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, 394 insertions, 0 deletions

diff --git a/drivers/base/power/wakeirq.c b/drivers/base/power/wakeirq.c
new file mode 100644
index 000000000..ad23f0fa5
--- /dev/null
+++ b/drivers/base/power/wakeirq.c
@@ -0,0 +1,394 @@
+// SPDX-License-Identifier: GPL-2.0
+/* Device wakeirq helper functions */
+#include <linux/device.h>
+#include <linux/interrupt.h>
+#include <linux/irq.h>
+#include <linux/slab.h>
+#include <linux/pm_runtime.h>
+#include <linux/pm_wakeirq.h>
+
+#include "power.h"
+
+/**
+ * dev_pm_attach_wake_irq - Attach device interrupt as a wake IRQ
+ * @dev: Device entry
+ * @wirq: Wake irq specific data
+ *
+ * Internal function to attach a dedicated wake-up interrupt as a wake IRQ.
+ */
+static int dev_pm_attach_wake_irq(struct device *dev, struct wake_irq *wirq)
+{
+	unsigned long flags;
+
+	if (!dev || !wirq)
+		return -EINVAL;
+
+	spin_lock_irqsave(&dev->power.lock, flags);
+	if (dev_WARN_ONCE(dev, dev->power.wakeirq,
+			  "wake irq already initialized\n")) {
+		spin_unlock_irqrestore(&dev->power.lock, flags);
+		return -EEXIST;
+	}
+
+	dev->power.wakeirq = wirq;
+	device_wakeup_attach_irq(dev, wirq);
+
+	spin_unlock_irqrestore(&dev->power.lock, flags);
+	return 0;
+}
+
+/**
+ * dev_pm_set_wake_irq - Attach device IO interrupt as wake IRQ
+ * @dev: Device entry
+ * @irq: Device IO interrupt
+ *
+ * Attach a device IO interrupt as a wake IRQ. The wake IRQ gets
+ * automatically configured for wake-up from suspend  based
+ * on the device specific sysfs wakeup entry. Typically called
+ * during driver probe after calling device_init_wakeup().
+ */
+int dev_pm_set_wake_irq(struct device *dev, int irq)
+{
+	struct wake_irq *wirq;
+	int err;
+
+	if (irq < 0)
+		return -EINVAL;
+
+	wirq = kzalloc_obj(*wirq);
+	if (!wirq)
+		return -ENOMEM;
+
+	wirq->dev = dev;
+	wirq->irq = irq;
+
+	err = dev_pm_attach_wake_irq(dev, wirq);
+	if (err)
+		kfree(wirq);
+
+	return err;
+}
+EXPORT_SYMBOL_GPL(dev_pm_set_wake_irq);
+
+/**
+ * dev_pm_clear_wake_irq - Detach a device IO interrupt wake IRQ
+ * @dev: Device entry
+ *
+ * Detach a device wake IRQ and free resources.
+ *
+ * Note that it's OK for drivers to call this without calling
+ * dev_pm_set_wake_irq() as all the driver instances may not have
+ * a wake IRQ configured. This avoid adding wake IRQ specific
+ * checks into the drivers.
+ */
+void dev_pm_clear_wake_irq(struct device *dev)
+{
+	struct wake_irq *wirq;
+	unsigned long flags;
+
+	spin_lock_irqsave(&dev->power.lock, flags);
+	wirq = dev->power.wakeirq;
+	if (!wirq) {
+		spin_unlock_irqrestore(&dev->power.lock, flags);
+		return;
+	}
+
+	device_wakeup_detach_irq(dev);
+	dev->power.wakeirq = NULL;
+	spin_unlock_irqrestore(&dev->power.lock, flags);
+
+	if (wirq->status & WAKE_IRQ_DEDICATED_ALLOCATED) {
+		free_irq(wirq->irq, wirq);
+		wirq->status &= ~WAKE_IRQ_DEDICATED_MASK;
+	}
+	kfree(wirq->name);
+	kfree(wirq);
+}
+EXPORT_SYMBOL_GPL(dev_pm_clear_wake_irq);
+
+static void devm_pm_clear_wake_irq(void *dev)
+{
+	dev_pm_clear_wake_irq(dev);
+}
+
+/**
+ * devm_pm_set_wake_irq - device-managed variant of dev_pm_set_wake_irq
+ * @dev: Device entry
+ * @irq: Device IO interrupt
+ *
+ *
+ * Attach a device IO interrupt as a wake IRQ, same with dev_pm_set_wake_irq,
+ * but the device will be auto clear wake capability on driver detach.
