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, 351 insertions, 0 deletions

diff --git a/net/l2tp/l2tp_debugfs.c b/net/l2tp/l2tp_debugfs.c
new file mode 100644
index 000000000..c12d5e1ee
--- /dev/null
+++ b/net/l2tp/l2tp_debugfs.c
@@ -0,0 +1,351 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
+/* L2TP subsystem debugfs
+ *
+ * Copyright (c) 2010 Katalix Systems Ltd
+ */
+
+#define pr_fmt(fmt) KBUILD_MODNAME ": " fmt
+
+#include <linux/module.h>
+#include <linux/skbuff.h>
+#include <linux/socket.h>
+#include <linux/hash.h>
+#include <linux/l2tp.h>
+#include <linux/in.h>
+#include <linux/etherdevice.h>
+#include <linux/spinlock.h>
+#include <linux/debugfs.h>
+#include <net/sock.h>
+#include <net/ip.h>
+#include <net/icmp.h>
+#include <net/udp.h>
+#include <net/inet_common.h>
+#include <net/inet_hashtables.h>
+#include <net/tcp_states.h>
+#include <net/protocol.h>
+#include <net/xfrm.h>
+#include <net/net_namespace.h>
+#include <net/netns/generic.h>
+
+#include "l2tp_core.h"
+
+static struct dentry *rootdir;
+
+struct l2tp_dfs_seq_data {
+	struct net	*net;
+	netns_tracker	ns_tracker;
+	unsigned long tkey;		/* lookup key of current tunnel */
+	unsigned long skey;		/* lookup key of current session */
+	struct l2tp_tunnel *tunnel;
+	struct l2tp_session *session;	/* NULL means get next tunnel */
+};
+
+static void l2tp_dfs_next_tunnel(struct l2tp_dfs_seq_data *pd)
+{
+	/* Drop reference taken during previous invocation */
+	if (pd->tunnel)
+		l2tp_tunnel_put(pd->tunnel);
+
+	pd->tunnel = l2tp_tunnel_get_next(pd->net, &pd->tkey);
+	pd->tkey++;
+}
+
+static void l2tp_dfs_next_session(struct l2tp_dfs_seq_data *pd)
+{
+	/* Drop reference taken during previous invocation */
+	if (pd->session)
+		l2tp_session_put(pd->session);
+
+	pd->session = l2tp_session_get_next(pd->net, pd->tunnel->sock,
+					    pd->tunnel->version,
+					    pd->tunnel->tunnel_id, &pd->skey);
+	pd->skey++;
+
+	if (!pd->session) {
+		pd->skey = 0;
+		l2tp_dfs_next_tunnel(pd);
+	}
+}
+
+static void *l2tp_dfs_seq_start(struct seq_file *m, loff_t *offs)
+{
+	struct l2tp_dfs_seq_data *pd = SEQ_START_TOKEN;
+	loff_t pos = *offs;
+
+	if (!pos)
+		goto out;
+
+	if (WARN_ON(!m->private)) {
+		pd = NULL;
+		goto out;
+	}
+	pd = m->private;
+
+	if (!pd->tunnel)
+		l2tp_dfs_next_tunnel(pd);
+	else
+		l2tp_dfs_next_session(pd);
+
+	/* NULL tunnel and session indicates end of list */
+	if (!pd->tunnel && !pd->session)
+		pd = NULL;
+
+out:
+	return pd;
+}
+
+static void *l2tp_dfs_seq_next(struct seq_file *m, void *v, loff_t *pos)
+{
+	(*pos)++;
+	return NULL;
+}
+
+static void l2tp_dfs_seq_stop(struct seq_file *p, void *v)
+{
+	struct l2tp_dfs_seq_data *pd = v;
+
+	if (!pd || pd == SEQ_START_TOKEN)
+		return;
+
+	/* Drop reference taken by last invocation of l2tp_dfs_next_session()
+	 * or l2tp_dfs_next_tunnel().
