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, 432 insertions, 0 deletions

diff --git a/net/nfc/rawsock.c b/net/nfc/rawsock.c
new file mode 100644
index 000000000..f7d7a599f
--- /dev/null
+++ b/net/nfc/rawsock.c
@@ -0,0 +1,432 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
+/*
+ * Copyright (C) 2011 Instituto Nokia de Tecnologia
+ *
+ * Authors:
+ *    Aloisio Almeida Jr <aloisio.almeida@openbossa.org>
+ *    Lauro Ramos Venancio <lauro.venancio@openbossa.org>
+ */
+
+#define pr_fmt(fmt) KBUILD_MODNAME ": %s: " fmt, __func__
+
+#include <net/tcp_states.h>
+#include <linux/nfc.h>
+#include <linux/export.h>
+#include <linux/kcov.h>
+
+#include "nfc.h"
+
+static struct nfc_sock_list raw_sk_list = {
+	.lock = __RW_LOCK_UNLOCKED(raw_sk_list.lock)
+};
+
+static void nfc_sock_link(struct nfc_sock_list *l, struct sock *sk)
+{
+	write_lock(&l->lock);
+	sk_add_node(sk, &l->head);
+	write_unlock(&l->lock);
+}
+
+static void nfc_sock_unlink(struct nfc_sock_list *l, struct sock *sk)
+{
+	write_lock(&l->lock);
+	sk_del_node_init(sk);
+	write_unlock(&l->lock);
+}
+
+static void rawsock_write_queue_purge(struct sock *sk)
+{
+	pr_debug("sk=%p\n", sk);
+
+	spin_lock_bh(&sk->sk_write_queue.lock);
+	__skb_queue_purge(&sk->sk_write_queue);
+	nfc_rawsock(sk)->tx_work_scheduled = false;
+	spin_unlock_bh(&sk->sk_write_queue.lock);
+}
+
+static void rawsock_report_error(struct sock *sk, int err)
+{
+	pr_debug("sk=%p err=%d\n", sk, err);
+
+	sk->sk_shutdown = SHUTDOWN_MASK;
+	sk->sk_err = -err;
+	sk_error_report(sk);
+
+	rawsock_write_queue_purge(sk);
+}
+
+static int rawsock_release(struct socket *sock)
+{
+	struct sock *sk = sock->sk;
+
+	pr_debug("sock=%p sk=%p\n", sock, sk);
+
+	if (!sk)
+		return 0;
+
+	if (sock->type == SOCK_RAW)
+		nfc_sock_unlink(&raw_sk_list, sk);
+
+	if (sk->sk_state == TCP_ESTABLISHED) {
+		/* Prevent rawsock_tx_work from starting new transmits and
+		 * wait for any in-progress work to finish.  This must happen
+		 * before the socket is orphaned to avoid a race where
+		 * rawsock_tx_work runs after the NCI device has been freed.
+		 */
+		sk->sk_shutdown |= SEND_SHUTDOWN;
+		cancel_work_sync(&nfc_rawsock(sk)->tx_work);
+		rawsock_write_queue_purge(sk);
+	}
+
+	sock_orphan(sk);
+	sock_put(sk);
+
+	return 0;
+}
+
+static int rawsock_connect(struct socket *sock, struct sockaddr_unsized *_addr,
+			   int len, int flags)
+{
+	struct sock *sk = sock->sk;
+	struct sockaddr_nfc *addr = (struct sockaddr_nfc *)_addr;
+	struct nfc_dev *dev;
+	int rc = 0;
+
+	pr_debug("sock=%p sk=%p flags=%d\n", sock, sk, flags);
+
+	if (!addr || len < sizeof(struct sockaddr_nfc) ||
+	    addr->sa_family != AF_NFC)
+		return -EINVAL;
+
+	pr_debug("addr dev_idx=%u target_idx=%u protocol=%u\n",
+		 addr->dev_idx, addr->target_idx, addr->nfc_protocol);
+
+	lock_sock(sk);
+
+	if (sock->state == SS_CONNECTED) {
+		rc = -EISCONN;
+		goto error;
+	}
+
+	dev = nfc_get_device(addr->dev_idx);
+	if (!dev) {
+		rc = -ENODEV;
+		goto error;
+	}
+
+	if (addr->target_idx > dev->target_next_idx - 1 ||
+	    addr->target_idx < dev->target_next_idx - dev->n_targets) {
+		rc = -EINVAL;
+		goto put_dev;
+	}
+
+	rc = nfc_activate_target(dev, addr->target_idx, addr->nfc_protocol);
+	if (rc)
+		goto put_dev;
+
+	nfc_rawsock(sk)->dev = dev;
+	nfc_rawsock(sk)->target_idx = addr->target_idx;
+	sock->state = SS_CONNECTED;
+	sk->sk_state = TCP_ESTABLISHED;
+	sk->sk_state_change(sk);
+
