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, 502 insertions, 0 deletions

diff --git a/fs/adfs/super.c b/fs/adfs/super.c
new file mode 100644
index 000000000..888aa81a6
--- /dev/null
+++ b/fs/adfs/super.c
@@ -0,0 +1,502 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ *  linux/fs/adfs/super.c
+ *
+ *  Copyright (C) 1997-1999 Russell King
+ */
+#include <linux/module.h>
+#include <linux/init.h>
+#include <linux/fs_parser.h>
+#include <linux/fs_context.h>
+#include <linux/mount.h>
+#include <linux/seq_file.h>
+#include <linux/slab.h>
+#include <linux/statfs.h>
+#include <linux/user_namespace.h>
+#include <linux/blkdev.h>
+#include "adfs.h"
+#include "dir_f.h"
+#include "dir_fplus.h"
+
+#define ADFS_SB_FLAGS SB_NOATIME
+
+#define ADFS_DEFAULT_OWNER_MASK S_IRWXU
+#define ADFS_DEFAULT_OTHER_MASK (S_IRWXG | S_IRWXO)
+
+void __adfs_error(struct super_block *sb, const char *function, const char *fmt, ...)
+{
+	struct va_format vaf;
+	va_list args;
+
+	va_start(args, fmt);
+	vaf.fmt = fmt;
+	vaf.va = &args;
+
+	printk(KERN_CRIT "ADFS-fs error (device %s)%s%s: %pV\n",
+		sb->s_id, function ? ": " : "",
+		function ? function : "", &vaf);
+
+	va_end(args);
+}
+
+void adfs_msg(struct super_block *sb, const char *pfx, const char *fmt, ...)
+{
+	struct va_format vaf;
+	va_list args;
+
+	va_start(args, fmt);
+	vaf.fmt = fmt;
+	vaf.va = &args;
+	printk("%sADFS-fs (%s): %pV\n", pfx, sb->s_id, &vaf);
+	va_end(args);
+}
+
+static int adfs_checkdiscrecord(struct adfs_discrecord *dr)
+{
+	unsigned int max_idlen;
+	int i;
+
+	/* sector size must be 256, 512 or 1024 bytes */
+	if (dr->log2secsize != 8 &&
+	    dr->log2secsize != 9 &&
+	    dr->log2secsize != 10)
+		return 1;
+
+	/* idlen must be at least log2secsize + 3 */
+	if (dr->idlen < dr->log2secsize + 3)
+		return 1;
+
+	/* we cannot have such a large disc that we
+	 * are unable to represent sector offsets in
+	 * 32 bits.  This works out at 2.0 TB.
+	 */
+	if (le32_to_cpu(dr->disc_size_high) >> dr->log2secsize)
+		return 1;
+
+	/*
+	 * Maximum idlen is limited to 16 bits for new directories by
+	 * the three-byte storage of an indirect disc address.  For
+	 * big directories, idlen must be no greater than 19 v2 [1.0]
+	 */
+	max_idlen = dr->format_version ? 19 : 16;
+	if (dr->idlen > max_idlen)
+		return 1;
+
+	/* reserved bytes should be zero */
+	for (i = 0; i < sizeof(dr->unused52); i++)
+		if (dr->unused52[i] != 0)
+			return 1;
+
+	return 0;
+}
+
+static void adfs_put_super(struct super_block *sb)
+{
+	adfs_free_map(sb);
+}
+
+static int adfs_show_options(struct seq_file *seq, struct dentry *root)
+{
+	struct adfs_sb_info *asb = ADFS_SB(root->d_sb);
+
+	if (!uid_eq(asb->s_uid, GLOBAL_ROOT_UID))
+		seq_printf(seq, ",uid=%u", from_kuid_munged(&init_user_ns, asb->s_uid));
+	if (!gid_eq(asb->s_gid, GLOBAL_ROOT_GID))
+		seq_printf(seq, ",gid=%u", from_kgid_munged(&init_user_ns, asb->s_gid));
+	if (asb->s_owner_mask != ADFS_DEFAULT_OWNER_MASK)
+		seq_printf(seq, ",ownmask=%o", asb->s_owner_mask);
+	if (asb->s_other_mask != ADFS_DEFAULT_OTHER_MASK)
