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, 290 insertions, 0 deletions

diff --git a/drivers/i2c/i2c-slave-testunit.c b/drivers/i2c/i2c-slave-testunit.c
new file mode 100644
index 000000000..4a1bf1aa5
--- /dev/null
+++ b/drivers/i2c/i2c-slave-testunit.c
@@ -0,0 +1,290 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ * I2C slave mode testunit
+ *
+ * Copyright (C) 2020 by Wolfram Sang, Sang Engineering <wsa@sang-engineering.com>
+ * Copyright (C) 2020 by Renesas Electronics Corporation
+ */
+
+#include <generated/utsrelease.h>
+#include <linux/bitops.h>
+#include <linux/completion.h>
+#include <linux/gpio/consumer.h>
+#include <linux/i2c.h>
+#include <linux/init.h>
+#include <linux/module.h>
+#include <linux/of.h>
+#include <linux/slab.h>
+#include <linux/workqueue.h>
+
+#define TU_VERSION_MAX_LENGTH 128
+
+enum testunit_cmds {
+	TU_CMD_READ_BYTES = 1,	/* save 0 for ABORT, RESET or similar */
+	TU_CMD_SMBUS_HOST_NOTIFY,
+	TU_CMD_SMBUS_BLOCK_PROC_CALL,
+	TU_CMD_GET_VERSION_WITH_REP_START,
+	TU_CMD_SMBUS_ALERT_REQUEST,
+	TU_NUM_CMDS
+};
+
+enum testunit_regs {
+	TU_REG_CMD,
+	TU_REG_DATAL,
+	TU_REG_DATAH,
+	TU_REG_DELAY,
+	TU_NUM_REGS
+};
+
+enum testunit_flags {
+	TU_FLAG_IN_PROCESS,
+	TU_FLAG_NACK,
+};
+
+struct testunit_data {
+	unsigned long flags;
+	u8 regs[TU_NUM_REGS];
+	u8 reg_idx;
+	u8 read_idx;
+	struct i2c_client *client;
+	struct delayed_work worker;
+	struct gpio_desc *gpio;
+	struct completion alert_done;
+};
+
+static char tu_version_info[] = "v" UTS_RELEASE "\n\0";
+
+static int i2c_slave_testunit_smbalert_cb(struct i2c_client *client,
+					  enum i2c_slave_event event, u8 *val)
+{
+	struct testunit_data *tu = i2c_get_clientdata(client);
+
+	switch (event) {
+	case I2C_SLAVE_READ_PROCESSED:
+		gpiod_set_value(tu->gpio, 0);
+		fallthrough;
+	case I2C_SLAVE_READ_REQUESTED:
+		*val = tu->regs[TU_REG_DATAL];
+		break;
+
+	case I2C_SLAVE_STOP:
+		complete(&tu->alert_done);
+		break;
+
+	case I2C_SLAVE_WRITE_REQUESTED:
+	case I2C_SLAVE_WRITE_RECEIVED:
+		return -EOPNOTSUPP;
+	}
+
+	return 0;
+}
+
+static int i2c_slave_testunit_slave_cb(struct i2c_client *client,
+				     enum i2c_slave_event event, u8 *val)
+{
+	struct testunit_data *tu = i2c_get_clientdata(client);
+	bool is_proc_call = tu->reg_idx == 3 && tu->regs[TU_REG_DATAL] == 1 &&
+			    tu->regs[TU_REG_CMD] == TU_CMD_SMBUS_BLOCK_PROC_CALL;
+	bool is_get_version = tu->reg_idx == 3 &&
+			      tu->regs[TU_REG_CMD] == TU_CMD_GET_VERSION_WITH_REP_START;
+	int ret = 0;
+
+	switch (event) {
+	case I2C_SLAVE_WRITE_REQUESTED:
+		if (test_bit(TU_FLAG_IN_PROCESS | TU_FLAG_NACK, &tu->flags)) {
+			ret = -EBUSY;
+			break;
+		}
+
+		memset(tu->regs, 0, TU_NUM_REGS);
+		tu->reg_idx = 0;
+		tu->read_idx = 0;
+		break;
+
+	case I2C_SLAVE_WRITE_RECEIVED:
+		if (test_bit(TU_FLAG_IN_PROCESS | TU_FLAG_NACK, &tu->flags)) {
+			ret = -EBUSY;
+			break;
+		}
+
+		if (tu->reg_idx < TU_NUM_REGS)
+			tu->regs[tu->reg_idx] = *val;
+		else
+			ret = -EMSGSIZE;
+
+		if (tu->reg_idx <= TU_NUM_REGS)
+			tu->reg_idx++;
+
+		/* TU_REG_CMD always written at this point */
+		if (tu->regs[TU_REG_CMD] >= TU_NUM_CMDS)
+			ret = -EINVAL;
+
+		break;
+
+	case I2C_SLAVE_STOP:
+		if (tu->reg_idx == TU_NUM_REGS) {
+			set_bit(TU_FLAG_IN_PROCESS, &tu->flags);
+			queue_delayed_work(system_dfl_long_wq, &tu->worker,
+					   msecs_to_jiffies(10 * tu->regs[TU_REG_DELAY]));
+		}
+
+		/*
+		 * Reset reg_idx to avoid that work gets queued again in case of
+		 * STOP after a following read message. But do not clear TU regs
+		 * here because we still need them in the workqueue!
