Re: [PATCH v4 11/14] digest_cache: Reset digest cache on file/directory change

["Jarkko Sakkinen" <jarkko@kernel.org>]

On Mon Apr 15, 2024 at 5:24 PM EEST, Roberto Sassu wrote:
> From: Roberto Sassu <roberto.sassu@huawei.com>
>
> Register six new LSM hooks, path_truncate, file_release, inode_unlink,
> inode_rename, inode_post_setxattr and inode_post_removexattr, to monitor
> digest lists/directory modifications.
>
> If an action affects a digest list or the parent directory, the new LSM
> hook implementations call digest_cache_reset_owner() to set the RESET bit
> (if unset) on the digest cache pointed to by dig_owner in the inode
> security blob. This will cause next calls to digest_cache_get() and
> digest_cache_create() to respectively put and clear dig_user and dig_owner,
> and request a new digest cache.
>
> If an action affects a file using a digest cache, the new LSM hook
> implementations call digest_cache_reset_user() to set the RESET_USER bit
> (if unset) on the digest cache pointed to by dig_user in the inode security
> blob. This will cause next calls to digest_cache_get() to put and clear
> dig_user, and retrieve the digest cache again.
>
> That does not affect other users of the old digest cache, since that one
> remains valid as long as the reference count is greater than zero. However,
> they will be notified in a subsequent patch about the reset, so that they
> can eventually request a new digest cache.
>
> Recreating a file digest cache means reading the digest list again and
> extracting the digests. Recreating a directory digest cache, instead, does
> not mean recreating the digest cache for directory entries, since those
> digest caches are likely already stored in the inode security blob. It
> would happen however for new files.
>
> Dig_owner reset for file digest caches is done on path_truncate, when a
> digest list is truncated (there is no inode_truncate, file_truncate does
> not catch operations through the truncate() system call), file_release,
> when a digest list opened for write is being closed, inode_unlink, when a
> digest list is removed, and inode_rename when a digest list is renamed.
>
> Dig_owner reset for directory digest caches is done on file_release, when a
> new digest list is written in the digest list directory, on inode_unlink,
> when a digest list is deleted from that directory, and finally on
> inode_rename, when a digest list is moved to/from that directory.
>
> Dig_user reset is always done on inode_post_setxattr and
> inode_post_removexattr, when the security.digest_list xattr is respectively
> set or removed from a file using a digest cache.
>
> With the exception of file_release, which will always be executed (cannot
> be denied), and inode_post_setxattr and inode_post_removexattr, which are
> executed after the actual operation, the other LSM hooks are not optimal,
> since the digest_cache LSM does not know whether or not the operation will
> be allowed also by other LSMs. If the operation is denied, the digest_cache
> LSM would do an unnecessary reset.
>
> Signed-off-by: Roberto Sassu <roberto.sassu@huawei.com>
> ---
>  security/digest_cache/Kconfig    |   1 +
>  security/digest_cache/Makefile   |   3 +-
>  security/digest_cache/dir.c      |   6 +
>  security/digest_cache/internal.h |  14 +++
>  security/digest_cache/main.c     |  19 +++
>  security/digest_cache/reset.c    | 197 +++++++++++++++++++++++++++++++
>  6 files changed, 239 insertions(+), 1 deletion(-)
>  create mode 100644 security/digest_cache/reset.c
>
> diff --git a/security/digest_cache/Kconfig b/security/digest_cache/Kconfig
> index cb4fa44e8f2a..54ba3a585073 100644
> --- a/security/digest_cache/Kconfig
> +++ b/security/digest_cache/Kconfig
> @@ -2,6 +2,7 @@
>  config SECURITY_DIGEST_CACHE
>  	bool "Digest_cache LSM"
>  	select TLV_PARSER
> +	select SECURITY_PATH
>  	default n
>  	help
>  	  This option enables an LSM maintaining a cache of digests
