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, 344 insertions, 0 deletions

diff --git a/drivers/peci/cpu.c b/drivers/peci/cpu.c
new file mode 100644
index 000000000..c9fc06c9f
--- /dev/null
+++ b/drivers/peci/cpu.c
@@ -0,0 +1,344 @@
+// SPDX-License-Identifier: GPL-2.0-only
+// Copyright (c) 2021 Intel Corporation
+
+#include <linux/auxiliary_bus.h>
+#include <linux/module.h>
+#include <linux/peci.h>
+#include <linux/peci-cpu.h>
+#include <linux/slab.h>
+
+#include "internal.h"
+
+/**
+ * peci_temp_read() - read the maximum die temperature from PECI target device
+ * @device: PECI device to which request is going to be sent
+ * @temp_raw: where to store the read temperature
+ *
+ * It uses GetTemp PECI command.
+ *
+ * Return: 0 if succeeded, other values in case errors.
+ */
+int peci_temp_read(struct peci_device *device, s16 *temp_raw)
+{
+	struct peci_request *req;
+
+	req = peci_xfer_get_temp(device);
+	if (IS_ERR(req))
+		return PTR_ERR(req);
+
+	*temp_raw = peci_request_temp_read(req);
+
+	peci_request_free(req);
+
+	return 0;
+}
+EXPORT_SYMBOL_NS_GPL(peci_temp_read, "PECI_CPU");
+
+/**
+ * peci_pcs_read() - read PCS register
+ * @device: PECI device to which request is going to be sent
+ * @index: PCS index
+ * @param: PCS parameter
+ * @data: where to store the read data
+ *
+ * It uses RdPkgConfig PECI command.
+ *
+ * Return: 0 if succeeded, other values in case errors.
+ */
+int peci_pcs_read(struct peci_device *device, u8 index, u16 param, u32 *data)
+{
+	struct peci_request *req;
+	int ret;
+
+	req = peci_xfer_pkg_cfg_readl(device, index, param);
+	if (IS_ERR(req))
+		return PTR_ERR(req);
+
+	ret = peci_request_status(req);
+	if (ret)
+		goto out_req_free;
+
+	*data = peci_request_data_readl(req);
+out_req_free:
+	peci_request_free(req);
+
+	return ret;
+}
+EXPORT_SYMBOL_NS_GPL(peci_pcs_read, "PECI_CPU");
+
+/**
+ * peci_pci_local_read() - read 32-bit memory location using raw address
+ * @device: PECI device to which request is going to be sent
+ * @bus: bus
+ * @dev: device
+ * @func: function
+ * @reg: register
+ * @data: where to store the read data
+ *
+ * It uses RdPCIConfigLocal PECI command.
+ *
+ * Return: 0 if succeeded, other values in case errors.
+ */
+int peci_pci_local_read(struct peci_device *device, u8 bus, u8 dev, u8 func,
+			u16 reg, u32 *data)
+{
+	struct peci_request *req;
+	int ret;
+
+	req = peci_xfer_pci_cfg_local_readl(device, bus, dev, func, reg);
+	if (IS_ERR(req))
+		return PTR_ERR(req);
+
+	ret = peci_request_status(req);
+	if (ret)
+		goto out_req_free;
+
+	*data = peci_request_data_readl(req);
+out_req_free:
+	peci_request_free(req);
+
+	return ret;
+}
+EXPORT_SYMBOL_NS_GPL(peci_pci_local_read, "PECI_CPU");
+
+/**
+ * peci_ep_pci_local_read() - read 32-bit memory location using raw address
+ * @device: PECI device to which request is going to be sent
+ * @seg: PCI segment
+ * @bus: bus
+ * @dev: device
+ * @func: function
+ * @reg: register
+ * @data: where to store the read data
+ *
+ * Like &peci_pci_local_read, but it uses RdEndpointConfig PECI command.
+ *
+ * Return: 0 if succeeded, other values in case errors.
+ */
+int peci_ep_pci_local_read(struct peci_device *device, u8 seg,
+			   u8 bus, u8 dev, u8 func, u16 reg, u32 *data)
+{
+	struct peci_request *req;
+	int ret;
+
+	req = peci_xfer_ep_pci_cfg_local_readl(device, seg, bus, dev, func, reg);
+	if (IS_ERR(req))
+		return PTR_ERR(req);
+
+	ret = peci_request_status(req);
+	if (ret)
+		goto out_req_free;
+
+	*data = peci_request_data_readl(req);
+out_req_free:
+	peci_request_free(req);
+
+	return ret;
+}
+EXPORT_SYMBOL_NS_GPL(peci_ep_pci_local_read, "PECI_CPU");
+
+/**
+ * peci_mmio_read() - read 32-bit memory location using 64-bit bar offset address
+ * @device: PECI device to which request is going to be sent
+ * @bar: PCI bar
+ * @seg: PCI segment
+ * @bus: bus
+ * @dev: device
+ * @func: function
+ * @address: 64-bit MMIO address
+ * @data: where to store the read data
+ *
+ * It uses RdEndpointConfig PECI command.
+ *
+ * Return: 0 if succeeded, other values in case errors.