+	 */
+	if (pd->session) {
+		l2tp_session_put(pd->session);
+		pd->session = NULL;
+	}
+	if (pd->tunnel) {
+		l2tp_tunnel_put(pd->tunnel);
+		pd->tunnel = NULL;
+	}
+}
+
+static void l2tp_dfs_seq_tunnel_show(struct seq_file *m, void *v)
+{
+	struct l2tp_tunnel *tunnel = v;
+	struct l2tp_session *session;
+	int session_count = 0;
+
+	rcu_read_lock_bh();
+	list_for_each_entry_rcu(session, &tunnel->session_list, list) {
+		/* Session ID of zero is a dummy/reserved value used by pppol2tp */
+		if (session->session_id == 0)
+			continue;
+
+		session_count++;
+	}
+	rcu_read_unlock_bh();
+
+	seq_printf(m, "\nTUNNEL %u peer %u", tunnel->tunnel_id, tunnel->peer_tunnel_id);
+	if (tunnel->sock) {
+		struct inet_sock *inet = inet_sk(tunnel->sock);
+
+#if IS_ENABLED(CONFIG_IPV6)
+		if (tunnel->sock->sk_family == AF_INET6) {
+			const struct ipv6_pinfo *np = inet6_sk(tunnel->sock);
+
+			seq_printf(m, " from %pI6c to %pI6c\n",
+				   &np->saddr, &tunnel->sock->sk_v6_daddr);
+		}
+#endif
+		if (tunnel->sock->sk_family == AF_INET)
+			seq_printf(m, " from %pI4 to %pI4\n",
+				   &inet->inet_saddr, &inet->inet_daddr);
+
+		if (tunnel->encap == L2TP_ENCAPTYPE_UDP)
+			seq_printf(m, " source port %hu, dest port %hu\n",
+				   ntohs(inet->inet_sport), ntohs(inet->inet_dport));
+	}
+	seq_printf(m, " L2TPv%d, %s\n", tunnel->version,
+		   tunnel->encap == L2TP_ENCAPTYPE_UDP ? "UDP" :
+		   tunnel->encap == L2TP_ENCAPTYPE_IP ? "IP" :
+		   "");
+	seq_printf(m, " %d sessions, refcnt %d/%d\n", session_count,
+		   tunnel->sock ? refcount_read(&tunnel->sock->sk_refcnt) : 0,
+		   refcount_read(&tunnel->ref_count));
+	seq_printf(m, " %08x tx %ld/%ld/%ld rx %ld/%ld/%ld\n",
+		   0,
+		   atomic_long_read(&tunnel->stats.tx_packets),
+		   atomic_long_read(&tunnel->stats.tx_bytes),
+		   atomic_long_read(&tunnel->stats.tx_errors),
+		   atomic_long_read(&tunnel->stats.rx_packets),
+		   atomic_long_read(&tunnel->stats.rx_bytes),
+		   atomic_long_read(&tunnel->stats.rx_errors));
+}
+
+static void l2tp_dfs_seq_session_show(struct seq_file *m, void *v)
+{
+	struct l2tp_session *session = v;
+
+	seq_printf(m, "  SESSION %u, peer %u, %s\n", session->session_id,
+		   session->peer_session_id,
+		   session->pwtype == L2TP_PWTYPE_ETH ? "ETH" :
+		   session->pwtype == L2TP_PWTYPE_PPP ? "PPP" :
+		   "");
+	if (session->send_seq || session->recv_seq)
+		seq_printf(m, "   nr %u, ns %u\n", session->nr, session->ns);
+	seq_printf(m, "   refcnt %d\n", refcount_read(&session->ref_count));
+	seq_printf(m, "   config 0/0/%c/%c/-/%s %08x %u\n",
+		   session->recv_seq ? 'R' : '-',
+		   session->send_seq ? 'S' : '-',
+		   session->lns_mode ? "LNS" : "LAC",
+		   0,
+		   jiffies_to_msecs(session->reorder_timeout));
+	seq_printf(m, "   offset 0 l2specific %hu/%d\n",
+		   session->l2specific_type, l2tp_get_l2specific_len(session));
+	if (session->cookie_len) {
+		seq_printf(m, "   cookie %02x%02x%02x%02x",
+			   session->cookie[0], session->cookie[1],
+			   session->cookie[2], session->cookie[3]);
+		if (session->cookie_len == 8)
+			seq_printf(m, "%02x%02x%02x%02x",
+				   session->cookie[4], session->cookie[5],
+				   session->cookie[6], session->cookie[7]);
+		seq_puts(m, "\n");
+	}
+	if (session->peer_cookie_len) {
+		seq_printf(m, "   peer cookie %02x%02x%02x%02x",
+			   session->peer_cookie[0], session->peer_cookie[1],
+			   session->peer_cookie[2], session->peer_cookie[3]);
+		if (session->peer_cookie_len == 8)
+			seq_printf(m, "%02x%02x%02x%02x",
+				   session->peer_cookie[4], session->peer_cookie[5],
