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, 278 insertions, 0 deletions

diff --git a/drivers/rtc/rtc-xgene.c b/drivers/rtc/rtc-xgene.c
new file mode 100644
index 000000000..6660b664e
--- /dev/null
+++ b/drivers/rtc/rtc-xgene.c
@@ -0,0 +1,278 @@
+// SPDX-License-Identifier: GPL-2.0+
+/*
+ * APM X-Gene SoC Real Time Clock Driver
+ *
+ * Copyright (c) 2014, Applied Micro Circuits Corporation
+ * Author: Rameshwar Prasad Sahu <rsahu@apm.com>
+ *         Loc Ho <lho@apm.com>
+ */
+
+#include <linux/clk.h>
+#include <linux/delay.h>
+#include <linux/init.h>
+#include <linux/io.h>
+#include <linux/module.h>
+#include <linux/of.h>
+#include <linux/platform_device.h>
+#include <linux/rtc.h>
+#include <linux/slab.h>
+
+/* RTC CSR Registers */
+#define RTC_CCVR		0x00
+#define RTC_CMR			0x04
+#define RTC_CLR			0x08
+#define RTC_CCR			0x0C
+#define  RTC_CCR_IE		BIT(0)
+#define  RTC_CCR_MASK		BIT(1)
+#define  RTC_CCR_EN		BIT(2)
+#define  RTC_CCR_WEN		BIT(3)
+#define RTC_STAT		0x10
+#define  RTC_STAT_BIT		BIT(0)
+#define RTC_RSTAT		0x14
+#define RTC_EOI			0x18
+#define RTC_VER			0x1C
+
+struct xgene_rtc_dev {
+	struct rtc_device *rtc;
+	void __iomem *csr_base;
+	struct clk *clk;
+	unsigned int irq_wake;
+	unsigned int irq_enabled;
+};
+
+static int xgene_rtc_read_time(struct device *dev, struct rtc_time *tm)
+{
+	struct xgene_rtc_dev *pdata = dev_get_drvdata(dev);
+
+	rtc_time64_to_tm(readl(pdata->csr_base + RTC_CCVR), tm);
+	return 0;
+}
+
+static int xgene_rtc_set_time(struct device *dev, struct rtc_time *tm)
+{
+	struct xgene_rtc_dev *pdata = dev_get_drvdata(dev);
+
+	/*
+	 * NOTE: After the following write, the RTC_CCVR is only reflected
+	 *       after the update cycle of 1 seconds.
+	 */
+	writel((u32)rtc_tm_to_time64(tm), pdata->csr_base + RTC_CLR);
+	readl(pdata->csr_base + RTC_CLR); /* Force a barrier */
+
+	return 0;
+}
+
+static int xgene_rtc_read_alarm(struct device *dev, struct rtc_wkalrm *alrm)
+{
+	struct xgene_rtc_dev *pdata = dev_get_drvdata(dev);
+
+	/* If possible, CMR should be read here */
+	rtc_time64_to_tm(0, &alrm->time);
+	alrm->enabled = readl(pdata->csr_base + RTC_CCR) & RTC_CCR_IE;
+
+	return 0;
+}
+
+static int xgene_rtc_alarm_irq_enable(struct device *dev, u32 enabled)
+{
+	struct xgene_rtc_dev *pdata = dev_get_drvdata(dev);
+	u32 ccr;
+
+	ccr = readl(pdata->csr_base + RTC_CCR);
+	if (enabled) {
+		ccr &= ~RTC_CCR_MASK;
+		ccr |= RTC_CCR_IE;
+	} else {
+		ccr &= ~RTC_CCR_IE;
+		ccr |= RTC_CCR_MASK;
+	}
+	writel(ccr, pdata->csr_base + RTC_CCR);
+
+	return 0;
+}
+
+static int xgene_rtc_alarm_irq_enabled(struct device *dev)
+{
+	struct xgene_rtc_dev *pdata = dev_get_drvdata(dev);
+
+	return readl(pdata->csr_base + RTC_CCR) & RTC_CCR_IE ? 1 : 0;
+}
+
+static int xgene_rtc_set_alarm(struct device *dev, struct rtc_wkalrm *alrm)
+{
+	struct xgene_rtc_dev *pdata = dev_get_drvdata(dev);
+
+	writel((u32)rtc_tm_to_time64(&alrm->time), pdata->csr_base + RTC_CMR);
+
+	xgene_rtc_alarm_irq_enable(dev, alrm->enabled);
+
+	return 0;
+}
+
+static const struct rtc_class_ops xgene_rtc_ops = {
+	.read_time	= xgene_rtc_read_time,
+	.set_time	= xgene_rtc_set_time,
+	.read_alarm	= xgene_rtc_read_alarm,
+	.set_alarm	= xgene_rtc_set_alarm,
+	.alarm_irq_enable = xgene_rtc_alarm_irq_enable,
+};
+
+static irqreturn_t xgene_rtc_interrupt(int irq, void *id)
+{
+	struct xgene_rtc_dev *pdata = id;
+
+	/* Check if interrupt asserted */
+	if (!(readl(pdata->csr_base + RTC_STAT) & RTC_STAT_BIT))
+		return IRQ_NONE;
+
+	/* Clear interrupt */
+	readl(pdata->csr_base + RTC_EOI);
+
+	rtc_update_irq(pdata->rtc, 1, RTC_IRQF | RTC_AF);
+