+ */
+int peci_mmio_read(struct peci_device *device, u8 bar, u8 seg,
+		   u8 bus, u8 dev, u8 func, u64 address, u32 *data)
+{
+	struct peci_request *req;
+	int ret;
+
+	req = peci_xfer_ep_mmio64_readl(device, bar, seg, bus, dev, func, address);
+	if (IS_ERR(req))
+		return PTR_ERR(req);
+
+	ret = peci_request_status(req);
+	if (ret)
+		goto out_req_free;
+
+	*data = peci_request_data_readl(req);
+out_req_free:
+	peci_request_free(req);
+
+	return ret;
+}
+EXPORT_SYMBOL_NS_GPL(peci_mmio_read, "PECI_CPU");
+
+static const char * const peci_adev_types[] = {
+	"cputemp",
+	"dimmtemp",
+};
+
+struct peci_cpu {
+	struct peci_device *device;
+	const struct peci_device_id *id;
+};
+
+static void adev_release(struct device *dev)
+{
+	struct auxiliary_device *adev = to_auxiliary_dev(dev);
+
+	kfree(adev->name);
+	kfree(adev);
+}
+
+static struct auxiliary_device *adev_alloc(struct peci_cpu *priv, int idx)
+{
+	struct peci_controller *controller = to_peci_controller(priv->device->dev.parent);
+	struct auxiliary_device *adev;
+	const char *name;
+	int ret;
+
+	adev = kzalloc_obj(*adev);
+	if (!adev)
+		return ERR_PTR(-ENOMEM);
+
+	name = kasprintf(GFP_KERNEL, "%s.%s", peci_adev_types[idx], (const char *)priv->id->data);
+	if (!name) {
+		ret = -ENOMEM;
+		goto free_adev;
+	}
+
+	adev->name = name;
+	adev->dev.parent = &priv->device->dev;
+	adev->dev.release = adev_release;
+	adev->id = (controller->id << 16) | (priv->device->addr);
+
+	ret = auxiliary_device_init(adev);
+	if (ret)
+		goto free_name;
+
+	return adev;
+
+free_name:
+	kfree(name);
+free_adev:
+	kfree(adev);
+	return ERR_PTR(ret);
+}
+
+static void unregister_adev(void *_adev)
+{
+	struct auxiliary_device *adev = _adev;
+
+	auxiliary_device_delete(adev);
+	auxiliary_device_uninit(adev);
+}
+
+static int devm_adev_add(struct device *dev, int idx)
+{
+	struct peci_cpu *priv = dev_get_drvdata(dev);
+	struct auxiliary_device *adev;
+	int ret;
+
+	adev = adev_alloc(priv, idx);
+	if (IS_ERR(adev))
+		return PTR_ERR(adev);
+
+	ret = auxiliary_device_add(adev);
+	if (ret) {
+		auxiliary_device_uninit(adev);
+		return ret;
+	}
+
+	ret = devm_add_action_or_reset(&priv->device->dev, unregister_adev, adev);
+	if (ret)
+		return ret;
+
+	return 0;
+}
+
+static void peci_cpu_add_adevices(struct peci_cpu *priv)
+{
+	struct device *dev = &priv->device->dev;
+	int ret, i;
+
+	for (i = 0; i < ARRAY_SIZE(peci_adev_types); i++) {
+		ret = devm_adev_add(dev, i);
+		if (ret) {
+			dev_warn(dev, "Failed to register PECI auxiliary: %s, ret = %d\n",
+				 peci_adev_types[i], ret);
+			continue;
+		}
+	}
+}
+
+static int
+peci_cpu_probe(struct peci_device *device, const struct peci_device_id *id)
+{
+	struct device *dev = &device->dev;
+	struct peci_cpu *priv;
+
+	priv = devm_kzalloc(dev, sizeof(*priv), GFP_KERNEL);
+	if (!priv)
+		return -ENOMEM;
+
+	dev_set_drvdata(dev, priv);
+	priv->device = device;
+	priv->id = id;
+
+	peci_cpu_add_adevices(priv);
+
+	return 0;
+}
+
+static const struct peci_device_id peci_cpu_device_ids[] = {
+	{ /* Haswell Xeon */
+		.x86_vfm = INTEL_HASWELL_X,
+		.data	= "hsx",
+	},
+	{ /* Broadwell Xeon */
+		.x86_vfm = INTEL_BROADWELL_X,
+		.data	= "bdx",
+	},
+	{ /* Broadwell Xeon D */
+		.x86_vfm = INTEL_BROADWELL_D,
+		.data	= "bdxd",
+	},
+	{ /* Skylake Xeon */
+		.x86_vfm = INTEL_SKYLAKE_X,
+		.data	= "skx",
+	},
+	{ /* Icelake Xeon */
+		.x86_vfm = INTEL_ICELAKE_X,
+		.data	= "icx",
+	},
+	{ /* Icelake Xeon D */
+		.x86_vfm = INTEL_ICELAKE_D,
+		.data	= "icxd",
+	},
+	{ /* Sapphire Rapids Xeon */
+		.x86_vfm = INTEL_SAPPHIRERAPIDS_X,
+		.data	= "spr",
+	},
+	{ /* Emerald Rapids Xeon */
+		.x86_vfm  = INTEL_EMERALDRAPIDS_X,
+		.data  = "emr",
+	},
+	{ }
+};
+MODULE_DEVICE_TABLE(peci, peci_cpu_device_ids);
+
+static struct peci_driver peci_cpu_driver = {
+	.probe		= peci_cpu_probe,
+	.id_table	= peci_cpu_device_ids,
+	.driver		= {
+		.name		= "peci-cpu",
+	},
+};
+module_peci_driver(peci_cpu_driver);
+
+MODULE_AUTHOR("Iwona Winiarska <iwona.winiarska@intel.com>");
+MODULE_DESCRIPTION("PECI CPU driver");
+MODULE_LICENSE("GPL");
+MODULE_IMPORT_NS("PECI");