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, 341 insertions, 0 deletions

diff --git a/drivers/pnp/driver.c b/drivers/pnp/driver.c
new file mode 100644
index 000000000..6d58b1465
--- /dev/null
+++ b/drivers/pnp/driver.c
@@ -0,0 +1,341 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * driver.c - device id matching, driver model, etc.
+ *
+ * Copyright 2002 Adam Belay <ambx1@neo.rr.com>
+ */
+
+#include <linux/string.h>
+#include <linux/list.h>
+#include <linux/module.h>
+#include <linux/ctype.h>
+#include <linux/slab.h>
+#include <linux/pnp.h>
+#include "base.h"
+
+static int compare_func(const char *ida, const char *idb)
+{
+	int i;
+
+	/* we only need to compare the last 4 chars */
+	for (i = 3; i < 7; i++) {
+		if (ida[i] != 'X' &&
+		    idb[i] != 'X' && toupper(ida[i]) != toupper(idb[i]))
+			return 0;
+	}
+	return 1;
+}
+
+int compare_pnp_id(struct pnp_id *pos, const char *id)
+{
+	if (!pos || !id || (strlen(id) != 7))
+		return 0;
+	if (memcmp(id, "ANYDEVS", 7) == 0)
+		return 1;
+	while (pos) {
+		if (memcmp(pos->id, id, 3) == 0)
+			if (compare_func(pos->id, id) == 1)
+				return 1;
+		pos = pos->next;
+	}
+	return 0;
+}
+
+static const struct pnp_device_id *match_device(const struct pnp_driver *drv,
+						struct pnp_dev *dev)
+{
+	const struct pnp_device_id *drv_id = drv->id_table;
+
+	if (!drv_id)
+		return NULL;
+
+	while (*drv_id->id) {
+		if (compare_pnp_id(dev->id, drv_id->id))
+			return drv_id;
+		drv_id++;
+	}
+	return NULL;
+}
+
+int pnp_device_attach(struct pnp_dev *pnp_dev)
+{
+	mutex_lock(&pnp_lock);
+	if (pnp_dev->status != PNP_READY) {
+		mutex_unlock(&pnp_lock);
+		return -EBUSY;
+	}
+	pnp_dev->status = PNP_ATTACHED;
+	mutex_unlock(&pnp_lock);
+	return 0;
+}
+EXPORT_SYMBOL(pnp_device_attach);
+
+void pnp_device_detach(struct pnp_dev *pnp_dev)
+{
+	mutex_lock(&pnp_lock);
+	if (pnp_dev->status == PNP_ATTACHED)
+		pnp_dev->status = PNP_READY;
+	mutex_unlock(&pnp_lock);
+}
+EXPORT_SYMBOL(pnp_device_detach);
+
+static int pnp_device_probe(struct device *dev)
+{
+	int error;
+	struct pnp_driver *pnp_drv;
+	struct pnp_dev *pnp_dev;
+	const struct pnp_device_id *dev_id = NULL;
+	pnp_dev = to_pnp_dev(dev);
+	pnp_drv = to_pnp_driver(dev->driver);
+
+	error = pnp_device_attach(pnp_dev);
+	if (error < 0)
+		return error;
+
+	if (pnp_dev->active == 0) {
+		if (!(pnp_drv->flags & PNP_DRIVER_RES_DO_NOT_CHANGE)) {
+			error = pnp_activate_dev(pnp_dev);
+			if (error < 0)
+				return error;
+		}
+	} else if ((pnp_drv->flags & PNP_DRIVER_RES_DISABLE)
+		   == PNP_DRIVER_RES_DISABLE) {
+		error = pnp_disable_dev(pnp_dev);
+		if (error < 0)
+			return error;
+	}
+	error = 0;
+	if (pnp_drv->probe) {
+		dev_id = match_device(pnp_drv, pnp_dev);
+		if (dev_id != NULL)
+			error = pnp_drv->probe(pnp_dev, dev_id);
+	}
+	if (error >= 0) {
+		pnp_dev->driver = pnp_drv;
+		error = 0;
+	} else
+		goto fail;
+
+	return error;
+
+fail:
+	pnp_device_detach(pnp_dev);
+	return error;
+}
+
+static void pnp_device_remove(struct device *dev)
+{
+	struct pnp_dev *pnp_dev = to_pnp_dev(dev);
+	struct pnp_driver *drv = pnp_dev->driver;
+
+	if (drv) {
+		if (drv->remove)
+			drv->remove(pnp_dev);
+		pnp_dev->driver = NULL;
+	}
+
+	if (pnp_dev->active &&
+	    (!drv || !(drv->flags & PNP_DRIVER_RES_DO_NOT_CHANGE)))
+		pnp_disable_dev(pnp_dev);
+
+	pnp_device_detach(pnp_dev);
+}
+
+static void pnp_device_shutdown(struct device *dev)
+{
+	struct pnp_dev *pnp_dev = to_pnp_dev(dev);
+	struct pnp_driver *drv = pnp_dev->driver;
+
+	if (drv && drv->shutdown)
+		drv->shutdown(pnp_dev);
+}
+
+static int pnp_uevent(const struct device *dev, struct kobj_uevent_env *env)
+{
+	struct pnp_id *pos;
+	const struct pnp_dev *pnp_dev = to_pnp_dev(dev);
+
+	if (!dev)
+		return -ENODEV;
+
+	pos = pnp_dev->id;
+	while (pos) {
+		if (add_uevent_var(env, "MODALIAS=pnp:d%s", pos->id))
+			return -ENOMEM;
