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, 210 insertions, 0 deletions

diff --git a/drivers/powercap/dtpm_devfreq.c b/drivers/powercap/dtpm_devfreq.c
new file mode 100644
index 000000000..fa71285de
--- /dev/null
+++ b/drivers/powercap/dtpm_devfreq.c
@@ -0,0 +1,210 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ * Copyright 2021 Linaro Limited
+ *
+ * Author: Daniel Lezcano <daniel.lezcano@linaro.org>
+ *
+ * The devfreq device combined with the energy model and the load can
+ * give an estimation of the power consumption as well as limiting the
+ * power.
+ *
+ */
+#define pr_fmt(fmt) KBUILD_MODNAME ": " fmt
+
+#include <linux/cpumask.h>
+#include <linux/devfreq.h>
+#include <linux/dtpm.h>
+#include <linux/energy_model.h>
+#include <linux/of.h>
+#include <linux/pm_qos.h>
+#include <linux/slab.h>
+#include <linux/units.h>
+
+struct dtpm_devfreq {
+	struct dtpm dtpm;
+	struct dev_pm_qos_request qos_req;
+	struct devfreq *devfreq;
+};
+
+static struct dtpm_devfreq *to_dtpm_devfreq(struct dtpm *dtpm)
+{
+	return container_of(dtpm, struct dtpm_devfreq, dtpm);
+}
+
+static int update_pd_power_uw(struct dtpm *dtpm)
+{
+	struct dtpm_devfreq *dtpm_devfreq = to_dtpm_devfreq(dtpm);
+	struct devfreq *devfreq = dtpm_devfreq->devfreq;
+	struct device *dev = devfreq->dev.parent;
+	struct em_perf_domain *pd = em_pd_get(dev);
+	struct em_perf_state *table;
+
+	rcu_read_lock();
+	table = em_perf_state_from_pd(pd);
+
+	dtpm->power_min = table[0].power;
+
+	dtpm->power_max = table[pd->nr_perf_states - 1].power;
+
+	rcu_read_unlock();
+	return 0;
+}
+
+static u64 set_pd_power_limit(struct dtpm *dtpm, u64 power_limit)
+{
+	struct dtpm_devfreq *dtpm_devfreq = to_dtpm_devfreq(dtpm);
+	struct devfreq *devfreq = dtpm_devfreq->devfreq;
+	struct device *dev = devfreq->dev.parent;
+	struct em_perf_domain *pd = em_pd_get(dev);
+	struct em_perf_state *table;
+	unsigned long freq;
+	int i;
+
+	rcu_read_lock();
+	table = em_perf_state_from_pd(pd);
+	for (i = 0; i < pd->nr_perf_states; i++) {
+		if (table[i].power > power_limit)
+			break;
+	}
+
+	freq = table[i - 1].frequency;
+	power_limit = table[i - 1].power;
+	rcu_read_unlock();
+
+	dev_pm_qos_update_request(&dtpm_devfreq->qos_req, freq);
+
+	return power_limit;
+}
+
+static void _normalize_load(struct devfreq_dev_status *status)
+{
+	if (status->total_time > 0xfffff) {
+		status->total_time >>= 10;
+		status->busy_time >>= 10;
+	}
+
+	status->busy_time <<= 10;
+	status->busy_time /= status->total_time ? : 1;
+
+	status->busy_time = status->busy_time ? : 1;
+	status->total_time = 1024;
+}
+
+static u64 get_pd_power_uw(struct dtpm *dtpm)
+{
+	struct dtpm_devfreq *dtpm_devfreq = to_dtpm_devfreq(dtpm);
+	struct devfreq *devfreq = dtpm_devfreq->devfreq;
+	struct device *dev = devfreq->dev.parent;
+	struct em_perf_domain *pd = em_pd_get(dev);
+	struct devfreq_dev_status status;
+	struct em_perf_state *table;
+	unsigned long freq;
+	u64 power = 0;
+	int i;
+
+	mutex_lock(&devfreq->lock);
+	status = devfreq->last_status;
+	mutex_unlock(&devfreq->lock);
+
+	freq = DIV_ROUND_UP(status.current_frequency, HZ_PER_KHZ);
+	_normalize_load(&status);
+
+	rcu_read_lock();
+	table = em_perf_state_from_pd(pd);
+	for (i = 0; i < pd->nr_perf_states; i++) {
+
+		if (table[i].frequency < freq)
+			continue;
+
+		power = table[i].power;
+		power *= status.busy_time;
+		power >>= 10;
+
+		break;
+	}
+	rcu_read_unlock();
+
+	return power;
+}
+
+static void pd_release(struct dtpm *dtpm)
+{
+	struct dtpm_devfreq *dtpm_devfreq = to_dtpm_devfreq(dtpm);
+
+	if (dev_pm_qos_request_active(&dtpm_devfreq->qos_req))
+		dev_pm_qos_remove_request(&dtpm_devfreq->qos_req);
+
+	kfree(dtpm_devfreq);
+}
+
+static struct dtpm_ops dtpm_ops = {
+	.set_power_uw = set_pd_power_limit,
+	.get_power_uw = get_pd_power_uw,
+	.update_power_uw = update_pd_power_uw,
+	.release = pd_release,
+};
+
+static int __dtpm_devfreq_setup(struct devfreq *devfreq, struct dtpm *parent)
+{
+	struct device *dev = devfreq->dev.parent;
+	struct dtpm_devfreq *dtpm_devfreq;
+	struct em_perf_domain *pd;
+	int ret = -ENOMEM;
+
+	pd = em_pd_get(dev);
+	if (!pd) {
+		ret = dev_pm_opp_of_register_em(dev, NULL);
+		if (ret) {
+			pr_err("No energy model available for '%s'\n", dev_name(dev));
+			return -EINVAL;
+		}
+	}
+
+	dtpm_devfreq = kzalloc_obj(*dtpm_devfreq);
+	if (!dtpm_devfreq)
+		return -ENOMEM;
+
+	dtpm_init(&dtpm_devfreq->dtpm, &dtpm_ops);
+
+	dtpm_devfreq->devfreq = devfreq;
+
+	ret = dtpm_register(dev_name(dev), &dtpm_devfreq->dtpm, parent);
+	if (ret) {
+		pr_err("Failed to register '%s': %d\n", dev_name(dev), ret);
+		kfree(dtpm_devfreq);
+		return ret;
+	}
+
+	ret = dev_pm_qos_add_request(dev, &dtpm_devfreq->qos_req,
+				     DEV_PM_QOS_MAX_FREQUENCY,
+				     PM_QOS_MAX_FREQUENCY_DEFAULT_VALUE);
+	if (ret < 0) {
+		pr_err("Failed to add QoS request: %d\n", ret);
+		goto out_dtpm_unregister;
+	}
+
+	dtpm_update_power(&dtpm_devfreq->dtpm);
+
+	return 0;
+
+out_dtpm_unregister:
+	dtpm_unregister(&dtpm_devfreq->dtpm);
+
+	return ret;
+}
+
+static int dtpm_devfreq_setup(struct dtpm *dtpm, struct device_node *np)
+{
+	struct devfreq *devfreq;
+
+	devfreq = devfreq_get_devfreq_by_node(np);
+	if (IS_ERR(devfreq))
+		return 0;
+
+	return __dtpm_devfreq_setup(devfreq, dtpm);
+}
+
+struct dtpm_subsys_ops dtpm_devfreq_ops = {
+	.name = KBUILD_MODNAME,
+	.setup = dtpm_devfreq_setup,
+};