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, 246 insertions, 0 deletions

diff --git a/drivers/phy/qualcomm/phy-qcom-usb-ss.c b/drivers/phy/qualcomm/phy-qcom-usb-ss.c
new file mode 100644
index 000000000..a3a6d3ce7
--- /dev/null
+++ b/drivers/phy/qualcomm/phy-qcom-usb-ss.c
@@ -0,0 +1,246 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * Copyright (c) 2012-2014,2017 The Linux Foundation. All rights reserved.
+ * Copyright (c) 2018-2020, Linaro Limited
+ */
+
+#include <linux/clk.h>
+#include <linux/delay.h>
+#include <linux/err.h>
+#include <linux/io.h>
+#include <linux/kernel.h>
+#include <linux/module.h>
+#include <linux/of.h>
+#include <linux/phy/phy.h>
+#include <linux/platform_device.h>
+#include <linux/regulator/consumer.h>
+#include <linux/reset.h>
+#include <linux/slab.h>
+
+#define PHY_CTRL0			0x6C
+#define PHY_CTRL1			0x70
+#define PHY_CTRL2			0x74
+#define PHY_CTRL4			0x7C
+
+/* PHY_CTRL bits */
+#define REF_PHY_EN			BIT(0)
+#define LANE0_PWR_ON			BIT(2)
+#define SWI_PCS_CLK_SEL			BIT(4)
+#define TST_PWR_DOWN			BIT(4)
+#define PHY_RESET			BIT(7)
+
+#define NUM_BULK_CLKS			3
+#define NUM_BULK_REGS			2
+
+struct ssphy_priv {
+	void __iomem *base;
+	struct device *dev;
+	struct reset_control *reset_com;
+	struct reset_control *reset_phy;
+	struct regulator_bulk_data regs[NUM_BULK_REGS];
+	struct clk_bulk_data clks[NUM_BULK_CLKS];
+	enum phy_mode mode;
+};
+
+static inline void qcom_ssphy_updatel(void __iomem *addr, u32 mask, u32 val)
+{
+	writel((readl(addr) & ~mask) | val, addr);
+}
+
+static int qcom_ssphy_do_reset(struct ssphy_priv *priv)
+{
+	int ret;
+
+	if (!priv->reset_com) {
+		qcom_ssphy_updatel(priv->base + PHY_CTRL1, PHY_RESET,
+				   PHY_RESET);
+		usleep_range(10, 20);
+		qcom_ssphy_updatel(priv->base + PHY_CTRL1, PHY_RESET, 0);
+	} else {
+		ret = reset_control_assert(priv->reset_com);
+		if (ret) {
+			dev_err(priv->dev, "Failed to assert reset com\n");
+			return ret;
+		}
+
+		ret = reset_control_assert(priv->reset_phy);
+		if (ret) {
+			dev_err(priv->dev, "Failed to assert reset phy\n");
+			return ret;
+		}
+
+		usleep_range(10, 20);
+
+		ret = reset_control_deassert(priv->reset_com);
+		if (ret) {
+			dev_err(priv->dev, "Failed to deassert reset com\n");
+			return ret;
+		}
+
+		ret = reset_control_deassert(priv->reset_phy);
+		if (ret) {
+			dev_err(priv->dev, "Failed to deassert reset phy\n");
+			return ret;
+		}
+	}
+
+	return 0;
+}
+
+static int qcom_ssphy_power_on(struct phy *phy)
+{
+	struct ssphy_priv *priv = phy_get_drvdata(phy);
+	int ret;
+
+	ret = regulator_bulk_enable(NUM_BULK_REGS, priv->regs);
+	if (ret)
+		return ret;
+
+	ret = clk_bulk_prepare_enable(NUM_BULK_CLKS, priv->clks);
+	if (ret)
+		goto err_disable_regulator;
+
+	ret = qcom_ssphy_do_reset(priv);
+	if (ret)
+		goto err_disable_clock;
+
+	writeb(SWI_PCS_CLK_SEL, priv->base + PHY_CTRL0);
+	qcom_ssphy_updatel(priv->base + PHY_CTRL4, LANE0_PWR_ON, LANE0_PWR_ON);
+	qcom_ssphy_updatel(priv->base + PHY_CTRL2, REF_PHY_EN, REF_PHY_EN);
+	qcom_ssphy_updatel(priv->base + PHY_CTRL4, TST_PWR_DOWN, 0);
+
+	return 0;
+err_disable_clock:
+	clk_bulk_disable_unprepare(NUM_BULK_CLKS, priv->clks);
+err_disable_regulator:
+	regulator_bulk_disable(NUM_BULK_REGS, priv->regs);
