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, 141 insertions, 0 deletions

diff --git a/net/wireless/wext-proc.c b/net/wireless/wext-proc.c
new file mode 100644
index 000000000..be6b2b695
--- /dev/null
+++ b/net/wireless/wext-proc.c
@@ -0,0 +1,141 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * This file implement the Wireless Extensions proc API.
+ *
+ * Authors :	Jean Tourrilhes - HPL - <jt@hpl.hp.com>
+ * Copyright (c) 1997-2007 Jean Tourrilhes, All Rights Reserved.
+ */
+
+/*
+ * The /proc/net/wireless file is a human readable user-space interface
+ * exporting various wireless specific statistics from the wireless devices.
+ * This is the most popular part of the Wireless Extensions ;-)
+ *
+ * This interface is a pure clone of /proc/net/dev (in net/core/dev.c).
+ * The content of the file is basically the content of "struct iw_statistics".
+ */
+
+#include <linux/module.h>
+#include <linux/proc_fs.h>
+#include <linux/seq_file.h>
+#include <linux/wireless.h>
+#include <linux/netdevice.h>
+#include <linux/rtnetlink.h>
+#include <net/iw_handler.h>
+#include <net/wext.h>
+
+
+static void wireless_seq_printf_stats(struct seq_file *seq,
+				      struct net_device *dev)
+{
+	/* Get stats from the driver */
+	struct iw_statistics *stats = get_wireless_stats(dev);
+	static struct iw_statistics nullstats = {};
+
+	/* show device if it's wireless regardless of current stats */
+	if (!stats) {
+#ifdef CONFIG_WIRELESS_EXT
+		if (dev->wireless_handlers)
+			stats = &nullstats;
+#endif
+#ifdef CONFIG_CFG80211
+		if (dev->ieee80211_ptr)
+			stats = &nullstats;
+#endif
+	}
+
+	if (stats) {
+		seq_printf(seq, "%6s: %04x  %3d%c  %3d%c  %3d%c  %6d %6d %6d "
+				"%6d %6d   %6d\n",
+			   dev->name, stats->status, stats->qual.qual,
+			   stats->qual.updated & IW_QUAL_QUAL_UPDATED
+			   ? '.' : ' ',
+			   ((__s32) stats->qual.level) -
+			   ((stats->qual.updated & IW_QUAL_DBM) ? 0x100 : 0),
+			   stats->qual.updated & IW_QUAL_LEVEL_UPDATED
+			   ? '.' : ' ',
+			   ((__s32) stats->qual.noise) -
+			   ((stats->qual.updated & IW_QUAL_DBM) ? 0x100 : 0),
+			   stats->qual.updated & IW_QUAL_NOISE_UPDATED
+			   ? '.' : ' ',
+			   stats->discard.nwid, stats->discard.code,
+			   stats->discard.fragment, stats->discard.retries,
+			   stats->discard.misc, stats->miss.beacon);
+
+		if (stats != &nullstats)
+			stats->qual.updated &= ~IW_QUAL_ALL_UPDATED;
+	}
+}
+
+/* ---------------------------------------------------------------- */
+/*
+ * Print info for /proc/net/wireless (print all entries)
+ */
+static int wireless_dev_seq_show(struct seq_file *seq, void *v)
+{
+	might_sleep();
+
+	if (v == SEQ_START_TOKEN)
+		seq_printf(seq, "Inter-| sta-|   Quality        |   Discarded "
+				"packets               | Missed | WE\n"
+				" face | tus | link level noise |  nwid  "
+				"crypt   frag  retry   misc | beacon | %d\n",
+			   WIRELESS_EXT);
+	else
+		wireless_seq_printf_stats(seq, v);
+	return 0;
+}
+
+static void *wireless_dev_seq_start(struct seq_file *seq, loff_t *pos)
+{
+	struct net *net = seq_file_net(seq);
+	loff_t off;
+	struct net_device *dev;
+
+	rtnl_lock();
+	if (!*pos)
+		return SEQ_START_TOKEN;
+
+	off = 1;
+	for_each_netdev(net, dev)
+		if (off++ == *pos)
+			return dev;
+	return NULL;
+}
+
+static void *wireless_dev_seq_next(struct seq_file *seq, void *v, loff_t *pos)
+{
+	struct net *net = seq_file_net(seq);
+
+	++*pos;
+
+	return v == SEQ_START_TOKEN ?
+		first_net_device(net) : next_net_device(v);
+}
+
+static void wireless_dev_seq_stop(struct seq_file *seq, void *v)
+{
+	rtnl_unlock();
+}
+
+static const struct seq_operations wireless_seq_ops = {
+	.start = wireless_dev_seq_start,
+	.next  = wireless_dev_seq_next,
+	.stop  = wireless_dev_seq_stop,
+	.show  = wireless_dev_seq_show,
+};
+
+int __net_init wext_proc_init(struct net *net)
+{
+	/* Create /proc/net/wireless entry */
+	if (!proc_create_net("wireless", 0444, net->proc_net,
+			&wireless_seq_ops, sizeof(struct seq_net_private)))
+		return -ENOMEM;
+
+	return 0;
+}
+
+void __net_exit wext_proc_exit(struct net *net)
+{
+	remove_proc_entry("wireless", net->proc_net);
+}