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, 258 insertions, 0 deletions

diff --git a/drivers/mfd/max8925-i2c.c b/drivers/mfd/max8925-i2c.c
new file mode 100644
index 000000000..ab19ff0c7
--- /dev/null
+++ b/drivers/mfd/max8925-i2c.c
@@ -0,0 +1,258 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ * I2C driver for Maxim MAX8925
+ *
+ * Copyright (C) 2009 Marvell International Ltd.
+ *	Haojian Zhuang <haojian.zhuang@marvell.com>
+ */
+#include <linux/kernel.h>
+#include <linux/init.h>
+#include <linux/platform_device.h>
+#include <linux/i2c.h>
+#include <linux/mfd/max8925.h>
+#include <linux/slab.h>
+
+#define RTC_I2C_ADDR		0x68
+#define ADC_I2C_ADDR		0x47
+
+static inline int max8925_read_device(struct i2c_client *i2c,
+				      int reg, int bytes, void *dest)
+{
+	int ret;
+
+	if (bytes > 1)
+		ret = i2c_smbus_read_i2c_block_data(i2c, reg, bytes, dest);
+	else {
+		ret = i2c_smbus_read_byte_data(i2c, reg);
+		if (ret < 0)
+			return ret;
+		*(unsigned char *)dest = (unsigned char)ret;
+	}
+	return ret;
+}
+
+static inline int max8925_write_device(struct i2c_client *i2c,
+				       int reg, int bytes, void *src)
+{
+	unsigned char buf[9];
+	int ret;
+
+	buf[0] = (unsigned char)reg;
+	memcpy(&buf[1], src, bytes);
+
+	ret = i2c_master_send(i2c, buf, bytes + 1);
+	if (ret < 0)
+		return ret;
+	return 0;
+}
+
+int max8925_reg_read(struct i2c_client *i2c, int reg)
+{
+	struct max8925_chip *chip = i2c_get_clientdata(i2c);
+	unsigned char data = 0;
+	int ret;
+
+	mutex_lock(&chip->io_lock);
+	ret = max8925_read_device(i2c, reg, 1, &data);
+	mutex_unlock(&chip->io_lock);
+
+	if (ret < 0)
+		return ret;
+	else
+		return (int)data;
+}
+EXPORT_SYMBOL(max8925_reg_read);
+
+int max8925_reg_write(struct i2c_client *i2c, int reg,
+		unsigned char data)
+{
+	struct max8925_chip *chip = i2c_get_clientdata(i2c);
+	int ret;
+
+	mutex_lock(&chip->io_lock);
+	ret = max8925_write_device(i2c, reg, 1, &data);
+	mutex_unlock(&chip->io_lock);
+
+	return ret;
+}
+EXPORT_SYMBOL(max8925_reg_write);
+
+int max8925_bulk_read(struct i2c_client *i2c, int reg,
+		int count, unsigned char *buf)
+{
+	struct max8925_chip *chip = i2c_get_clientdata(i2c);
+	int ret;
+
+	mutex_lock(&chip->io_lock);
+	ret = max8925_read_device(i2c, reg, count, buf);
+	mutex_unlock(&chip->io_lock);
+
+	return ret;
+}
+EXPORT_SYMBOL(max8925_bulk_read);
+
+int max8925_bulk_write(struct i2c_client *i2c, int reg,
+		int count, unsigned char *buf)
+{
+	struct max8925_chip *chip = i2c_get_clientdata(i2c);
+	int ret;
+
+	mutex_lock(&chip->io_lock);
+	ret = max8925_write_device(i2c, reg, count, buf);
+	mutex_unlock(&chip->io_lock);
+
+	return ret;
+}
+EXPORT_SYMBOL(max8925_bulk_write);
+
+int max8925_set_bits(struct i2c_client *i2c, int reg,
+		unsigned char mask, unsigned char data)
+{
+	struct max8925_chip *chip = i2c_get_clientdata(i2c);
+	unsigned char value;
+	int ret;
+
+	mutex_lock(&chip->io_lock);
+	ret = max8925_read_device(i2c, reg, 1, &value);
+	if (ret < 0)
+		goto out;
+	value &= ~mask;
+	value |= data;
+	ret = max8925_write_device(i2c, reg, 1, &value);
+out:
+	mutex_unlock(&chip->io_lock);
+	return ret;
+}
+EXPORT_SYMBOL(max8925_set_bits);
+
+
+static const struct i2c_device_id max8925_id_table[] = {
+	{ "max8925" },
+	{ }
