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, 366 insertions, 0 deletions

diff --git a/drivers/rapidio/rio-sysfs.c b/drivers/rapidio/rio-sysfs.c
new file mode 100644
index 000000000..0949c869b
--- /dev/null
+++ b/drivers/rapidio/rio-sysfs.c
@@ -0,0 +1,366 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
+/*
+ * RapidIO sysfs attributes and support
+ *
+ * Copyright 2005 MontaVista Software, Inc.
+ * Matt Porter <mporter@kernel.crashing.org>
+ */
+
+#include <linux/kernel.h>
+#include <linux/rio.h>
+#include <linux/rio_drv.h>
+#include <linux/stat.h>
+#include <linux/capability.h>
+
+#include "rio.h"
+
+/* Sysfs support */
+#define rio_config_attr(field, format_string)					\
+static ssize_t								\
+field##_show(struct device *dev, struct device_attribute *attr, char *buf)			\
+{									\
+	struct rio_dev *rdev = to_rio_dev(dev);				\
+									\
+	return sprintf(buf, format_string, rdev->field);		\
+}									\
+static DEVICE_ATTR_RO(field);
+
+rio_config_attr(did, "0x%04x\n");
+rio_config_attr(vid, "0x%04x\n");
+rio_config_attr(device_rev, "0x%08x\n");
+rio_config_attr(asm_did, "0x%04x\n");
+rio_config_attr(asm_vid, "0x%04x\n");
+rio_config_attr(asm_rev, "0x%04x\n");
+rio_config_attr(destid, "0x%04x\n");
+rio_config_attr(hopcount, "0x%02x\n");
+
+static ssize_t routes_show(struct device *dev, struct device_attribute *attr, char *buf)
+{
+	struct rio_dev *rdev = to_rio_dev(dev);
+	char *str = buf;
+	int i;
+
+	for (i = 0; i < RIO_MAX_ROUTE_ENTRIES(rdev->net->hport->sys_size);
+			i++) {
+		if (rdev->rswitch->route_table[i] == RIO_INVALID_ROUTE)
+			continue;
+		str +=
+		    sprintf(str, "%04x %02x\n", i,
+			    rdev->rswitch->route_table[i]);
+	}
+
+	return (str - buf);
+}
+static DEVICE_ATTR_RO(routes);
+
+static ssize_t lprev_show(struct device *dev,
+			  struct device_attribute *attr, char *buf)
+{
+	struct rio_dev *rdev = to_rio_dev(dev);
+
+	return sprintf(buf, "%s\n",
+			(rdev->prev) ? rio_name(rdev->prev) : "root");
+}
+static DEVICE_ATTR_RO(lprev);
+
+static ssize_t lnext_show(struct device *dev,
+			  struct device_attribute *attr, char *buf)
+{
+	struct rio_dev *rdev = to_rio_dev(dev);
+	char *str = buf;
+	int i;
+
+	if (rdev->pef & RIO_PEF_SWITCH) {
+		for (i = 0; i < RIO_GET_TOTAL_PORTS(rdev->swpinfo); i++) {
+			if (rdev->rswitch->nextdev[i])
+				str += sprintf(str, "%s\n",
+					rio_name(rdev->rswitch->nextdev[i]));
+			else
+				str += sprintf(str, "null\n");
+		}
+	}
+
+	return str - buf;
+}
+static DEVICE_ATTR_RO(lnext);
+
+static ssize_t modalias_show(struct device *dev,
+			     struct device_attribute *attr, char *buf)
+{
+	struct rio_dev *rdev = to_rio_dev(dev);
+
+	return sprintf(buf, "rapidio:v%04Xd%04Xav%04Xad%04X\n",
+		       rdev->vid, rdev->did, rdev->asm_vid, rdev->asm_did);
+}
+static DEVICE_ATTR_RO(modalias);
+
+static struct attribute *rio_dev_attrs[] = {
