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, 304 insertions, 0 deletions

diff --git a/fs/cachefiles/xattr.c b/fs/cachefiles/xattr.c
new file mode 100644
index 000000000..c70bf67e5
--- /dev/null
+++ b/fs/cachefiles/xattr.c
@@ -0,0 +1,304 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
+/* CacheFiles extended attribute management
+ *
+ * Copyright (C) 2021 Red Hat, Inc. All Rights Reserved.
+ * Written by David Howells (dhowells@redhat.com)
+ */
+
+#include <linux/module.h>
+#include <linux/sched.h>
+#include <linux/file.h>
+#include <linux/fs.h>
+#include <linux/fsnotify.h>
+#include <linux/quotaops.h>
+#include <linux/xattr.h>
+#include <linux/slab.h>
+#include <linux/unaligned.h>
+#include "internal.h"
+
+#define CACHEFILES_COOKIE_TYPE_DATA 1
+
+struct cachefiles_xattr {
+	__be64	object_size;	/* Actual size of the object */
+	__be64	zero_point;	/* Size after which server has no data not written by us */
+	__u8	type;		/* Type of object */
+	__u8	content;	/* Content presence (enum cachefiles_content) */
+	__u8	data[];		/* netfs coherency data */
+} __packed;
+
+static const char cachefiles_xattr_cache[] =
+	XATTR_USER_PREFIX "CacheFiles.cache";
+
+struct cachefiles_vol_xattr {
+	__be32	reserved;	/* Reserved, should be 0 */
+	__u8	data[];		/* netfs volume coherency data */
+} __packed;
+
+/*
+ * set the state xattr on a cache file
+ */
+int cachefiles_set_object_xattr(struct cachefiles_object *object)
+{
+	struct cachefiles_xattr *buf;
+	struct dentry *dentry;
+	struct file *file = object->file;
+	unsigned int len = object->cookie->aux_len;
+	int ret;
+
+	if (!file)
+		return -ESTALE;
+	dentry = file->f_path.dentry;
+
+	_enter("%x,#%d", object->debug_id, len);
+
+	buf = kmalloc(sizeof(struct cachefiles_xattr) + max(len, sizeof(__be64)), GFP_KERNEL);
+	if (!buf)
+		return -ENOMEM;
+
+	buf->object_size	= cpu_to_be64(object->cookie->object_size);
+	buf->zero_point		= 0;
+	buf->type		= CACHEFILES_COOKIE_TYPE_DATA;
+	buf->content		= object->content_info;
+	if (test_bit(FSCACHE_COOKIE_LOCAL_WRITE, &object->cookie->flags))
+		buf->content	= CACHEFILES_CONTENT_DIRTY;
+	put_unaligned_be64(0, (__be64 *)buf->data);
+	if (len > 0)
+		memcpy(buf->data, fscache_get_aux(object->cookie), len);
+
+	ret = cachefiles_inject_write_error();
+	if (ret == 0) {
+		ret = mnt_want_write_file(file);
+		if (ret == 0) {
+			ret = vfs_setxattr(&nop_mnt_idmap, dentry,
+					   cachefiles_xattr_cache, buf,
+					   sizeof(struct cachefiles_xattr) + len, 0);
+			mnt_drop_write_file(file);
+		}
+	}
+	if (ret < 0) {
+		trace_cachefiles_vfs_error(object, file_inode(file), ret,
+					   cachefiles_trace_setxattr_error);
+		trace_cachefiles_coherency(object, file_inode(file)->i_ino,
+					   buf->data, buf->content,
+					   cachefiles_coherency_set_fail);
+		if (ret != -ENOMEM)
+			cachefiles_io_error_obj(
+				object,
+				"Failed to set xattr with error %d", ret);
+	} else {
+		trace_cachefiles_coherency(object, file_inode(file)->i_ino,
+					   buf->data, buf->content,
+					   cachefiles_coherency_set_ok);
+	}
+
+	kfree(buf);
+	_leave(" = %d", ret);
+	return ret;
+}
+
+/*
+ * check the consistency between the backing cache and the FS-Cache cookie
+ */