> diff --git a/security/digest_cache/Makefile b/security/digest_cache/Makefile
> index e417da0383ab..3d5e600a2c45 100644
> --- a/security/digest_cache/Makefile
> +++ b/security/digest_cache/Makefile
> @@ -4,7 +4,8 @@
>  
>  obj-$(CONFIG_SECURITY_DIGEST_CACHE) += digest_cache.o
>  
> -digest_cache-y := main.o secfs.o htable.o populate.o modsig.o verif.o dir.o
> +digest_cache-y := main.o secfs.o htable.o populate.o modsig.o verif.o dir.o \
> +		  reset.o
>  
>  digest_cache-y += parsers/tlv.o
>  digest_cache-y += parsers/rpm.o
> diff --git a/security/digest_cache/dir.c b/security/digest_cache/dir.c
> index a7d203c15386..937177660242 100644
> --- a/security/digest_cache/dir.c
> +++ b/security/digest_cache/dir.c
> @@ -148,6 +148,12 @@ digest_cache_dir_lookup_digest(struct dentry *dentry,
>  
>  	list_for_each_entry(dir_entry, &digest_cache->dir_entries, list) {
>  		mutex_lock(&dir_entry->digest_cache_mutex);
> +		if (dir_entry->digest_cache &&
> +		    test_bit(RESET, &dir_entry->digest_cache->flags)) {
> +			digest_cache_put(dir_entry->digest_cache);
> +			dir_entry->digest_cache = NULL;
> +		}
> +
>  		if (!dir_entry->digest_cache) {
>  			cache = digest_cache_create(dentry, digest_list_path,
>  						    digest_cache->path_str,
> diff --git a/security/digest_cache/internal.h b/security/digest_cache/internal.h
> index c13b35f6b2c0..c816929c4743 100644
> --- a/security/digest_cache/internal.h
> +++ b/security/digest_cache/internal.h
> @@ -18,6 +18,8 @@
>  #define INVALID			1	/* Digest cache marked as invalid. */
>  #define IS_DIR			2	/* Digest cache created from dir. */
>  #define DIR_PREFETCH		3	/* Prefetching requested for dir. */
> +#define RESET			4	/* Digest cache to be recreated. */
> +#define RESET_USER		5	/* Dig_user pointer to be released. */
>  
>  /**
>   * struct readdir_callback - Structure to store information for dir iteration
> @@ -247,4 +249,16 @@ digest_cache_dir_lookup_filename(struct dentry *dentry,
>  				 char *filename);
>  void digest_cache_dir_free(struct digest_cache *digest_cache);
>  
> +/* reset.c */
> +int digest_cache_path_truncate(const struct path *path);
> +void digest_cache_file_release(struct file *file);
> +int digest_cache_inode_unlink(struct inode *dir, struct dentry *dentry);
> +int digest_cache_inode_rename(struct inode *old_dir, struct dentry *old_dentry,
> +			      struct inode *new_dir, struct dentry *new_dentry);
> +void digest_cache_inode_post_setxattr(struct dentry *dentry, const char *name,
> +				      const void *value, size_t size,
> +				      int flags);
> +void digest_cache_inode_post_removexattr(struct dentry *dentry,
> +					 const char *name);
> +
>  #endif /* _DIGEST_CACHE_INTERNAL_H */
> diff --git a/security/digest_cache/main.c b/security/digest_cache/main.c
> index a5616fd07c1d..ce3518a33c80 100644
> --- a/security/digest_cache/main.c
> +++ b/security/digest_cache/main.c
> @@ -169,6 +169,11 @@ struct digest_cache *digest_cache_create(struct dentry *dentry,
>  
>  	/* Serialize check and assignment of dig_owner. */
>  	mutex_lock(&dig_sec->dig_owner_mutex);
> +	if (dig_sec->dig_owner && test_bit(RESET, &dig_sec->dig_owner->flags)) {
> +		digest_cache_put(dig_sec->dig_owner);
> +		dig_sec->dig_owner = NULL;
> +	}
> +
>  	if (dig_sec->dig_owner) {
>  		/* Increment ref. count for reference returned to the caller. */
>  		digest_cache = digest_cache_ref(dig_sec->dig_owner);
> @@ -403,6 +408,13 @@ struct digest_cache *digest_cache_get(struct dentry *dentry)
>  
>  	/* Serialize accesses to inode for which the digest cache is used. */
>  	mutex_lock(&dig_sec->dig_user_mutex);
> +	if (dig_sec->dig_user &&
> +	    (test_bit(RESET, &dig_sec->dig_user->flags) ||
> +	     test_bit(RESET_USER, &dig_sec->dig_user->flags))) {
> +		digest_cache_put(dig_sec->dig_user);
> +		dig_sec->dig_user = NULL;
> +	}
> +
>  	if (!dig_sec->dig_user) {
>  		down_read(&default_path_sem);
>  		/* Consume extra reference from digest_cache_create(). */
> @@ -491,6 +503,13 @@ static void digest_cache_inode_free_security(struct inode *inode)
