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, 200 insertions, 0 deletions

diff --git a/drivers/soc/fsl/rcpm.c b/drivers/soc/fsl/rcpm.c
new file mode 100644
index 000000000..06bd94b29
--- /dev/null
+++ b/drivers/soc/fsl/rcpm.c
@@ -0,0 +1,200 @@
+// SPDX-License-Identifier: GPL-2.0
+//
+// rcpm.c - Freescale QorIQ RCPM driver
+//
+// Copyright 2019-2020 NXP
+//
+// Author: Ran Wang <ran.wang_1@nxp.com>
+
+#include <linux/init.h>
+#include <linux/module.h>
+#include <linux/platform_device.h>
+#include <linux/of_address.h>
+#include <linux/slab.h>
+#include <linux/suspend.h>
+#include <linux/kernel.h>
+#include <linux/acpi.h>
+
+#define RCPM_WAKEUP_CELL_MAX_SIZE	7
+
+struct rcpm {
+	unsigned int	wakeup_cells;
+	void __iomem	*ippdexpcr_base;
+	bool		little_endian;
+};
+
+#define  SCFG_SPARECR8	0x051c
+
+static void copy_ippdexpcr1_setting(u32 val)
+{
+	struct device_node *np;
+	void __iomem *regs;
+	u32 reg_val;
+
+	np = of_find_compatible_node(NULL, NULL, "fsl,ls1021a-scfg");
+	if (!np)
+		return;
+
+	regs = of_iomap(np, 0);
+	of_node_put(np);
+	if (!regs)
+		return;
+
+	reg_val = ioread32be(regs + SCFG_SPARECR8);
+	iowrite32be(val | reg_val, regs + SCFG_SPARECR8);
+
+	iounmap(regs);
+}
+
+/**
+ * rcpm_pm_prepare - performs device-level tasks associated with power
+ * management, such as programming related to the wakeup source control.
+ * @dev: Device to handle.
+ *
+ */
+static int rcpm_pm_prepare(struct device *dev)
+{
+	int i, ret, idx;
+	void __iomem *base;
+	struct wakeup_source	*ws;
+	struct rcpm		*rcpm;
+	struct device_node	*np = dev->of_node;
+	u32 value[RCPM_WAKEUP_CELL_MAX_SIZE + 1];
+	u32 setting[RCPM_WAKEUP_CELL_MAX_SIZE] = {0};
+
+	rcpm = dev_get_drvdata(dev);
+	if (!rcpm)
+		return -EINVAL;
+
+	base = rcpm->ippdexpcr_base;
+	idx = wakeup_sources_read_lock();
+
+	/* Begin with first registered wakeup source */
+	for_each_wakeup_source(ws) {
+
+		/* skip object which is not attached to device */
+		if (!ws->dev || !ws->dev->parent)
+			continue;
+
+		ret = device_property_read_u32_array(ws->dev->parent,
+				"fsl,rcpm-wakeup", value,
+				rcpm->wakeup_cells + 1);
+
+		if (ret)
+			continue;
+
+		/*
+		 * For DT mode, would handle devices with "fsl,rcpm-wakeup"
+		 * pointing to the current RCPM node.
+		 *
+		 * For ACPI mode, currently we assume there is only one
+		 * RCPM controller existing.
+		 */
+		if (is_of_node(dev->fwnode))
+			if (np->phandle != value[0])
+				continue;
+
+		/* Property "#fsl,rcpm-wakeup-cells" of rcpm node defines the
+		 * number of IPPDEXPCR register cells, and "fsl,rcpm-wakeup"
+		 * of wakeup source IP contains an integer array: <phandle to
+		 * RCPM node, IPPDEXPCR0 setting, IPPDEXPCR1 setting,
+		 * IPPDEXPCR2 setting, etc>.
+		 *
+		 * So we will go thought them to collect setting data.
+		 */
+		for (i = 0; i < rcpm->wakeup_cells; i++)
+			setting[i] |= value[i + 1];
+	}
+
+	wakeup_sources_read_unlock(idx);
+
+	/* Program all IPPDEXPCRn once */
+	for (i = 0; i < rcpm->wakeup_cells; i++) {
+		u32 tmp = setting[i];
+		void __iomem *address = base + i * 4;
+
+		if (!tmp)
+			continue;
+
+		/* We can only OR related bits */
+		if (rcpm->little_endian) {
+			tmp |= ioread32(address);
+			iowrite32(tmp, address);
+		} else {
+			tmp |= ioread32be(address);
+			iowrite32be(tmp, address);
+		}
+		/*
+		 * Workaround of errata A-008646 on SoC LS1021A:
+		 * There is a bug of register ippdexpcr1.
+		 * Reading configuration register RCPM_IPPDEXPCR1
+		 * always return zero. So save ippdexpcr1's value
+		 * to register SCFG_SPARECR8.And the value of
+		 * ippdexpcr1 will be read from SCFG_SPARECR8.
+		 */
+		if (dev_of_node(dev) && (i == 1))
+			if (of_device_is_compatible(np, "fsl,ls1021a-rcpm"))
+				copy_ippdexpcr1_setting(tmp);
+	}
+
+	return 0;
+}
+
+static const struct dev_pm_ops rcpm_pm_ops = {
+	.prepare =  rcpm_pm_prepare,
+};
+
+static int rcpm_probe(struct platform_device *pdev)
+{
+	struct device	*dev = &pdev->dev;
+	struct rcpm	*rcpm;
+	int ret;
+
+	rcpm = devm_kzalloc(dev, sizeof(*rcpm), GFP_KERNEL);
+	if (!rcpm)
+		return -ENOMEM;
+
+	rcpm->ippdexpcr_base = devm_platform_ioremap_resource(pdev, 0);
+	if (IS_ERR(rcpm->ippdexpcr_base)) {
+		ret =  PTR_ERR(rcpm->ippdexpcr_base);
+		return ret;
+	}
+
+	rcpm->little_endian = device_property_read_bool(
+			&pdev->dev, "little-endian");
+
+	ret = device_property_read_u32(&pdev->dev,
+			"#fsl,rcpm-wakeup-cells", &rcpm->wakeup_cells);
+	if (ret)
+		return ret;
+
+	dev_set_drvdata(&pdev->dev, rcpm);
+
+	return 0;
+}
+
+static const struct of_device_id rcpm_of_match[] = {
+	{ .compatible = "fsl,qoriq-rcpm-2.1+", },
+	{}
+};
+MODULE_DEVICE_TABLE(of, rcpm_of_match);
+
+#ifdef CONFIG_ACPI
+static const struct acpi_device_id rcpm_acpi_ids[] = {
+	{"NXP0015",},
+	{ }
+};
+MODULE_DEVICE_TABLE(acpi, rcpm_acpi_ids);
+#endif
+
+static struct platform_driver rcpm_driver = {
+	.driver = {
+		.name = "rcpm",
+		.of_match_table = rcpm_of_match,
+		.acpi_match_table = ACPI_PTR(rcpm_acpi_ids),
+		.pm	= &rcpm_pm_ops,
+	},
+	.probe = rcpm_probe,
+};
+
+module_platform_driver(rcpm_driver);