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/clocksource/timer-goldfish.c b/drivers/clocksource/timer-goldfish.c
new file mode 100644
index 000000000..3fe12c7ee
--- /dev/null
+++ b/drivers/clocksource/timer-goldfish.c
@@ -0,0 +1,153 @@
+// SPDX-License-Identifier: GPL-2.0
+
+#include <linux/interrupt.h>
+#include <linux/ioport.h>
+#include <linux/clocksource.h>
+#include <linux/clockchips.h>
+#include <linux/module.h>
+#include <linux/slab.h>
+#include <linux/goldfish.h>
+#include <clocksource/timer-goldfish.h>
+
+struct goldfish_timer {
+	struct clocksource cs;
+	struct clock_event_device ced;
+	struct resource res;
+	void __iomem *base;
+};
+
+static struct goldfish_timer *ced_to_gf(struct clock_event_device *ced)
+{
+	return container_of(ced, struct goldfish_timer, ced);
+}
+
+static struct goldfish_timer *cs_to_gf(struct clocksource *cs)
+{
+	return container_of(cs, struct goldfish_timer, cs);
+}
+
+static u64 goldfish_timer_read(struct clocksource *cs)
+{
+	struct goldfish_timer *timerdrv = cs_to_gf(cs);
+	void __iomem *base = timerdrv->base;
+	u32 time_low, time_high;
+	u64 ticks;
+
+	/*
+	 * time_low: get low bits of current time and update time_high
+	 * time_high: get high bits of time at last time_low read
+	 */
+	time_low = gf_ioread32(base + TIMER_TIME_LOW);
+	time_high = gf_ioread32(base + TIMER_TIME_HIGH);
+
+	ticks = ((u64)time_high << 32) | time_low;
+
+	return ticks;
+}
+
+static int goldfish_timer_set_oneshot(struct clock_event_device *evt)
+{
+	struct goldfish_timer *timerdrv = ced_to_gf(evt);
+	void __iomem *base = timerdrv->base;
+
+	gf_iowrite32(0, base + TIMER_ALARM_HIGH);
+	gf_iowrite32(0, base + TIMER_ALARM_LOW);
+	gf_iowrite32(1, base + TIMER_IRQ_ENABLED);
+
+	return 0;
+}
+
+static int goldfish_timer_shutdown(struct clock_event_device *evt)
+{
+	struct goldfish_timer *timerdrv = ced_to_gf(evt);
+	void __iomem *base = timerdrv->base;
+
+	gf_iowrite32(0, base + TIMER_IRQ_ENABLED);
+
+	return 0;
+}
+
+static int goldfish_timer_next_event(unsigned long delta,
+				     struct clock_event_device *evt)
+{
+	struct goldfish_timer *timerdrv = ced_to_gf(evt);
+	void __iomem *base = timerdrv->base;
+	u64 now;
+
+	now = goldfish_timer_read(&timerdrv->cs);
+
+	now += delta;
+
+	gf_iowrite32(upper_32_bits(now), base + TIMER_ALARM_HIGH);
+	gf_iowrite32(lower_32_bits(now), base + TIMER_ALARM_LOW);
+
+	return 0;
+}
+
+static irqreturn_t goldfish_timer_irq(int irq, void *dev_id)
+{
+	struct goldfish_timer *timerdrv = dev_id;
+	struct clock_event_device *evt = &timerdrv->ced;
+	void __iomem *base = timerdrv->base;
+
+	gf_iowrite32(1, base + TIMER_CLEAR_INTERRUPT);
+
+	evt->event_handler(evt);
+
+	return IRQ_HANDLED;
+}
+
+int __init goldfish_timer_init(int irq, void __iomem *base)
+{
+	struct goldfish_timer *timerdrv;
+	int ret;
+
+	timerdrv = kzalloc_obj(*timerdrv);
+	if (!timerdrv)
+		return -ENOMEM;
+
+	timerdrv->base = base;
+
+	timerdrv->ced = (struct clock_event_device){
+		.name			= "goldfish_timer",
+		.features		= CLOCK_EVT_FEAT_ONESHOT,
+		.set_state_shutdown	= goldfish_timer_shutdown,
+		.set_state_oneshot      = goldfish_timer_set_oneshot,
+		.set_next_event		= goldfish_timer_next_event,
+	};
+
+	timerdrv->res = (struct resource){
+		.name  = "goldfish_timer",
+		.start = (unsigned long)base,
+		.end   = (unsigned long)base + 0xfff,
+	};
+
+	ret = request_resource(&iomem_resource, &timerdrv->res);
+	if (ret) {
+		pr_err("Cannot allocate '%s' resource\n", timerdrv->res.name);
+		return ret;
+	}
+
+	timerdrv->cs = (struct clocksource){
+		.name		= "goldfish_timer",
+		.rating		= 400,
+		.read		= goldfish_timer_read,
+		.mask		= CLOCKSOURCE_MASK(64),
+		.flags		= 0,
+		.max_idle_ns	= LONG_MAX,
+	};
+
+	clocksource_register_hz(&timerdrv->cs, NSEC_PER_SEC);
+
+	ret = request_irq(irq, goldfish_timer_irq, IRQF_TIMER,
+			  "goldfish_timer", timerdrv);
+	if (ret) {
+		pr_err("Couldn't register goldfish-timer interrupt\n");
+		return ret;
+	}
+
+	clockevents_config_and_register(&timerdrv->ced, NSEC_PER_SEC,
+					1, 0xffffffff);
+
+	return 0;
+}