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, 237 insertions, 0 deletions

diff --git a/drivers/rtc/rtc-rtd119x.c b/drivers/rtc/rtc-rtd119x.c
new file mode 100644
index 000000000..85a138db8
--- /dev/null
+++ b/drivers/rtc/rtc-rtd119x.c
@@ -0,0 +1,237 @@
+// SPDX-License-Identifier: GPL-2.0+
+/*
+ * Realtek RTD129x RTC
+ *
+ * Copyright (c) 2017 Andreas Färber
+ */
+
+#include <linux/clk.h>
+#include <linux/io.h>
+#include <linux/module.h>
+#include <linux/of.h>
+#include <linux/of_address.h>
+#include <linux/platform_device.h>
+#include <linux/rtc.h>
+#include <linux/spinlock.h>
+
+#define RTD_RTCSEC		0x00
+#define RTD_RTCMIN		0x04
+#define RTD_RTCHR		0x08
+#define RTD_RTCDATE1		0x0c
+#define RTD_RTCDATE2		0x10
+#define RTD_RTCACR		0x28
+#define RTD_RTCEN		0x2c
+#define RTD_RTCCR		0x30
+
+#define RTD_RTCSEC_RTCSEC_MASK		0x7f
+
+#define RTD_RTCMIN_RTCMIN_MASK		0x3f
+
+#define RTD_RTCHR_RTCHR_MASK		0x1f
+
+#define RTD_RTCDATE1_RTCDATE1_MASK	0xff
+
+#define RTD_RTCDATE2_RTCDATE2_MASK	0x7f
+
+#define RTD_RTCACR_RTCPWR		BIT(7)
+
+#define RTD_RTCEN_RTCEN_MASK		0xff
+
+#define RTD_RTCCR_RTCRST		BIT(6)
+
+struct rtd119x_rtc {
+	void __iomem *base;
+	struct clk *clk;
+	struct rtc_device *rtcdev;
+	unsigned int base_year;
+};
+
+static inline int rtd119x_rtc_days_in_year(int year)
+{
+	return 365 + (is_leap_year(year) ? 1 : 0);
+}
+
+static void rtd119x_rtc_reset(struct device *dev)
+{
+	struct rtd119x_rtc *data = dev_get_drvdata(dev);
+	u32 val;
+
+	val = readl_relaxed(data->base + RTD_RTCCR);
+	val |= RTD_RTCCR_RTCRST;
+	writel_relaxed(val, data->base + RTD_RTCCR);
+
+	val &= ~RTD_RTCCR_RTCRST;
+	writel(val, data->base + RTD_RTCCR);
+}
+
+static void rtd119x_rtc_set_enabled(struct device *dev, bool enable)
+{
+	struct rtd119x_rtc *data = dev_get_drvdata(dev);
+	u32 val;
+
+	val = readl_relaxed(data->base + RTD_RTCEN);
+	if (enable) {
+		if ((val & RTD_RTCEN_RTCEN_MASK) == 0x5a)
+			return;
+		writel_relaxed(0x5a, data->base + RTD_RTCEN);
+	} else {
+		writel_relaxed(0, data->base + RTD_RTCEN);
+	}
+}
+
+static int rtd119x_rtc_read_time(struct device *dev, struct rtc_time *tm)
+{
+	struct rtd119x_rtc *data = dev_get_drvdata(dev);
+	s32 day;
+	u32 sec;
+	unsigned int year;
+	int tries = 0;
+
+	while (true) {
+		tm->tm_sec = (readl_relaxed(data->base + RTD_RTCSEC) & RTD_RTCSEC_RTCSEC_MASK) >> 1;
+		tm->tm_min  = readl_relaxed(data->base + RTD_RTCMIN) & RTD_RTCMIN_RTCMIN_MASK;
+		tm->tm_hour = readl_relaxed(data->base + RTD_RTCHR) & RTD_RTCHR_RTCHR_MASK;
+		day  =  readl_relaxed(data->base + RTD_RTCDATE1) & RTD_RTCDATE1_RTCDATE1_MASK;
+		day |= (readl_relaxed(data->base + RTD_RTCDATE2) & RTD_RTCDATE2_RTCDATE2_MASK) << 8;
+		sec  = (readl_relaxed(data->base + RTD_RTCSEC) & RTD_RTCSEC_RTCSEC_MASK) >> 1;
+		tries++;
+
+		if (sec == tm->tm_sec)
+			break;
+
+		if (tries >= 3)
+			return -EINVAL;
+	}
+	if (tries > 1)
+		dev_dbg(dev, "%s: needed %i tries\n", __func__, tries);
+
+	year = data->base_year;
+	while (day >= rtd119x_rtc_days_in_year(year)) {
+		day -= rtd119x_rtc_days_in_year(year);
+		year++;
+	}
+	tm->tm_year = year - 1900;
+	tm->tm_yday = day;
+
+	tm->tm_mon = 0;
