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, 245 insertions, 0 deletions

diff --git a/net/llc/llc_proc.c b/net/llc/llc_proc.c
new file mode 100644
index 000000000..4804a08c2
--- /dev/null
+++ b/net/llc/llc_proc.c
@@ -0,0 +1,245 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * proc_llc.c - proc interface for LLC
+ *
+ * Copyright (c) 2001 by Jay Schulist <jschlst@samba.org>
+ *		 2002-2003 by Arnaldo Carvalho de Melo <acme@conectiva.com.br>
+ */
+
+#include <linux/init.h>
+#include <linux/kernel.h>
+#include <linux/proc_fs.h>
+#include <linux/errno.h>
+#include <linux/seq_file.h>
+#include <linux/export.h>
+#include <net/net_namespace.h>
+#include <net/sock.h>
+#include <net/llc.h>
+#include <net/llc_c_ac.h>
+#include <net/llc_c_ev.h>
+#include <net/llc_c_st.h>
+#include <net/llc_conn.h>
+
+static void llc_ui_format_mac(struct seq_file *seq, const u8 *addr)
+{
+	seq_printf(seq, "%pM", addr);
+}
+
+static struct sock *llc_get_sk_idx(loff_t pos)
+{
+	struct llc_sap *sap;
+	struct sock *sk = NULL;
+	int i;
+
+	list_for_each_entry_rcu(sap, &llc_sap_list, node) {
+		spin_lock_bh(&sap->sk_lock);
+		for (i = 0; i < LLC_SK_LADDR_HASH_ENTRIES; i++) {
+			struct hlist_nulls_head *head = &sap->sk_laddr_hash[i];
+			struct hlist_nulls_node *node;
+
+			sk_nulls_for_each(sk, node, head) {
+				if (!pos)
+					goto found; /* keep the lock */
+				--pos;
+			}
+		}
+		spin_unlock_bh(&sap->sk_lock);
+	}
+	sk = NULL;
+found:
+	return sk;
+}
+
+static void *llc_seq_start(struct seq_file *seq, loff_t *pos) __acquires(RCU)
+{
+	loff_t l = *pos;
+
+	rcu_read_lock_bh();
+	return l ? llc_get_sk_idx(--l) : SEQ_START_TOKEN;
+}
+
+static struct sock *laddr_hash_next(struct llc_sap *sap, int bucket)
+{
+	struct hlist_nulls_node *node;
+	struct sock *sk = NULL;
+
+	while (++bucket < LLC_SK_LADDR_HASH_ENTRIES)
+		sk_nulls_for_each(sk, node, &sap->sk_laddr_hash[bucket])
+			goto out;
+
+out:
+	return sk;
+}
+
+static void *llc_seq_next(struct seq_file *seq, void *v, loff_t *pos)
+{
+	struct sock* sk, *next;
+	struct llc_sock *llc;
+	struct llc_sap *sap;
+
+	++*pos;
+	if (v == SEQ_START_TOKEN) {
+		sk = llc_get_sk_idx(0);
+		goto out;
+	}
+	sk = v;
+	next = sk_nulls_next(sk);
+	if (next) {
+		sk = next;
+		goto out;
+	}
+	llc = llc_sk(sk);
+	sap = llc->sap;
+	sk = laddr_hash_next(sap, llc_sk_laddr_hashfn(sap, &llc->laddr));
+	if (sk)
+		goto out;
+	spin_unlock_bh(&sap->sk_lock);
+	list_for_each_entry_continue_rcu(sap, &llc_sap_list, node) {
+		spin_lock_bh(&sap->sk_lock);
+		sk = laddr_hash_next(sap, -1);
+		if (sk)
+			break; /* keep the lock */
+		spin_unlock_bh(&sap->sk_lock);
+	}
+out:
+	return sk;
+}
+
+static void llc_seq_stop(struct seq_file *seq, void *v)
+{
+	if (v && v != SEQ_START_TOKEN) {
+		struct sock *sk = v;
+		struct llc_sock *llc = llc_sk(sk);
+		struct llc_sap *sap = llc->sap;
+
+		spin_unlock_bh(&sap->sk_lock);
+	}
+	rcu_read_unlock_bh();
+}
+
+static int llc_seq_socket_show(struct seq_file *seq, void *v)
+{
+	struct sock* sk;
+	struct llc_sock *llc;
+
+	if (v == SEQ_START_TOKEN) {
+		seq_puts(seq, "SKt Mc local_mac_sap        remote_mac_sap   "
+			      "    tx_queue rx_queue st uid link\n");
+		goto out;
+	}
+	sk = v;
