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, 97 insertions, 0 deletions

diff --git a/include/kunit/test-bug.h b/include/kunit/test-bug.h
new file mode 100644
index 000000000..99869029f
--- /dev/null
+++ b/include/kunit/test-bug.h
@@ -0,0 +1,97 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+/*
+ * KUnit API providing hooks for non-test code to interact with tests.
+ *
+ * Copyright (C) 2020, Google LLC.
+ * Author: Uriel Guajardo <urielguajardo@google.com>
+ */
+
+#ifndef _KUNIT_TEST_BUG_H
+#define _KUNIT_TEST_BUG_H
+
+#include <linux/stddef.h> /* for NULL */
+#include <linux/types.h>  /* for bool */
+
+#if IS_ENABLED(CONFIG_KUNIT)
+
+#include <linux/jump_label.h> /* For static branch */
+#include <linux/sched.h>
+
+/* Static key if KUnit is running any tests. */
+DECLARE_STATIC_KEY_FALSE(kunit_running);
+
+/* Hooks table: a table of function pointers filled in when kunit loads */
+extern struct kunit_hooks_table {
+	__printf(3, 4) void (*fail_current_test)(const char*, int, const char*, ...);
+	void *(*get_static_stub_address)(struct kunit *test, void *real_fn_addr);
+	bool (*is_suppressed_warning)(bool count);
+} kunit_hooks;
+
+/**
+ * kunit_get_current_test() - Return a pointer to the currently running
+ *			      KUnit test.
+ *
+ * If a KUnit test is running in the current task, returns a pointer to its
+ * associated struct kunit. This pointer can then be passed to any KUnit
+ * function or assertion. If no test is running (or a test is running in a
+ * different task), returns NULL.
+ *
+ * This function is safe to call even when KUnit is disabled. If CONFIG_KUNIT
+ * is not enabled, it will compile down to nothing and will return quickly no
+ * test is running.
+ */
+static inline struct kunit *kunit_get_current_test(void)
+{
+	if (!static_branch_unlikely(&kunit_running))
+		return NULL;
+
+	return current->kunit_test;
+}
+
+
+/**
+ * kunit_fail_current_test() - If a KUnit test is running, fail it.
+ *
+ * If a KUnit test is running in the current task, mark that test as failed.
+ */
+#define kunit_fail_current_test(fmt, ...) do {					\
+		if (static_branch_unlikely(&kunit_running)) {			\
+			/* Guaranteed to be non-NULL when kunit_running true*/	\
+			kunit_hooks.fail_current_test(__FILE__, __LINE__,	\
+						  fmt, ##__VA_ARGS__);		\
+		}								\
+	} while (0)
+
+/**
+ * kunit_is_suppressed_warning() - Check if warnings are being suppressed
+ *                                 by the current KUnit test.
+ * @count: if true, increment the suppression counter on match.
+ *
+ * Returns true if the current task has active warning suppression.
+ * Uses the kunit_running static branch for zero overhead when no tests run.
+ *
+ * A single WARN*() may traverse multiple call sites in the warning path
+ * (e.g., __warn_printk() and __report_bug()). Pass @count = true at the
+ * primary suppression point to count each warning exactly once, and
+ * @count = false at secondary points to suppress output without
+ * inflating the count.
+ */
+static inline bool kunit_is_suppressed_warning(bool count)
+{
+	if (!static_branch_unlikely(&kunit_running))
+		return false;
+
+	return kunit_hooks.is_suppressed_warning &&
+	       kunit_hooks.is_suppressed_warning(count);
+}
+
+#else
+
+static inline struct kunit *kunit_get_current_test(void) { return NULL; }
+static inline bool kunit_is_suppressed_warning(bool count) { return false; }
+
+#define kunit_fail_current_test(fmt, ...) do {} while (0)
+
+#endif
+
+#endif /* _KUNIT_TEST_BUG_H */