+	while (day >= rtc_month_days(tm->tm_mon, year)) {
+		day -= rtc_month_days(tm->tm_mon, year);
+		tm->tm_mon++;
+	}
+	tm->tm_mday = day + 1;
+
+	return 0;
+}
+
+static int rtd119x_rtc_set_time(struct device *dev, struct rtc_time *tm)
+{
+	struct rtd119x_rtc *data = dev_get_drvdata(dev);
+	unsigned int day;
+	int i;
+
+	if (1900 + tm->tm_year < data->base_year)
+		return -EINVAL;
+
+	day = 0;
+	for (i = data->base_year; i < 1900 + tm->tm_year; i++)
+		day += rtd119x_rtc_days_in_year(i);
+
+	day += tm->tm_yday;
+	if (day > 0x7fff)
+		return -EINVAL;
+
+	rtd119x_rtc_set_enabled(dev, false);
+
+	writel_relaxed((tm->tm_sec << 1) & RTD_RTCSEC_RTCSEC_MASK, data->base + RTD_RTCSEC);
+	writel_relaxed(tm->tm_min & RTD_RTCMIN_RTCMIN_MASK, data->base + RTD_RTCMIN);
+	writel_relaxed(tm->tm_hour & RTD_RTCHR_RTCHR_MASK, data->base + RTD_RTCHR);
+	writel_relaxed(day & RTD_RTCDATE1_RTCDATE1_MASK, data->base + RTD_RTCDATE1);
+	writel_relaxed((day >> 8) & RTD_RTCDATE2_RTCDATE2_MASK, data->base + RTD_RTCDATE2);
+
+	rtd119x_rtc_set_enabled(dev, true);
+
+	return 0;
+}
+
+static const struct rtc_class_ops rtd119x_rtc_ops = {
+	.read_time	= rtd119x_rtc_read_time,
+	.set_time	= rtd119x_rtc_set_time,
+};
+
+static const struct of_device_id rtd119x_rtc_dt_ids[] = {
+	 { .compatible = "realtek,rtd1295-rtc" },
+	 { }
+};
+
+static int rtd119x_rtc_probe(struct platform_device *pdev)
+{
+	struct rtd119x_rtc *data;
+	u32 val;
+	int ret;
+
+	data = devm_kzalloc(&pdev->dev, sizeof(*data), GFP_KERNEL);
+	if (!data)
+		return -ENOMEM;
+
+	platform_set_drvdata(pdev, data);
+	data->base_year = 2014;
+
+	data->base = devm_platform_ioremap_resource(pdev, 0);
+	if (IS_ERR(data->base))
+		return PTR_ERR(data->base);
+
+	data->clk = of_clk_get(pdev->dev.of_node, 0);
+	if (IS_ERR(data->clk))
+		return PTR_ERR(data->clk);
+
+	ret = clk_prepare_enable(data->clk);
+	if (ret) {
+		clk_put(data->clk);
+		return ret;
+	}
+
+	val = readl_relaxed(data->base + RTD_RTCACR);
+	if (!(val & RTD_RTCACR_RTCPWR)) {
+		writel_relaxed(RTD_RTCACR_RTCPWR, data->base + RTD_RTCACR);
+
+		rtd119x_rtc_reset(&pdev->dev);
+
+		writel_relaxed(0, data->base + RTD_RTCMIN);
+		writel_relaxed(0, data->base + RTD_RTCHR);
+		writel_relaxed(0, data->base + RTD_RTCDATE1);
+		writel_relaxed(0, data->base + RTD_RTCDATE2);
+	}
+
+	rtd119x_rtc_set_enabled(&pdev->dev, true);
+
+	data->rtcdev = devm_rtc_device_register(&pdev->dev, "rtc",
+						&rtd119x_rtc_ops, THIS_MODULE);
+	if (IS_ERR(data->rtcdev)) {
+		dev_err(&pdev->dev, "failed to register rtc device");
+		clk_disable_unprepare(data->clk);
+		clk_put(data->clk);
+		return PTR_ERR(data->rtcdev);
+	}
+
+	return 0;
+}
+
+static void rtd119x_rtc_remove(struct platform_device *pdev)
+{
+	struct rtd119x_rtc *data = platform_get_drvdata(pdev);
+
+	rtd119x_rtc_set_enabled(&pdev->dev, false);
+
+	clk_disable_unprepare(data->clk);
+	clk_put(data->clk);
+}
+
+static struct platform_driver rtd119x_rtc_driver = {
+	.probe = rtd119x_rtc_probe,
+	.remove = rtd119x_rtc_remove,
+	.driver = {
+		.name = "rtd1295-rtc",
+		.of_match_table	= rtd119x_rtc_dt_ids,
+	},
+};
+builtin_platform_driver(rtd119x_rtc_driver);