+ */
+int devm_pm_set_wake_irq(struct device *dev, int irq)
+{
+	int ret;
+
+	ret = dev_pm_set_wake_irq(dev, irq);
+	if (ret)
+		return ret;
+
+	return devm_add_action_or_reset(dev, devm_pm_clear_wake_irq, dev);
+}
+EXPORT_SYMBOL_GPL(devm_pm_set_wake_irq);
+
+/**
+ * handle_threaded_wake_irq - Handler for dedicated wake-up interrupts
+ * @irq: Device specific dedicated wake-up interrupt
+ * @_wirq: Wake IRQ data
+ *
+ * Some devices have a separate wake-up interrupt in addition to the
+ * device IO interrupt. The wake-up interrupt signals that a device
+ * should be woken up from it's idle state. This handler uses device
+ * specific pm_runtime functions to wake the device, and then it's
+ * up to the device to do whatever it needs to. Note that as the
+ * device may need to restore context and start up regulators, we
+ * use a threaded IRQ.
+ *
+ * Also note that we are not resending the lost device interrupts.
+ * We assume that the wake-up interrupt just needs to wake-up the
+ * device, and then device's pm_runtime_resume() can deal with the
+ * situation.
+ */
+static irqreturn_t handle_threaded_wake_irq(int irq, void *_wirq)
+{
+	struct wake_irq *wirq = _wirq;
+	int res;
+
+	/* Maybe abort suspend? */
+	if (irqd_is_wakeup_set(irq_get_irq_data(irq))) {
+		pm_wakeup_event(wirq->dev, 0);
+
+		return IRQ_HANDLED;
+	}
+
+	/* We don't want RPM_ASYNC or RPM_NOWAIT here */
+	res = pm_runtime_resume(wirq->dev);
+	if (res < 0)
+		dev_warn(wirq->dev,
+			 "wake IRQ with no resume: %i\n", res);
+
+	return IRQ_HANDLED;
+}
+
+static int __dev_pm_set_dedicated_wake_irq(struct device *dev, int irq, unsigned int flag)
+{
+	struct wake_irq *wirq;
+	int err;
+
+	if (irq < 0)
+		return -EINVAL;
+
+	wirq = kzalloc_obj(*wirq);
+	if (!wirq)
+		return -ENOMEM;
+
+	wirq->name = kasprintf(GFP_KERNEL, "%s:wakeup", dev_name(dev));
+	if (!wirq->name) {
+		err = -ENOMEM;
+		goto err_free;
+	}
+
+	wirq->dev = dev;
+	wirq->irq = irq;
+
+	/* Prevent deferred spurious wakeirqs with disable_irq_nosync() */
+	irq_set_status_flags(irq, IRQ_DISABLE_UNLAZY);
+
+	/*
+	 * Consumer device may need to power up and restore state
+	 * so we use a threaded irq.
+	 */
+	err = request_threaded_irq(irq, NULL, handle_threaded_wake_irq,
+				   IRQF_ONESHOT | IRQF_NO_AUTOEN,
+				   wirq->name, wirq);
+	if (err)
+		goto err_free_name;
+
+	err = dev_pm_attach_wake_irq(dev, wirq);
+	if (err)
+		goto err_free_irq;
+
+	wirq->status = WAKE_IRQ_DEDICATED_ALLOCATED | flag;
+
+	return err;
+
+err_free_irq:
+	free_irq(irq, wirq);
+err_free_name:
+	kfree(wirq->name);
+err_free:
+	kfree(wirq);
+
+	return err;
+}
+
+/**
+ * dev_pm_set_dedicated_wake_irq - Request a dedicated wake-up interrupt
+ * @dev: Device entry
+ * @irq: Device wake-up interrupt
+ *
+ * Unless your hardware has separate wake-up interrupts in addition
+ * to the device IO interrupts, you don't need this.
+ *
+ * Sets up a threaded interrupt handler for a device that has
+ * a dedicated wake-up interrupt in addition to the device IO
+ * interrupt.
+ */
+int dev_pm_set_dedicated_wake_irq(struct device *dev, int irq)
+{
+	return __dev_pm_set_dedicated_wake_irq(dev, irq, 0);
+}
+EXPORT_SYMBOL_GPL(dev_pm_set_dedicated_wake_irq);
+
+/**
+ * dev_pm_set_dedicated_wake_irq_reverse - Request a dedicated wake-up interrupt
+ *                                         with reverse enable ordering
+ * @dev: Device entry
+ * @irq: Device wake-up interrupt
+ *
+ * Unless your hardware has separate wake-up interrupts in addition
+ * to the device IO interrupts, you don't need this.
+ *
+ * Sets up a threaded interrupt handler for a device that has a dedicated
+ * wake-up interrupt in addition to the device IO interrupt. It sets
+ * the status of WAKE_IRQ_DEDICATED_REVERSE to tell rpm_suspend()
+ * to enable dedicated wake-up interrupt after running the runtime suspend
+ * callback for @dev.
+ */
+int dev_pm_set_dedicated_wake_irq_reverse(struct device *dev, int irq)
+{
+	return __dev_pm_set_dedicated_wake_irq(dev, irq, WAKE_IRQ_DEDICATED_REVERSE);
+}
+EXPORT_SYMBOL_GPL(dev_pm_set_dedicated_wake_irq_reverse);
+
+/**
+ * dev_pm_enable_wake_irq_check - Checks and enables wake-up interrupt
+ * @dev: Device
+ * @can_change_status: Can change wake-up interrupt status
+ *
+ * Enables wakeirq conditionally. We need to enable wake-up interrupt
+ * lazily on the first rpm_suspend(). This is needed as the consumer device
+ * starts in RPM_SUSPENDED state, and the first pm_runtime_get() would
+ * otherwise try to disable already disabled wakeirq. The wake-up interrupt
+ * starts disabled with IRQ_NOAUTOEN set.