+		seq_printf(seq, ",othmask=%o", asb->s_other_mask);
+	if (asb->s_ftsuffix != 0)
+		seq_printf(seq, ",ftsuffix=%u", asb->s_ftsuffix);
+
+	return 0;
+}
+
+enum {Opt_uid, Opt_gid, Opt_ownmask, Opt_othmask, Opt_ftsuffix};
+
+static const struct fs_parameter_spec adfs_param_spec[] = {
+	fsparam_uid	("uid",		Opt_uid),
+	fsparam_gid	("gid",		Opt_gid),
+	fsparam_u32oct	("ownmask",	Opt_ownmask),
+	fsparam_u32oct	("othmask",	Opt_othmask),
+	fsparam_u32	("ftsuffix",	Opt_ftsuffix),
+	{}
+};
+
+static int adfs_parse_param(struct fs_context *fc, struct fs_parameter *param)
+{
+	struct adfs_sb_info *asb = fc->s_fs_info;
+	struct fs_parse_result result;
+	int opt;
+
+	opt = fs_parse(fc, adfs_param_spec, param, &result);
+	if (opt < 0)
+		return opt;
+
+	switch (opt) {
+	case Opt_uid:
+		asb->s_uid = result.uid;
+		break;
+	case Opt_gid:
+		asb->s_gid = result.gid;
+		break;
+	case Opt_ownmask:
+		asb->s_owner_mask = result.uint_32;
+		break;
+	case Opt_othmask:
+		asb->s_other_mask = result.uint_32;
+		break;
+	case Opt_ftsuffix:
+		asb->s_ftsuffix = result.uint_32;
+		break;
+	default:
+		return -EINVAL;
+	}
+	return 0;
+}
+
+static int adfs_reconfigure(struct fs_context *fc)
+{
+	struct adfs_sb_info *new_asb = fc->s_fs_info;
+	struct adfs_sb_info *asb = ADFS_SB(fc->root->d_sb);
+
+	sync_filesystem(fc->root->d_sb);
+	fc->sb_flags |= ADFS_SB_FLAGS;
+
+	/* Structure copy newly parsed options */
+	*asb = *new_asb;
+
+	return 0;
+}
+
+static int adfs_statfs(struct dentry *dentry, struct kstatfs *buf)
+{
+	struct super_block *sb = dentry->d_sb;
+	struct adfs_sb_info *sbi = ADFS_SB(sb);
+	u64 id = huge_encode_dev(sb->s_bdev->bd_dev);
+
+	adfs_map_statfs(sb, buf);
+
+	buf->f_type    = ADFS_SUPER_MAGIC;
+	buf->f_namelen = sbi->s_namelen;
+	buf->f_bsize   = sb->s_blocksize;
+	buf->f_ffree   = (long)(buf->f_bfree * buf->f_files) / (long)buf->f_blocks;
+	buf->f_fsid    = u64_to_fsid(id);
+
+	return 0;
+}
+
+static struct kmem_cache *adfs_inode_cachep;
+
+static struct inode *adfs_alloc_inode(struct super_block *sb)
+{
+	struct adfs_inode_info *ei;
+	ei = alloc_inode_sb(sb, adfs_inode_cachep, GFP_KERNEL);
+	if (!ei)
+		return NULL;
+	return &ei->vfs_inode;
+}
+
+static void adfs_free_inode(struct inode *inode)
+{
+	kmem_cache_free(adfs_inode_cachep, ADFS_I(inode));
+}
+
+static int adfs_drop_inode(struct inode *inode)
+{
+	/* always drop inodes if we are read-only */
+	return !IS_ENABLED(CONFIG_ADFS_FS_RW) || IS_RDONLY(inode);
+}
+
+static void init_once(void *foo)
+{
+	struct adfs_inode_info *ei = (struct adfs_inode_info *) foo;
+
+	inode_init_once(&ei->vfs_inode);
+}
+
+static int __init init_inodecache(void)
+{
+	adfs_inode_cachep = kmem_cache_create("adfs_inode_cache",
+					     sizeof(struct adfs_inode_info),
+					     0, (SLAB_RECLAIM_ACCOUNT|
+						SLAB_ACCOUNT),
+					     init_once);
+	if (adfs_inode_cachep == NULL)
+		return -ENOMEM;
+	return 0;
+}
+
+static void destroy_inodecache(void)
+{
+	/*
+	 * Make sure all delayed rcu free inodes are flushed before we
+	 * destroy cache.
