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, 411 insertions, 0 deletions

diff --git a/net/ipv6/mip6.c b/net/ipv6/mip6.c
new file mode 100644
index 000000000..6a16a5bd0
--- /dev/null
+++ b/net/ipv6/mip6.c
@@ -0,0 +1,411 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
+/*
+ * Copyright (C)2003-2006 Helsinki University of Technology
+ * Copyright (C)2003-2006 USAGI/WIDE Project
+ */
+/*
+ * Authors:
+ *	Noriaki TAKAMIYA @USAGI
+ *	Masahide NAKAMURA @USAGI
+ */
+
+#define pr_fmt(fmt) KBUILD_MODNAME ": " fmt
+
+#include <linux/module.h>
+#include <linux/skbuff.h>
+#include <linux/time.h>
+#include <linux/ipv6.h>
+#include <linux/icmpv6.h>
+#include <net/sock.h>
+#include <net/ipv6.h>
+#include <net/ip6_checksum.h>
+#include <net/rawv6.h>
+#include <net/xfrm.h>
+#include <net/mip6.h>
+
+static inline unsigned int calc_padlen(unsigned int len, unsigned int n)
+{
+	return (n - len + 16) & 0x7;
+}
+
+static inline void *mip6_padn(__u8 *data, __u8 padlen)
+{
+	if (!data)
+		return NULL;
+	if (padlen == 1) {
+		data[0] = IPV6_TLV_PAD1;
+	} else if (padlen > 1) {
+		data[0] = IPV6_TLV_PADN;
+		data[1] = padlen - 2;
+		if (padlen > 2)
+			memset(data+2, 0, data[1]);
+	}
+	return data + padlen;
+}
+
+static inline void mip6_param_prob(struct sk_buff *skb, u8 code, int pos)
+{
+	icmpv6_send(skb, ICMPV6_PARAMPROB, code, pos);
+}
+
+static int mip6_mh_len(int type)
+{
+	int len = 0;
+
+	switch (type) {
+	case IP6_MH_TYPE_BRR:
+		len = 0;
+		break;
+	case IP6_MH_TYPE_HOTI:
+	case IP6_MH_TYPE_COTI:
+	case IP6_MH_TYPE_BU:
+	case IP6_MH_TYPE_BACK:
+		len = 1;
+		break;
+	case IP6_MH_TYPE_HOT:
+	case IP6_MH_TYPE_COT:
+	case IP6_MH_TYPE_BERROR:
+		len = 2;
+		break;
+	}
+	return len;
+}
+
+static int mip6_mh_filter(struct sock *sk, struct sk_buff *skb)
+{
+	struct ip6_mh _hdr;
+	const struct ip6_mh *mh;
+
+	mh = skb_header_pointer(skb, skb_transport_offset(skb),
+				sizeof(_hdr), &_hdr);
+	if (!mh)
+		return -1;
+
+	if (((mh->ip6mh_hdrlen + 1) << 3) > skb->len)
+		return -1;
+
+	if (mh->ip6mh_hdrlen < mip6_mh_len(mh->ip6mh_type)) {
+		net_dbg_ratelimited("mip6: MH message too short: %d vs >=%d\n",
+				    mh->ip6mh_hdrlen,
+				    mip6_mh_len(mh->ip6mh_type));
+		mip6_param_prob(skb, 0, offsetof(struct ip6_mh, ip6mh_hdrlen) +
+				skb_network_header_len(skb));
+		return -1;
+	}
+
+	if (mh->ip6mh_proto != IPPROTO_NONE) {
+		net_dbg_ratelimited("mip6: MH invalid payload proto = %d\n",
+				    mh->ip6mh_proto);
+		mip6_param_prob(skb, 0, offsetof(struct ip6_mh, ip6mh_proto) +
+				skb_network_header_len(skb));
+		return -1;
+	}
+
+	return 0;
+}
+
+struct mip6_report_rate_limiter {
+	spinlock_t lock;
+	ktime_t stamp;
+	int iif;
+	struct in6_addr src;
+	struct in6_addr dst;
+};
+
+static struct mip6_report_rate_limiter mip6_report_rl = {
+	.lock = __SPIN_LOCK_UNLOCKED(mip6_report_rl.lock)
+};
+
+static int mip6_destopt_input(struct xfrm_state *x, struct sk_buff *skb)
+{
+	const struct ipv6hdr *iph = ipv6_hdr(skb);
+	struct ipv6_destopt_hdr *destopt = (struct ipv6_destopt_hdr *)skb->data;
+	int err = destopt->nexthdr;
+
+	spin_lock(&x->lock);
+	if (!ipv6_addr_equal(&iph->saddr, (struct in6_addr *)x->coaddr) &&
+	    !ipv6_addr_any((struct in6_addr *)x->coaddr))
+		err = -ENOENT;
+	spin_unlock(&x->lock);
+
+	return err;
+}
+
+/* Destination Option Header is inserted.
