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, 281 insertions, 0 deletions

diff --git a/net/8021q/vlanproc.c b/net/8021q/vlanproc.c
new file mode 100644
index 000000000..0e424e089
--- /dev/null
+++ b/net/8021q/vlanproc.c
@@ -0,0 +1,281 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
+/******************************************************************************
+ * vlanproc.c	VLAN Module. /proc filesystem interface.
+ *
+ *		This module is completely hardware-independent and provides
+ *		access to the router using Linux /proc filesystem.
+ *
+ * Author:	Ben Greear, <greearb@candelatech.com> coppied from wanproc.c
+ *               by: Gene Kozin	<genek@compuserve.com>
+ *
+ * Copyright:	(c) 1998 Ben Greear
+ *
+ * ============================================================================
+ * Jan 20, 1998        Ben Greear     Initial Version
+ *****************************************************************************/
+
+#define pr_fmt(fmt) KBUILD_MODNAME ": " fmt
+
+#include <linux/module.h>
+#include <linux/errno.h>
+#include <linux/kernel.h>
+#include <linux/string.h>
+#include <linux/proc_fs.h>
+#include <linux/seq_file.h>
+#include <linux/fs.h>
+#include <linux/netdevice.h>
+#include <linux/if_vlan.h>
+#include <net/net_namespace.h>
+#include <net/netns/generic.h>
+#include "vlanproc.h"
+#include "vlan.h"
+
+/****** Function Prototypes *************************************************/
+
+/* Methods for preparing data for reading proc entries */
+static int vlan_seq_show(struct seq_file *seq, void *v);
+static void *vlan_seq_start(struct seq_file *seq, loff_t *pos);
+static void *vlan_seq_next(struct seq_file *seq, void *v, loff_t *pos);
+static void vlan_seq_stop(struct seq_file *seq, void *);
+static int vlandev_seq_show(struct seq_file *seq, void *v);
+
+/*
+ *	Global Data
+ */
+
+
+/*
+ *	Names of the proc directory entries
+ */
+
+static const char name_root[]	 = "vlan";
+static const char name_conf[]	 = "config";
+
+/*
+ *	Structures for interfacing with the /proc filesystem.
+ *	VLAN creates its own directory /proc/net/vlan with the following
+ *	entries:
+ *	config		device status/configuration
+ *	<device>	entry for each  device
+ */
+
+/*
+ *	Generic /proc/net/vlan/<file> file and inode operations
+ */
+
+static const struct seq_operations vlan_seq_ops = {
+	.start = vlan_seq_start,
+	.next = vlan_seq_next,
+	.stop = vlan_seq_stop,
+	.show = vlan_seq_show,
+};
+
+/*
+ * Proc filesystem directory entries.
+ */
+
+/* Strings */
+static const char *const vlan_name_type_str[VLAN_NAME_TYPE_HIGHEST] = {
+    [VLAN_NAME_TYPE_RAW_PLUS_VID]        = "VLAN_NAME_TYPE_RAW_PLUS_VID",
+    [VLAN_NAME_TYPE_PLUS_VID_NO_PAD]	 = "VLAN_NAME_TYPE_PLUS_VID_NO_PAD",
+    [VLAN_NAME_TYPE_RAW_PLUS_VID_NO_PAD] = "VLAN_NAME_TYPE_RAW_PLUS_VID_NO_PAD",
+    [VLAN_NAME_TYPE_PLUS_VID]		 = "VLAN_NAME_TYPE_PLUS_VID",
+};
+/*
+ *	Interface functions
+ */
+
+/*
+ *	Clean up /proc/net/vlan entries
+ */
+
+void vlan_proc_cleanup(struct net *net)
+{
+	struct vlan_net *vn = net_generic(net, vlan_net_id);
+
+	if (vn->proc_vlan_conf)
+		remove_proc_entry(name_conf, vn->proc_vlan_dir);
+
+	if (vn->proc_vlan_dir)
+		remove_proc_entry(name_root, net->proc_net);
+
+	/* Dynamically added entries should be cleaned up as their vlan_device
+	 * is removed, so we should not have to take care of it here...
+	 */
+}
+
+/*
+ *	Create /proc/net/vlan entries
+ */
+
+int __net_init vlan_proc_init(struct net *net)
+{
+	struct vlan_net *vn = net_generic(net, vlan_net_id);
+
+	vn->proc_vlan_dir = proc_net_mkdir(net, name_root, net->proc_net);
+	if (!vn->proc_vlan_dir)
+		goto err;
+
+	vn->proc_vlan_conf = proc_create_net(name_conf, S_IFREG | 0600,
+			vn->proc_vlan_dir, &vlan_seq_ops,
+			sizeof(struct seq_net_private));
+	if (!vn->proc_vlan_conf)
+		goto err;
+	return 0;
+
+err:
+	pr_err("can't create entry in proc filesystem!\n");
+	vlan_proc_cleanup(net);
+	return -ENOBUFS;
+}
+
+/*
+ *	Add directory entry for VLAN device.
+ */
+
+int vlan_proc_add_dev(struct net_device *vlandev)
+{
+	struct vlan_dev_priv *vlan = vlan_dev_priv(vlandev);
+	struct vlan_net *vn = net_generic(dev_net(vlandev), vlan_net_id);
+
+	if (!strcmp(vlandev->name, name_conf))
+		return -EINVAL;
+	vlan->dent = proc_create_single_data(vlandev->name, S_IFREG | 0600,
+			vn->proc_vlan_dir, vlandev_seq_show, vlandev);
+	if (!vlan->dent)
+		return -ENOBUFS;
+	return 0;
+}
+
+/*
+ *	Delete directory entry for VLAN device.
