Re: [PATCH 3/5] ext4: Implement ci comparison using fscrypt_name

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



Eric Biggers <ebiggers@xxxxxxxxxx> writes:

> On Tue, Mar 29, 2022 at 12:11:04PM -0400, Gabriel Krisman Bertazi wrote:
>> Eric Biggers <ebiggers@xxxxxxxxxx> writes:
>> 
>> > On Mon, Mar 21, 2022 at 11:00:02PM -0400, Gabriel Krisman Bertazi wrote:
>> >> By using fscrypt_name here, we can hide most of the caching casefold
>> >> logic from ext4.  The condition in ext4_match is now quite redundant,
>> >> but this is addressed in the next patch.
>> >> 
>> >> Signed-off-by: Gabriel Krisman Bertazi <krisman@xxxxxxxxxxxxx>
>> >> ---
>> >>  fs/ext4/namei.c         | 26 ++++++++++++--------------
>> >>  include/linux/fscrypt.h |  4 ++++
>> >>  2 files changed, 16 insertions(+), 14 deletions(-)
>> >> 
>> >> diff --git a/fs/ext4/namei.c b/fs/ext4/namei.c
>> >> index 8976e5a28c73..71b4b05fae89 100644
>> >> --- a/fs/ext4/namei.c
>> >> +++ b/fs/ext4/namei.c
>> >> @@ -1321,10 +1321,9 @@ static void dx_insert_block(struct dx_frame *frame, u32 hash, ext4_lblk_t block)
>> >>  /**
>> >>   * ext4_ci_compare() - Match (case-insensitive) a name with a dirent.
>> >>   * @parent: Inode of the parent of the dentry.
>> >> - * @name: name under lookup.
>> >> + * @fname: name under lookup.
>> >>   * @de_name: Dirent name.
>> >>   * @de_name_len: dirent name length.
>> >> - * @quick: whether @name is already casefolded.
>> >>   *
>> >>   * Test whether a case-insensitive directory entry matches the filename
>> >>   * being searched.  If quick is set, the @name being looked up is
>> >> @@ -1333,8 +1332,9 @@ static void dx_insert_block(struct dx_frame *frame, u32 hash, ext4_lblk_t block)
>> >>   * Return: > 0 if the directory entry matches, 0 if it doesn't match, or
>> >>   * < 0 on error.
>> >>   */
>> >> -static int ext4_ci_compare(const struct inode *parent, const struct qstr *name,
>> >> -			   u8 *de_name, size_t de_name_len, bool quick)
>> >> +static int ext4_ci_compare(const struct inode *parent,
>> >> +			   const struct fscrypt_name *fname,
>> >> +			   u8 *de_name, size_t de_name_len)
>> >>  {
>> >>  	const struct super_block *sb = parent->i_sb;
>> >>  	const struct unicode_map *um = sb->s_encoding;
>> >> @@ -1357,10 +1357,10 @@ static int ext4_ci_compare(const struct inode *parent, const struct qstr *name,
>> >>  		entry.len = decrypted_name.len;
>> >>  	}
>> >>  
>> >> -	if (quick)
>> >> -		ret = utf8_strncasecmp_folded(um, name, &entry);
>> >> +	if (fname->cf_name.name)
>> >> +		ret = utf8_strncasecmp_folded(um, &fname->cf_name, &entry);
>> >>  	else
>> >> -		ret = utf8_strncasecmp(um, name, &entry);
>> >> +		ret = utf8_strncasecmp(um, fname->usr_fname, &entry);
>> >>  
>> >>  	if (!ret)
>> >>  		match = true;
>> >> @@ -1370,8 +1370,8 @@ static int ext4_ci_compare(const struct inode *parent, const struct qstr *name,
>> >>  		 * the names have invalid characters.
>> >>  		 */
>> >>  		ret = 0;
>> >> -		match = ((name->len == entry.len) &&
>> >> -			 !memcmp(name->name, entry.name, entry.len));
>> >> +		match = ((fname->usr_fname->len == entry.len) &&
>> >> +			 !memcmp(fname->usr_fname->name, entry.name, entry.len));
>> >>  	}
>> >>  
>> >>  out:
>> >> @@ -1440,6 +1440,8 @@ static bool ext4_match(struct inode *parent,
>> >>  #endif
>> >>  
>> >>  #if IS_ENABLED(CONFIG_UNICODE)
>> >> +	f.cf_name = fname->cf_name;
>> >> +
>> >>  	if (parent->i_sb->s_encoding && IS_CASEFOLDED(parent) &&
>> >>  	    (!IS_ENCRYPTED(parent) || fscrypt_has_encryption_key(parent))) {
>> >>  		if (fname->cf_name.name) {
>> >> @@ -1451,13 +1453,9 @@ static bool ext4_match(struct inode *parent,
>> >>  					return false;
>> >>  				}
>> >>  			}
>> >> -			ret = ext4_ci_compare(parent, &fname->cf_name, de->name,
>> >> -					      de->name_len, true);
>> >> -		} else {
>> >> -			ret = ext4_ci_compare(parent, fname->usr_fname,
>> >> -					      de->name, de->name_len, false);
>> >>  		}
>> >>  
>> >> +		ret = ext4_ci_compare(parent, &f, de->name, de->name_len);
>> >>  		if (ret < 0) {
>> >>  			/*
>> >>  			 * Treat comparison errors as not a match.  The
>> >> diff --git a/include/linux/fscrypt.h b/include/linux/fscrypt.h
>> >> index 91ea9477e9bd..5dc4b3c805e4 100644
>> >> --- a/include/linux/fscrypt.h
>> >> +++ b/include/linux/fscrypt.h
>> >> @@ -36,6 +36,10 @@ struct fscrypt_name {
>> >>  	u32 minor_hash;
>> >>  	struct fscrypt_str crypto_buf;
>> >>  	bool is_nokey_name;
>> >> +
>> >> +#ifdef CONFIG_UNICODE
>> >> +	struct qstr cf_name;
>> >> +#endif
>> >>  };
>> >>  
>> >
>> > This seems like the wrong approach.  struct fscrypt_name shouldn't have fields
>> > that aren't used by the fs/crypto/ layer.
>> >
>> > Did you check what f2fs does?  It has a struct f2fs_filename to represent
>> > everything f2fs needs to know about a filename, and it only uses
>> > struct fscrypt_name when communicating with the fs/crypto/ layer.
>> >
>> > struct ext4_filename already exists.  Couldn't you use that here?
>> 
>> Hi Eric,
>> 
>> The reason I'm not using struct ext4_filename here is because I'm trying
>> to make this generic, so this function can be shared across filesystems
>> implementing casefold.  Since the fscrypt_name abstraction is used for
>> case-sensitive comparison, I was trying to reuse that type for
>> case-insensitive as well.  It seemed unnecessary to define a generic
>> casefold_name type just for passing the cf_name and disk_name to this
>> function, considering that fscrypt_name is already initialized by
>> ext4_match.
>> 
>
> Which function, specifically, are you trying to share across filesystems?
> Do you have patches that show what your end goal is?

ext4_ci_compare/f2fs_match_ci_name :)

Let me follow up with a v2 that merges them so it makes more sense.

-- 
Gabriel Krisman Bertazi



[Index of Archives]     [Reiser Filesystem Development]     [Ceph FS]     [Kernel Newbies]     [Security]     [Netfilter]     [Bugtraq]     [Linux FS]     [Yosemite National Park]     [MIPS Linux]     [ARM Linux]     [Linux Security]     [Linux RAID]     [Samba]     [Device Mapper]     [Linux Media]

  Powered by Linux