+int cachefiles_check_auxdata(struct cachefiles_object *object, struct file *file)
+{
+	struct cachefiles_xattr *buf;
+	struct dentry *dentry = file->f_path.dentry;
+	unsigned int len = object->cookie->aux_len, tlen;
+	const void *p = fscache_get_aux(object->cookie);
+	enum cachefiles_coherency_trace why;
+	ssize_t xlen;
+	int ret = -ESTALE;
+
+	tlen = sizeof(struct cachefiles_xattr) + len;
+	buf = kmalloc(sizeof(struct cachefiles_xattr) + max(len, sizeof(__be64)), GFP_KERNEL);
+	if (!buf)
+		return -ENOMEM;
+	put_unaligned_be64(0, (__be64 *)buf->data);
+
+	xlen = cachefiles_inject_read_error();
+	if (xlen == 0)
+		xlen = vfs_getxattr(&nop_mnt_idmap, dentry, cachefiles_xattr_cache, buf, tlen);
+	if (xlen != tlen) {
+		if (xlen < 0) {
+			ret = xlen;
+			trace_cachefiles_vfs_error(object, file_inode(file), xlen,
+						   cachefiles_trace_getxattr_error);
+		}
+		if (xlen == -EIO)
+			cachefiles_io_error_obj(
+				object,
+				"Failed to read aux with error %zd", xlen);
+		why = cachefiles_coherency_check_xattr;
+		goto out;
+	}
+
+	if (buf->type != CACHEFILES_COOKIE_TYPE_DATA) {
+		why = cachefiles_coherency_check_type;
+	} else if (memcmp(buf->data, p, len) != 0) {
+		why = cachefiles_coherency_check_aux;
+	} else if (be64_to_cpu(buf->object_size) != object->cookie->object_size) {
+		why = cachefiles_coherency_check_objsize;
+	} else if (buf->content == CACHEFILES_CONTENT_DIRTY) {
+		// TODO: Begin conflict resolution
+		pr_warn("Dirty object in cache\n");
+		why = cachefiles_coherency_check_dirty;
+	} else {
+		why = cachefiles_coherency_check_ok;
+		ret = 0;
+	}
+
+out:
+	trace_cachefiles_coherency(object, file_inode(file)->i_ino,
+				   buf->data, buf->content, why);
+	kfree(buf);
+	return ret;
+}
+
+/*
+ * remove the object's xattr to mark it stale
+ */
+int cachefiles_remove_object_xattr(struct cachefiles_cache *cache,
+				   struct cachefiles_object *object,
+				   struct dentry *dentry)
+{
+	int ret;
+
+	ret = cachefiles_inject_remove_error();
+	if (ret == 0) {
+		ret = mnt_want_write(cache->mnt);
+		if (ret == 0) {
+			ret = vfs_removexattr(&nop_mnt_idmap, dentry,
+					      cachefiles_xattr_cache);
+			mnt_drop_write(cache->mnt);
+		}
+	}
+	if (ret < 0) {
+		trace_cachefiles_vfs_error(object, d_inode(dentry), ret,
+					   cachefiles_trace_remxattr_error);
+		if (ret == -ENOENT || ret == -ENODATA)
+			ret = 0;
+		else if (ret != -ENOMEM)
+			cachefiles_io_error(cache,
+					    "Can't remove xattr from %llu"
+					    " (error %d)",
+					    d_backing_inode(dentry)->i_ino, -ret);
+	}
+
+	_leave(" = %d", ret);
+	return ret;
+}
+
+/*
+ * Stick a marker on the cache object to indicate that it's dirty.
+ */
+void cachefiles_prepare_to_write(struct fscache_cookie *cookie)
+{
+	const struct cred *saved_cred;
+	struct cachefiles_object *object = cookie->cache_priv;
+	struct cachefiles_cache *cache = object->volume->cache;
+
+	_enter("c=%08x", object->cookie->debug_id);
+
+	if (!test_bit(CACHEFILES_OBJECT_USING_TMPFILE, &object->flags)) {
+		cachefiles_begin_secure(cache, &saved_cred);
+		cachefiles_set_object_xattr(object);
+		cachefiles_end_secure(cache, saved_cred);
+	}
+}
+
+/*
+ * Set the state xattr on a volume directory.
