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, 131 insertions, 0 deletions

diff --git a/drivers/peci/internal.h b/drivers/peci/internal.h
new file mode 100644
index 000000000..99924a118
--- /dev/null
+++ b/drivers/peci/internal.h
@@ -0,0 +1,131 @@
+/* SPDX-License-Identifier: GPL-2.0-only */
+/* Copyright (c) 2018-2021 Intel Corporation */
+
+#ifndef __PECI_INTERNAL_H
+#define __PECI_INTERNAL_H
+
+#include <linux/device.h>
+#include <linux/types.h>
+
+struct peci_controller;
+struct attribute_group;
+struct peci_device;
+struct peci_request;
+
+/* PECI CPU address range 0x30-0x37 */
+#define PECI_BASE_ADDR		0x30
+#define PECI_DEVICE_NUM_MAX	8
+
+struct peci_request *peci_request_alloc(struct peci_device *device, u8 tx_len, u8 rx_len);
+void peci_request_free(struct peci_request *req);
+
+int peci_request_status(struct peci_request *req);
+
+u64 peci_request_dib_read(struct peci_request *req);
+s16 peci_request_temp_read(struct peci_request *req);
+
+u8 peci_request_data_readb(struct peci_request *req);
+u16 peci_request_data_readw(struct peci_request *req);
+u32 peci_request_data_readl(struct peci_request *req);
+u64 peci_request_data_readq(struct peci_request *req);
+
+struct peci_request *peci_xfer_get_dib(struct peci_device *device);
+struct peci_request *peci_xfer_get_temp(struct peci_device *device);
+
+struct peci_request *peci_xfer_pkg_cfg_readb(struct peci_device *device, u8 index, u16 param);
+struct peci_request *peci_xfer_pkg_cfg_readw(struct peci_device *device, u8 index, u16 param);
+struct peci_request *peci_xfer_pkg_cfg_readl(struct peci_device *device, u8 index, u16 param);
+struct peci_request *peci_xfer_pkg_cfg_readq(struct peci_device *device, u8 index, u16 param);
+
+struct peci_request *peci_xfer_pci_cfg_local_readb(struct peci_device *device,
+						   u8 bus, u8 dev, u8 func, u16 reg);
+struct peci_request *peci_xfer_pci_cfg_local_readw(struct peci_device *device,
+						   u8 bus, u8 dev, u8 func, u16 reg);
+struct peci_request *peci_xfer_pci_cfg_local_readl(struct peci_device *device,
+						   u8 bus, u8 dev, u8 func, u16 reg);
+
+struct peci_request *peci_xfer_ep_pci_cfg_local_readb(struct peci_device *device, u8 seg,
+						      u8 bus, u8 dev, u8 func, u16 reg);
+struct peci_request *peci_xfer_ep_pci_cfg_local_readw(struct peci_device *device, u8 seg,
+						      u8 bus, u8 dev, u8 func, u16 reg);
+struct peci_request *peci_xfer_ep_pci_cfg_local_readl(struct peci_device *device, u8 seg,
+						      u8 bus, u8 dev, u8 func, u16 reg);
+
+struct peci_request *peci_xfer_ep_pci_cfg_readb(struct peci_device *device, u8 seg,
+						u8 bus, u8 dev, u8 func, u16 reg);
+struct peci_request *peci_xfer_ep_pci_cfg_readw(struct peci_device *device, u8 seg,
+						u8 bus, u8 dev, u8 func, u16 reg);
+struct peci_request *peci_xfer_ep_pci_cfg_readl(struct peci_device *device, u8 seg,
+						u8 bus, u8 dev, u8 func, u16 reg);
+
+struct peci_request *peci_xfer_ep_mmio32_readl(struct peci_device *device, u8 bar, u8 seg,
+					       u8 bus, u8 dev, u8 func, u64 offset);
+
+struct peci_request *peci_xfer_ep_mmio64_readl(struct peci_device *device, u8 bar, u8 seg,
+					       u8 bus, u8 dev, u8 func, u64 offset);
+/**
+ * struct peci_device_id - PECI device data to match
+ * @data: pointer to driver private data specific to device
+ * @x86_vfm: device vendor-family-model
+ */
+struct peci_device_id {
+	const void *data;
+	u32 x86_vfm;
+};
+
+extern const struct device_type peci_device_type;
+extern const struct attribute_group *peci_device_groups[];
+
+int peci_device_create(struct peci_controller *controller, u8 addr);
+void peci_device_destroy(struct peci_device *device);
+
+extern const struct bus_type peci_bus_type;
+extern const struct attribute_group *peci_bus_groups[];
+
+/**
+ * struct peci_driver - PECI driver
+ * @driver: inherit device driver
+ * @probe: probe callback
+ * @remove: remove callback
+ * @id_table: PECI device match table to decide which device to bind
+ */
+struct peci_driver {
+	struct device_driver driver;
+	int (*probe)(struct peci_device *device, const struct peci_device_id *id);
+	void (*remove)(struct peci_device *device);
+	const struct peci_device_id *id_table;
+};
+
+#define to_peci_driver(__drv)	container_of_const(__drv, struct peci_driver, driver)
+
+int __peci_driver_register(struct peci_driver *driver, struct module *owner,
+			   const char *mod_name);
+/**
+ * peci_driver_register() - register PECI driver
+ * @driver: the driver to be registered
+ *
+ * PECI drivers that don't need to do anything special in module init should
+ * use the convenience "module_peci_driver" macro instead
+ *
+ * Return: zero on success, else a negative error code.
+ */
+#define peci_driver_register(driver) \
+	__peci_driver_register(driver, THIS_MODULE, KBUILD_MODNAME)
+void peci_driver_unregister(struct peci_driver *driver);
+
+/**
+ * module_peci_driver() - helper macro for registering a modular PECI driver
+ * @__peci_driver: peci_driver struct
+ *
+ * Helper macro for PECI drivers which do not do anything special in module
+ * init/exit. This eliminates a lot of boilerplate. Each module may only
+ * use this macro once, and calling it replaces module_init() and module_exit()
+ */
+#define module_peci_driver(__peci_driver) \
+	module_driver(__peci_driver, peci_driver_register, peci_driver_unregister)
+
+extern const struct device_type peci_controller_type;
+
+int peci_controller_scan_devices(struct peci_controller *controller);
+
+#endif /* __PECI_INTERNAL_H */