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, 153 insertions, 0 deletions

diff --git a/drivers/clk/clk-divider_test.c b/drivers/clk/clk-divider_test.c
new file mode 100644
index 000000000..8ee161412
--- /dev/null
+++ b/drivers/clk/clk-divider_test.c
@@ -0,0 +1,153 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * KUnit tests for clk_divider_bestdiv()
+ */
+#include <kunit/test.h>
+#include <linux/clk.h>
+#include <linux/clk-provider.h>
+#include <linux/limits.h>
+#include <linux/units.h>
+
+#define PARENT_RATE_1GHZ	GIGA
+#define PARENT_RATE_2GHZ	(2 * GIGA)
+#define PARENT_RATE_4GHZ	(4 * GIGA)
+
+KUNIT_DEFINE_ACTION_WRAPPER(clk_hw_unregister_fixed_rate_wrapper,
+			    clk_hw_unregister_fixed_rate, struct clk_hw *);
+KUNIT_DEFINE_ACTION_WRAPPER(clk_hw_unregister_divider_wrapper,
+			    clk_hw_unregister_divider, struct clk_hw *);
+KUNIT_DEFINE_ACTION_WRAPPER(clk_hw_unregister_mux_wrapper,
+			    clk_hw_unregister_mux, struct clk_hw *);
+
+static const struct clk_div_table bestdiv_table[] = {
+	{ .val = 0, .div = 2 },
+	{ .val = 1, .div = 4 },
+	{ .val = 2, .div = 8 },
+	{ /* sentinel */ }
+};
+
+/*
+ * Test that clk_round_rate(clk, ULONG_MAX) returns the maximum achievable
+ * rate for a divider clock.
+ */
+static void clk_divider_bestdiv_ulong_max_returns_max_rate(struct kunit *test)
+{
+	struct clk_hw *parent_hw, *div_hw;
+	unsigned long rate;
+	u32 *fake_reg;
+
+	fake_reg = kunit_kzalloc(test, sizeof(*fake_reg), GFP_KERNEL);
+	KUNIT_ASSERT_NOT_ERR_OR_NULL(test, fake_reg);
+
+	parent_hw = clk_hw_register_fixed_rate(NULL, "bestdiv-parent",
+					       NULL, 0, PARENT_RATE_1GHZ);
+	KUNIT_ASSERT_NOT_ERR_OR_NULL(test, parent_hw);
+	KUNIT_ASSERT_EQ(test, 0,
+			kunit_add_action_or_reset(test, clk_hw_unregister_fixed_rate_wrapper,
+						  parent_hw));
+
+	div_hw = clk_hw_register_divider_table(NULL, "bestdiv-div",
+					       "bestdiv-parent",
+					       CLK_SET_RATE_PARENT,
+					       (void __iomem __force *)fake_reg,
+					       0, 2, 0, bestdiv_table, NULL);
+	KUNIT_ASSERT_NOT_ERR_OR_NULL(test, div_hw);
+	KUNIT_ASSERT_EQ(test, 0,
+			kunit_add_action_or_reset(test, clk_hw_unregister_divider_wrapper,
+						  div_hw));
+
+	/*
+	 * ULONG_MAX is the canonical way to probe the maximum rate a clock
+	 * can produce.
+	 */
+	rate = clk_hw_round_rate(div_hw, ULONG_MAX);
+	KUNIT_EXPECT_EQ(test, rate, PARENT_RATE_1GHZ / 2);
+}
+
+/*
+ * Test that clk_round_rate(clk, ULONG_MAX) returns the correct maximum rate
+ * when a mux clock sits between a divider and its parent candidates.