+	llc = llc_sk(sk);
+
+	/* FIXME: check if the address is multicast */
+	seq_printf(seq, "%2X  %2X ", sk->sk_type, 0);
+
+	if (llc->dev)
+		llc_ui_format_mac(seq, llc->dev->dev_addr);
+	else {
+		u8 addr[6] = {0,0,0,0,0,0};
+		llc_ui_format_mac(seq, addr);
+	}
+	seq_printf(seq, "@%02X ", llc->sap->laddr.lsap);
+	llc_ui_format_mac(seq, llc->daddr.mac);
+	seq_printf(seq, "@%02X %8d %8d %2d %3u %4d\n", llc->daddr.lsap,
+		   sk_wmem_alloc_get(sk),
+		   sk_rmem_alloc_get(sk) - llc->copied_seq,
+		   sk->sk_state,
+		   from_kuid_munged(seq_user_ns(seq), sk_uid(sk)),
+		   llc->link);
+out:
+	return 0;
+}
+
+static const char *const llc_conn_state_names[] = {
+	[LLC_CONN_STATE_ADM] =        "adm",
+	[LLC_CONN_STATE_SETUP] =      "setup",
+	[LLC_CONN_STATE_NORMAL] =     "normal",
+	[LLC_CONN_STATE_BUSY] =       "busy",
+	[LLC_CONN_STATE_REJ] =        "rej",
+	[LLC_CONN_STATE_AWAIT] =      "await",
+	[LLC_CONN_STATE_AWAIT_BUSY] = "await_busy",
+	[LLC_CONN_STATE_AWAIT_REJ] =  "await_rej",
+	[LLC_CONN_STATE_D_CONN]	=     "d_conn",
+	[LLC_CONN_STATE_RESET] =      "reset",
+	[LLC_CONN_STATE_ERROR] =      "error",
+	[LLC_CONN_STATE_TEMP] =       "temp",
+};
+
+static int llc_seq_core_show(struct seq_file *seq, void *v)
+{
+	struct sock* sk;
+	struct llc_sock *llc;
+
+	if (v == SEQ_START_TOKEN) {
+		seq_puts(seq, "Connection list:\n"
+			      "dsap state      retr txw rxw pf ff sf df rs cs "
+			      "tack tpfc trs tbs blog busr\n");
+		goto out;
+	}
+	sk = v;
+	llc = llc_sk(sk);
+
+	seq_printf(seq, " %02X  %-10s %3d  %3d %3d %2d %2d %2d %2d %2d %2d "
+			"%4d %4d %3d %3d %4d %4d\n",
+		   llc->daddr.lsap, llc_conn_state_names[llc->state],
+		   llc->retry_count, llc->k, llc->rw, llc->p_flag, llc->f_flag,
+		   llc->s_flag, llc->data_flag, llc->remote_busy_flag,
+		   llc->cause_flag, timer_pending(&llc->ack_timer.timer),
+		   timer_pending(&llc->pf_cycle_timer.timer),
+		   timer_pending(&llc->rej_sent_timer.timer),
+		   timer_pending(&llc->busy_state_timer.timer),
+		   !!sk->sk_backlog.tail, sock_owned_by_user_nocheck(sk));
+out:
+	return 0;
+}
+
+static const struct seq_operations llc_seq_socket_ops = {
+	.start  = llc_seq_start,
+	.next   = llc_seq_next,
+	.stop   = llc_seq_stop,
+	.show   = llc_seq_socket_show,
+};
+
+static const struct seq_operations llc_seq_core_ops = {
+	.start  = llc_seq_start,
+	.next   = llc_seq_next,
+	.stop   = llc_seq_stop,
+	.show   = llc_seq_core_show,
+};
+
+static struct proc_dir_entry *llc_proc_dir;
+
+int __init llc_proc_init(void)
+{
+	int rc = -ENOMEM;
+	struct proc_dir_entry *p;
+
+	llc_proc_dir = proc_mkdir("llc", init_net.proc_net);
+	if (!llc_proc_dir)
+		goto out;
+
+	p = proc_create_seq("socket", 0444, llc_proc_dir, &llc_seq_socket_ops);
+	if (!p)
+		goto out_socket;
+
+	p = proc_create_seq("core", 0444, llc_proc_dir, &llc_seq_core_ops);
+	if (!p)
+		goto out_core;
+
+	rc = 0;
+out:
+	return rc;
+out_core:
+	remove_proc_entry("socket", llc_proc_dir);
+out_socket:
+	remove_proc_entry("llc", init_net.proc_net);
+	goto out;
+}
+
+void llc_proc_exit(void)
+{
+	remove_proc_entry("socket", llc_proc_dir);
+	remove_proc_entry("core", llc_proc_dir);
+	remove_proc_entry("llc", init_net.proc_net);
+}