+ * IP Header's src address is replaced with Home Address Option in
+ * Destination Option Header.
+ */
+static int mip6_destopt_output(struct xfrm_state *x, struct sk_buff *skb)
+{
+	struct ipv6hdr *iph;
+	struct ipv6_destopt_hdr *dstopt;
+	struct ipv6_destopt_hao *hao;
+	u8 nexthdr;
+	int len;
+
+	skb_push(skb, -skb_network_offset(skb));
+	iph = ipv6_hdr(skb);
+
+	nexthdr = *skb_mac_header(skb);
+	*skb_mac_header(skb) = IPPROTO_DSTOPTS;
+
+	dstopt = (struct ipv6_destopt_hdr *)skb_transport_header(skb);
+	dstopt->nexthdr = nexthdr;
+
+	hao = mip6_padn((char *)(dstopt + 1),
+			calc_padlen(sizeof(*dstopt), 6));
+
+	hao->type = IPV6_TLV_HAO;
+	BUILD_BUG_ON(sizeof(*hao) != 18);
+	hao->length = sizeof(*hao) - 2;
+
+	len = ((char *)hao - (char *)dstopt) + sizeof(*hao);
+
+	memcpy(&hao->addr, &iph->saddr, sizeof(hao->addr));
+	spin_lock_bh(&x->lock);
+	memcpy(&iph->saddr, x->coaddr, sizeof(iph->saddr));
+	spin_unlock_bh(&x->lock);
+
+	WARN_ON(len != x->props.header_len);
+	dstopt->hdrlen = (x->props.header_len >> 3) - 1;
+
+	return 0;
+}
+
+static inline int mip6_report_rl_allow(ktime_t stamp,
+				       const struct in6_addr *dst,
+				       const struct in6_addr *src, int iif)
+{
+	int allow = 0;
+
+	spin_lock_bh(&mip6_report_rl.lock);
+	if (mip6_report_rl.stamp != stamp ||
+	    mip6_report_rl.iif != iif ||
+	    !ipv6_addr_equal(&mip6_report_rl.src, src) ||
+	    !ipv6_addr_equal(&mip6_report_rl.dst, dst)) {
+		mip6_report_rl.stamp = stamp;
+		mip6_report_rl.iif = iif;
+		mip6_report_rl.src = *src;
+		mip6_report_rl.dst = *dst;
+		allow = 1;
+	}
+	spin_unlock_bh(&mip6_report_rl.lock);
+	return allow;
+}
+
+static int mip6_destopt_reject(struct xfrm_state *x, struct sk_buff *skb,
+			       const struct flowi *fl)
+{
+	struct net *net = xs_net(x);
+	struct inet6_skb_parm *opt = (struct inet6_skb_parm *)skb->cb;
+	const struct flowi6 *fl6 = &fl->u.ip6;
+	struct ipv6_destopt_hao *hao = NULL;
+	struct xfrm_selector sel;
+	int offset;
+	ktime_t stamp;
+	int err = 0;
+
+	if (unlikely(fl6->flowi6_proto == IPPROTO_MH &&
+		     fl6->fl6_mh_type <= IP6_MH_TYPE_MAX))
+		goto out;
+
+	if (likely(opt->dsthao)) {
+		offset = ipv6_find_tlv(skb, opt->dsthao, IPV6_TLV_HAO);
+		if (likely(offset >= 0))
+			hao = (struct ipv6_destopt_hao *)
+					(skb_network_header(skb) + offset);
+	}
+
+	stamp = skb_get_ktime(skb);
+
+	if (!mip6_report_rl_allow(stamp, &ipv6_hdr(skb)->daddr,