+ */
+bool cachefiles_set_volume_xattr(struct cachefiles_volume *volume)
+{
+	struct cachefiles_vol_xattr *buf;
+	unsigned int len = volume->vcookie->coherency_len;
+	const void *p = volume->vcookie->coherency;
+	struct dentry *dentry = volume->dentry;
+	int ret;
+
+	_enter("%x,#%d", volume->vcookie->debug_id, len);
+
+	len += sizeof(*buf);
+	buf = kmalloc(len, GFP_KERNEL);
+	if (!buf)
+		return false;
+	buf->reserved = cpu_to_be32(0);
+	memcpy(buf->data, p, volume->vcookie->coherency_len);
+
+	ret = cachefiles_inject_write_error();
+	if (ret == 0) {
+		ret = mnt_want_write(volume->cache->mnt);
+		if (ret == 0) {
+			ret = vfs_setxattr(&nop_mnt_idmap, dentry,
+					   cachefiles_xattr_cache,
+					   buf, len, 0);
+			mnt_drop_write(volume->cache->mnt);
+		}
+	}
+	if (ret < 0) {
+		trace_cachefiles_vfs_error(NULL, d_inode(dentry), ret,
+					   cachefiles_trace_setxattr_error);
+		trace_cachefiles_vol_coherency(volume, d_inode(dentry)->i_ino,
+					       cachefiles_coherency_vol_set_fail);
+		if (ret != -ENOMEM)
+			cachefiles_io_error(
+				volume->cache, "Failed to set xattr with error %d", ret);
+	} else {
+		trace_cachefiles_vol_coherency(volume, d_inode(dentry)->i_ino,
+					       cachefiles_coherency_vol_set_ok);
+	}
+
+	kfree(buf);
+	_leave(" = %d", ret);
+	return ret == 0;
+}
+
+/*
+ * Check the consistency between the backing cache and the volume cookie.
+ */
+int cachefiles_check_volume_xattr(struct cachefiles_volume *volume)
+{
+	struct cachefiles_vol_xattr *buf;
+	struct dentry *dentry = volume->dentry;
+	unsigned int len = volume->vcookie->coherency_len;
+	const void *p = volume->vcookie->coherency;
+	enum cachefiles_coherency_trace why;
+	ssize_t xlen;
+	int ret = -ESTALE;
+
+	_enter("");
+
+	len += sizeof(*buf);
+	buf = kmalloc(len, GFP_KERNEL);
+	if (!buf)
+		return -ENOMEM;
+
+	xlen = cachefiles_inject_read_error();
+	if (xlen == 0)
+		xlen = vfs_getxattr(&nop_mnt_idmap, dentry, cachefiles_xattr_cache, buf, len);
+	if (xlen != len) {
+		if (xlen < 0) {
+			ret = xlen;
+			trace_cachefiles_vfs_error(NULL, d_inode(dentry), xlen,
+						   cachefiles_trace_getxattr_error);
+			if (xlen == -EIO)
+				cachefiles_io_error(
+					volume->cache,
+					"Failed to read xattr with error %zd", xlen);
+		}
+		why = cachefiles_coherency_vol_check_xattr;
+	} else if (buf->reserved != cpu_to_be32(0)) {
+		why = cachefiles_coherency_vol_check_resv;
+	} else if (memcmp(buf->data, p, len - sizeof(*buf)) != 0) {
+		why = cachefiles_coherency_vol_check_cmp;
+	} else {
+		why = cachefiles_coherency_vol_check_ok;
+		ret = 0;
+	}
+
+	trace_cachefiles_vol_coherency(volume, d_inode(dentry)->i_ino, why);
+	kfree(buf);
+	_leave(" = %d", ret);
+	return ret;
+}