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, 226 insertions, 0 deletions

diff --git a/net/ipv6/fou6.c b/net/ipv6/fou6.c
new file mode 100644
index 000000000..4b659ca60
--- /dev/null
+++ b/net/ipv6/fou6.c
@@ -0,0 +1,226 @@
+// SPDX-License-Identifier: GPL-2.0-only
+#include <linux/module.h>
+#include <linux/errno.h>
+#include <linux/socket.h>
+#include <linux/skbuff.h>
+#include <linux/ip.h>
+#include <linux/udp.h>
+#include <linux/icmpv6.h>
+#include <linux/types.h>
+#include <linux/kernel.h>
+#include <net/fou.h>
+#include <net/ip.h>
+#include <net/ip6_tunnel.h>
+#include <net/ip6_checksum.h>
+#include <net/protocol.h>
+#include <net/udp.h>
+#include <net/udp_tunnel.h>
+
+#if IS_ENABLED(CONFIG_IPV6_FOU_TUNNEL)
+
+static void fou6_build_udp(struct sk_buff *skb, struct ip_tunnel_encap *e,
+			   struct flowi6 *fl6, u8 *protocol, __be16 sport)
+{
+	struct udphdr *uh;
+
+	skb_push(skb, sizeof(struct udphdr));
+	skb_reset_transport_header(skb);
+
+	uh = udp_hdr(skb);
+
+	uh->dest = e->dport;
+	uh->source = sport;
+	udp_set_len(uh, skb->len);
+	udp6_set_csum(!(e->flags & TUNNEL_ENCAP_FLAG_CSUM6), skb,
+		      &fl6->saddr, &fl6->daddr, skb->len);
+
+	*protocol = IPPROTO_UDP;
+}
+
+static int fou6_build_header(struct sk_buff *skb, struct ip_tunnel_encap *e,
+			     u8 *protocol, struct flowi6 *fl6)
+{
+	__be16 sport;
+	int err;
+	int type = e->flags & TUNNEL_ENCAP_FLAG_CSUM6 ?
+		SKB_GSO_UDP_TUNNEL_CSUM : SKB_GSO_UDP_TUNNEL;
+
+	err = __fou_build_header(skb, e, protocol, &sport, type);
+	if (err)
+		return err;
+
+	fou6_build_udp(skb, e, fl6, protocol, sport);
+
+	return 0;
+}
+
+static int gue6_build_header(struct sk_buff *skb, struct ip_tunnel_encap *e,
+			     u8 *protocol, struct flowi6 *fl6)
+{
+	__be16 sport;
+	int err;
+	int type = e->flags & TUNNEL_ENCAP_FLAG_CSUM6 ?
+		SKB_GSO_UDP_TUNNEL_CSUM : SKB_GSO_UDP_TUNNEL;
+
+	err = __gue_build_header(skb, e, protocol, &sport, type);
+	if (err)
+		return err;
+
+	fou6_build_udp(skb, e, fl6, protocol, sport);
+
+	return 0;
+}
+
+static int gue6_err_proto_handler(int proto, struct sk_buff *skb,
+				  struct inet6_skb_parm *opt,
+				  u8 type, u8 code, int offset, __be32 info)
+{
+	const struct inet6_protocol *ipprot;
+
+	ipprot = rcu_dereference(inet6_protos[proto]);
+	if (ipprot && ipprot->err_handler) {
+		if (!ipprot->err_handler(skb, opt, type, code, offset, info))
+			return 0;
+	}
+
+	return -ENOENT;
+}
+
+static int gue6_err(struct sk_buff *skb, struct inet6_skb_parm *opt,
+		    u8 type, u8 code, int offset, __be32 info)
+{
+	int transport_offset = skb_transport_offset(skb);
+	struct guehdr *guehdr;
+	size_t len, optlen;
+	int ret;
+
+	len = sizeof(struct udphdr) + sizeof(struct guehdr);
+	if (!pskb_may_pull(skb, transport_offset + len))
+		return -EINVAL;
+
+	guehdr = (struct guehdr *)&udp_hdr(skb)[1];