+				  hao ? &hao->addr : &ipv6_hdr(skb)->saddr,
+				  opt->iif))
+		goto out;
+
+	memset(&sel, 0, sizeof(sel));
+	memcpy(&sel.daddr, (xfrm_address_t *)&ipv6_hdr(skb)->daddr,
+	       sizeof(sel.daddr));
+	sel.prefixlen_d = 128;
+	memcpy(&sel.saddr, (xfrm_address_t *)&ipv6_hdr(skb)->saddr,
+	       sizeof(sel.saddr));
+	sel.prefixlen_s = 128;
+	sel.family = AF_INET6;
+	sel.proto = fl6->flowi6_proto;
+	sel.dport = xfrm_flowi_dport(fl, &fl6->uli);
+	if (sel.dport)
+		sel.dport_mask = htons(~0);
+	sel.sport = xfrm_flowi_sport(fl, &fl6->uli);
+	if (sel.sport)
+		sel.sport_mask = htons(~0);
+	sel.ifindex = fl6->flowi6_oif;
+
+	err = km_report(net, IPPROTO_DSTOPTS, &sel,
+			(hao ? (xfrm_address_t *)&hao->addr : NULL));
+
+ out:
+	return err;
+}
+
+static int mip6_destopt_init_state(struct xfrm_state *x, struct netlink_ext_ack *extack)
+{
+	if (x->id.spi) {
+		NL_SET_ERR_MSG(extack, "SPI must be 0");
+		return -EINVAL;
+	}
+	if (x->props.mode != XFRM_MODE_ROUTEOPTIMIZATION) {
+		NL_SET_ERR_MSG(extack, "XFRM mode must be XFRM_MODE_ROUTEOPTIMIZATION");
+		return -EINVAL;
+	}
+
+	x->props.header_len = sizeof(struct ipv6_destopt_hdr) +
+		calc_padlen(sizeof(struct ipv6_destopt_hdr), 6) +
+		sizeof(struct ipv6_destopt_hao);
+	WARN_ON(x->props.header_len != 24);
+
+	return 0;
+}
+
+/*
+ * Do nothing about destroying since it has no specific operation for
+ * destination options header unlike IPsec protocols.
+ */
+static void mip6_destopt_destroy(struct xfrm_state *x)
+{
+}
+
+static const struct xfrm_type mip6_destopt_type = {
+	.owner		= THIS_MODULE,
+	.proto		= IPPROTO_DSTOPTS,
+	.flags		= XFRM_TYPE_NON_FRAGMENT | XFRM_TYPE_LOCAL_COADDR,
+	.init_state	= mip6_destopt_init_state,
+	.destructor	= mip6_destopt_destroy,
+	.input		= mip6_destopt_input,
+	.output		= mip6_destopt_output,
+	.reject		= mip6_destopt_reject,
+};
+
+static int mip6_rthdr_input(struct xfrm_state *x, struct sk_buff *skb)
+{
+	const struct ipv6hdr *iph = ipv6_hdr(skb);
+	struct rt2_hdr *rt2 = (struct rt2_hdr *)skb->data;
+	int err = rt2->rt_hdr.nexthdr;
+
+	spin_lock(&x->lock);
+	if (!ipv6_addr_equal(&iph->daddr, (struct in6_addr *)x->coaddr) &&
+	    !ipv6_addr_any((struct in6_addr *)x->coaddr))
+		err = -ENOENT;
+	spin_unlock(&x->lock);
+
+	return err;
+}
+
+/* Routing Header type 2 is inserted.
+ * IP Header's dst address is replaced with Routing Header's Home Address.