+	release_sock(sk);
+	return 0;
+
+put_dev:
+	nfc_put_device(dev);
+error:
+	release_sock(sk);
+	return rc;
+}
+
+static int rawsock_add_header(struct sk_buff *skb)
+{
+	*(u8 *)skb_push(skb, NFC_HEADER_SIZE) = 0;
+
+	return 0;
+}
+
+static void rawsock_data_exchange_complete(void *context, struct sk_buff *skb,
+					   int err)
+{
+	struct sock *sk = (struct sock *) context;
+
+	BUG_ON(in_hardirq());
+
+	pr_debug("sk=%p err=%d\n", sk, err);
+
+	if (err)
+		goto error;
+
+	err = rawsock_add_header(skb);
+	if (err)
+		goto error_skb;
+
+	err = sock_queue_rcv_skb(sk, skb);
+	if (err)
+		goto error_skb;
+
+	spin_lock_bh(&sk->sk_write_queue.lock);
+	if (!skb_queue_empty(&sk->sk_write_queue))
+		schedule_work(&nfc_rawsock(sk)->tx_work);
+	else
+		nfc_rawsock(sk)->tx_work_scheduled = false;
+	spin_unlock_bh(&sk->sk_write_queue.lock);
+
+	sock_put(sk);
+	return;
+
+error_skb:
+	kfree_skb(skb);
+
+error:
+	rawsock_report_error(sk, err);
+	sock_put(sk);
+}
+
+static void rawsock_tx_work(struct work_struct *work)
+{
+	struct sock *sk = to_rawsock_sk(work);
+	struct nfc_dev *dev = nfc_rawsock(sk)->dev;
+	u32 target_idx = nfc_rawsock(sk)->target_idx;
+	struct sk_buff *skb;
+	int rc;
+
+	pr_debug("sk=%p target_idx=%u\n", sk, target_idx);
+
+	if (sk->sk_shutdown & SEND_SHUTDOWN) {
+		rawsock_write_queue_purge(sk);
+		return;
+	}
+
+	skb = skb_dequeue(&sk->sk_write_queue);
+	kcov_remote_start_common(skb_get_kcov_handle(skb));
+
+	sock_hold(sk);
+	rc = nfc_data_exchange(dev, target_idx, skb,
+			       rawsock_data_exchange_complete, sk);
+	if (rc) {
+		rawsock_report_error(sk, rc);
+		sock_put(sk);
+	}
+	kcov_remote_stop();
+}
+
+static int rawsock_sendmsg(struct socket *sock, struct msghdr *msg, size_t len)
+{
+	struct sock *sk = sock->sk;
+	struct nfc_dev *dev = nfc_rawsock(sk)->dev;
+	struct sk_buff *skb;
+	int rc;
+
+	pr_debug("sock=%p sk=%p len=%zu\n", sock, sk, len);
+
+	if (msg->msg_namelen)
+		return -EOPNOTSUPP;
+
+	if (sock->state != SS_CONNECTED)
+		return -ENOTCONN;
+
+	skb = nfc_alloc_send_skb(dev, sk, msg->msg_flags, len, &rc);
+	if (skb == NULL)
+		return rc;
+
+	rc = memcpy_from_msg(skb_put(skb, len), msg, len);
+	if (rc < 0) {
+		kfree_skb(skb);
+		return rc;
+	}
+
+	spin_lock_bh(&sk->sk_write_queue.lock);
+	__skb_queue_tail(&sk->sk_write_queue, skb);
+	if (!nfc_rawsock(sk)->tx_work_scheduled) {
+		schedule_work(&nfc_rawsock(sk)->tx_work);
+		nfc_rawsock(sk)->tx_work_scheduled = true;
+	}
+	spin_unlock_bh(&sk->sk_write_queue.lock);
+
+	return len;
+}
+
+static int rawsock_recvmsg(struct socket *sock, struct msghdr *msg, size_t len,
+			   int flags)
+{
+	struct sock *sk = sock->sk;
+	struct sk_buff *skb;
+	int copied;
+	int rc;
+
+	pr_debug("sock=%p sk=%p len=%zu flags=%d\n", sock, sk, len, flags);
+
+	skb = skb_recv_datagram(sk, flags, &rc);
+	if (!skb)
+		return rc;
+
+	copied = skb->len;
+	if (len < copied) {
+		msg->msg_flags |= MSG_TRUNC;
+		copied = len;
+	}
+
+	rc = skb_copy_datagram_msg(skb, 0, msg, copied);
+
+	skb_free_datagram(sk, skb);
+
+	return rc ? : copied;
+}
+
+static const struct proto_ops rawsock_ops = {
+	.family         = PF_NFC,
+	.owner          = THIS_MODULE,
+	.release        = rawsock_release,
+	.bind           = sock_no_bind,
+	.connect        = rawsock_connect,
+	.socketpair     = sock_no_socketpair,
+	.accept         = sock_no_accept,
+	.getname        = sock_no_getname,
+	.poll           = datagram_poll,
+	.ioctl          = sock_no_ioctl,
