From a5f7a5bb3b7f28ba7e4fa246775b29a0e5537255 Mon Sep 17 00:00:00 2001 From: Jinke Han Date: Tue, 8 Sep 2026 15:37:42 +0800 Subject: x86/kprobes: Fix crash when probing CS CALL instructions 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) Signed-off-by: Jinke Han Signed-off-by: Ingo Molnar Reviewed-by: Masami Hiramatsu (Google) Acked-by: Yafang Shao Acked-by: Borislav Petkov Cc: Peter Zijlstra Link: https://patch.msgid.link/20260908073742.GA10517@didi-ThinkCentre-M920t-N000 --- drivers/rpmsg/rpmsg_internal.h | 96 ++++++++++++++++++++++++++++++++++++++++++ 1 file changed, 96 insertions(+) create mode 100644 drivers/rpmsg/rpmsg_internal.h (limited to 'drivers/rpmsg/rpmsg_internal.h') diff --git a/drivers/rpmsg/rpmsg_internal.h b/drivers/rpmsg/rpmsg_internal.h new file mode 100644 index 000000000..a8b7065fd --- /dev/null +++ b/drivers/rpmsg/rpmsg_internal.h @@ -0,0 +1,96 @@ +/* SPDX-License-Identifier: GPL-2.0 */ +/* + * remote processor messaging bus internals + * + * Copyright (C) 2011 Texas Instruments, Inc. + * Copyright (C) 2011 Google, Inc. + * + * Ohad Ben-Cohen + * Brian Swetland + */ + +#ifndef __RPMSG_INTERNAL_H__ +#define __RPMSG_INTERNAL_H__ + +#include +#include + +#define to_rpmsg_device(d) container_of(d, struct rpmsg_device, dev) +#define to_rpmsg_driver(d) container_of_const(d, struct rpmsg_driver, drv) + +extern const struct class rpmsg_class; + +/** + * struct rpmsg_device_ops - indirection table for the rpmsg_device operations + * @create_channel: create backend-specific channel, optional + * @release_channel: release backend-specific channel, optional + * @create_ept: create backend-specific endpoint, required + * @announce_create: announce presence of new channel, optional + * @announce_destroy: announce destruction of channel, optional + * + * Indirection table for the operations that a rpmsg backend should implement. + * @announce_create and @announce_destroy are optional as the backend might + * advertise new channels implicitly by creating the endpoints. + */ +struct rpmsg_device_ops { + struct rpmsg_device *(*create_channel)(struct rpmsg_device *rpdev, + struct rpmsg_channel_info *chinfo); + int (*release_channel)(struct rpmsg_device *rpdev, + struct rpmsg_channel_info *chinfo); + struct rpmsg_endpoint *(*create_ept)(struct rpmsg_device *rpdev, + rpmsg_rx_cb_t cb, void *priv, + struct rpmsg_channel_info chinfo); + + int (*announce_create)(struct rpmsg_device *rpdev); + int (*announce_destroy)(struct rpmsg_device *rpdev); +}; + +/** + * struct rpmsg_endpoint_ops - indirection table for rpmsg_endpoint operations + * @destroy_ept: see @rpmsg_destroy_ept(), required + * @send: see @rpmsg_send(), required + * @sendto: see @rpmsg_sendto(), optional + * @trysend: see @rpmsg_trysend(), required + * @trysendto: see @rpmsg_trysendto(), optional + * @poll: see @rpmsg_poll(), optional + * @set_flow_control: see @rpmsg_set_flow_control(), optional + * @get_mtu: see @rpmsg_get_mtu(), optional + * + * Indirection table for the operations that a rpmsg backend should implement. + * In addition to @destroy_ept, the backend must at least implement @send and + * @trysend, while the variants sending data off-channel are optional. + */ +struct rpmsg_endpoint_ops { + void (*destroy_ept)(struct rpmsg_endpoint *ept); + + int (*send)(struct rpmsg_endpoint *ept, const void *data, int len); + int (*sendto)(struct rpmsg_endpoint *ept, const void *data, int len, u32 dst); + + int (*trysend)(struct rpmsg_endpoint *ept, const void *data, int len); + int (*trysendto)(struct rpmsg_endpoint *ept, const void *data, int len, u32 dst); + __poll_t (*poll)(struct rpmsg_endpoint *ept, struct file *filp, + poll_table *wait); + int (*set_flow_control)(struct rpmsg_endpoint *ept, bool pause, u32 dst); + ssize_t (*get_mtu)(struct rpmsg_endpoint *ept); +}; + +struct device *rpmsg_find_device(struct device *parent, + struct rpmsg_channel_info *chinfo); + +struct rpmsg_device *rpmsg_create_channel(struct rpmsg_device *rpdev, + struct rpmsg_channel_info *chinfo); +int rpmsg_release_channel(struct rpmsg_device *rpdev, + struct rpmsg_channel_info *chinfo); +/** + * rpmsg_ctrldev_register_device() - register a char device for control based on rpdev + * @rpdev: prepared rpdev to be used for creating endpoints + * + * This function wraps rpmsg_register_device() preparing the rpdev for use as + * basis for the rpmsg chrdev. + */ +static inline int rpmsg_ctrldev_register_device(struct rpmsg_device *rpdev) +{ + return rpmsg_register_device_override(rpdev, "rpmsg_ctrl"); +} + +#endif -- cgit v1.3.1