+ *
+ * Topology:
+ *
+ *   [fixed 4 GHz] --\
+ *                    +--> [mux CLK_SET_RATE_PARENT] --> [div {2,4,8} CLK_SET_RATE_PARENT]
+ *   [fixed 2 GHz] --/
+ *
+ */
+static void clk_divider_bestdiv_mux_ulong_max_returns_max_rate(struct kunit *test)
+{
+	static const char * const mux_parents[] = {
+		"bestdiv-mux-parent-a",
+		"bestdiv-mux-parent-b",
+	};
+	struct clk_hw *parent_a_hw, *parent_b_hw, *mux_hw, *div_hw;
+	u32 *fake_reg_mux, *fake_reg_div;
+	unsigned long rate;
+
+	fake_reg_mux = kunit_kzalloc(test, sizeof(*fake_reg_mux), GFP_KERNEL);
+	KUNIT_ASSERT_NOT_ERR_OR_NULL(test, fake_reg_mux);
+
+	fake_reg_div = kunit_kzalloc(test, sizeof(*fake_reg_div), GFP_KERNEL);
+	KUNIT_ASSERT_NOT_ERR_OR_NULL(test, fake_reg_div);
+
+	/* Higher-rate parent: the mux should select this for ULONG_MAX. */
+	parent_a_hw = clk_hw_register_fixed_rate(NULL, "bestdiv-mux-parent-a",
+						 NULL, 0, PARENT_RATE_4GHZ);
+	KUNIT_ASSERT_NOT_ERR_OR_NULL(test, parent_a_hw);
+	KUNIT_ASSERT_EQ(test, 0,
+			kunit_add_action_or_reset(test, clk_hw_unregister_fixed_rate_wrapper,
+						  parent_a_hw));
+
+	/* Lower-rate parent: should not be selected. */
+	parent_b_hw = clk_hw_register_fixed_rate(NULL, "bestdiv-mux-parent-b",
+						 NULL, 0, PARENT_RATE_2GHZ);
+	KUNIT_ASSERT_NOT_ERR_OR_NULL(test, parent_b_hw);
+	KUNIT_ASSERT_EQ(test, 0,
+			kunit_add_action_or_reset(test, clk_hw_unregister_fixed_rate_wrapper,
+						  parent_b_hw));
+
+	/*
+	 * 1-bit mux register selects between the two parents.
+	 * CLK_SET_RATE_PARENT allows the divider's rate request to
+	 * propagate into clk_mux_determine_rate().
+	 */
+	mux_hw = clk_hw_register_mux(NULL, "bestdiv-mux",
+				     mux_parents, ARRAY_SIZE(mux_parents),
+				     CLK_SET_RATE_PARENT,
+				     (void __iomem __force *)fake_reg_mux,
+				     0, 1, 0, NULL);
+	KUNIT_ASSERT_NOT_ERR_OR_NULL(test, mux_hw);
+	KUNIT_ASSERT_EQ(test, 0,
+			kunit_add_action_or_reset(test, clk_hw_unregister_mux_wrapper,
+						  mux_hw));
+
+	div_hw = clk_hw_register_divider_table(NULL, "bestdiv-mux-div",
+					       "bestdiv-mux",
+					       CLK_SET_RATE_PARENT,
+					       (void __iomem __force *)fake_reg_div,
+					       0, 2, 0, bestdiv_table, NULL);
+	KUNIT_ASSERT_NOT_ERR_OR_NULL(test, div_hw);
+	KUNIT_ASSERT_EQ(test, 0,
+			kunit_add_action_or_reset(test, clk_hw_unregister_divider_wrapper,
+						  div_hw));
+
+	rate = clk_hw_round_rate(div_hw, ULONG_MAX);
+	KUNIT_EXPECT_EQ(test, rate, PARENT_RATE_4GHZ / 2);
+}
+
+static struct kunit_case clk_divider_bestdiv_test_cases[] = {
+	KUNIT_CASE(clk_divider_bestdiv_ulong_max_returns_max_rate),
+	KUNIT_CASE(clk_divider_bestdiv_mux_ulong_max_returns_max_rate),
+	{}
+};
+
+static struct kunit_suite clk_divider_bestdiv_test_suite = {
+	.name = "clk_divider_bestdiv",
+	.test_cases = clk_divider_bestdiv_test_cases,
+};
+
+kunit_test_suite(clk_divider_bestdiv_test_suite);
+
+MODULE_DESCRIPTION("KUnit tests for clk divider");
+MODULE_LICENSE("GPL");