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, 195 insertions, 0 deletions

diff --git a/net/ethtool/privflags.c b/net/ethtool/privflags.c
new file mode 100644
index 000000000..46a4d2a43
--- /dev/null
+++ b/net/ethtool/privflags.c
@@ -0,0 +1,195 @@
+// SPDX-License-Identifier: GPL-2.0-only
+
+#include "bitset.h"
+#include "common.h"
+#include "netlink.h"
+
+struct privflags_req_info {
+	struct ethnl_req_info		base;
+};
+
+struct privflags_reply_data {
+	struct ethnl_reply_data		base;
+	const char			(*priv_flag_names)[ETH_GSTRING_LEN];
+	unsigned int			n_priv_flags;
+	u32				priv_flags;
+};
+
+#define PRIVFLAGS_REPDATA(__reply_base) \
+	container_of(__reply_base, struct privflags_reply_data, base)
+
+const struct nla_policy ethnl_privflags_get_policy[] = {
+	[ETHTOOL_A_PRIVFLAGS_HEADER]		=
+		NLA_POLICY_NESTED(ethnl_header_policy),
+};
+
+static int ethnl_get_priv_flags_info(struct net_device *dev,
+				     unsigned int *count,
+				     const char (**names)[ETH_GSTRING_LEN])
+{
+	const struct ethtool_ops *ops = dev->ethtool_ops;
+	int nflags;
+
+	nflags = ops->get_sset_count(dev, ETH_SS_PRIV_FLAGS);
+	if (nflags < 0)
+		return nflags;
+
+	if (names) {
+		*names = kcalloc(nflags, ETH_GSTRING_LEN, GFP_KERNEL);
+		if (!*names)
+			return -ENOMEM;
+		ops->get_strings(dev, ETH_SS_PRIV_FLAGS, (u8 *)*names);
+	}
+
+	/* We can pass more than 32 private flags to userspace via netlink but
+	 * we cannot get more with ethtool_ops::get_priv_flags(). Note that we
+	 * must not adjust nflags before allocating the space for flag names
+	 * as the buffer must be large enough for all flags.
+	 */
+	if (WARN_ONCE(nflags > 32,
+		      "device %s reports more than 32 private flags (%d)\n",
+		      netdev_name(dev), nflags))
+		nflags = 32;
+	*count = nflags;
+
+	return 0;
+}
+
+static int privflags_prepare_data(const struct ethnl_req_info *req_base,
+				  struct ethnl_reply_data *reply_base,
+				  const struct genl_info *info)
+{
+	struct privflags_reply_data *data = PRIVFLAGS_REPDATA(reply_base);
+	struct net_device *dev = reply_base->dev;
+	const char (*names)[ETH_GSTRING_LEN];
+	const struct ethtool_ops *ops;
+	unsigned int nflags;
+	int ret;
+
+	ops = dev->ethtool_ops;
+	if (!ops->get_priv_flags || !ops->get_sset_count || !ops->get_strings)
+		return -EOPNOTSUPP;
+	ret = ethnl_ops_begin(dev);
+	if (ret < 0)
+		return ret;
+
+	ret = ethnl_get_priv_flags_info(dev, &nflags, &names);
+	if (ret < 0)
+		goto out_ops;
+	data->priv_flags = ops->get_priv_flags(dev);
+	data->priv_flag_names = names;
+	data->n_priv_flags = nflags;
+
+out_ops:
+	ethnl_ops_complete(dev);
+	return ret;
+}
+
+static int privflags_reply_size(const struct ethnl_req_info *req_base,
+				const struct ethnl_reply_data *reply_base)
+{
+	const struct privflags_reply_data *data = PRIVFLAGS_REPDATA(reply_base);
+	bool compact = req_base->flags & ETHTOOL_FLAG_COMPACT_BITSETS;
+	const u32 all_flags = ~(u32)0 >> (32 - data->n_priv_flags);