+ */
+void vlan_proc_rem_dev(struct net_device *vlandev)
+{
+	/** NOTE:  This will consume the memory pointed to by dent, it seems. */
+	proc_remove(vlan_dev_priv(vlandev)->dent);
+	vlan_dev_priv(vlandev)->dent = NULL;
+}
+
+/****** Proc filesystem entry points ****************************************/
+
+/*
+ * The following few functions build the content of /proc/net/vlan/config
+ */
+
+static void *vlan_seq_from_index(struct seq_file *seq, loff_t *pos)
+{
+	unsigned long ifindex = *pos;
+	struct net_device *dev;
+
+	for_each_netdev_dump(seq_file_net(seq), dev, ifindex) {
+		if (!is_vlan_dev(dev))
+			continue;
+		*pos = dev->ifindex;
+		return dev;
+	}
+	return NULL;
+}
+
+static void *vlan_seq_start(struct seq_file *seq, loff_t *pos)
+	__acquires(rcu)
+{
+	rcu_read_lock();
+	if (*pos == 0)
+		return SEQ_START_TOKEN;
+
+	return vlan_seq_from_index(seq, pos);
+}
+
+static void *vlan_seq_next(struct seq_file *seq, void *v, loff_t *pos)
+{
+	++*pos;
+	return vlan_seq_from_index(seq, pos);
+}
+
+static void vlan_seq_stop(struct seq_file *seq, void *v)
+	__releases(rcu)
+{
+	rcu_read_unlock();
+}
+
+static int vlan_seq_show(struct seq_file *seq, void *v)
+{
+	struct net *net = seq_file_net(seq);
+	struct vlan_net *vn = net_generic(net, vlan_net_id);
+
+	if (v == SEQ_START_TOKEN) {
+		const char *nmtype = NULL;
+
+		seq_puts(seq, "VLAN Dev name	 | VLAN ID\n");
+
+		if (vn->name_type < ARRAY_SIZE(vlan_name_type_str))
+		    nmtype =  vlan_name_type_str[vn->name_type];
+
+		seq_printf(seq, "Name-Type: %s\n",
+			   nmtype ? nmtype :  "UNKNOWN");
+	} else {
+		const struct net_device *vlandev = v;
+		const struct vlan_dev_priv *vlan = vlan_dev_priv(vlandev);
+
+		seq_printf(seq, "%-15s| %d  | %s\n",  vlandev->name,
+			   vlan->vlan_id,    vlan->real_dev->name);
+	}
+	return 0;
+}
+
+static int vlandev_seq_show(struct seq_file *seq, void *offset)
+{
+	struct net_device *vlandev = (struct net_device *) seq->private;
+	const struct vlan_dev_priv *vlan = vlan_dev_priv(vlandev);
+	struct rtnl_link_stats64 temp;
+	const struct rtnl_link_stats64 *stats;
+	static const char fmt64[] = "%30s %12llu\n";
+	int i;
+
+	if (!is_vlan_dev(vlandev))
+		return 0;
+
+	stats = dev_get_stats(vlandev, &temp);
+	seq_printf(seq,
+		   "%s  VID: %d	 REORDER_HDR: %i  dev->priv_flags: %x\n",
+		   vlandev->name, vlan->vlan_id,
+		   (int)(vlan->flags & 1), (u32)vlandev->priv_flags);
+
+	seq_printf(seq, fmt64, "total frames received", stats->rx_packets);
+	seq_printf(seq, fmt64, "total bytes received", stats->rx_bytes);
+	seq_printf(seq, fmt64, "Broadcast/Multicast Rcvd", stats->multicast);
+	seq_puts(seq, "\n");
+	seq_printf(seq, fmt64, "total frames transmitted", stats->tx_packets);
+	seq_printf(seq, fmt64, "total bytes transmitted", stats->tx_bytes);
+	seq_printf(seq, "Device: %s", vlan->real_dev->name);
+	/* now show all PRIORITY mappings relating to this VLAN */
+	seq_printf(seq, "\nINGRESS priority mappings: "
+			"0:%u  1:%u  2:%u  3:%u  4:%u  5:%u  6:%u 7:%u\n",
+		   vlan->ingress_priority_map[0],
+		   vlan->ingress_priority_map[1],
+		   vlan->ingress_priority_map[2],
+		   vlan->ingress_priority_map[3],
+		   vlan->ingress_priority_map[4],
+		   vlan->ingress_priority_map[5],
+		   vlan->ingress_priority_map[6],
+		   vlan->ingress_priority_map[7]);
+
+	seq_printf(seq, " EGRESS priority mappings: ");
+	rcu_read_lock();
+	for (i = 0; i < 16; i++) {
+		const struct vlan_priority_tci_mapping *mp =
+			rcu_dereference(vlan->egress_priority_map[i]);
+		while (mp) {
+			u16 vlan_qos = READ_ONCE(mp->vlan_qos);
+
+			seq_printf(seq, "%u:%d ",
+				   mp->priority, ((vlan_qos >> 13) & 0x7));
+			mp = rcu_dereference(mp->next);
+		}
+	}
+	rcu_read_unlock();
+	seq_puts(seq, "\n");
+
+	return 0;
+}