+	 */
+	rcu_barrier();
+	kmem_cache_destroy(adfs_inode_cachep);
+}
+
+static const struct super_operations adfs_sops = {
+	.alloc_inode	= adfs_alloc_inode,
+	.free_inode	= adfs_free_inode,
+	.drop_inode	= adfs_drop_inode,
+	.write_inode	= adfs_write_inode,
+	.put_super	= adfs_put_super,
+	.statfs		= adfs_statfs,
+	.show_options	= adfs_show_options,
+};
+
+static int adfs_probe(struct super_block *sb, unsigned int offset, int silent,
+		      int (*validate)(struct super_block *sb,
+				      struct buffer_head *bh,
+				      struct adfs_discrecord **bhp))
+{
+	struct adfs_sb_info *asb = ADFS_SB(sb);
+	struct adfs_discrecord *dr;
+	struct buffer_head *bh;
+	unsigned int blocksize = BLOCK_SIZE;
+	int ret, try;
+
+	for (try = 0; try < 2; try++) {
+		/* try to set the requested block size */
+		if (sb->s_blocksize != blocksize &&
+		    !sb_set_blocksize(sb, blocksize)) {
+			if (!silent)
+				adfs_msg(sb, KERN_ERR,
+					 "error: unsupported blocksize");
+			return -EINVAL;
+		}
+
+		/* read the buffer */
+		bh = sb_bread(sb, offset >> sb->s_blocksize_bits);
+		if (!bh) {
+			adfs_msg(sb, KERN_ERR,
+				 "error: unable to read block %u, try %d",
+				 offset >> sb->s_blocksize_bits, try);
+			return -EIO;
+		}
+
+		/* validate it */
+		ret = validate(sb, bh, &dr);
+		if (ret) {
+			brelse(bh);
+			return ret;
+		}
+
+		/* does the block size match the filesystem block size? */
+		blocksize = 1 << dr->log2secsize;
+		if (sb->s_blocksize == blocksize) {
+			asb->s_map = adfs_read_map(sb, dr);
+			brelse(bh);
+			return PTR_ERR_OR_ZERO(asb->s_map);
+		}
+
+		brelse(bh);
+	}
+
+	return -EIO;
+}
+
+static int adfs_validate_bblk(struct super_block *sb, struct buffer_head *bh,
+			      struct adfs_discrecord **drp)
+{
+	struct adfs_discrecord *dr;
+	unsigned char *b_data;
+
+	b_data = bh->b_data + (ADFS_DISCRECORD % sb->s_blocksize);
+	if (adfs_checkbblk(b_data))
+		return -EILSEQ;
+
+	/* Do some sanity checks on the ADFS disc record */
+	dr = (struct adfs_discrecord *)(b_data + ADFS_DR_OFFSET);
+	if (adfs_checkdiscrecord(dr))
+		return -EILSEQ;
+
+	if ((dr->nzones | dr->nzones_high << 8) == 0)
+		return -EILSEQ;
+
+	*drp = dr;
+	return 0;
+}
+
+static int adfs_validate_dr0(struct super_block *sb, struct buffer_head *bh,
+			      struct adfs_discrecord **drp)
+{
+	struct adfs_discrecord *dr;
+
+	/* Do some sanity checks on the ADFS disc record */
+	dr = (struct adfs_discrecord *)(bh->b_data + 4);
+	if (adfs_checkdiscrecord(dr) || dr->nzones_high || dr->nzones != 1)
+		return -EILSEQ;
+
+	*drp = dr;
+	return 0;
+}
+
+static int adfs_fill_super(struct super_block *sb, struct fs_context *fc)
+{
+	struct adfs_discrecord *dr;
+	struct object_info root_obj;
+	struct adfs_sb_info *asb = sb->s_fs_info;
+	struct inode *root;
+	int ret = -EINVAL;
+	int silent = fc->sb_flags & SB_SILENT;
+
+	sb->s_flags |= ADFS_SB_FLAGS;
+
+	sb->s_fs_info = asb;
+	sb->s_magic = ADFS_SUPER_MAGIC;
+	sb->s_time_gran = 10000000;
+
+	/* Try to probe the filesystem boot block */
+	ret = adfs_probe(sb, ADFS_DISCRECORD, 1, adfs_validate_bblk);
+	if (ret == -EILSEQ)
+		ret = adfs_probe(sb, 0, silent, adfs_validate_dr0);
+	if (ret == -EILSEQ) {
+		if (!silent)
+			adfs_msg(sb, KERN_ERR,
+				 "error: can't find an ADFS filesystem on dev %s.",
+				 sb->s_id);
+		ret = -EINVAL;
+	}
+	if (ret)
+		return ret;
+
+	/* set up enough so that we can read an inode */
+	sb->s_op = &adfs_sops;
+
+	dr = adfs_map_discrecord(asb->s_map);
+
+	root_obj.parent_id = root_obj.indaddr = le32_to_cpu(dr->root);
+	root_obj.name_len  = 0;
+	/* Set root object date as 01 Jan 1987 00:00:00 */
+	root_obj.loadaddr  = 0xfff0003f;
+	root_obj.execaddr  = 0xec22c000;
+	root_obj.size	   = ADFS_NEWDIR_SIZE;
+	root_obj.attr	   = ADFS_NDA_DIRECTORY   | ADFS_NDA_OWNER_READ |
+			     ADFS_NDA_OWNER_WRITE | ADFS_NDA_PUBLIC_READ;
+
+	/*
+	 * If this is a F+ disk with variable length directories,
+	 * get the root_size from the disc record.