>  static struct security_hook_list digest_cache_hooks[] __ro_after_init = {
>  	LSM_HOOK_INIT(inode_alloc_security, digest_cache_inode_alloc_security),
>  	LSM_HOOK_INIT(inode_free_security, digest_cache_inode_free_security),
> +	LSM_HOOK_INIT(path_truncate, digest_cache_path_truncate),
> +	LSM_HOOK_INIT(file_release, digest_cache_file_release),
> +	LSM_HOOK_INIT(inode_unlink, digest_cache_inode_unlink),
> +	LSM_HOOK_INIT(inode_rename, digest_cache_inode_rename),
> +	LSM_HOOK_INIT(inode_post_setxattr, digest_cache_inode_post_setxattr),
> +	LSM_HOOK_INIT(inode_post_removexattr,
> +		      digest_cache_inode_post_removexattr),
>  };
>  
>  /**
> diff --git a/security/digest_cache/reset.c b/security/digest_cache/reset.c
> new file mode 100644
> index 000000000000..e07222b0e0e8
> --- /dev/null
> +++ b/security/digest_cache/reset.c
> @@ -0,0 +1,197 @@
> +// SPDX-License-Identifier: GPL-2.0
> +/*
> + * Copyright (C) 2023-2024 Huawei Technologies Duesseldorf GmbH
> + *
> + * Author: Roberto Sassu <roberto.sassu@huawei.com>
> + *
> + * Reset digest cache on digest lists/directory modifications.
> + */
> +
> +#define pr_fmt(fmt) "DIGEST CACHE: "fmt
> +#include "internal.h"
> +
> +/**
> + * digest_cache_reset_owner - Reset dig_owner
> + * @inode: Inode of the digest list/directory containing the digest list
> + * @reason: Reason for reset
> + *
> + * This function sets the RESET bit of the digest cache pointed to by dig_owner
> + * (if unset), so that digest_cache_get() and digest_cache_create() respectively
> + * release and clear dig_user and dig_owner in the inode security blob. This
> + * causes new callers of digest_cache_get() to get a new digest cache.
> + */
> +static void digest_cache_reset_owner(struct inode *inode, const char *reason)
> +{
> +	struct digest_cache_security *dig_sec;
> +
> +	dig_sec = digest_cache_get_security(inode);
> +	if (unlikely(!dig_sec))
> +		return;
> +
> +	mutex_lock(&dig_sec->dig_owner_mutex);
> +	if (dig_sec->dig_owner &&
> +	    !test_and_set_bit(RESET, &dig_sec->dig_owner->flags))
> +		pr_debug("Resetting %s (dig_owner), reason: %s\n",
> +			 dig_sec->dig_owner->path_str, reason);
> +	mutex_unlock(&dig_sec->dig_owner_mutex);
> +}
> +
> +/**
> + * digest_cache_reset_user - Reset dig_user
> + * @inode: Inode of the file using the digest cache
> + * @reason: Reason for reset
> + *
> + * This function sets the RESET_USER bit (if unset), so that digest_cache_get()
> + * clears the dig_user pointer in the inode security blob and determines again
> + * the digest list inode to get a digest cache from.
> + */
> +static void digest_cache_reset_user(struct inode *inode, const char *reason)
> +{
> +	struct digest_cache_security *dig_sec;
> +
> +	dig_sec = digest_cache_get_security(inode);
> +	if (unlikely(!dig_sec))
> +		return;
> +
> +	mutex_lock(&dig_sec->dig_user_mutex);
> +	if (dig_sec->dig_user &&
> +	    !test_and_set_bit(RESET_USER, &dig_sec->dig_user->flags))
> +		pr_debug("Resetting %s (dig_user), reason: %s\n",
> +			 dig_sec->dig_user->path_str, reason);
> +	mutex_unlock(&dig_sec->dig_user_mutex);
> +}
> +
> +/**
> + * digest_cache_path_truncate - A file is being truncated
> + * @path: File path
> + *
> + * This function is called when a file is being truncated. If the inode is a
> + * digest list, it resets the inode dig_owner, to force rebuilding the digest
> + * cache.
> + *
> + * Return: Zero.
> + */
> +int digest_cache_path_truncate(const struct path *path)
> +{
> +	struct inode *inode = d_backing_inode(path->dentry);
> +
> +	if (!S_ISREG(inode->i_mode))
> +		return 0;
> +
> +	digest_cache_reset_owner(inode, "file_truncate");
> +	return 0;
> +}
> +
> +/**
> + * digest_cache_file_release - Last reference of a file desc is being released
> + * @file: File descriptor
> + *
> + * This function is called when the last reference of a file descriptor is
> + * being released. If the inode is a regular file and was opened for write, or