+
+	switch (guehdr->version) {
+	case 0: /* Full GUE header present */
+		break;
+	case 1: {
+		/* Direct encasulation of IPv4 or IPv6 */
+		skb_set_transport_header(skb, -(int)sizeof(struct icmp6hdr));
+
+		switch (((struct iphdr *)guehdr)->version) {
+		case 4:
+			ret = gue6_err_proto_handler(IPPROTO_IPIP, skb, opt,
+						     type, code, offset, info);
+			goto out;
+		case 6:
+			ret = gue6_err_proto_handler(IPPROTO_IPV6, skb, opt,
+						     type, code, offset, info);
+			goto out;
+		default:
+			ret = -EOPNOTSUPP;
+			goto out;
+		}
+	}
+	default: /* Undefined version */
+		return -EOPNOTSUPP;
+	}
+
+	if (guehdr->control)
+		return -ENOENT;
+
+	optlen = guehdr->hlen << 2;
+
+	if (!pskb_may_pull(skb, transport_offset + len + optlen))
+		return -EINVAL;
+
+	guehdr = (struct guehdr *)&udp_hdr(skb)[1];
+	if (validate_gue_flags(guehdr, optlen))
+		return -EINVAL;
+
+	/* Handling exceptions for direct UDP encapsulation in GUE would lead to
+	 * recursion. Besides, this kind of encapsulation can't even be
+	 * configured currently. Discard this.
+	 */
+	if (guehdr->proto_ctype == IPPROTO_UDP)
+		return -EOPNOTSUPP;
+
+	skb_set_transport_header(skb, -(int)sizeof(struct icmp6hdr));
+	ret = gue6_err_proto_handler(guehdr->proto_ctype, skb,
+				     opt, type, code, offset, info);
+
+out:
+	skb_set_transport_header(skb, transport_offset);
+	return ret;
+}
+
+
+static const struct ip6_tnl_encap_ops fou_ip6tun_ops = {
+	.encap_hlen = fou_encap_hlen,
+	.build_header = fou6_build_header,
+	.err_handler = gue6_err,
+};
+
+static const struct ip6_tnl_encap_ops gue_ip6tun_ops = {
+	.encap_hlen = gue_encap_hlen,
+	.build_header = gue6_build_header,
+	.err_handler = gue6_err,
+};
+
+static int ip6_tnl_encap_add_fou_ops(void)
+{
+	int ret;
+
+	ret = ip6_tnl_encap_add_ops(&fou_ip6tun_ops, TUNNEL_ENCAP_FOU);
+	if (ret < 0) {
+		pr_err("can't add fou6 ops\n");
+		return ret;
+	}
+
+	ret = ip6_tnl_encap_add_ops(&gue_ip6tun_ops, TUNNEL_ENCAP_GUE);
+	if (ret < 0) {
+		pr_err("can't add gue6 ops\n");
+		ip6_tnl_encap_del_ops(&fou_ip6tun_ops, TUNNEL_ENCAP_FOU);
+		return ret;
+	}
+
+	return 0;
+}
+
+static void ip6_tnl_encap_del_fou_ops(void)
+{
+	ip6_tnl_encap_del_ops(&fou_ip6tun_ops, TUNNEL_ENCAP_FOU);
+	ip6_tnl_encap_del_ops(&gue_ip6tun_ops, TUNNEL_ENCAP_GUE);
+}
+
+#else
+
+static int ip6_tnl_encap_add_fou_ops(void)
+{
+	return 0;
+}
+
+static void ip6_tnl_encap_del_fou_ops(void)
+{
+}
+
+#endif
+
+static int __init fou6_init(void)
+{
+	int ret;
+
+	ret = ip6_tnl_encap_add_fou_ops();
+
+	return ret;
+}
+
+static void __exit fou6_fini(void)
+{
+	ip6_tnl_encap_del_fou_ops();
+}
+
+module_init(fou6_init);
+module_exit(fou6_fini);
+MODULE_AUTHOR("Tom Herbert <therbert@google.com>");
+MODULE_LICENSE("GPL");
+MODULE_DESCRIPTION("Foo over UDP (IPv6)");