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

3 files changed, 165 insertions, 0 deletions

diff --git a/net/nsh/Kconfig b/net/nsh/Kconfig
new file mode 100644
index 000000000..84f2a2823
--- /dev/null
+++ b/net/nsh/Kconfig
@@ -0,0 +1,10 @@
+# SPDX-License-Identifier: GPL-2.0-only
+menuconfig NET_NSH
+	tristate "Network Service Header (NSH) protocol"
+	default n
+	help
+	  Network Service Header is an implementation of Service Function
+	  Chaining (RFC 7665). The current implementation in Linux supports
+	  only MD type 1 and only with the openvswitch module.
+
+	  If unsure, say N.
diff --git a/net/nsh/Makefile b/net/nsh/Makefile
new file mode 100644
index 000000000..de34a615a
--- /dev/null
+++ b/net/nsh/Makefile
@@ -0,0 +1,2 @@
+# SPDX-License-Identifier: GPL-2.0-only
+obj-$(CONFIG_NET_NSH) += nsh.o
diff --git a/net/nsh/nsh.c b/net/nsh/nsh.c
new file mode 100644
index 000000000..bfb775806
--- /dev/null
+++ b/net/nsh/nsh.c
@@ -0,0 +1,153 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ * Network Service Header
+ *
+ * Copyright (c) 2017 Red Hat, Inc. -- Jiri Benc <jbenc@redhat.com>
+ */
+
+#include <linux/module.h>
+#include <linux/netdevice.h>
+#include <linux/skbuff.h>
+#include <net/gso.h>
+#include <net/nsh.h>
+#include <net/tun_proto.h>
+
+int nsh_push(struct sk_buff *skb, const struct nshhdr *pushed_nh)
+{
+	struct nshhdr *nh;
+	size_t length = nsh_hdr_len(pushed_nh);
+	u8 next_proto;
+
+	if (skb->mac_len) {
+		next_proto = TUN_P_ETHERNET;
+	} else {
+		next_proto = tun_p_from_eth_p(skb->protocol);
+		if (!next_proto)
+			return -EAFNOSUPPORT;
+	}
+
+	/* Add the NSH header */
+	if (skb_cow_head(skb, length) < 0)
+		return -ENOMEM;
+
+	skb_push(skb, length);
+	nh = (struct nshhdr *)(skb->data);
+	memcpy(nh, pushed_nh, length);
+	nh->np = next_proto;
+	skb_postpush_rcsum(skb, nh, length);
+
+	skb->protocol = htons(ETH_P_NSH);
+	skb_reset_mac_header(skb);
+	skb_reset_network_header(skb);
+	skb_reset_mac_len(skb);
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(nsh_push);
+
+int nsh_pop(struct sk_buff *skb)
+{
+	struct nshhdr *nh;
+	size_t length;
+	__be16 inner_proto;
+
+	if (!pskb_may_pull(skb, NSH_BASE_HDR_LEN))
+		return -ENOMEM;
+	nh = (struct nshhdr *)(skb->data);
+	length = nsh_hdr_len(nh);
+	if (length < NSH_BASE_HDR_LEN)
+		return -EINVAL;
+	inner_proto = tun_p_to_eth_p(nh->np);
+	if (!pskb_may_pull(skb, length))
+		return -ENOMEM;
+
+	if (!inner_proto)
+		return -EAFNOSUPPORT;
+
+	skb_pull_rcsum(skb, length);
+	skb_reset_mac_header(skb);
+	skb_reset_network_header(skb);
+	skb_reset_mac_len(skb);
+	skb->protocol = inner_proto;
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(nsh_pop);
+
+static struct sk_buff *nsh_gso_segment(struct sk_buff *skb,
+				       netdev_features_t features)
+{
+	unsigned int outer_hlen, mac_len, nsh_len;
+	struct sk_buff *segs = ERR_PTR(-EINVAL);
+	u16 mac_offset = skb->mac_header;
+	__be16 outer_proto, proto;
+
+	skb_reset_network_header(skb);
+
+	outer_proto = skb->protocol;
+	outer_hlen = skb_mac_header_len(skb);
+	mac_len = skb->mac_len;
+
+	if (unlikely(!pskb_may_pull(skb, NSH_BASE_HDR_LEN)))
+		goto out;
+	nsh_len = nsh_hdr_len(nsh_hdr(skb));
+	if (nsh_len < NSH_BASE_HDR_LEN)
+		goto out;
+	if (unlikely(!pskb_may_pull(skb, nsh_len)))
+		goto out;
+
+	proto = tun_p_to_eth_p(nsh_hdr(skb)->np);
+	if (!proto)
+		goto out;
+
+	__skb_pull(skb, nsh_len);
+
+	skb_reset_mac_header(skb);
+	skb->mac_len = proto == htons(ETH_P_TEB) ? ETH_HLEN : 0;
+	skb->protocol = proto;
+
+	features &= NETIF_F_SG;
+	segs = skb_mac_gso_segment(skb, features);
+	if (IS_ERR_OR_NULL(segs)) {
+		skb_gso_error_unwind(skb, htons(ETH_P_NSH), nsh_len,
+				     mac_offset, mac_len);
+		goto out;
+	}
+
+	for (skb = segs; skb; skb = skb->next) {
+		skb->protocol = outer_proto;
+		__skb_push(skb, nsh_len + outer_hlen);
+		skb_reset_mac_header(skb);
+		skb_set_network_header(skb, outer_hlen);
+		skb->mac_len = mac_len;
+	}
+
+out:
+	return segs;
+}
+
+static struct packet_offload nsh_packet_offload __read_mostly = {
+	.type = htons(ETH_P_NSH),
+	.priority = 15,
+	.callbacks = {
+		.gso_segment = nsh_gso_segment,
+	},
+};
+
+static int __init nsh_init_module(void)
+{
+	dev_add_offload(&nsh_packet_offload);
+	return 0;
+}
+
+static void __exit nsh_cleanup_module(void)
+{
+	dev_remove_offload(&nsh_packet_offload);
+}
+
+module_init(nsh_init_module);
+module_exit(nsh_cleanup_module);
+
+MODULE_AUTHOR("Jiri Benc <jbenc@redhat.com>");
+MODULE_DESCRIPTION("NSH protocol");
+MODULE_LICENSE("GPL v2");