+	 */
+	if (dr->format_version) {
+		root_obj.size = le32_to_cpu(dr->root_size);
+		asb->s_dir     = &adfs_fplus_dir_ops;
+		asb->s_namelen = ADFS_FPLUS_NAME_LEN;
+	} else {
+		asb->s_dir     = &adfs_f_dir_ops;
+		asb->s_namelen = ADFS_F_NAME_LEN;
+	}
+	/*
+	 * ,xyz hex filetype suffix may be added by driver
+	 * to files that have valid RISC OS filetype
+	 */
+	if (asb->s_ftsuffix)
+		asb->s_namelen += 4;
+
+	set_default_d_op(sb, &adfs_dentry_operations);
+	root = adfs_iget(sb, &root_obj);
+	sb->s_root = d_make_root(root);
+	if (!sb->s_root) {
+		adfs_free_map(sb);
+		adfs_error(sb, "get root inode failed\n");
+		return -EIO;
+	}
+	return 0;
+}
+
+static int adfs_get_tree(struct fs_context *fc)
+{
+	return get_tree_bdev(fc, adfs_fill_super);
+}
+
+static void adfs_free_fc(struct fs_context *fc)
+{
+	struct adfs_context *asb = fc->s_fs_info;
+
+	kfree(asb);
+}
+
+static const struct fs_context_operations adfs_context_ops = {
+	.parse_param	= adfs_parse_param,
+	.get_tree	= adfs_get_tree,
+	.reconfigure	= adfs_reconfigure,
+	.free		= adfs_free_fc,
+};
+
+static int adfs_init_fs_context(struct fs_context *fc)
+{
+	struct adfs_sb_info *asb;
+
+	asb = kzalloc_obj(struct adfs_sb_info);
+	if (!asb)
+		return -ENOMEM;
+
+	if (fc->purpose == FS_CONTEXT_FOR_RECONFIGURE) {
+		struct super_block *sb = fc->root->d_sb;
+		struct adfs_sb_info *old_asb = ADFS_SB(sb);
+
+		/* structure copy existing options before parsing */
+		*asb = *old_asb;
+	} else {
+		/* set default options */
+		asb->s_uid = GLOBAL_ROOT_UID;
+		asb->s_gid = GLOBAL_ROOT_GID;
+		asb->s_owner_mask = ADFS_DEFAULT_OWNER_MASK;
+		asb->s_other_mask = ADFS_DEFAULT_OTHER_MASK;
+		asb->s_ftsuffix = 0;
+	}
+
+	fc->ops = &adfs_context_ops;
+	fc->s_fs_info = asb;
+
+	return 0;
+}
+
+static void adfs_kill_sb(struct super_block *sb)
+{
+	struct adfs_sb_info *asb = ADFS_SB(sb);
+
+	kill_block_super(sb);
+
+	kfree_rcu(asb, rcu);
+}
+
+static struct file_system_type adfs_fs_type = {
+	.owner		= THIS_MODULE,
+	.name		= "adfs",
+	.kill_sb	= adfs_kill_sb,
+	.fs_flags	= FS_REQUIRES_DEV,
+	.init_fs_context = adfs_init_fs_context,
+	.parameters	= adfs_param_spec,
+};
+MODULE_ALIAS_FS("adfs");
+
+static int __init init_adfs_fs(void)
+{
+	int err = init_inodecache();
+	if (err)
+		goto out1;
+	err = register_filesystem(&adfs_fs_type);
+	if (err)
+		goto out;
+	return 0;
+out:
+	destroy_inodecache();
+out1:
+	return err;
+}
+
+static void __exit exit_adfs_fs(void)
+{
+	unregister_filesystem(&adfs_fs_type);
+	destroy_inodecache();
+}
+
+module_init(init_adfs_fs)
+module_exit(exit_adfs_fs)
+MODULE_DESCRIPTION("Acorn Disc Filing System");
+MODULE_LICENSE("GPL");