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, 234 insertions, 0 deletions

diff --git a/include/asm-generic/uaccess.h b/include/asm-generic/uaccess.h
new file mode 100644
index 000000000..4569045e7
--- /dev/null
+++ b/include/asm-generic/uaccess.h
@@ -0,0 +1,234 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+#ifndef __ASM_GENERIC_UACCESS_H
+#define __ASM_GENERIC_UACCESS_H
+
+/*
+ * User space memory access functions, these should work
+ * on any machine that has kernel and user data in the same
+ * address space, e.g. all NOMMU machines.
+ */
+#include <linux/string.h>
+#include <asm-generic/access_ok.h>
+
+#ifdef CONFIG_UACCESS_MEMCPY
+#include <linux/unaligned.h>
+
+static __always_inline int
+__get_user_fn(size_t size, const void __user *from, void *to)
+{
+	BUILD_BUG_ON(!__builtin_constant_p(size));
+
+	switch (size) {
+	case 1:
+		*(u8 *)to = *((u8 __force *)from);
+		return 0;
+	case 2:
+		*(u16 *)to = get_unaligned((u16 __force *)from);
+		return 0;
+	case 4:
+		*(u32 *)to = get_unaligned((u32 __force *)from);
+		return 0;
+	case 8:
+		*(u64 *)to = get_unaligned((u64 __force *)from);
+		return 0;
+	default:
+		BUILD_BUG();
+		return 0;
+	}
+
+}
+#define __get_user_fn(sz, u, k)	__get_user_fn(sz, u, k)
+
+static __always_inline int
+__put_user_fn(size_t size, void __user *to, void *from)
+{
+	BUILD_BUG_ON(!__builtin_constant_p(size));
+
+	switch (size) {
+	case 1:
+		*(u8 __force *)to = *(u8 *)from;
+		return 0;
+	case 2:
+		put_unaligned(*(u16 *)from, (u16 __force *)to);
+		return 0;
+	case 4:
+		put_unaligned(*(u32 *)from, (u32 __force *)to);
+		return 0;
+	case 8:
+		put_unaligned(*(u64 *)from, (u64 __force *)to);
+		return 0;
+	default:
+		BUILD_BUG();
+		return 0;
+	}
+}
+#define __put_user_fn(sz, u, k)	__put_user_fn(sz, u, k)
+
+#define __get_kernel_nofault(dst, src, type, err_label)			\
+do {									\
+	*((type *)dst) = get_unaligned((type *)(src));			\
+	if (0) /* make sure the label looks used to the compiler */	\
+		goto err_label;						\
+} while (0)
+
+#define __put_kernel_nofault(dst, src, type, err_label)			\
+do {									\
+	put_unaligned(*((type *)src), (type *)(dst));			\
+	if (0) /* make sure the label looks used to the compiler */	\
+		goto err_label;						\
+} while (0)
+
+static inline __must_check unsigned long
+raw_copy_from_user(void *to, const void __user * from, unsigned long n)
+{
+	memcpy(to, (const void __force *)from, n);
+	return 0;
+}
+
+static inline __must_check unsigned long
+raw_copy_to_user(void __user *to, const void *from, unsigned long n)
+{
+	memcpy((void __force *)to, from, n);
+	return 0;
+}
+#define INLINE_COPY_USER
+#endif /* CONFIG_UACCESS_MEMCPY */
+
+/*
+ * These are the main single-value transfer routines.  They automatically
+ * use the right size if we just have the right pointer type.
+ * This version just falls back to copy_{from,to}_user, which should
+ * provide a fast-path for small values.
