x86/kprobes: Fix crash when probing CS CALL instructionsgrafted

When using eBPF to probe CS CALL instructions within a function,
a crash can be triggered.

The eBPF tool probes offset 257 of the __hrtimer_run_queues()
function:

<__hrtimer_run_queues+249>:  nopl   0x0(%rax,%rax,1)
<__hrtimer_run_queues+254>:  mov    %r14,%rdi
<__hrtimer_run_queues+257>:  cs call <__x86_indirect_thunk_r12>
<__hrtimer_run_queues+263>:  mov    %eax,%r12d
<__hrtimer_run_queues+266>:  xchg   %ax,%ax
<__hrtimer_run_queues+268>:  mov    %r13,%rdi

Which triggers this crash:

  BUG: unable to handle page fault for address: 00000000000f41c9
  #PF: supervisor write access in kernel mode
  #PF: error_code(0x0002) - not-present page
  PGD 0 P4D 0
  Oops: 0002 [#1] SMP NOPTI
  CPU: 1 PID: 0 Comm: swapper/1 Kdump: loaded Tainted: P
  RIP: 0010:__hrtimer_run_queues+0x106/0x230

Note that __hrtimer_run_queues+0x106 is __hrtimer_run_queues+262, which is
at the 6th byte of the above CS CALL instruction. Since the CS CALL
instruction occupies 6 bytes, the exception occurred in the middle of that
call instruction.

The root cause is that when using eBPF tools to probe in the middle of a
function, a kprobe with INT3 is used as the underlying implementation.

During single-step emulation of the original CALL instruction,
int3_emulate_call() assumes that the probed CALL instruction is 5 bytes
long. However, the actual CS-prefixed CALL instruction occupies 6 bytes,
so it constructs an incorrect exception return address. When the CPU
returns from the kprobe handler, the next instruction to be executed is at
the address of the last byte of that CS CALL instruction. Coincidentally,
starting from that address, the CPU fetches and decodes a completely
different instruction, which ultimately triggers a kernel crash.

Fix the issue by using the actual instruction length obtained from
the instruction decoder when constructing the exception return
address, rather than relying on the hardcoded CALL_INSN_SIZE macro.

[ mingo: Refined the changelog ]

Fixes: 6256e668b7af ("x86/kprobes: Use int3 instead of debug trap for single-step")
Suggested-by: Masami Hiramatsu (Google) <mhiramat@kernel.org>
Signed-off-by: Jinke Han <jinkehan@didiglobal.com>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Reviewed-by: Masami Hiramatsu (Google) <mhiramat@kernel.org>
Acked-by: Yafang Shao <laoar.shao@gmail.com>
Acked-by: Borislav Petkov <bp@alien8.de>
Cc: Peter Zijlstra <peterz@infradead.org>
Link: https://patch.msgid.link/20260908073742.GA10517@didi-ThinkCentre-M920t-N000

