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, 189 insertions, 0 deletions

diff --git a/drivers/clocksource/timer-orion.c b/drivers/clocksource/timer-orion.c
new file mode 100644
index 000000000..61f1e27fc
--- /dev/null
+++ b/drivers/clocksource/timer-orion.c
@@ -0,0 +1,189 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * Marvell Orion SoC timer handling.
+ *
+ * Sebastian Hesselbarth <sebastian.hesselbarth@gmail.com>
+ *
+ * Timer 0 is used as free-running clocksource, while timer 1 is
+ * used as clock_event_device.
+ */
+
+#include <linux/kernel.h>
+#include <linux/bitops.h>
+#include <linux/clk.h>
+#include <linux/clockchips.h>
+#include <linux/delay.h>
+#include <linux/interrupt.h>
+#include <linux/of_address.h>
+#include <linux/of_irq.h>
+#include <linux/spinlock.h>
+#include <linux/sched_clock.h>
+
+#define TIMER_CTRL		0x00
+#define  TIMER0_EN		BIT(0)
+#define  TIMER0_RELOAD_EN	BIT(1)
+#define  TIMER1_EN		BIT(2)
+#define  TIMER1_RELOAD_EN	BIT(3)
+#define TIMER0_RELOAD		0x10
+#define TIMER0_VAL		0x14
+#define TIMER1_RELOAD		0x18
+#define TIMER1_VAL		0x1c
+
+#define ORION_ONESHOT_MIN	1
+#define ORION_ONESHOT_MAX	0xfffffffe
+
+static void __iomem *timer_base;
+
+static unsigned long notrace orion_read_timer(void)
+{
+	return ~readl(timer_base + TIMER0_VAL);
+}
+
+static struct delay_timer orion_delay_timer = {
+	.read_current_timer = orion_read_timer,
+};
+
+static void __init orion_delay_timer_init(unsigned long rate)
+{
+	orion_delay_timer.freq = rate;
+	register_current_timer_delay(&orion_delay_timer);
+}
+
+/*
+ * Free-running clocksource handling.
+ */
+static u64 notrace orion_read_sched_clock(void)
+{
+	return ~readl(timer_base + TIMER0_VAL);
+}
+
+/*
+ * Clockevent handling.
+ */
+static u32 ticks_per_jiffy;
+
+static int orion_clkevt_next_event(unsigned long delta,
+				   struct clock_event_device *dev)
+{
+	/* setup and enable one-shot timer */
+	writel(delta, timer_base + TIMER1_VAL);
+	atomic_io_modify(timer_base + TIMER_CTRL,
+		TIMER1_RELOAD_EN | TIMER1_EN, TIMER1_EN);
+
+	return 0;
+}
+
+static int orion_clkevt_shutdown(struct clock_event_device *dev)
+{
+	/* disable timer */
+	atomic_io_modify(timer_base + TIMER_CTRL,
+			 TIMER1_RELOAD_EN | TIMER1_EN, 0);
+	return 0;
+}
+
+static int orion_clkevt_set_periodic(struct clock_event_device *dev)
+{
+	/* setup and enable periodic timer at 1/HZ intervals */
+	writel(ticks_per_jiffy - 1, timer_base + TIMER1_RELOAD);
+	writel(ticks_per_jiffy - 1, timer_base + TIMER1_VAL);
+	atomic_io_modify(timer_base + TIMER_CTRL,
+			 TIMER1_RELOAD_EN | TIMER1_EN,
+			 TIMER1_RELOAD_EN | TIMER1_EN);
+	return 0;
+}
+
+static struct clock_event_device orion_clkevt = {
+	.name			= "orion_event",
+	.features		= CLOCK_EVT_FEAT_ONESHOT |
+				  CLOCK_EVT_FEAT_PERIODIC,
+	.shift			= 32,
+	.rating			= 300,
+	.set_next_event		= orion_clkevt_next_event,
+	.set_state_shutdown	= orion_clkevt_shutdown,
+	.set_state_periodic	= orion_clkevt_set_periodic,
+	.set_state_oneshot	= orion_clkevt_shutdown,
+	.tick_resume		= orion_clkevt_shutdown,
+};
+
+static irqreturn_t orion_clkevt_irq_handler(int irq, void *dev_id)
+{
+	orion_clkevt.event_handler(&orion_clkevt);
+	return IRQ_HANDLED;
+}
+
+static int __init orion_timer_init(struct device_node *np)
+{
+	unsigned long rate;
+	struct clk *clk;
+	int irq, ret;
+
+	/* timer registers are shared with watchdog timer */
+	timer_base = of_iomap(np, 0);
+	if (!timer_base) {
+		pr_err("%pOFn: unable to map resource\n", np);
+		return -ENXIO;
+	}
+
+	clk = of_clk_get(np, 0);
+	if (IS_ERR(clk)) {
+		pr_err("%pOFn: unable to get clk\n", np);
+		return PTR_ERR(clk);
+	}
+
+	ret = clk_prepare_enable(clk);
+	if (ret) {
+		pr_err("Failed to prepare clock\n");
+		return ret;
+	}
+
+	/* we are only interested in timer1 irq */
+	irq = irq_of_parse_and_map(np, 1);
+	if (irq <= 0) {
+		pr_err("%pOFn: unable to parse timer1 irq\n", np);
+		ret = -EINVAL;
+		goto out_unprep_clk;
+	}
+
+	rate = clk_get_rate(clk);
+
+	/* setup timer0 as free-running clocksource */
+	writel(~0, timer_base + TIMER0_VAL);
+	writel(~0, timer_base + TIMER0_RELOAD);
+	atomic_io_modify(timer_base + TIMER_CTRL,
+		TIMER0_RELOAD_EN | TIMER0_EN,
+		TIMER0_RELOAD_EN | TIMER0_EN);
+
+	ret = clocksource_mmio_init(timer_base + TIMER0_VAL,
+				    "orion_clocksource", rate, 300, 32,
+				    clocksource_mmio_readl_down);
+	if (ret) {
+		pr_err("Failed to initialize mmio timer\n");
+		goto out_unprep_clk;
+	}
+
+	sched_clock_register(orion_read_sched_clock, 32, rate);
+
+	/* setup timer1 as clockevent timer */
+	ret = request_irq(irq, orion_clkevt_irq_handler, IRQF_TIMER,
+			  "orion_event", NULL);
+	if (ret) {
+		pr_err("%pOFn: unable to setup irq\n", np);
+		goto out_unprep_clk;
+	}
+
+	ticks_per_jiffy = (clk_get_rate(clk) + HZ/2) / HZ;
+	orion_clkevt.cpumask = cpumask_of(0);
+	orion_clkevt.irq = irq;
+	clockevents_config_and_register(&orion_clkevt, rate,
+					ORION_ONESHOT_MIN, ORION_ONESHOT_MAX);
+
+
+	orion_delay_timer_init(rate);
+
+	return 0;
+
+out_unprep_clk:
+	clk_disable_unprepare(clk);
+	return ret;
+}
+TIMER_OF_DECLARE(orion_timer, "marvell,orion-timer", orion_timer_init);