EXCHANGE, the
cond_exchange structure is supposed to carry the attributes of the
*new* dentry (since it is used to authorize moving new_dentry to the
old location). However, line 412 reads:

    vfsuid = i_uid_into_vfsuid(idmap, d_backing_inode(old_dentry));

This fetches the uid of old_dentry instead of new_dentry. As a result,
the RENAME_EXCHANGE permission check uses the wrong file owner, which
can allow a rename that should be denied (if old_dentry's owner has
more privileges) or deny one that should be allowed.

Note that cond_exchange.mode on the line above correctly uses
new_dentry. Only the uid lookup is wrong.

Fix by changing old_dentry to new_dentry in the i_uid_into_vfsuid call.

Fixes: 5e26a01e56fd ("apparmor: use type safe idmapping helpers")
Reviewed-by: Georgia Garcia 
Signed-off-by: Dudu Lu 
Signed-off-by: John Johansen

apparmor: fix unpack_tags to properly return error in failure cases

2026-04-23T03:02:46+00:00

error is initialized to -EPROTO but set by some of the internal
functions, unfortunately the last two checks assume error is set to
-EPROTO already for the failure case. Ensure it is by setting it
before these checks.

Fixes: 3d28e2397af7a ("apparmor: add support loading per permission tagging")
Reported-by: Dan Carpenter 
Signed-off-by: John Johansen

apparmor: fix dfa size check

2026-04-23T02:28:14+00:00

AppArmor dfas need a minimum of two states to be valid. State 0 is the
default trap state, and State 1 the default start state. When verifying
the dfa ensure that this is the case.

Fixes: c27c6bd2c4d6b ("apparmor: ensure that dfa state tables have entries")
Signed-off-by: John Johansen

apparmor: Use sysfs_emit in param_get_{audit,mode}

2026-04-22T17:57:52+00:00

Replace sprintf() with sysfs_emit() in param_get_audit() and
param_get_mode(). sysfs_emit() is preferred for formatting sysfs output
because it provides safer bounds checking.  Add terminating newlines as
suggested by checkpatch.

Signed-off-by: Thorsten Blum 
Signed-off-by: John Johansen