+ */
+#define __put_user(x, ptr) \
+({								\
+	__typeof__(*(ptr)) __x = (x);				\
+	int __pu_err = -EFAULT;					\
+        __chk_user_ptr(ptr);                                    \
+	switch (sizeof (*(ptr))) {				\
+	case 1:							\
+	case 2:							\
+	case 4:							\
+	case 8:							\
+		__pu_err = __put_user_fn(sizeof (*(ptr)),	\
+					 ptr, &__x);		\
+		break;						\
+	default:						\
+		__put_user_bad();				\
+		break;						\
+	 }							\
+	__pu_err;						\
+})
+
+#define put_user(x, ptr)					\
+({								\
+	void __user *__p = (ptr);				\
+	might_fault();						\
+	access_ok(__p, sizeof(*ptr)) ?		\
+		__put_user((x), ((__typeof__(*(ptr)) __user *)__p)) :	\
+		-EFAULT;					\
+})
+
+#ifndef __put_user_fn
+
+static inline int __put_user_fn(size_t size, void __user *ptr, void *x)
+{
+	return unlikely(raw_copy_to_user(ptr, x, size)) ? -EFAULT : 0;
+}
+
+#define __put_user_fn(sz, u, k)	__put_user_fn(sz, u, k)
+
+#endif
+
+extern int __put_user_bad(void) __attribute__((noreturn));
+
+#define __get_user(x, ptr)					\
+({								\
+	int __gu_err = -EFAULT;					\
+	__chk_user_ptr(ptr);					\
+	switch (sizeof(*(ptr))) {				\
+	case 1: {						\
+		unsigned char __x = 0;				\
+		__gu_err = __get_user_fn(sizeof (*(ptr)),	\
+					 ptr, &__x);		\
+		(x) = *(__force __typeof__(*(ptr)) *) &__x;	\
+		break;						\
+	};							\
+	case 2: {						\
+		unsigned short __x = 0;				\
+		__gu_err = __get_user_fn(sizeof (*(ptr)),	\
+					 ptr, &__x);		\
+		(x) = *(__force __typeof__(*(ptr)) *) &__x;	\
+		break;						\
+	};							\
+	case 4: {						\
+		unsigned int __x = 0;				\
+		__gu_err = __get_user_fn(sizeof (*(ptr)),	\
+					 ptr, &__x);		\
+		(x) = *(__force __typeof__(*(ptr)) *) &__x;	\
+		break;						\
+	};							\
+	case 8: {						\
+		unsigned long long __x = 0;			\
+		__gu_err = __get_user_fn(sizeof (*(ptr)),	\
+					 ptr, &__x);		\
+		(x) = *(__force __typeof__(*(ptr)) *) &__x;	\
+		break;						\
+	};							\
+	default:						\
+		__get_user_bad();				\
+		break;						\
+	}							\
+	__gu_err;						\
+})
+
+#define get_user(x, ptr)					\
+({								\
+	const void __user *__p = (ptr);				\
+	might_fault();						\
+	access_ok(__p, sizeof(*ptr)) ?		\
+		__get_user((x), (__typeof__(*(ptr)) __user *)__p) :\
+		((x) = (__typeof__(*(ptr)))0,-EFAULT);		\
+})
+
+#ifndef __get_user_fn
+static inline int __get_user_fn(size_t size, const void __user *ptr, void *x)
+{
+	return unlikely(raw_copy_from_user(x, ptr, size)) ? -EFAULT : 0;
+}
+
+#define __get_user_fn(sz, u, k)	__get_user_fn(sz, u, k)
+
+#endif
+
+extern int __get_user_bad(void) __attribute__((noreturn));
+
+/*
+ * Zero Userspace
+ */
+#ifndef __clear_user
+static inline __must_check unsigned long
+__clear_user(void __user *to, unsigned long n)
+{
+	memset((void __force *)to, 0, n);
+	return 0;
+}
+#endif
+
+static inline __must_check unsigned long
+clear_user(void __user *to, unsigned long n)
+{
+	might_fault();
+	if (!access_ok(to, n))
+		return n;
+
+	return __clear_user(to, n);
+}
+
+#include <asm/extable.h>
+
+__must_check long strncpy_from_user(char *dst, const char __user *src,
+				    long count);
+__must_check long strnlen_user(const char __user *src, long n);
+
+#endif /* __ASM_GENERIC_UACCESS_H */