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, 118 insertions, 0 deletions

diff --git a/lib/extable.c b/lib/extable.c
new file mode 100644
index 000000000..9c9f40bd2
--- /dev/null
+++ b/lib/extable.c
@@ -0,0 +1,118 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
+/*
+ * Derived from arch/ppc/mm/extable.c and arch/i386/mm/extable.c.
+ *
+ * Copyright (C) 2004 Paul Mackerras, IBM Corp.
+ */
+
+#include <linux/bsearch.h>
+#include <linux/module.h>
+#include <linux/init.h>
+#include <linux/sort.h>
+#include <linux/uaccess.h>
+#include <linux/extable.h>
+
+#ifndef ARCH_HAS_RELATIVE_EXTABLE
+#define ex_to_insn(x)	((x)->insn)
+#else
+static inline unsigned long ex_to_insn(const struct exception_table_entry *x)
+{
+	return (unsigned long)&x->insn + x->insn;
+}
+#endif
+
+#ifndef ARCH_HAS_RELATIVE_EXTABLE
+#define swap_ex		NULL
+#else
+static void swap_ex(void *a, void *b, int size)
+{
+	struct exception_table_entry *x = a, *y = b, tmp;
+	int delta = b - a;
+
+	tmp = *x;
+	x->insn = y->insn + delta;
+	y->insn = tmp.insn - delta;
+
+#ifdef swap_ex_entry_fixup
+	swap_ex_entry_fixup(x, y, tmp, delta);
+#else
+	x->fixup = y->fixup + delta;
+	y->fixup = tmp.fixup - delta;
+#endif
+}
+#endif /* ARCH_HAS_RELATIVE_EXTABLE */
+
+/*
+ * The exception table needs to be sorted so that the binary
+ * search that we use to find entries in it works properly.
+ * This is used both for the kernel exception table and for
+ * the exception tables of modules that get loaded.
+ */
+static int cmp_ex_sort(const void *a, const void *b)
+{
+	const struct exception_table_entry *x = a, *y = b;
+
+	/* avoid overflow */
+	if (ex_to_insn(x) > ex_to_insn(y))
+		return 1;
+	if (ex_to_insn(x) < ex_to_insn(y))
+		return -1;
+	return 0;
+}
+
+void sort_extable(struct exception_table_entry *start,
+		  struct exception_table_entry *finish)
+{
+	sort(start, finish - start, sizeof(struct exception_table_entry),
+	     cmp_ex_sort, swap_ex);
+}
+
+#ifdef CONFIG_MODULES
+/*
+ * If the exception table is sorted, any referring to the module init
+ * will be at the beginning or the end.
+ */
+void trim_init_extable(struct module *m)
+{
+	/*trim the beginning*/
+	while (m->num_exentries &&
+	       within_module_init(ex_to_insn(&m->extable[0]), m)) {
+		m->extable++;
+		m->num_exentries--;
+	}
+	/*trim the end*/
+	while (m->num_exentries &&
+	       within_module_init(ex_to_insn(&m->extable[m->num_exentries - 1]),
+				  m))
+		m->num_exentries--;
+}
+#endif /* CONFIG_MODULES */
+
+static int cmp_ex_search(const void *key, const void *elt)
+{
+	const struct exception_table_entry *_elt = elt;
+	unsigned long _key = *(unsigned long *)key;
+
+	/* avoid overflow */
+	if (_key > ex_to_insn(_elt))
+		return 1;
+	if (_key < ex_to_insn(_elt))
+		return -1;
+	return 0;
+}
+
+/*
+ * Search one exception table for an entry corresponding to the
+ * given instruction address, and return the address of the entry,
+ * or NULL if none is found.
+ * We use a binary search, and thus we assume that the table is
+ * already sorted.
+ */
+const struct exception_table_entry *
+search_extable(const struct exception_table_entry *base,
+	       const size_t num,
+	       unsigned long value)
+{
+	return bsearch(&value, base, num,
+		       sizeof(struct exception_table_entry), cmp_ex_search);
+}