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, 165 insertions, 0 deletions

diff --git a/drivers/leds/trigger/ledtrig-input-events.c b/drivers/leds/trigger/ledtrig-input-events.c
new file mode 100644
index 000000000..d057b2a23
--- /dev/null
+++ b/drivers/leds/trigger/ledtrig-input-events.c
@@ -0,0 +1,165 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ * Input Events LED trigger
+ *
+ * Copyright (C) 2024 Hans de Goede <hansg@kernel.org>
+ */
+
+#include <linux/input.h>
+#include <linux/jiffies.h>
+#include <linux/leds.h>
+#include <linux/module.h>
+#include <linux/moduleparam.h>
+#include <linux/slab.h>
+#include <linux/spinlock.h>
+#include <linux/workqueue.h>
+#include "../leds.h"
+
+static unsigned long led_off_delay_ms = 5000;
+module_param(led_off_delay_ms, ulong, 0644);
+MODULE_PARM_DESC(led_off_delay_ms,
+	"Specify delay in ms for turning LEDs off after last input event");
+
+static struct input_events_data {
+	struct delayed_work work;
+	spinlock_t lock;
+	/* To avoid repeatedly setting the brightness while there are events */
+	bool led_on;
+	unsigned long led_off_time;
+} input_events_data;
+
+static struct led_trigger *input_events_led_trigger;
+
+static void led_input_events_work(struct work_struct *work)
+{
+	struct input_events_data *data =
+		container_of(work, struct input_events_data, work.work);
+
+	spin_lock_irq(&data->lock);
+
+	/*
+	 * This time_after_eq() check avoids a race where this work starts
+	 * running before a new event pushed led_off_time back.
+	 */
+	if (time_after_eq(jiffies, data->led_off_time)) {
+		led_trigger_event(input_events_led_trigger, LED_OFF);
+		data->led_on = false;
+	}
+
+	spin_unlock_irq(&data->lock);
+}
+
+static void input_events_event(struct input_handle *handle, unsigned int type,
+			       unsigned int code, int val)
+{
+	struct input_events_data *data = &input_events_data;
+	unsigned long led_off_delay = msecs_to_jiffies(led_off_delay_ms);
+	unsigned long flags;
+
+	spin_lock_irqsave(&data->lock, flags);
+
+	if (!data->led_on) {
+		led_trigger_event(input_events_led_trigger, LED_FULL);
+		data->led_on = true;
+	}
+	data->led_off_time = jiffies + led_off_delay;
+
+	spin_unlock_irqrestore(&data->lock, flags);
+
+	mod_delayed_work(system_percpu_wq, &data->work, led_off_delay);
+}
+
+static int input_events_connect(struct input_handler *handler, struct input_dev *dev,
+				const struct input_device_id *id)
+{
+	struct input_handle *handle;
+	int ret;
+
+	handle = kzalloc_obj(*handle);
+	if (!handle)
+		return -ENOMEM;
+
+	handle->dev = dev;
+	handle->handler = handler;
+	handle->name = KBUILD_MODNAME;
+
+	ret = input_register_handle(handle);
+	if (ret)
+		goto err_free_handle;
+
+	ret = input_open_device(handle);
+	if (ret)
+		goto err_unregister_handle;
+
+	return 0;
+
+err_unregister_handle:
+	input_unregister_handle(handle);
+err_free_handle:
+	kfree(handle);
+	return ret;
+}
+
+static void input_events_disconnect(struct input_handle *handle)
+{
+	input_close_device(handle);
+	input_unregister_handle(handle);
+	kfree(handle);
+}
+
+static const struct input_device_id input_events_ids[] = {
+	{
+		.flags = INPUT_DEVICE_ID_MATCH_EVBIT,
+		.evbit = { BIT_MASK(EV_KEY) },
+	},
+	{
+		.flags = INPUT_DEVICE_ID_MATCH_EVBIT,
+		.evbit = { BIT_MASK(EV_REL) },
+	},
+	{
+		.flags = INPUT_DEVICE_ID_MATCH_EVBIT,
+		.evbit = { BIT_MASK(EV_ABS) },
+	},
+	{ }
+};
+
+static struct input_handler input_events_handler = {
+	.name = KBUILD_MODNAME,
+	.event = input_events_event,
+	.connect = input_events_connect,
+	.disconnect = input_events_disconnect,
+	.id_table = input_events_ids,
+};
+
+static int __init input_events_init(void)
+{
+	int ret;
+
+	INIT_DELAYED_WORK(&input_events_data.work, led_input_events_work);
+	spin_lock_init(&input_events_data.lock);
+
+	led_trigger_register_simple("input-events", &input_events_led_trigger);
+
+	ret = input_register_handler(&input_events_handler);
+	if (ret) {
+		led_trigger_unregister_simple(input_events_led_trigger);
+		return ret;
+	}
+
+	return 0;
+}
+
+static void __exit input_events_exit(void)
+{
+	input_unregister_handler(&input_events_handler);
+	cancel_delayed_work_sync(&input_events_data.work);
+	led_trigger_unregister_simple(input_events_led_trigger);
+}
+
+module_init(input_events_init);
+module_exit(input_events_exit);
+
+MODULE_AUTHOR("Hans de Goede <hansg@kernel.org>");
+MODULE_DESCRIPTION("Input Events LED trigger");
+MODULE_LICENSE("GPL");
+MODULE_ALIAS("ledtrig:input-events");