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, 166 insertions, 0 deletions

diff --git a/drivers/acpi/thermal_lib.c b/drivers/acpi/thermal_lib.c
new file mode 100644
index 000000000..f81591927
--- /dev/null
+++ b/drivers/acpi/thermal_lib.c
@@ -0,0 +1,166 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * Copyright 2023 Linaro Limited
+ * Copyright 2023 Intel Corporation
+ *
+ * Library routines for retrieving trip point temperature values from the
+ * platform firmware via ACPI.
+ */
+#include <linux/acpi.h>
+#include <linux/units.h>
+#include <linux/thermal.h>
+#include "internal.h"
+
+/*
+ * Minimum temperature for full military grade is 218°K (-55°C) and
+ * max temperature is 448°K (175°C). We can consider those values as
+ * the boundaries for the [trips] temperature returned by the
+ * firmware. Any values out of these boundaries may be considered
+ * bogus and we can assume the firmware has no data to provide.
+ */
+#define TEMP_MIN_DECIK	2180ULL
+#define TEMP_MAX_DECIK	4480ULL
+
+static int acpi_trip_temp(struct acpi_device *adev, char *obj_name,
+			  int *ret_temp)
+{
+	unsigned long long temp;
+	acpi_status status;
+
+	status = acpi_evaluate_integer(adev->handle, obj_name, NULL, &temp);
+	if (ACPI_FAILURE(status)) {
+		acpi_handle_debug(adev->handle, "%s evaluation failed\n", obj_name);
+		return -ENODATA;
+	}
+
+	if (temp >= TEMP_MIN_DECIK && temp <= TEMP_MAX_DECIK) {
+		*ret_temp = temp;
+	} else {
+		acpi_handle_debug(adev->handle, "%s result %llu out of range\n",
+				  obj_name, temp);
+		*ret_temp = THERMAL_TEMP_INVALID;
+	}
+
+	return 0;
+}
+
+int acpi_active_trip_temp(struct acpi_device *adev, int id, int *ret_temp)
+{
+	char obj_name[] = {'_', 'A', 'C', '0' + id, '\0'};
+
+	if (id < 0 || id > 9)
+		return -EINVAL;
+
+	return acpi_trip_temp(adev, obj_name, ret_temp);
+}
+EXPORT_SYMBOL_NS_GPL(acpi_active_trip_temp, "ACPI_THERMAL");
+
+int acpi_passive_trip_temp(struct acpi_device *adev, int *ret_temp)
+{
+	return acpi_trip_temp(adev, "_PSV", ret_temp);
+}
+EXPORT_SYMBOL_NS_GPL(acpi_passive_trip_temp, "ACPI_THERMAL");
+
+int acpi_hot_trip_temp(struct acpi_device *adev, int *ret_temp)
+{
+	return acpi_trip_temp(adev, "_HOT", ret_temp);
+}
+EXPORT_SYMBOL_NS_GPL(acpi_hot_trip_temp, "ACPI_THERMAL");
+
+int acpi_critical_trip_temp(struct acpi_device *adev, int *ret_temp)
+{
+	return acpi_trip_temp(adev, "_CRT", ret_temp);
+}
+EXPORT_SYMBOL_NS_GPL(acpi_critical_trip_temp, "ACPI_THERMAL");
+
+static int thermal_temp(int error, int temp_decik, int *ret_temp)
+{
+	if (error)
+		return error;
+
+	if (temp_decik == THERMAL_TEMP_INVALID)
+		*ret_temp = THERMAL_TEMP_INVALID;
+	else
+		*ret_temp = deci_kelvin_to_millicelsius(temp_decik);
+
+	return 0;
+}
+
+/**
+ * thermal_acpi_active_trip_temp - Retrieve active trip point temperature
+ * @adev: Target thermal zone ACPI device object.
+ * @id: Active cooling level (0 - 9).
+ * @ret_temp: Address to store the retrieved temperature value on success.
+ *
+ * Evaluate the _ACx object for the thermal zone represented by @adev to obtain
+ * the temperature of the active cooling trip point corresponding to the active
+ * cooling level given by @id.
+ *
+ * Return 0 on success or a negative error value on failure.
+ */
+int thermal_acpi_active_trip_temp(struct acpi_device *adev, int id, int *ret_temp)
+{
+	int temp_decik = 0;
+	int ret = acpi_active_trip_temp(adev, id, &temp_decik);
+
+	return thermal_temp(ret, temp_decik, ret_temp);
+}
+EXPORT_SYMBOL_GPL(thermal_acpi_active_trip_temp);
+
+/**
+ * thermal_acpi_passive_trip_temp - Retrieve passive trip point temperature
+ * @adev: Target thermal zone ACPI device object.
+ * @ret_temp: Address to store the retrieved temperature value on success.
+ *
+ * Evaluate the _PSV object for the thermal zone represented by @adev to obtain
+ * the temperature of the passive cooling trip point.
+ *
+ * Return 0 on success or -ENODATA on failure.
+ */
+int thermal_acpi_passive_trip_temp(struct acpi_device *adev, int *ret_temp)
+{
+	int temp_decik = 0;
+	int ret = acpi_passive_trip_temp(adev, &temp_decik);
+
+	return thermal_temp(ret, temp_decik, ret_temp);
+}
+EXPORT_SYMBOL_GPL(thermal_acpi_passive_trip_temp);
+
+/**
+ * thermal_acpi_hot_trip_temp - Retrieve hot trip point temperature
+ * @adev: Target thermal zone ACPI device object.
+ * @ret_temp: Address to store the retrieved temperature value on success.
+ *
+ * Evaluate the _HOT object for the thermal zone represented by @adev to obtain
+ * the temperature of the trip point at which the system is expected to be put
+ * into the S4 sleep state.
+ *
+ * Return 0 on success or -ENODATA on failure.
+ */
+int thermal_acpi_hot_trip_temp(struct acpi_device *adev, int *ret_temp)
+{
+	int temp_decik = 0;
+	int ret = acpi_hot_trip_temp(adev, &temp_decik);
+
+	return thermal_temp(ret, temp_decik, ret_temp);
+}
+EXPORT_SYMBOL_GPL(thermal_acpi_hot_trip_temp);
+
+/**
+ * thermal_acpi_critical_trip_temp - Retrieve critical trip point temperature
+ * @adev: Target thermal zone ACPI device object.
+ * @ret_temp: Address to store the retrieved temperature value on success.
+ *
+ * Evaluate the _CRT object for the thermal zone represented by @adev to obtain
+ * the temperature of the critical cooling trip point.
+ *
+ * Return 0 on success or -ENODATA on failure.
+ */
+int thermal_acpi_critical_trip_temp(struct acpi_device *adev, int *ret_temp)
+{
+	int temp_decik = 0;
+	int ret = acpi_critical_trip_temp(adev, &temp_decik);
+
+	return thermal_temp(ret, temp_decik, ret_temp);
+}
+EXPORT_SYMBOL_GPL(thermal_acpi_critical_trip_temp);