+	return IRQ_HANDLED;
+}
+
+static int xgene_rtc_probe(struct platform_device *pdev)
+{
+	struct xgene_rtc_dev *pdata;
+	int ret;
+	int irq;
+
+	pdata = devm_kzalloc(&pdev->dev, sizeof(*pdata), GFP_KERNEL);
+	if (!pdata)
+		return -ENOMEM;
+	platform_set_drvdata(pdev, pdata);
+
+	pdata->csr_base = devm_platform_ioremap_resource(pdev, 0);
+	if (IS_ERR(pdata->csr_base))
+		return PTR_ERR(pdata->csr_base);
+
+	pdata->rtc = devm_rtc_allocate_device(&pdev->dev);
+	if (IS_ERR(pdata->rtc))
+		return PTR_ERR(pdata->rtc);
+
+	irq = platform_get_irq(pdev, 0);
+	if (irq < 0)
+		return irq;
+	ret = devm_request_irq(&pdev->dev, irq, xgene_rtc_interrupt, 0,
+			       dev_name(&pdev->dev), pdata);
+	if (ret) {
+		dev_err(&pdev->dev, "Could not request IRQ\n");
+		return ret;
+	}
+
+	pdata->clk = devm_clk_get(&pdev->dev, NULL);
+	if (IS_ERR(pdata->clk)) {
+		dev_err(&pdev->dev, "Couldn't get the clock for RTC\n");
+		return -ENODEV;
+	}
+	ret = clk_prepare_enable(pdata->clk);
+	if (ret)
+		return ret;
+
+	/* Turn on the clock and the crystal */
+	writel(RTC_CCR_EN, pdata->csr_base + RTC_CCR);
+
+	ret = device_init_wakeup(&pdev->dev, true);
+	if (ret) {
+		clk_disable_unprepare(pdata->clk);
+		return ret;
+	}
+
+	pdata->rtc->ops = &xgene_rtc_ops;
+	pdata->rtc->range_max = U32_MAX;
+
+	ret = devm_rtc_register_device(pdata->rtc);
+	if (ret) {
+		clk_disable_unprepare(pdata->clk);
+		return ret;
+	}
+
+	return 0;
+}
+
+static void xgene_rtc_remove(struct platform_device *pdev)
+{
+	struct xgene_rtc_dev *pdata = platform_get_drvdata(pdev);
+
+	xgene_rtc_alarm_irq_enable(&pdev->dev, 0);
+	device_init_wakeup(&pdev->dev, false);
+	clk_disable_unprepare(pdata->clk);
+}
+
+static int __maybe_unused xgene_rtc_suspend(struct device *dev)
+{
+	struct platform_device *pdev = to_platform_device(dev);
+	struct xgene_rtc_dev *pdata = platform_get_drvdata(pdev);
+	int irq;
+
+	irq = platform_get_irq(pdev, 0);
+
+	/*
+	 * If this RTC alarm will be used for waking the system up,
+	 * don't disable it of course. Else we just disable the alarm
+	 * and await suspension.
+	 */
+	if (device_may_wakeup(&pdev->dev)) {
+		if (!enable_irq_wake(irq))
+			pdata->irq_wake = 1;
+	} else {
+		pdata->irq_enabled = xgene_rtc_alarm_irq_enabled(dev);
+		xgene_rtc_alarm_irq_enable(dev, 0);
+		clk_disable_unprepare(pdata->clk);
+	}
+	return 0;
+}
+
+static int __maybe_unused xgene_rtc_resume(struct device *dev)
+{
+	struct platform_device *pdev = to_platform_device(dev);
+	struct xgene_rtc_dev *pdata = platform_get_drvdata(pdev);
+	int irq;
+	int rc;
+
+	irq = platform_get_irq(pdev, 0);
+
+	if (device_may_wakeup(&pdev->dev)) {
+		if (pdata->irq_wake) {
+			disable_irq_wake(irq);
+			pdata->irq_wake = 0;
+		}
+	} else {
+		rc = clk_prepare_enable(pdata->clk);
+		if (rc) {
+			dev_err(dev, "Unable to enable clock error %d\n", rc);
+			return rc;
+		}
+		xgene_rtc_alarm_irq_enable(dev, pdata->irq_enabled);
+	}
+
+	return 0;
+}
+
+static SIMPLE_DEV_PM_OPS(xgene_rtc_pm_ops, xgene_rtc_suspend, xgene_rtc_resume);
+
+#ifdef CONFIG_OF
+static const struct of_device_id xgene_rtc_of_match[] = {
+	{.compatible = "apm,xgene-rtc" },
+	{ }
+};
+MODULE_DEVICE_TABLE(of, xgene_rtc_of_match);
+#endif
+
+static struct platform_driver xgene_rtc_driver = {
+	.probe		= xgene_rtc_probe,
+	.remove		= xgene_rtc_remove,
+	.driver		= {
+		.name	= "xgene-rtc",
+		.pm = &xgene_rtc_pm_ops,
+		.of_match_table	= of_match_ptr(xgene_rtc_of_match),
+	},
+};
+
+module_platform_driver(xgene_rtc_driver);
+
+MODULE_DESCRIPTION("APM X-Gene SoC RTC driver");
+MODULE_AUTHOR("Rameshwar Sahu <rsahu@apm.com>");
+MODULE_LICENSE("GPL");