+		pos = pos->next;
+	}
+
+	return 0;
+}
+
+static int pnp_bus_match(struct device *dev, const struct device_driver *drv)
+{
+	struct pnp_dev *pnp_dev = to_pnp_dev(dev);
+	const struct pnp_driver *pnp_drv = to_pnp_driver(drv);
+
+	if (match_device(pnp_drv, pnp_dev) == NULL)
+		return 0;
+	return 1;
+}
+
+static int __pnp_bus_suspend(struct device *dev, pm_message_t state)
+{
+	struct pnp_dev *pnp_dev = to_pnp_dev(dev);
+	struct pnp_driver *pnp_drv = pnp_dev->driver;
+	int error;
+
+	if (!pnp_drv)
+		return 0;
+
+	if (pnp_drv->driver.pm && pnp_drv->driver.pm->suspend) {
+		error = pnp_drv->driver.pm->suspend(dev);
+		suspend_report_result(dev, pnp_drv->driver.pm->suspend, error);
+		if (error)
+			return error;
+	}
+
+	if (pnp_drv->suspend) {
+		error = pnp_drv->suspend(pnp_dev, state);
+		if (error)
+			return error;
+	}
+
+	/* can_write is necessary to be able to re-start the device on resume */
+	if (pnp_can_disable(pnp_dev) && pnp_can_write(pnp_dev)) {
+		error = pnp_stop_dev(pnp_dev);
+		if (error)
+			return error;
+	}
+
+	if (pnp_can_suspend(pnp_dev))
+		pnp_dev->protocol->suspend(pnp_dev, state);
+	return 0;
+}
+
+static int pnp_bus_suspend(struct device *dev)
+{
+	return __pnp_bus_suspend(dev, PMSG_SUSPEND);
+}
+
+static int pnp_bus_freeze(struct device *dev)
+{
+	return __pnp_bus_suspend(dev, PMSG_FREEZE);
+}
+
+static int pnp_bus_poweroff(struct device *dev)
+{
+	return __pnp_bus_suspend(dev, PMSG_HIBERNATE);
+}
+
+static int pnp_bus_resume(struct device *dev)
+{
+	struct pnp_dev *pnp_dev = to_pnp_dev(dev);
+	struct pnp_driver *pnp_drv = pnp_dev->driver;
+	int error;
+
+	if (!pnp_drv)
+		return 0;
+
+	if (pnp_dev->protocol->resume) {
+		error = pnp_dev->protocol->resume(pnp_dev);
+		if (error)
+			return error;
+	}
+
+	if (pnp_can_write(pnp_dev)) {
+		error = pnp_start_dev(pnp_dev);
+		if (error)
+			return error;
+	}
+
+	if (pnp_drv->driver.pm && pnp_drv->driver.pm->resume) {
+		error = pnp_drv->driver.pm->resume(dev);
+		if (error)
+			return error;
+	}
+
+	if (pnp_drv->resume) {
+		error = pnp_drv->resume(pnp_dev);
+		if (error)
+			return error;
+	}
+
+	return 0;
+}
+
+static const struct dev_pm_ops pnp_bus_dev_pm_ops = {
+	/* Suspend callbacks */
+	.suspend = pnp_bus_suspend,
+	.resume = pnp_bus_resume,
+	/* Hibernate callbacks */
+	.freeze = pnp_bus_freeze,
+	.thaw = pnp_bus_resume,
+	.poweroff = pnp_bus_poweroff,
+	.restore = pnp_bus_resume,
+};
+
+const struct bus_type pnp_bus_type = {
+	.name    = "pnp",
+	.match   = pnp_bus_match,
+	.uevent  = pnp_uevent,
+	.probe   = pnp_device_probe,
+	.remove  = pnp_device_remove,
+	.shutdown = pnp_device_shutdown,
+	.pm	 = &pnp_bus_dev_pm_ops,
+	.dev_groups = pnp_dev_groups,
+};
+
+bool dev_is_pnp(const struct device *dev)
+{
+	return dev->bus == &pnp_bus_type;
+}
+EXPORT_SYMBOL_GPL(dev_is_pnp);
+
+int pnp_register_driver(struct pnp_driver *drv)
+{
+	drv->driver.name = drv->name;
+	drv->driver.bus = &pnp_bus_type;
+
+	return driver_register(&drv->driver);
+}
+EXPORT_SYMBOL(pnp_register_driver);
+
+void pnp_unregister_driver(struct pnp_driver *drv)
+{
+	driver_unregister(&drv->driver);
+}
+EXPORT_SYMBOL(pnp_unregister_driver);
+
+/**
+ * pnp_add_id - adds an EISA id to the specified device
+ * @dev: pointer to the desired device
+ * @id: pointer to an EISA id string
+ */
+struct pnp_id *pnp_add_id(struct pnp_dev *dev, const char *id)
+{
+	struct pnp_id *dev_id, *ptr;
+
+	dev_id = kzalloc_obj(struct pnp_id);
+	if (!dev_id)
+		return NULL;
+
+	dev_id->id[0] = id[0];
+	dev_id->id[1] = id[1];
+	dev_id->id[2] = id[2];
+	dev_id->id[3] = tolower(id[3]);
+	dev_id->id[4] = tolower(id[4]);
+	dev_id->id[5] = tolower(id[5]);
+	dev_id->id[6] = tolower(id[6]);
+	dev_id->id[7] = '\0';
+
+	dev_id->next = NULL;
+	ptr = dev->id;
+	while (ptr && ptr->next)
+		ptr = ptr->next;
+	if (ptr)
+		ptr->next = dev_id;
+	else
+		dev->id = dev_id;
+
+	return dev_id;
+}