1 files changed, 176 insertions, 0 deletions

diff --git a/net/x25/x25_timer.c b/net/x25/x25_timer.c
new file mode 100644
index 000000000..7896cd43f
--- /dev/null
+++ b/net/x25/x25_timer.c
@@ -0,0 +1,176 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
+/*
+ *	X.25 Packet Layer release 002
+ *
+ *	This is ALPHA test software. This code may break your machine,
+ *	randomly fail to work with new releases, misbehave and/or generally
+ *	screw up. It might even work.
+ *
+ *	This code REQUIRES 2.1.15 or higher
+ *
+ *	History
+ *	X.25 001	Jonathan Naylor	Started coding.
+ *	X.25 002	Jonathan Naylor	New timer architecture.
+ *					Centralised disconnection processing.
+ */
+
+#include <linux/errno.h>
+#include <linux/jiffies.h>
+#include <linux/timer.h>
+#include <net/sock.h>
+#include <net/tcp_states.h>
+#include <net/x25.h>
+
+static void x25_heartbeat_expiry(struct timer_list *t);
+static void x25_timer_expiry(struct timer_list *t);
+
+void x25_init_timers(struct sock *sk)
+{
+	struct x25_sock *x25 = x25_sk(sk);
+
+	timer_setup(&x25->timer, x25_timer_expiry, 0);
+
+	/* initialized by sock_init_data */
+	sk->sk_timer.function = x25_heartbeat_expiry;
+}
+
+void x25_start_heartbeat(struct sock *sk)
+{
+	sk_reset_timer(sk, &sk->sk_timer, jiffies + 5 * HZ);
+}
+
+void x25_stop_heartbeat(struct sock *sk)
+{
+	sk_stop_timer(sk, &sk->sk_timer);
+}
+
+void x25_start_t2timer(struct sock *sk)
+{
+	struct x25_sock *x25 = x25_sk(sk);
+
+	sk_reset_timer(sk, &x25->timer, jiffies + x25->t2);
+}
+
+void x25_start_t21timer(struct sock *sk)
+{
+	struct x25_sock *x25 = x25_sk(sk);
+
+	sk_reset_timer(sk, &x25->timer, jiffies + x25->t21);
+}
+
+void x25_start_t22timer(struct sock *sk)
+{
+	struct x25_sock *x25 = x25_sk(sk);
+
+	sk_reset_timer(sk, &x25->timer, jiffies + x25->t22);
+}
+
+void x25_start_t23timer(struct sock *sk)
+{
+	struct x25_sock *x25 = x25_sk(sk);
+
+	sk_reset_timer(sk, &x25->timer, jiffies + x25->t23);
+}
+
+void x25_stop_timer(struct sock *sk)
+{
+	sk_stop_timer(sk, &x25_sk(sk)->timer);
+}
+
+unsigned long x25_display_timer(struct sock *sk)
+{
+	struct x25_sock *x25 = x25_sk(sk);
+
+	if (!timer_pending(&x25->timer))
+		return 0;
+
+	return x25->timer.expires - jiffies;
+}
+
+static void x25_heartbeat_expiry(struct timer_list *t)
+{
+	struct sock *sk = timer_container_of(sk, t, sk_timer);
+
+	bh_lock_sock(sk);
+	if (sock_owned_by_user(sk)) /* can currently only occur in state 3 */
+		goto restart_heartbeat;
+
+	switch (x25_sk(sk)->state) {
+
+		case X25_STATE_0:
+			/*
+			 * Magic here: If we listen() and a new link dies
+			 * before it is accepted() it isn't 'dead' so doesn't
+			 * get removed.
+			 */
+			if (sock_flag(sk, SOCK_DESTROY) ||
+			    (sk->sk_state == TCP_LISTEN &&
+			     sock_flag(sk, SOCK_DEAD))) {
+				bh_unlock_sock(sk);
+				x25_destroy_socket_from_timer(sk);
+				goto out;
+			}
+			break;
+
+		case X25_STATE_3:
+			/*
+			 * Check for the state of the receive buffer.
+			 */
+			x25_check_rbuf(sk);
+			break;
+	}
+restart_heartbeat:
+	/* Do not rearm once __x25_destroy_socket() has unlinked the socket:
+	 * it is past its cancel point and owns the teardown from there on.
+	 */
+	if (sk_hashed(sk))
+		x25_start_heartbeat(sk);
+	bh_unlock_sock(sk);
+out:
+	sock_put(sk);
+}
+
+/*
+ *	Timer has expired, it may have been T2, T21, T22, or T23. We can tell
+ *	by the state machine state.
+ */
+static inline void x25_do_timer_expiry(struct sock * sk)
+{
+	struct x25_sock *x25 = x25_sk(sk);
+
+	switch (x25->state) {
+
+		case X25_STATE_3:	/* T2 */
+			if (x25->condition & X25_COND_ACK_PENDING) {
+				x25->condition &= ~X25_COND_ACK_PENDING;
+				x25_enquiry_response(sk);
+			}
+			break;
+
+		case X25_STATE_1:	/* T21 */
+		case X25_STATE_4:	/* T22 */
+			x25_write_internal(sk, X25_CLEAR_REQUEST);
+			x25->state = X25_STATE_2;
+			x25_start_t23timer(sk);
+			break;
+
+		case X25_STATE_2:	/* T23 */
+			x25_disconnect(sk, ETIMEDOUT, 0, 0);
+			break;
+	}
+}
+
+static void x25_timer_expiry(struct timer_list *t)
+{
+	struct x25_sock *x25 = timer_container_of(x25, t, timer);
+	struct sock *sk = &x25->sk;
+
+	bh_lock_sock(sk);
+	if (sock_owned_by_user(sk)) { /* can currently only occur in state 3 */
+		if (x25_sk(sk)->state == X25_STATE_3)
+			x25_start_t2timer(sk);
+	} else
+		x25_do_timer_expiry(sk);
+	bh_unlock_sock(sk);
+	sock_put(sk);
+}