+
+	return ethnl_bitset32_size(&data->priv_flags, &all_flags,
+				   data->n_priv_flags,
+				   data->priv_flag_names, compact);
+}
+
+static int privflags_fill_reply(struct sk_buff *skb,
+				const struct ethnl_req_info *req_base,
+				const struct ethnl_reply_data *reply_base)
+{
+	const struct privflags_reply_data *data = PRIVFLAGS_REPDATA(reply_base);
+	bool compact = req_base->flags & ETHTOOL_FLAG_COMPACT_BITSETS;
+	const u32 all_flags = ~(u32)0 >> (32 - data->n_priv_flags);
+
+	return ethnl_put_bitset32(skb, ETHTOOL_A_PRIVFLAGS_FLAGS,
+				  &data->priv_flags, &all_flags,
+				  data->n_priv_flags, data->priv_flag_names,
+				  compact);
+}
+
+static void privflags_cleanup_data(struct ethnl_reply_data *reply_data)
+{
+	struct privflags_reply_data *data = PRIVFLAGS_REPDATA(reply_data);
+
+	kfree(data->priv_flag_names);
+}
+
+/* PRIVFLAGS_SET */
+
+const struct nla_policy ethnl_privflags_set_policy[] = {
+	[ETHTOOL_A_PRIVFLAGS_HEADER]		=
+		NLA_POLICY_NESTED(ethnl_header_policy),
+	[ETHTOOL_A_PRIVFLAGS_FLAGS]		= { .type = NLA_NESTED },
+};
+
+static int
+ethnl_set_privflags_validate(struct ethnl_req_info *req_info,
+			     struct genl_info *info)
+{
+	const struct ethtool_ops *ops = req_info->dev->ethtool_ops;
+
+	if (!info->attrs[ETHTOOL_A_PRIVFLAGS_FLAGS])
+		return -EINVAL;
+
+	if (!ops->get_priv_flags || !ops->set_priv_flags ||
+	    !ops->get_sset_count || !ops->get_strings)
+		return -EOPNOTSUPP;
+	return 1;
+}
+
+static int
+ethnl_set_privflags(struct ethnl_req_info *req_info, struct genl_info *info)
+{
+	const char (*names)[ETH_GSTRING_LEN] = NULL;
+	struct net_device *dev = req_info->dev;
+	struct nlattr **tb = info->attrs;
+	unsigned int nflags;
+	bool mod = false;
+	bool compact;
+	u32 flags;
+	int ret;
+
+	ret = ethnl_bitset_is_compact(tb[ETHTOOL_A_PRIVFLAGS_FLAGS], &compact);
+	if (ret < 0)
+		return ret;
+
+	ret = ethnl_get_priv_flags_info(dev, &nflags, compact ? NULL : &names);
+	if (ret < 0)
+		return ret;
+	flags = dev->ethtool_ops->get_priv_flags(dev);
+
+	ret = ethnl_update_bitset32(&flags, nflags,
+				    tb[ETHTOOL_A_PRIVFLAGS_FLAGS], names,
+				    info->extack, &mod);
+	if (ret < 0 || !mod)
+		goto out_free;
+	ret = dev->ethtool_ops->set_priv_flags(dev, flags);
+	if (ret < 0)
+		goto out_free;
+	ret = 1;
+
+out_free:
+	kfree(names);
+	return ret;
+}
+
+const struct ethnl_request_ops ethnl_privflags_request_ops = {
+	.request_cmd		= ETHTOOL_MSG_PRIVFLAGS_GET,
+	.reply_cmd		= ETHTOOL_MSG_PRIVFLAGS_GET_REPLY,
+	.hdr_attr		= ETHTOOL_A_PRIVFLAGS_HEADER,
+	.req_info_size		= sizeof(struct privflags_req_info),
+	.reply_data_size	= sizeof(struct privflags_reply_data),
+
+	.prepare_data		= privflags_prepare_data,
+	.reply_size		= privflags_reply_size,
+	.fill_reply		= privflags_fill_reply,
+	.cleanup_data		= privflags_cleanup_data,
+
+	.set_validate		= ethnl_set_privflags_validate,
+	.set			= ethnl_set_privflags,
+	.set_ntf_cmd		= ETHTOOL_MSG_PRIVFLAGS_NTF,
+};