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, 159 insertions, 0 deletions

diff --git a/drivers/mfd/88pm80x.c b/drivers/mfd/88pm80x.c
new file mode 100644
index 000000000..bbc1a87f0
--- /dev/null
+++ b/drivers/mfd/88pm80x.c
@@ -0,0 +1,159 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ * I2C driver for Marvell 88PM80x
+ *
+ * Copyright (C) 2012 Marvell International Ltd.
+ * Haojian Zhuang <haojian.zhuang@marvell.com>
+ * Joseph(Yossi) Hanin <yhanin@marvell.com>
+ * Qiao Zhou <zhouqiao@marvell.com>
+ */
+#include <linux/kernel.h>
+#include <linux/module.h>
+#include <linux/i2c.h>
+#include <linux/mfd/88pm80x.h>
+#include <linux/slab.h>
+#include <linux/uaccess.h>
+#include <linux/err.h>
+
+/* 88pm80x chips have same definition for chip id register. */
+#define PM80X_CHIP_ID			(0x00)
+#define PM80X_CHIP_ID_NUM(x)		(((x) >> 5) & 0x7)
+#define PM80X_CHIP_ID_REVISION(x)	((x) & 0x1F)
+
+struct pm80x_chip_mapping {
+	unsigned int	id;
+	int		type;
+};
+
+static struct pm80x_chip_mapping chip_mapping[] = {
+	/* 88PM800 chip id number */
+	{0x3,	CHIP_PM800},
+	/* 88PM805 chip id number */
+	{0x0,	CHIP_PM805},
+	/* 88PM860 chip id number */
+	{0x4,	CHIP_PM860},
+};
+
+/*
+ * workaround: some registers needed by pm805 are defined in pm800, so
+ * need to use this global variable to maintain the relation between
+ * pm800 and pm805. would remove it after HW chip fixes the issue.
+ */
+static struct pm80x_chip *g_pm80x_chip;
+
+const struct regmap_config pm80x_regmap_config = {
+	.reg_bits = 8,
+	.val_bits = 8,
+};
+EXPORT_SYMBOL_GPL(pm80x_regmap_config);
+
+
+int pm80x_init(struct i2c_client *client)
+{
+	struct pm80x_chip *chip;
+	struct regmap *map;
+	unsigned int val;
+	int i, ret = 0;
+
+	chip =
+	    devm_kzalloc(&client->dev, sizeof(struct pm80x_chip), GFP_KERNEL);
+	if (!chip)
+		return -ENOMEM;
+
+	map = devm_regmap_init_i2c(client, &pm80x_regmap_config);
+	if (IS_ERR(map)) {
+		ret = PTR_ERR(map);
+		dev_err(&client->dev, "Failed to allocate register map: %d\n",
+			ret);
+		return ret;
+	}
+
+	chip->client = client;
+	chip->regmap = map;
+
+	chip->irq = client->irq;
+
+	chip->dev = &client->dev;
+	i2c_set_clientdata(chip->client, chip);
+
+	ret = regmap_read(chip->regmap, PM80X_CHIP_ID, &val);
+	if (ret < 0) {
+		dev_err(chip->dev, "Failed to read CHIP ID: %d\n", ret);
+		return ret;
+	}
+
+	for (i = 0; i < ARRAY_SIZE(chip_mapping); i++) {
+		if (chip_mapping[i].id == PM80X_CHIP_ID_NUM(val)) {
+			chip->type = chip_mapping[i].type;
+			break;
+		}
+	}
+
+	if (i == ARRAY_SIZE(chip_mapping)) {
+		dev_err(chip->dev,
+			"Failed to detect Marvell 88PM800:ChipID[0x%x]\n", val);
+		return -EINVAL;
+	}
+
+	device_init_wakeup(&client->dev, 1);
+
+	/*
+	 * workaround: set g_pm80x_chip to the first probed chip. if the
+	 * second chip is probed, just point to the companion to each
+	 * other so that pm805 can access those specific register. would
+	 * remove it after HW chip fixes the issue.
+	 */
+	if (!g_pm80x_chip)
+		g_pm80x_chip = chip;
+	else {
+		chip->companion = g_pm80x_chip->client;
+		g_pm80x_chip->companion = chip->client;
+	}
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(pm80x_init);
+
+int pm80x_deinit(void)
+{
+	/*
+	 * workaround: clear the dependency between pm800 and pm805.
+	 * would remove it after HW chip fixes the issue.
+	 */
+	if (g_pm80x_chip->companion)
+		g_pm80x_chip->companion = NULL;
+	else
+		g_pm80x_chip = NULL;
+	return 0;
+}
+EXPORT_SYMBOL_GPL(pm80x_deinit);
+
+static int pm80x_suspend(struct device *dev)
+{
+	struct i2c_client *client = to_i2c_client(dev);
+	struct pm80x_chip *chip = i2c_get_clientdata(client);
+
+	if (chip && chip->wu_flag)
+		if (device_may_wakeup(chip->dev))
+			enable_irq_wake(chip->irq);
+
+	return 0;
+}
+
+static int pm80x_resume(struct device *dev)
+{
+	struct i2c_client *client = to_i2c_client(dev);
+	struct pm80x_chip *chip = i2c_get_clientdata(client);
+
+	if (chip && chip->wu_flag)
+		if (device_may_wakeup(chip->dev))
+			disable_irq_wake(chip->irq);
+
+	return 0;
+}
+
+EXPORT_GPL_SIMPLE_DEV_PM_OPS(pm80x_pm_ops, pm80x_suspend, pm80x_resume);
+
+MODULE_DESCRIPTION("I2C Driver for Marvell 88PM80x");
+MODULE_AUTHOR("Qiao Zhou <zhouqiao@marvell.com>");
+MODULE_LICENSE("GPL");