+ */
+static int mip6_rthdr_output(struct xfrm_state *x, struct sk_buff *skb)
+{
+	struct ipv6hdr *iph;
+	struct rt2_hdr *rt2;
+	u8 nexthdr;
+
+	skb_push(skb, -skb_network_offset(skb));
+	iph = ipv6_hdr(skb);
+
+	nexthdr = *skb_mac_header(skb);
+	*skb_mac_header(skb) = IPPROTO_ROUTING;
+
+	rt2 = (struct rt2_hdr *)skb_transport_header(skb);
+	rt2->rt_hdr.nexthdr = nexthdr;
+	rt2->rt_hdr.hdrlen = (x->props.header_len >> 3) - 1;
+	rt2->rt_hdr.type = IPV6_SRCRT_TYPE_2;
+	rt2->rt_hdr.segments_left = 1;
+	memset(&rt2->reserved, 0, sizeof(rt2->reserved));
+
+	WARN_ON(rt2->rt_hdr.hdrlen != 2);
+
+	memcpy(&rt2->addr, &iph->daddr, sizeof(rt2->addr));
+	spin_lock_bh(&x->lock);
+	memcpy(&iph->daddr, x->coaddr, sizeof(iph->daddr));
+	spin_unlock_bh(&x->lock);
+
+	return 0;
+}
+
+static int mip6_rthdr_init_state(struct xfrm_state *x, struct netlink_ext_ack *extack)
+{
+	if (x->id.spi) {
+		NL_SET_ERR_MSG(extack, "SPI must be 0");
+		return -EINVAL;
+	}
+	if (x->props.mode != XFRM_MODE_ROUTEOPTIMIZATION) {
+		NL_SET_ERR_MSG(extack, "XFRM mode must be XFRM_MODE_ROUTEOPTIMIZATION");
+		return -EINVAL;
+	}
+
+	x->props.header_len = sizeof(struct rt2_hdr);
+
+	return 0;
+}
+
+/*
+ * Do nothing about destroying since it has no specific operation for routing
+ * header type 2 unlike IPsec protocols.
+ */
+static void mip6_rthdr_destroy(struct xfrm_state *x)
+{
+}
+
+static const struct xfrm_type mip6_rthdr_type = {
+	.owner		= THIS_MODULE,
+	.proto		= IPPROTO_ROUTING,
+	.flags		= XFRM_TYPE_NON_FRAGMENT | XFRM_TYPE_REMOTE_COADDR,
+	.init_state	= mip6_rthdr_init_state,
+	.destructor	= mip6_rthdr_destroy,
+	.input		= mip6_rthdr_input,
+	.output		= mip6_rthdr_output,
+};
+
+static int __init mip6_init(void)
+{
+	pr_info("Mobile IPv6\n");
+
+	if (xfrm_register_type(&mip6_destopt_type, AF_INET6) < 0) {
+		pr_info("%s: can't add xfrm type(destopt)\n", __func__);
+		goto mip6_destopt_xfrm_fail;
+	}
+	if (xfrm_register_type(&mip6_rthdr_type, AF_INET6) < 0) {
+		pr_info("%s: can't add xfrm type(rthdr)\n", __func__);
+		goto mip6_rthdr_xfrm_fail;
+	}
+	if (rawv6_mh_filter_register(mip6_mh_filter) < 0) {
+		pr_info("%s: can't add rawv6 mh filter\n", __func__);
+		goto mip6_rawv6_mh_fail;
+	}
+
+
+	return 0;
+
+ mip6_rawv6_mh_fail:
+	xfrm_unregister_type(&mip6_rthdr_type, AF_INET6);
+ mip6_rthdr_xfrm_fail:
+	xfrm_unregister_type(&mip6_destopt_type, AF_INET6);
+ mip6_destopt_xfrm_fail:
+	return -EAGAIN;
+}
+
+static void __exit mip6_fini(void)
+{
+	if (rawv6_mh_filter_unregister(mip6_mh_filter) < 0)
+		pr_info("%s: can't remove rawv6 mh filter\n", __func__);
+	xfrm_unregister_type(&mip6_rthdr_type, AF_INET6);
+	xfrm_unregister_type(&mip6_destopt_type, AF_INET6);
+}
+
+module_init(mip6_init);
+module_exit(mip6_fini);
+
+MODULE_DESCRIPTION("IPv6 Mobility driver");
+MODULE_LICENSE("GPL");
+MODULE_ALIAS_XFRM_TYPE(AF_INET6, XFRM_PROTO_DSTOPTS);
+MODULE_ALIAS_XFRM_TYPE(AF_INET6, XFRM_PROTO_ROUTING);