+		 */
+		tu->reg_idx = 0;
+
+		clear_bit(TU_FLAG_NACK, &tu->flags);
+		break;
+
+	case I2C_SLAVE_READ_PROCESSED:
+		/* Advance until we reach the NUL character */
+		if (is_get_version && tu_version_info[tu->read_idx] != 0)
+			tu->read_idx++;
+		else if (is_proc_call && tu->regs[TU_REG_DATAH])
+			tu->regs[TU_REG_DATAH]--;
+
+		fallthrough;
+
+	case I2C_SLAVE_READ_REQUESTED:
+		if (is_get_version)
+			*val = tu_version_info[tu->read_idx];
+		else if (is_proc_call)
+			*val = tu->regs[TU_REG_DATAH];
+		else
+			*val = test_bit(TU_FLAG_IN_PROCESS, &tu->flags) ?
+					tu->regs[TU_REG_CMD] : 0;
+		break;
+	}
+
+	/* If an error occurred somewhen, we NACK everything until next STOP */
+	if (ret)
+		set_bit(TU_FLAG_NACK, &tu->flags);
+
+	return ret;
+}
+
+static void i2c_slave_testunit_work(struct work_struct *work)
+{
+	struct testunit_data *tu = container_of(work, struct testunit_data, worker.work);
+	unsigned long time_left;
+	struct i2c_msg msg;
+	u8 msgbuf[256];
+	u16 orig_addr;
+	int ret = 0;
+
+	msg.addr = I2C_CLIENT_END;
+	msg.buf = msgbuf;
+
+	switch (tu->regs[TU_REG_CMD]) {
+	case TU_CMD_READ_BYTES:
+		msg.addr = tu->regs[TU_REG_DATAL];
+		msg.flags = I2C_M_RD;
+		msg.len = tu->regs[TU_REG_DATAH];
+		break;
+
+	case TU_CMD_SMBUS_HOST_NOTIFY:
+		msg.addr = 0x08;
+		msg.flags = 0;
+		msg.len = 3;
+		msgbuf[0] = tu->client->addr;
+		msgbuf[1] = tu->regs[TU_REG_DATAL];
+		msgbuf[2] = tu->regs[TU_REG_DATAH];
+		break;
+
+	case TU_CMD_SMBUS_ALERT_REQUEST:
+		if (!tu->gpio) {
+			ret = -ENOENT;
+			break;
+		}
+		i2c_slave_unregister(tu->client);
+		orig_addr = tu->client->addr;
+		tu->client->addr = 0x0c;
+		ret = i2c_slave_register(tu->client, i2c_slave_testunit_smbalert_cb);
+		if (ret)
+			goto out_smbalert;
+
+		reinit_completion(&tu->alert_done);
+		gpiod_set_value(tu->gpio, 1);
+		time_left = wait_for_completion_timeout(&tu->alert_done, HZ);
+		if (!time_left)
+			ret = -ETIMEDOUT;
+
+		i2c_slave_unregister(tu->client);
+out_smbalert:
+		tu->client->addr = orig_addr;
+		i2c_slave_register(tu->client, i2c_slave_testunit_slave_cb);
+		break;
+
+	default:
+		break;
+	}
+
+	if (msg.addr != I2C_CLIENT_END) {
+		ret = i2c_transfer(tu->client->adapter, &msg, 1);
+		/* convert '0 msgs transferred' to errno */
+		ret = (ret == 0) ? -EIO : ret;
+	}
+
+	if (ret < 0)
+		dev_err(&tu->client->dev, "CMD%02X failed (%d)\n", tu->regs[TU_REG_CMD], ret);
+
+	clear_bit(TU_FLAG_IN_PROCESS, &tu->flags);
+}
+
+static int i2c_slave_testunit_probe(struct i2c_client *client)
+{
+	struct testunit_data *tu;
+
+	tu = devm_kzalloc(&client->dev, sizeof(struct testunit_data), GFP_KERNEL);
+	if (!tu)
+		return -ENOMEM;
+
+	tu->client = client;
+	i2c_set_clientdata(client, tu);
+	init_completion(&tu->alert_done);
+	INIT_DELAYED_WORK(&tu->worker, i2c_slave_testunit_work);
+
+	tu->gpio = devm_gpiod_get_index_optional(&client->dev, NULL, 0, GPIOD_OUT_LOW);
+	if (IS_ERR(tu->gpio))
+		return PTR_ERR(tu->gpio);
+
+	if (gpiod_cansleep(tu->gpio)) {
+		dev_err(&client->dev, "GPIO access which may sleep is not allowed\n");
+		return -EDEADLK;
+	}
+
+	if (sizeof(tu_version_info) > TU_VERSION_MAX_LENGTH)
+		tu_version_info[TU_VERSION_MAX_LENGTH - 1] = 0;
+
+	return i2c_slave_register(client, i2c_slave_testunit_slave_cb);
+};
+
+static void i2c_slave_testunit_remove(struct i2c_client *client)
+{
+	struct testunit_data *tu = i2c_get_clientdata(client);
+
+	cancel_delayed_work_sync(&tu->worker);
+	i2c_slave_unregister(client);
+}
+
+static const struct i2c_device_id i2c_slave_testunit_id[] = {
+	{ .name = "slave-testunit" },
+	{ }
+};
+MODULE_DEVICE_TABLE(i2c, i2c_slave_testunit_id);
+
+static struct i2c_driver i2c_slave_testunit_driver = {
+	.driver = {
+		.name = "i2c-slave-testunit",
+	},
+	.probe = i2c_slave_testunit_probe,
+	.remove = i2c_slave_testunit_remove,
+	.id_table = i2c_slave_testunit_id,
+};
+module_i2c_driver(i2c_slave_testunit_driver);
+
+MODULE_AUTHOR("Wolfram Sang <wsa@sang-engineering.com>");
+MODULE_DESCRIPTION("I2C slave mode test unit");
+MODULE_LICENSE("GPL v2");