+
+	return ret;
+}
+
+static int qcom_ssphy_power_off(struct phy *phy)
+{
+	struct ssphy_priv *priv = phy_get_drvdata(phy);
+
+	qcom_ssphy_updatel(priv->base + PHY_CTRL4, LANE0_PWR_ON, 0);
+	qcom_ssphy_updatel(priv->base + PHY_CTRL2, REF_PHY_EN, 0);
+	qcom_ssphy_updatel(priv->base + PHY_CTRL4, TST_PWR_DOWN, TST_PWR_DOWN);
+
+	clk_bulk_disable_unprepare(NUM_BULK_CLKS, priv->clks);
+	regulator_bulk_disable(NUM_BULK_REGS, priv->regs);
+
+	return 0;
+}
+
+static int qcom_ssphy_init_clock(struct ssphy_priv *priv)
+{
+	priv->clks[0].id = "ref";
+	priv->clks[1].id = "ahb";
+	priv->clks[2].id = "pipe";
+
+	return devm_clk_bulk_get(priv->dev, NUM_BULK_CLKS, priv->clks);
+}
+
+static int qcom_ssphy_init_regulator(struct ssphy_priv *priv)
+{
+	int ret;
+
+	priv->regs[0].supply = "vdd";
+	priv->regs[1].supply = "vdda1p8";
+	ret = devm_regulator_bulk_get(priv->dev, NUM_BULK_REGS, priv->regs);
+	if (ret) {
+		if (ret != -EPROBE_DEFER)
+			dev_err(priv->dev, "Failed to get regulators\n");
+		return ret;
+	}
+
+	return ret;
+}
+
+static int qcom_ssphy_init_reset(struct ssphy_priv *priv)
+{
+	priv->reset_com = devm_reset_control_get_optional_exclusive(priv->dev, "com");
+	if (IS_ERR(priv->reset_com)) {
+		dev_err(priv->dev, "Failed to get reset control com\n");
+		return PTR_ERR(priv->reset_com);
+	}
+
+	if (priv->reset_com) {
+		/* if reset_com is present, reset_phy is no longer optional */
+		priv->reset_phy = devm_reset_control_get_exclusive(priv->dev, "phy");
+		if (IS_ERR(priv->reset_phy)) {
+			dev_err(priv->dev, "Failed to get reset control phy\n");
+			return PTR_ERR(priv->reset_phy);
+		}
+	}
+
+	return 0;
+}
+
+static const struct phy_ops qcom_ssphy_ops = {
+	.power_off = qcom_ssphy_power_off,
+	.power_on = qcom_ssphy_power_on,
+	.owner = THIS_MODULE,
+};
+
+static int qcom_ssphy_probe(struct platform_device *pdev)
+{
+	struct device *dev = &pdev->dev;
+	struct phy_provider *provider;
+	struct ssphy_priv *priv;
+	struct phy *phy;
+	int ret;
+
+	priv = devm_kzalloc(dev, sizeof(struct ssphy_priv), GFP_KERNEL);
+	if (!priv)
+		return -ENOMEM;
+
+	priv->dev = dev;
+	priv->mode = PHY_MODE_INVALID;
+
+	priv->base = devm_platform_ioremap_resource(pdev, 0);
+	if (IS_ERR(priv->base))
+		return PTR_ERR(priv->base);
+
+	ret = qcom_ssphy_init_clock(priv);
+	if (ret)
+		return ret;
+
+	ret = qcom_ssphy_init_reset(priv);
+	if (ret)
+		return ret;
+
+	ret = qcom_ssphy_init_regulator(priv);
+	if (ret)
+		return ret;
+
+	phy = devm_phy_create(dev, dev->of_node, &qcom_ssphy_ops);
+	if (IS_ERR(phy)) {
+		dev_err(dev, "Failed to create the SS phy\n");
+		return PTR_ERR(phy);
+	}
+
+	phy_set_drvdata(phy, priv);
+
+	provider = devm_of_phy_provider_register(dev, of_phy_simple_xlate);
+
+	return PTR_ERR_OR_ZERO(provider);
+}
+
+static const struct of_device_id qcom_ssphy_match[] = {
+	{ .compatible = "qcom,usb-ss-28nm-phy", },
+	{ },
+};
+MODULE_DEVICE_TABLE(of, qcom_ssphy_match);
+
+static struct platform_driver qcom_ssphy_driver = {
+	.probe		= qcom_ssphy_probe,
+	.driver = {
+		.name	= "qcom-usb-ssphy",
+		.of_match_table = qcom_ssphy_match,
+	},
+};
+module_platform_driver(qcom_ssphy_driver);
+
+MODULE_DESCRIPTION("Qualcomm SuperSpeed USB PHY driver");
+MODULE_LICENSE("GPL v2");