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, 259 insertions, 0 deletions

diff --git a/net/hsr/hsr_slave.c b/net/hsr/hsr_slave.c
new file mode 100644
index 000000000..a546f70f9
--- /dev/null
+++ b/net/hsr/hsr_slave.c
@@ -0,0 +1,259 @@
+// SPDX-License-Identifier: GPL-2.0
+/* Copyright 2011-2014 Autronica Fire and Security AS
+ *
+ * Author(s):
+ *	2011-2014 Arvid Brodin, arvid.brodin@alten.se
+ *
+ * Frame handler other utility functions for HSR and PRP.
+ */
+
+#include "hsr_slave.h"
+#include <linux/etherdevice.h>
+#include <linux/if_arp.h>
+#include <linux/if_vlan.h>
+#include "hsr_main.h"
+#include "hsr_device.h"
+#include "hsr_forward.h"
+#include "hsr_framereg.h"
+
+bool hsr_invalid_dan_ingress_frame(__be16 protocol)
+{
+	return (protocol != htons(ETH_P_PRP) && protocol != htons(ETH_P_HSR));
+}
+
+static rx_handler_result_t hsr_handle_frame(struct sk_buff **pskb)
+{
+	struct sk_buff *skb = *pskb;
+	struct hsr_port *port;
+	struct hsr_priv *hsr;
+	__be16 protocol;
+
+	/* Packets from dev_loopback_xmit() do not have L2 header, bail out */
+	if (unlikely(skb->pkt_type == PACKET_LOOPBACK))
+		return RX_HANDLER_PASS;
+
+	if (!skb_mac_header_was_set(skb)) {
+		WARN_ONCE(1, "%s: skb invalid", __func__);
+		return RX_HANDLER_PASS;
+	}
+
+	port = hsr_port_get_rcu(skb->dev);
+	if (!port)
+		goto finish_pass;
+	hsr = port->hsr;
+
+	if (hsr_addr_is_self(port->hsr, eth_hdr(skb)->h_source)) {
+		/* Directly kill frames sent by ourselves */
+		kfree_skb(skb);
+		goto finish_consume;
+	}
+
+	/* For HSR, only tagged frames are expected (unless the device offloads
+	 * HSR tag removal), but for PRP there could be non tagged frames as
+	 * well from Single attached nodes (SANs).
+	 */
+	protocol = eth_hdr(skb)->h_proto;
+
+	if (!(port->dev->features & NETIF_F_HW_HSR_TAG_RM) &&
+	    port->type != HSR_PT_INTERLINK &&
+	    hsr->proto_ops->invalid_dan_ingress_frame &&
+	    hsr->proto_ops->invalid_dan_ingress_frame(protocol))
+		goto finish_pass;
+
+	skb_push(skb, ETH_HLEN);
+	skb_reset_mac_header(skb);
+	if ((!hsr->prot_version && protocol == htons(ETH_P_PRP)) ||
+	    protocol == htons(ETH_P_HSR)) {
+		if (!pskb_may_pull(skb, ETH_HLEN + HSR_HLEN)) {
+			kfree_skb(skb);
+			goto finish_consume;
+		}
+
+		skb_set_network_header(skb, ETH_HLEN + HSR_HLEN);
+	}
+	skb_reset_mac_len(skb);
+
+	/* Only the frames received over the interlink port will assign a
+	 * sequence number and require synchronisation vs other sender.
+	 */
+	if (port->type == HSR_PT_INTERLINK) {
+		spin_lock_bh(&hsr->seqnr_lock);
+		hsr_forward_skb(skb, port);
+		spin_unlock_bh(&hsr->seqnr_lock);
+	} else {
+		hsr_forward_skb(skb, port);
+	}
+
+finish_consume:
+	return RX_HANDLER_CONSUMED;
+
+finish_pass:
+	return RX_HANDLER_PASS;
+}
+
+bool hsr_port_exists(const struct net_device *dev)
+{
+	return rcu_access_pointer(dev->rx_handler) == hsr_handle_frame;
+}
+
+static int hsr_check_dev_ok(struct net_device *dev,
+			    struct netlink_ext_ack *extack)
+{
+	/* Don't allow HSR on non-ethernet like devices */
+	if ((dev->flags & IFF_LOOPBACK) || dev->type != ARPHRD_ETHER ||
+	    dev->addr_len != ETH_ALEN) {
+		NL_SET_ERR_MSG_MOD(extack, "Cannot use loopback or non-ethernet device as HSR slave.");
+		return -EINVAL;
+	}
+
+	/* Don't allow enslaving hsr devices */
+	if (is_hsr_master(dev)) {
+		NL_SET_ERR_MSG_MOD(extack,
+				   "Cannot create trees of HSR devices.");
+		return -EINVAL;
+	}
+
+	if (hsr_port_exists(dev)) {
+		NL_SET_ERR_MSG_MOD(extack,
+				   "This device is already a HSR slave.");
+		return -EINVAL;
+	}
+
+	if (is_vlan_dev(dev)) {
+		NL_SET_ERR_MSG_MOD(extack, "HSR on top of VLAN is not yet supported in this driver.");
+		return -EINVAL;
+	}
+
+	if (dev->priv_flags & IFF_DONT_BRIDGE) {
+		NL_SET_ERR_MSG_MOD(extack,
+				   "This device does not support bridging.");
+		return -EOPNOTSUPP;
+	}
+
+	/* HSR over bonded devices has not been tested, but I'm not sure it
+	 * won't work...
