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, 190 insertions, 0 deletions

diff --git a/fs/configfs/item.c b/fs/configfs/item.c
new file mode 100644
index 000000000..c378b5cbf
--- /dev/null
+++ b/fs/configfs/item.c
@@ -0,0 +1,190 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
+/*
+ * item.c - library routines for handling generic config items
+ *
+ * Based on kobject:
+ *	kobject is Copyright (c) 2002-2003 Patrick Mochel
+ *
+ * configfs Copyright (C) 2005 Oracle.  All rights reserved.
+ *
+ * Please see the file Documentation/filesystems/configfs.rst for
+ * critical information about using the config_item interface.
+ */
+
+#include <linux/string.h>
+#include <linux/module.h>
+#include <linux/stat.h>
+#include <linux/slab.h>
+
+#include <linux/configfs.h>
+
+
+static inline struct config_item *to_item(struct list_head *entry)
+{
+	return container_of(entry, struct config_item, ci_entry);
+}
+
+/* Evil kernel */
+static void config_item_release(struct kref *kref);
+
+/**
+ *	config_item_init - initialize item.
+ *	@item:	item in question.
+ */
+static void config_item_init(struct config_item *item)
+{
+	kref_init(&item->ci_kref);
+	INIT_LIST_HEAD(&item->ci_entry);
+}
+
+/**
+ *	config_item_set_name - Set the name of an item
+ *	@item:	item.
+ *	@fmt:  The vsnprintf()'s format string.
+ *
+ *	If strlen(name) >= CONFIGFS_ITEM_NAME_LEN, then use a
+ *	dynamically allocated string that @item->ci_name points to.
+ *	Otherwise, use the static @item->ci_namebuf array.
+ */
+int config_item_set_name(struct config_item *item, const char *fmt, ...)
+{
+	int limit = CONFIGFS_ITEM_NAME_LEN;
+	int need;
+	va_list args;
+	char *name;
+
+	/*
+	 * First, try the static array
+	 */
+	va_start(args, fmt);
+	need = vsnprintf(item->ci_namebuf, limit, fmt, args);
+	va_end(args);
+	if (need < limit)
+		name = item->ci_namebuf;
+	else {
+		va_start(args, fmt);
+		name = kvasprintf(GFP_KERNEL, fmt, args);
+		va_end(args);
+		if (!name)
+			return -ENOMEM;
+	}
+
+	/* Free the old name, if necessary. */
+	if (item->ci_name && item->ci_name != item->ci_namebuf)
+		kfree(item->ci_name);
+
+	/* Now, set the new name */
+	item->ci_name = name;
+	return 0;
+}
+EXPORT_SYMBOL(config_item_set_name);
+
+void config_item_init_type_name(struct config_item *item,
+				const char *name,
+				const struct config_item_type *type)
+{
+	config_item_set_name(item, "%s", name);
+	item->ci_type = type;
+	config_item_init(item);
+}
+EXPORT_SYMBOL(config_item_init_type_name);
+
+void config_group_init_type_name(struct config_group *group, const char *name,
+			 const struct config_item_type *type)
+{
+	config_item_set_name(&group->cg_item, "%s", name);
+	group->cg_item.ci_type = type;
+	config_group_init(group);
+}
+EXPORT_SYMBOL(config_group_init_type_name);
+
+struct config_item *config_item_get(struct config_item *item)
+{
+	if (item)
+		kref_get(&item->ci_kref);
+	return item;
+}
+EXPORT_SYMBOL(config_item_get);
+
+struct config_item *config_item_get_unless_zero(struct config_item *item)
+{
+	if (item && kref_get_unless_zero(&item->ci_kref))
+		return item;
+	return NULL;
+}
+EXPORT_SYMBOL(config_item_get_unless_zero);
+
+static void config_item_cleanup(struct config_item *item)
+{
+	const struct config_item_type *t = item->ci_type;
+	struct config_group *s = item->ci_group;
+	struct config_item *parent = item->ci_parent;
+
+	pr_debug("config_item %s: cleaning up\n", config_item_name(item));
+	if (item->ci_name != item->ci_namebuf)
+		kfree(item->ci_name);
+	item->ci_name = NULL;
+	if (t && t->ct_item_ops && t->ct_item_ops->release)
+		t->ct_item_ops->release(item);
+	if (s)
+		config_group_put(s);
+	if (parent)
+		config_item_put(parent);
+}
+
+static void config_item_release(struct kref *kref)
+{
+	config_item_cleanup(container_of(kref, struct config_item, ci_kref));
+}
+
+/**
+ *	config_item_put - decrement refcount for item.
+ *	@item:	item.
+ *
+ *	Decrement the refcount, and if 0, call config_item_cleanup().
+ */
+void config_item_put(struct config_item *item)
+{
+	if (item)
+		kref_put(&item->ci_kref, config_item_release);
+}
+EXPORT_SYMBOL(config_item_put);
+
+/**
+ *	config_group_init - initialize a group for use
+ *	@group:	config_group
+ */
+void config_group_init(struct config_group *group)
+{
+	config_item_init(&group->cg_item);
+	INIT_LIST_HEAD(&group->cg_children);
+	INIT_LIST_HEAD(&group->default_groups);
+}
+EXPORT_SYMBOL(config_group_init);
+
+/**
+ *	config_group_find_item - search for item in group.
+ *	@group:	group we're looking in.
+ *	@name:	item's name.
+ *
+ *	Iterate over @group->cg_list, looking for a matching config_item.
+ *	If matching item is found take a reference and return the item.
+ *	Caller must have locked group via @group->cg_subsys->su_mtx.
+ */
+struct config_item *config_group_find_item(struct config_group *group,
+					   const char *name)
+{
+	struct list_head *entry;
+	struct config_item *ret = NULL;
+
+	list_for_each(entry, &group->cg_children) {
+		struct config_item *item = to_item(entry);
+		if (config_item_name(item) &&
+		    !strcmp(config_item_name(item), name)) {
+			ret = config_item_get(item);
+			break;
+		}
+	}
+	return ret;
+}
+EXPORT_SYMBOL(config_group_find_item);