+				   session->peer_cookie[6], session->peer_cookie[7]);
+		seq_puts(m, "\n");
+	}
+
+	seq_printf(m, "   %u/%u tx %ld/%ld/%ld rx %ld/%ld/%ld\n",
+		   session->nr, session->ns,
+		   atomic_long_read(&session->stats.tx_packets),
+		   atomic_long_read(&session->stats.tx_bytes),
+		   atomic_long_read(&session->stats.tx_errors),
+		   atomic_long_read(&session->stats.rx_packets),
+		   atomic_long_read(&session->stats.rx_bytes),
+		   atomic_long_read(&session->stats.rx_errors));
+
+	if (session->show)
+		session->show(m, session);
+}
+
+static int l2tp_dfs_seq_show(struct seq_file *m, void *v)
+{
+	struct l2tp_dfs_seq_data *pd = v;
+
+	/* display header on line 1 */
+	if (v == SEQ_START_TOKEN) {
+		seq_puts(m, "TUNNEL ID, peer ID from IP to IP\n");
+		seq_puts(m, " L2TPv2/L2TPv3, UDP/IP\n");
+		seq_puts(m, " sessions session-count, refcnt refcnt/sk->refcnt\n");
+		seq_puts(m, " debug tx-pkts/bytes/errs rx-pkts/bytes/errs\n");
+		seq_puts(m, "  SESSION ID, peer ID, PWTYPE\n");
+		seq_puts(m, "   refcnt cnt\n");
+		seq_puts(m, "   offset OFFSET l2specific TYPE/LEN\n");
+		seq_puts(m, "   [ cookie ]\n");
+		seq_puts(m, "   [ peer cookie ]\n");
+		seq_puts(m, "   config mtu/mru/rcvseq/sendseq/dataseq/lns debug reorderto\n");
+		seq_puts(m, "   nr/ns tx-pkts/bytes/errs rx-pkts/bytes/errs\n");
+		goto out;
+	}
+
+	if (!pd->session)
+		l2tp_dfs_seq_tunnel_show(m, pd->tunnel);
+	else
+		l2tp_dfs_seq_session_show(m, pd->session);
+
+out:
+	return 0;
+}
+
+static const struct seq_operations l2tp_dfs_seq_ops = {
+	.start		= l2tp_dfs_seq_start,
+	.next		= l2tp_dfs_seq_next,
+	.stop		= l2tp_dfs_seq_stop,
+	.show		= l2tp_dfs_seq_show,
+};
+
+static int l2tp_dfs_seq_open(struct inode *inode, struct file *file)
+{
+	struct l2tp_dfs_seq_data *pd;
+	struct seq_file *seq;
+	int rc = -ENOMEM;
+
+	pd = kzalloc_obj(*pd);
+	if (!pd)
+		goto out;
+
+	/* Derive the network namespace from the pid opening the
+	 * file.
+	 */
+	pd->net = get_net_ns_by_pid(current->pid);
+	if (IS_ERR(pd->net)) {
+		rc = PTR_ERR(pd->net);
+		goto err_free_pd;
+	}
+	netns_tracker_alloc(pd->net, &pd->ns_tracker, GFP_KERNEL);
+	rc = seq_open(file, &l2tp_dfs_seq_ops);
+	if (rc)
+		goto err_free_net;
+
+	seq = file->private_data;
+	seq->private = pd;
+
+out:
+	return rc;
+
+err_free_net:
+	put_net_track(pd->net, &pd->ns_tracker);
+err_free_pd:
+	kfree(pd);
+	goto out;
+}
+
+static int l2tp_dfs_seq_release(struct inode *inode, struct file *file)
+{
+	struct l2tp_dfs_seq_data *pd;
+	struct seq_file *seq;
+
+	seq = file->private_data;
+	pd = seq->private;
+	if (pd->session)
+		l2tp_session_put(pd->session);
+	if (pd->tunnel)
+		l2tp_tunnel_put(pd->tunnel);
+	if (pd->net)
+		put_net_track(pd->net, &pd->ns_tracker);
+	kfree(pd);
+	seq_release(inode, file);
+
+	return 0;
+}
+
+static const struct file_operations l2tp_dfs_fops = {
+	.owner		= THIS_MODULE,
+	.open		= l2tp_dfs_seq_open,
+	.read		= seq_read,
+	.llseek		= seq_lseek,
+	.release	= l2tp_dfs_seq_release,
+};
+
+static int __init l2tp_debugfs_init(void)
+{
+	rootdir = debugfs_create_dir("l2tp", NULL);
+
+	debugfs_create_file("tunnels", 0600, rootdir, NULL, &l2tp_dfs_fops);
+
+	pr_info("L2TP debugfs support\n");
+
+	return 0;
+}
+
+static void __exit l2tp_debugfs_exit(void)
+{
+	debugfs_remove_recursive(rootdir);
+}
+
+module_init(l2tp_debugfs_init);
+module_exit(l2tp_debugfs_exit);
+
+MODULE_LICENSE("GPL");
+MODULE_AUTHOR("James Chapman <jchapman@katalix.com>");
+MODULE_DESCRIPTION("L2TP debugfs driver");
+MODULE_VERSION("1.0");