+	&dev_attr_did.attr,
+	&dev_attr_vid.attr,
+	&dev_attr_device_rev.attr,
+	&dev_attr_asm_did.attr,
+	&dev_attr_asm_vid.attr,
+	&dev_attr_asm_rev.attr,
+	&dev_attr_lprev.attr,
+	&dev_attr_destid.attr,
+	&dev_attr_modalias.attr,
+
+	/* Switch-only attributes */
+	&dev_attr_routes.attr,
+	&dev_attr_lnext.attr,
+	&dev_attr_hopcount.attr,
+	NULL,
+};
+
+static ssize_t
+rio_read_config(struct file *filp, struct kobject *kobj,
+		const struct bin_attribute *bin_attr,
+		char *buf, loff_t off, size_t count)
+{
+	struct rio_dev *dev = to_rio_dev(kobj_to_dev(kobj));
+	unsigned int size = 0x100;
+	loff_t init_off = off;
+	u8 *data = (u8 *) buf;
+
+	/* Several chips lock up trying to read undefined config space */
+	if (capable(CAP_SYS_ADMIN))
+		size = RIO_MAINT_SPACE_SZ;
+
+	if (off >= size)
+		return 0;
+	if (off + count > size) {
+		size -= off;
+		count = size;
+	} else {
+		size = count;
+	}
+
+	if ((off & 1) && size) {
+		u8 val;
+		rio_read_config_8(dev, off, &val);
+		data[off - init_off] = val;
+		off++;
+		size--;
+	}
+
+	if ((off & 3) && size > 2) {
+		u16 val;
+		rio_read_config_16(dev, off, &val);
+		data[off - init_off] = (val >> 8) & 0xff;
+		data[off - init_off + 1] = val & 0xff;
+		off += 2;
+		size -= 2;
+	}
+
+	while (size > 3) {
+		u32 val;
+		rio_read_config_32(dev, off, &val);
+		data[off - init_off] = (val >> 24) & 0xff;
+		data[off - init_off + 1] = (val >> 16) & 0xff;
+		data[off - init_off + 2] = (val >> 8) & 0xff;
+		data[off - init_off + 3] = val & 0xff;
+		off += 4;
+		size -= 4;
+	}
+
+	if (size >= 2) {
+		u16 val;
+		rio_read_config_16(dev, off, &val);
+		data[off - init_off] = (val >> 8) & 0xff;
+		data[off - init_off + 1] = val & 0xff;
+		off += 2;
+		size -= 2;
+	}
+
+	if (size > 0) {
+		u8 val;
+		rio_read_config_8(dev, off, &val);
+		data[off - init_off] = val;
+		off++;
+		--size;
+	}
+
+	return count;
+}
+
+static ssize_t
+rio_write_config(struct file *filp, struct kobject *kobj,
+		 const struct bin_attribute *bin_attr,
+		 char *buf, loff_t off, size_t count)
+{
+	struct rio_dev *dev = to_rio_dev(kobj_to_dev(kobj));
+	unsigned int size = count;
+	loff_t init_off = off;
+	u8 *data = (u8 *) buf;
+
+	if (off >= RIO_MAINT_SPACE_SZ)
+		return 0;
+	if (off + count > RIO_MAINT_SPACE_SZ) {
+		size = RIO_MAINT_SPACE_SZ - off;
+		count = size;
+	}
+
+	if ((off & 1) && size) {
+		rio_write_config_8(dev, off, data[off - init_off]);
+		off++;
+		size--;
+	}
+
+	if ((off & 3) && (size > 2)) {
+		u16 val = data[off - init_off + 1];
+		val |= (u16) data[off - init_off] << 8;
+		rio_write_config_16(dev, off, val);
+		off += 2;
+		size -= 2;
+	}
+
+	while (size > 3) {
+		u32 val = data[off - init_off + 3];
+		val |= (u32) data[off - init_off + 2] << 8;
+		val |= (u32) data[off - init_off + 1] << 16;
+		val |= (u32) data[off - init_off] << 24;
+		rio_write_config_32(dev, off, val);
+		off += 4;
+		size -= 4;
+	}