+	 */
+
+	return 0;
+}
+
+/* Setup device to be added to the HSR bridge. */
+static int hsr_portdev_setup(struct hsr_priv *hsr, struct net_device *dev,
+			     struct hsr_port *port,
+			     struct netlink_ext_ack *extack)
+
+{
+	struct netdev_lag_upper_info lag_upper_info;
+	struct net_device *hsr_dev;
+	struct hsr_port *master;
+	int res;
+
+	/* Don't use promiscuous mode for offload since L2 frame forward
+	 * happens at the offloaded hardware. The interlink port never
+	 * gets forwarding offload (RedBox forwarding to/from it is done
+	 * by this driver), so it still needs promiscuous mode to receive
+	 * frames addressed to hsr_dev's MAC rather than its own.
+	 */
+	if (!port->hsr->fwd_offloaded || port->type == HSR_PT_INTERLINK) {
+		res = dev_set_promiscuity(dev, 1);
+		if (res)
+			return res;
+	}
+
+	master = hsr_port_get_hsr(hsr, HSR_PT_MASTER);
+	hsr_dev = master->dev;
+
+	lag_upper_info.tx_type = NETDEV_LAG_TX_TYPE_BROADCAST;
+	lag_upper_info.hash_type = NETDEV_LAG_HASH_UNKNOWN;
+	res = netdev_master_upper_dev_link(dev, hsr_dev, NULL, &lag_upper_info, extack);
+	if (res)
+		goto fail_upper_dev_link;
+
+	res = netdev_rx_handler_register(dev, hsr_handle_frame, port);
+	if (res)
+		goto fail_rx_handler;
+	dev_disable_lro(dev);
+
+	return 0;
+
+fail_rx_handler:
+	netdev_upper_dev_unlink(dev, hsr_dev);
+fail_upper_dev_link:
+	if (!port->hsr->fwd_offloaded || port->type == HSR_PT_INTERLINK)
+		dev_set_promiscuity(dev, -1);
+
+	return res;
+}
+
+int hsr_add_port(struct hsr_priv *hsr, struct net_device *dev,
+		 enum hsr_port_type type, struct netlink_ext_ack *extack)
+{
+	struct hsr_port *port, *master;
+	int res;
+
+	if (type != HSR_PT_MASTER) {
+		res = hsr_check_dev_ok(dev, extack);
+		if (res)
+			return res;
+	}
+
+	port = hsr_port_get_hsr(hsr, type);
+	if (port)
+		return -EBUSY;	/* This port already exists */
+
+	port = kzalloc_obj(*port);
+	if (!port)
+		return -ENOMEM;
+
+	port->hsr = hsr;
+	port->dev = dev;
+	port->type = type;
+	ether_addr_copy(port->original_macaddress, dev->dev_addr);
+
+	list_add_tail_rcu(&port->port_list, &hsr->ports);
+
+	if (type != HSR_PT_MASTER) {
+		res = hsr_portdev_setup(hsr, dev, port, extack);
+		if (res)
+			goto fail_dev_setup;
+	}
+
+	master = hsr_port_get_hsr(hsr, HSR_PT_MASTER);
+	netdev_update_features(master->dev);
+	dev_set_mtu(master->dev, hsr_get_max_mtu(hsr));
+
+	return 0;
+
+fail_dev_setup:
+	list_del_rcu(&port->port_list);
+	kfree_rcu(port, rcu);
+	return res;
+}
+
+void hsr_del_port(struct hsr_port *port)
+{
+	struct hsr_priv *hsr;
+	struct hsr_port *master;
+
+	hsr = port->hsr;
+	master = hsr_port_get_hsr(hsr, HSR_PT_MASTER);
+	list_del_rcu(&port->port_list);
+
+	if (port != master) {
+		netdev_update_features(master->dev);
+		dev_set_mtu(master->dev, hsr_get_max_mtu(hsr));
+		netdev_rx_handler_unregister(port->dev);
+		if (!port->hsr->fwd_offloaded || port->type == HSR_PT_INTERLINK)
+			dev_set_promiscuity(port->dev, -1);
+		if (port->type == HSR_PT_SLAVE_A || port->type == HSR_PT_SLAVE_B)
+			vlan_vids_del_by_dev(port->dev, master->dev);
+		netdev_upper_dev_unlink(port->dev, master->dev);
+		if (hsr->prot_version == PRP_V1 &&
+		    port->type == HSR_PT_SLAVE_B) {
+			eth_hw_addr_set(port->dev, port->original_macaddress);
+			call_netdevice_notifiers(NETDEV_CHANGEADDR, port->dev);
+		}
+	}
+
+	kfree_rcu(port, rcu);
+}