+	.listen         = sock_no_listen,
+	.shutdown       = sock_no_shutdown,
+	.sendmsg        = rawsock_sendmsg,
+	.recvmsg        = rawsock_recvmsg,
+	.mmap           = sock_no_mmap,
+};
+
+static const struct proto_ops rawsock_raw_ops = {
+	.family         = PF_NFC,
+	.owner          = THIS_MODULE,
+	.release        = rawsock_release,
+	.bind           = sock_no_bind,
+	.connect        = sock_no_connect,
+	.socketpair     = sock_no_socketpair,
+	.accept         = sock_no_accept,
+	.getname        = sock_no_getname,
+	.poll           = datagram_poll,
+	.ioctl          = sock_no_ioctl,
+	.listen         = sock_no_listen,
+	.shutdown       = sock_no_shutdown,
+	.sendmsg        = sock_no_sendmsg,
+	.recvmsg        = rawsock_recvmsg,
+	.mmap           = sock_no_mmap,
+};
+
+static void rawsock_destruct(struct sock *sk)
+{
+	pr_debug("sk=%p\n", sk);
+
+	if (sk->sk_state == TCP_ESTABLISHED) {
+		nfc_deactivate_target(nfc_rawsock(sk)->dev,
+				      nfc_rawsock(sk)->target_idx,
+				      NFC_TARGET_MODE_IDLE);
+		nfc_put_device(nfc_rawsock(sk)->dev);
+	}
+
+	skb_queue_purge(&sk->sk_receive_queue);
+
+	if (!sock_flag(sk, SOCK_DEAD)) {
+		pr_err("Freeing alive NFC raw socket %p\n", sk);
+		return;
+	}
+}
+
+static int rawsock_create(struct net *net, struct socket *sock,
+			  const struct nfc_protocol *nfc_proto, int kern)
+{
+	struct sock *sk;
+
+	pr_debug("sock=%p\n", sock);
+
+	if ((sock->type != SOCK_SEQPACKET) && (sock->type != SOCK_RAW))
+		return -ESOCKTNOSUPPORT;
+
+	if (sock->type == SOCK_RAW) {
+		if (!ns_capable(net->user_ns, CAP_NET_RAW))
+			return -EPERM;
+		sock->ops = &rawsock_raw_ops;
+	} else {
+		sock->ops = &rawsock_ops;
+	}
+
+	sk = sk_alloc(net, PF_NFC, GFP_ATOMIC, nfc_proto->proto, kern);
+	if (!sk)
+		return -ENOMEM;
+
+	sock_init_data(sock, sk);
+	sk->sk_protocol = nfc_proto->id;
+	sk->sk_destruct = rawsock_destruct;
+	sock->state = SS_UNCONNECTED;
+	if (sock->type == SOCK_RAW)
+		nfc_sock_link(&raw_sk_list, sk);
+	else {
+		INIT_WORK(&nfc_rawsock(sk)->tx_work, rawsock_tx_work);
+		nfc_rawsock(sk)->tx_work_scheduled = false;
+	}
+
+	return 0;
+}
+
+void nfc_send_to_raw_sock(struct nfc_dev *dev, struct sk_buff *skb,
+			  u8 payload_type, u8 direction)
+{
+	struct sk_buff *skb_copy = NULL, *nskb;
+	struct sock *sk;
+	u8 *data;
+
+	read_lock(&raw_sk_list.lock);
+
+	sk_for_each(sk, &raw_sk_list.head) {
+		if (!skb_copy) {
+			skb_copy = __pskb_copy_fclone(skb, NFC_RAW_HEADER_SIZE,
+						      GFP_ATOMIC, true);
+			if (!skb_copy)
+				continue;
+
+			data = skb_push(skb_copy, NFC_RAW_HEADER_SIZE);
+
+			data[0] = dev ? dev->idx : 0xFF;
+			data[1] = direction & 0x01;
+			data[1] |= (payload_type << 1);
+		}
+
+		nskb = skb_clone(skb_copy, GFP_ATOMIC);
+		if (!nskb)
+			continue;
+
+		if (sock_queue_rcv_skb(sk, nskb))
+			kfree_skb(nskb);
+	}
+
+	read_unlock(&raw_sk_list.lock);
+
+	kfree_skb(skb_copy);
+}
+EXPORT_SYMBOL(nfc_send_to_raw_sock);
+
+static struct proto rawsock_proto = {
+	.name     = "NFC_RAW",
+	.owner    = THIS_MODULE,
+	.obj_size = sizeof(struct nfc_rawsock),
+};
+
+static const struct nfc_protocol rawsock_nfc_proto = {
+	.id	  = NFC_SOCKPROTO_RAW,
+	.proto    = &rawsock_proto,
+	.owner    = THIS_MODULE,
+	.create   = rawsock_create
+};
+
+int __init rawsock_init(void)
+{
+	int rc;
+
+	rc = nfc_proto_register(&rawsock_nfc_proto);
+
+	return rc;
+}
+
+void rawsock_exit(void)
+{
+	nfc_proto_unregister(&rawsock_nfc_proto);
+}