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, 174 insertions, 0 deletions

diff --git a/drivers/rtc/rtc-ds1216.c b/drivers/rtc/rtc-ds1216.c
new file mode 100644
index 000000000..7eeb3f359
--- /dev/null
+++ b/drivers/rtc/rtc-ds1216.c
@@ -0,0 +1,174 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ * Dallas DS1216 RTC driver
+ *
+ * Copyright (c) 2007 Thomas Bogendoerfer
+ *
+ */
+
+#include <linux/module.h>
+#include <linux/rtc.h>
+#include <linux/platform_device.h>
+#include <linux/bcd.h>
+#include <linux/slab.h>
+
+struct ds1216_regs {
+	u8 tsec;
+	u8 sec;
+	u8 min;
+	u8 hour;
+	u8 wday;
+	u8 mday;
+	u8 month;
+	u8 year;
+};
+
+#define DS1216_HOUR_1224	(1 << 7)
+#define DS1216_HOUR_AMPM	(1 << 5)
+
+struct ds1216_priv {
+	struct rtc_device *rtc;
+	void __iomem *ioaddr;
+};
+
+static const u8 magic[] = {
+	0xc5, 0x3a, 0xa3, 0x5c, 0xc5, 0x3a, 0xa3, 0x5c
+};
+
+/*
+ * Read the 64 bit we'd like to have - It a series
+ * of 64 bits showing up in the LSB of the base register.
+ *
+ */
+static void ds1216_read(u8 __iomem *ioaddr, u8 *buf)
+{
+	unsigned char c;
+	int i, j;
+
+	for (i = 0; i < 8; i++) {
+		c = 0;
+		for (j = 0; j < 8; j++)
+			c |= (readb(ioaddr) & 0x1) << j;
+		buf[i] = c;
+	}
+}
+
+static void ds1216_write(u8 __iomem *ioaddr, const u8 *buf)
+{
+	unsigned char c;
+	int i, j;
+
+	for (i = 0; i < 8; i++) {
+		c = buf[i];
+		for (j = 0; j < 8; j++) {
+			writeb(c, ioaddr);
+			c = c >> 1;
+		}
+	}
+}
+
+static void ds1216_switch_ds_to_clock(u8 __iomem *ioaddr)
+{
+	/* Reset magic pointer */
+	readb(ioaddr);
+	/* Write 64 bit magic to DS1216 */
+	ds1216_write(ioaddr, magic);
+}
+
+static int ds1216_rtc_read_time(struct device *dev, struct rtc_time *tm)
+{
+	struct ds1216_priv *priv = dev_get_drvdata(dev);
+	struct ds1216_regs regs;
+
+	ds1216_switch_ds_to_clock(priv->ioaddr);
+	ds1216_read(priv->ioaddr, (u8 *)&regs);
+
+	tm->tm_sec = bcd2bin(regs.sec);
+	tm->tm_min = bcd2bin(regs.min);
+	if (regs.hour & DS1216_HOUR_1224) {
+		/* AM/PM mode */
+		tm->tm_hour = bcd2bin(regs.hour & 0x1f);
+		if (regs.hour & DS1216_HOUR_AMPM)
+			tm->tm_hour += 12;
+	} else
+		tm->tm_hour = bcd2bin(regs.hour & 0x3f);
+	tm->tm_wday = (regs.wday & 7) - 1;
+	tm->tm_mday = bcd2bin(regs.mday & 0x3f);
+	tm->tm_mon = bcd2bin(regs.month & 0x1f);
+	tm->tm_year = bcd2bin(regs.year);
+	if (tm->tm_year < 70)
+		tm->tm_year += 100;
+
+	return 0;
+}
+
+static int ds1216_rtc_set_time(struct device *dev, struct rtc_time *tm)
+{
+	struct ds1216_priv *priv = dev_get_drvdata(dev);
+	struct ds1216_regs regs;
+
+	ds1216_switch_ds_to_clock(priv->ioaddr);
+	ds1216_read(priv->ioaddr, (u8 *)&regs);
+
+	regs.tsec = 0; /* clear 0.1 and 0.01 seconds */
+	regs.sec = bin2bcd(tm->tm_sec);
+	regs.min = bin2bcd(tm->tm_min);
+	regs.hour &= DS1216_HOUR_1224;
+	if (regs.hour && tm->tm_hour > 12) {
+		regs.hour |= DS1216_HOUR_AMPM;
+		tm->tm_hour -= 12;
+	}
+	regs.hour |= bin2bcd(tm->tm_hour);
+	regs.wday &= ~7;
+	regs.wday |= tm->tm_wday;
+	regs.mday = bin2bcd(tm->tm_mday);
+	regs.month = bin2bcd(tm->tm_mon);
+	regs.year = bin2bcd(tm->tm_year % 100);
+
+	ds1216_switch_ds_to_clock(priv->ioaddr);
+	ds1216_write(priv->ioaddr, (u8 *)&regs);
+	return 0;
+}
+
+static const struct rtc_class_ops ds1216_rtc_ops = {
+	.read_time	= ds1216_rtc_read_time,
+	.set_time	= ds1216_rtc_set_time,
+};
+
+static int __init ds1216_rtc_probe(struct platform_device *pdev)
+{
+	struct ds1216_priv *priv;
+	u8 dummy[8];
+
+	priv = devm_kzalloc(&pdev->dev, sizeof(*priv), GFP_KERNEL);
+	if (!priv)
+		return -ENOMEM;
+
+	platform_set_drvdata(pdev, priv);
+
+	priv->ioaddr = devm_platform_ioremap_resource(pdev, 0);
+	if (IS_ERR(priv->ioaddr))
+		return PTR_ERR(priv->ioaddr);
+
+	priv->rtc = devm_rtc_device_register(&pdev->dev, "ds1216",
+					&ds1216_rtc_ops, THIS_MODULE);
+	if (IS_ERR(priv->rtc))
+		return PTR_ERR(priv->rtc);
+
+	/* dummy read to get clock into a known state */
+	ds1216_read(priv->ioaddr, dummy);
+	return 0;
+}
+
+static struct platform_driver ds1216_rtc_platform_driver = {
+	.driver		= {
+		.name	= "rtc-ds1216",
+	},
+};
+
+module_platform_driver_probe(ds1216_rtc_platform_driver, ds1216_rtc_probe);
+
+MODULE_AUTHOR("Thomas Bogendoerfer <tsbogend@alpha.franken.de>");
+MODULE_DESCRIPTION("DS1216 RTC driver");
+MODULE_LICENSE("GPL");
+MODULE_ALIAS("platform:rtc-ds1216");