+ *
+ * Should be called from rpm_suspend(), rpm_resume(),
+ * pm_runtime_force_suspend() or pm_runtime_force_resume().
+ * Caller must hold &dev->power.lock or disable runtime PM to change
+ * wirq->status.
+ */
+void dev_pm_enable_wake_irq_check(struct device *dev,
+				  bool can_change_status)
+{
+	struct wake_irq *wirq = dev->power.wakeirq;
+
+	if (!wirq || !(wirq->status & WAKE_IRQ_DEDICATED_MASK))
+		return;
+
+	if (likely(wirq->status & WAKE_IRQ_DEDICATED_MANAGED)) {
+		goto enable;
+	} else if (can_change_status) {
+		wirq->status |= WAKE_IRQ_DEDICATED_MANAGED;
+		goto enable;
+	}
+
+	return;
+
+enable:
+	if (!can_change_status || !(wirq->status & WAKE_IRQ_DEDICATED_REVERSE)) {
+		enable_irq(wirq->irq);
+		wirq->status |= WAKE_IRQ_DEDICATED_ENABLED;
+	}
+}
+
+/**
+ * dev_pm_disable_wake_irq_check - Checks and disables wake-up interrupt
+ * @dev: Device
+ * @cond_disable: if set, also check WAKE_IRQ_DEDICATED_REVERSE
+ *
+ * Disables wake-up interrupt conditionally based on status.
+ * Should be called from rpm_suspend(), rpm_resume(),
+ * pm_runtime_force_suspend() or pm_runtime_force_resume().
+ */
+void dev_pm_disable_wake_irq_check(struct device *dev, bool cond_disable)
+{
+	struct wake_irq *wirq = dev->power.wakeirq;
+
+	if (!wirq || !(wirq->status & WAKE_IRQ_DEDICATED_MASK))
+		return;
+
+	if (cond_disable && (wirq->status & WAKE_IRQ_DEDICATED_REVERSE))
+		return;
+
+	if (wirq->status & WAKE_IRQ_DEDICATED_MANAGED) {
+		wirq->status &= ~WAKE_IRQ_DEDICATED_ENABLED;
+		disable_irq_nosync(wirq->irq);
+	}
+}
+
+/**
+ * dev_pm_enable_wake_irq_complete - enable wake IRQ not enabled before
+ * @dev: Device using the wake IRQ
+ *
+ * Enable wake IRQ conditionally based on status, mainly used if want to
+ * enable wake IRQ after running ->runtime_suspend() which depends on
+ * WAKE_IRQ_DEDICATED_REVERSE.
+ *
+ * Should be called from rpm_suspend() or pm_runtime_force_suspend().
+ */
+void dev_pm_enable_wake_irq_complete(struct device *dev)
+{
+	struct wake_irq *wirq = dev->power.wakeirq;
+
+	if (!wirq || !(wirq->status & WAKE_IRQ_DEDICATED_MASK))
+		return;
+
+	if (wirq->status & WAKE_IRQ_DEDICATED_MANAGED &&
+	    wirq->status & WAKE_IRQ_DEDICATED_REVERSE) {
+		enable_irq(wirq->irq);
+		wirq->status |= WAKE_IRQ_DEDICATED_ENABLED;
+	}
+}
+
+/**
+ * dev_pm_arm_wake_irq - Arm device wake-up
+ * @wirq: Device wake-up interrupt
+ *
+ * Sets up the wake-up event conditionally based on the
+ * device_may_wake().
+ */
+void dev_pm_arm_wake_irq(struct wake_irq *wirq)
+{
+	if (!wirq)
+		return;
+
+	if (device_may_wakeup(wirq->dev)) {
+		if (wirq->status & WAKE_IRQ_DEDICATED_ALLOCATED &&
+		    !(wirq->status & WAKE_IRQ_DEDICATED_ENABLED))
+			enable_irq(wirq->irq);
+
+		enable_irq_wake(wirq->irq);
+	}
+}
+
+/**
+ * dev_pm_disarm_wake_irq - Disarm device wake-up
+ * @wirq: Device wake-up interrupt
+ *
+ * Clears up the wake-up event conditionally based on the
+ * device_may_wake().
+ */
+void dev_pm_disarm_wake_irq(struct wake_irq *wirq)
+{
+	if (!wirq)
+		return;
+
+	if (device_may_wakeup(wirq->dev)) {
+		disable_irq_wake(wirq->irq);
+
+		if (wirq->status & WAKE_IRQ_DEDICATED_ALLOCATED &&
+		    !(wirq->status & WAKE_IRQ_DEDICATED_ENABLED))
+			disable_irq_nosync(wirq->irq);
+	}
+}