> + * the parent inode is the digest list directory and the file was created, it
> + * resets the inode dig_owner, to force rebuilding the digest cache.
> + */
> +void digest_cache_file_release(struct file *file)
> +{
> +	struct inode *dir = d_backing_inode(file_dentry(file)->d_parent);
> +
> +	if (!S_ISREG(file_inode(file)->i_mode) || !(file->f_mode & FMODE_WRITE))
> +		return;
> +
> +	digest_cache_reset_owner(file_inode(file), "file_file_release");
> +	if (file->f_mode & FMODE_CREATED)
> +		digest_cache_reset_owner(dir, "dir_file_release");
> +}
> +
> +/**
> + * digest_cache_inode_unlink - An inode is being removed
> + * @dir: Inode of the affected directory
> + * @dentry: Dentry of the inode being removed
> + *
> + * This function is called when an existing inode is being removed. If the
> + * inode is a digest list, or the parent inode is the digest list directory and
> + * the inode is a regular file, it resets the affected inode dig_owner, to force
> + * rebuilding the digest cache.
> + *
> + * Return: Zero.
> + */
> +int digest_cache_inode_unlink(struct inode *dir, struct dentry *dentry)
> +{
> +	struct inode *inode = d_backing_inode(dentry);
> +
> +	if (!S_ISREG(inode->i_mode))
> +		return 0;
> +
> +	digest_cache_reset_owner(inode, "file_unlink");
> +	digest_cache_reset_owner(dir, "dir_unlink");
> +	return 0;
> +}
> +
> +/**
> + * digest_cache_inode_rename - An inode is being renamed
> + * @old_dir: Inode of the directory containing the inode being renamed
> + * @old_dentry: Dentry of the inode being renamed
> + * @new_dir: Directory where the inode will be placed into
> + * @new_dentry: Dentry of the inode after being renamed
> + *
> + * This function is called when an existing inode is being moved from a
> + * directory to another (rename). If the inode is a digest list, or that inode
> + * is moved from/to the digest list directory, it resets the affected inode
> + * dig_owner, to force rebuilding the digest cache.
> + *
> + * Return: Zero.
> + */
> +int digest_cache_inode_rename(struct inode *old_dir, struct dentry *old_dentry,
> +			      struct inode *new_dir, struct dentry *new_dentry)
> +{
> +	struct inode *old_inode = d_backing_inode(old_dentry);
> +
> +	if (!S_ISREG(old_inode->i_mode))
> +		return 0;
> +
> +	digest_cache_reset_owner(old_inode, "file_rename");
> +	digest_cache_reset_owner(old_dir, "dir_rename_from");
> +	digest_cache_reset_owner(new_dir, "dir_rename_to");
> +	return 0;
> +}
> +
> +/**
> + * digest_cache_inode_post_setxattr() - An xattr was set
> + * @dentry: file
> + * @name: xattr name
> + * @value: xattr value
> + * @size: size of xattr value
> + * @flags: flags
> + *
> + * This function is called after an xattr was set on an existing inode. If the
> + * inode points to a digest cache and the xattr set is security.digest_list, it
> + * resets dig_user in the inode security blob, to force retrieving a fresh
> + * digest cache.
> + */
> +void digest_cache_inode_post_setxattr(struct dentry *dentry, const char *name,
> +				      const void *value, size_t size, int flags)
> +{
> +	if (strcmp(name, XATTR_NAME_DIGEST_LIST))
> +		return;
> +
> +	digest_cache_reset_user(d_backing_inode(dentry), "file_setxattr");
> +}
> +
> +/**
> + * digest_cache_inode_post_removexattr() - An xattr was removed
> + * @dentry: file
> + * @name: xattr name
> + *
> + * This function is called after an xattr was removed from an existing inode.
> + * If the inode points to a digest cache and the xattr removed is
> + * security.digest_list, it resets dig_user in the inode security blob, to
> + * force retrieving a fresh digest cache.
> + */
> +void digest_cache_inode_post_removexattr(struct dentry *dentry,
> +					 const char *name)

nit:

void digest_cache_inode_post_removexattr(struct dentry *dentry, const char *name)

would be fine too

> +{
> +	if (strcmp(name, XATTR_NAME_DIGEST_LIST))
> +		return;
> +
> +	digest_cache_reset_user(d_backing_inode(dentry), "file_removexattr");
> +}


BR, Jarkko