+};
+
+static int max8925_dt_init(struct device_node *np, struct device *dev,
+			   struct max8925_platform_data *pdata)
+{
+	int ret;
+
+	ret = of_property_read_u32(np, "maxim,tsc-irq", &pdata->tsc_irq);
+	if (ret) {
+		dev_err(dev, "Not found maxim,tsc-irq property\n");
+		return -EINVAL;
+	}
+	return 0;
+}
+
+static int max8925_probe(struct i2c_client *client)
+{
+	struct max8925_platform_data *pdata = dev_get_platdata(&client->dev);
+	struct max8925_chip *chip;
+	struct device_node *node = client->dev.of_node;
+
+	if (node && !pdata) {
+		/* parse DT to get platform data */
+		pdata = devm_kzalloc(&client->dev,
+				     sizeof(struct max8925_platform_data),
+				     GFP_KERNEL);
+		if (!pdata)
+			return -ENOMEM;
+
+		if (max8925_dt_init(node, &client->dev, pdata))
+			return -EINVAL;
+	} else if (!pdata) {
+		pr_info("%s: platform data is missing\n", __func__);
+		return -EINVAL;
+	}
+
+	chip = devm_kzalloc(&client->dev,
+			    sizeof(struct max8925_chip), GFP_KERNEL);
+	if (chip == NULL)
+		return -ENOMEM;
+	chip->i2c = client;
+	chip->dev = &client->dev;
+	i2c_set_clientdata(client, chip);
+	mutex_init(&chip->io_lock);
+
+	chip->rtc = i2c_new_dummy_device(chip->i2c->adapter, RTC_I2C_ADDR);
+	if (IS_ERR(chip->rtc)) {
+		dev_err(chip->dev, "Failed to allocate I2C device for RTC\n");
+		return PTR_ERR(chip->rtc);
+	}
+	i2c_set_clientdata(chip->rtc, chip);
+
+	chip->adc = i2c_new_dummy_device(chip->i2c->adapter, ADC_I2C_ADDR);
+	if (IS_ERR(chip->adc)) {
+		dev_err(chip->dev, "Failed to allocate I2C device for ADC\n");
+		i2c_unregister_device(chip->rtc);
+		return PTR_ERR(chip->adc);
+	}
+	i2c_set_clientdata(chip->adc, chip);
+
+	device_init_wakeup(&client->dev, 1);
+
+	max8925_device_init(chip, pdata);
+
+	return 0;
+}
+
+static void max8925_remove(struct i2c_client *client)
+{
+	struct max8925_chip *chip = i2c_get_clientdata(client);
+
+	max8925_device_exit(chip);
+	device_init_wakeup(&client->dev, false);
+	i2c_unregister_device(chip->adc);
+	i2c_unregister_device(chip->rtc);
+}
+
+static int max8925_suspend(struct device *dev)
+{
+	struct i2c_client *client = to_i2c_client(dev);
+	struct max8925_chip *chip = i2c_get_clientdata(client);
+
+	if (device_may_wakeup(dev) && chip->wakeup_flag)
+		enable_irq_wake(chip->core_irq);
+	return 0;
+}
+
+static int max8925_resume(struct device *dev)
+{
+	struct i2c_client *client = to_i2c_client(dev);
+	struct max8925_chip *chip = i2c_get_clientdata(client);
+
+	if (device_may_wakeup(dev) && chip->wakeup_flag)
+		disable_irq_wake(chip->core_irq);
+	return 0;
+}
+
+static DEFINE_SIMPLE_DEV_PM_OPS(max8925_pm_ops,
+				max8925_suspend, max8925_resume);
+
+static const struct of_device_id max8925_dt_ids[] = {
+	{ .compatible = "maxim,max8925", },
+	{},
+};
+
+static struct i2c_driver max8925_driver = {
+	.driver	= {
+		.name	= "max8925",
+		.pm     = pm_sleep_ptr(&max8925_pm_ops),
+		.of_match_table = max8925_dt_ids,
+	},
+	.probe		= max8925_probe,
+	.remove		= max8925_remove,
+	.id_table	= max8925_id_table,
+};
+
+static int __init max8925_i2c_init(void)
+{
+	int ret;
+
+	ret = i2c_add_driver(&max8925_driver);
+	if (ret != 0)
+		pr_err("Failed to register MAX8925 I2C driver: %d\n", ret);
+
+	return ret;
+}
+subsys_initcall(max8925_i2c_init);