+
+	if (size >= 2) {
+		u16 val = data[off - init_off + 1];
+		val |= (u16) data[off - init_off] << 8;
+		rio_write_config_16(dev, off, val);
+		off += 2;
+		size -= 2;
+	}
+
+	if (size) {
+		rio_write_config_8(dev, off, data[off - init_off]);
+		off++;
+		--size;
+	}
+
+	return count;
+}
+
+static const struct bin_attribute rio_config_attr = {
+	.attr = {
+		 .name = "config",
+		 .mode = S_IRUGO | S_IWUSR,
+		 },
+	.size = RIO_MAINT_SPACE_SZ,
+	.read = rio_read_config,
+	.write = rio_write_config,
+};
+
+static const struct bin_attribute *const rio_dev_bin_attrs[] = {
+	&rio_config_attr,
+	NULL,
+};
+
+static umode_t rio_dev_is_attr_visible(struct kobject *kobj,
+				       struct attribute *attr, int n)
+{
+	struct rio_dev *rdev = to_rio_dev(kobj_to_dev(kobj));
+	umode_t mode = attr->mode;
+
+	if (!(rdev->pef & RIO_PEF_SWITCH) &&
+	    (attr == &dev_attr_routes.attr ||
+	     attr == &dev_attr_lnext.attr ||
+	     attr == &dev_attr_hopcount.attr)) {
+		/*
+		 * Hide switch-specific attributes for a non-switch device.
+		 */
+		mode = 0;
+	}
+
+	return mode;
+}
+
+static const struct attribute_group rio_dev_group = {
+	.attrs		= rio_dev_attrs,
+	.is_visible	= rio_dev_is_attr_visible,
+	.bin_attrs	= rio_dev_bin_attrs,
+};
+
+const struct attribute_group *rio_dev_groups[] = {
+	&rio_dev_group,
+	NULL,
+};
+
+static ssize_t scan_store(const struct bus_type *bus, const char *buf, size_t count)
+{
+	long val;
+	int rc;
+
+	if (kstrtol(buf, 0, &val) < 0)
+		return -EINVAL;
+
+	if (val == RIO_MPORT_ANY) {
+		rc = rio_init_mports();
+		goto exit;
+	}
+
+	if (val < 0 || val >= RIO_MAX_MPORTS)
+		return -EINVAL;
+
+	rc = rio_mport_scan((int)val);
+exit:
+	if (!rc)
+		rc = count;
+
+	return rc;
+}
+static BUS_ATTR_WO(scan);
+
+static struct attribute *rio_bus_attrs[] = {
+	&bus_attr_scan.attr,
+	NULL,
+};
+
+static const struct attribute_group rio_bus_group = {
+	.attrs = rio_bus_attrs,
+};
+
+const struct attribute_group *rio_bus_groups[] = {
+	&rio_bus_group,
+	NULL,
+};
+
+static ssize_t
+port_destid_show(struct device *dev, struct device_attribute *attr,
+		 char *buf)
+{
+	struct rio_mport *mport = to_rio_mport(dev);
+
+	if (mport)
+		return sprintf(buf, "0x%04x\n", mport->host_deviceid);
+	else
+		return -ENODEV;
+}
+static DEVICE_ATTR_RO(port_destid);
+
+static ssize_t sys_size_show(struct device *dev, struct device_attribute *attr,
+			   char *buf)
+{
+	struct rio_mport *mport = to_rio_mport(dev);
+
+	if (mport)
+		return sprintf(buf, "%u\n", mport->sys_size);
+	else
+		return -ENODEV;
+}
+static DEVICE_ATTR_RO(sys_size);
+
+static struct attribute *rio_mport_attrs[] = {
+	&dev_attr_port_destid.attr,
+	&dev_attr_sys_size.attr,
+	NULL,
+};
+
+static const struct attribute_group rio_mport_group = {
+	.attrs = rio_mport_attrs,
+};
+
+const struct attribute_group *rio_mport_groups[] = {
+	&rio_mport_group,
+	NULL,
+};