Re: [PATCH v10 2/4] fs: Add standard casefolding support

[Eric Biggers <ebiggers@kernel.org>]

On Tue, Jul 07, 2020 at 04:31:21AM -0700, Daniel Rosenberg wrote:
> +/*
> + * Determine if the name of a dentry should be casefolded. It does not make
> + * sense to casefold the no-key token of an encrypted filename.
> + *
> + * Return: if names will need casefolding
> + */
> +static bool needs_casefold(const struct inode *dir, const struct dentry *dentry)
> +{
> +	return IS_CASEFOLDED(dir) && dir->i_sb->s_encoding &&
> +			!(dentry->d_flags & DCACHE_ENCRYPTED_NAME);
> +}
> +
[...]
> +/**
> + * generic_ci_d_hash - generic d_hash implementation for casefolding filesystems
> + * @dentry:	dentry whose name we are hashing
> + * @str:	qstr of name whose hash we should fill in
> + *
> + * Return: 0 if hash was successful, or -ERRNO
> + */
> +int generic_ci_d_hash(const struct dentry *dentry, struct qstr *str)
> +{
> +	const struct inode *inode = READ_ONCE(dentry->d_inode);
> +	struct super_block *sb = dentry->d_sb;
> +	const struct unicode_map *um = sb->s_encoding;
> +	int ret = 0;
> +
> +	if (!inode || !needs_casefold(inode, dentry))
> +		return 0;
> +
> +	ret = utf8_casefold_hash(um, dentry, str);
> +	if (ret < 0)
> +		goto err;
> +
> +	return 0;
> +err:
> +	if (sb_has_strict_encoding(sb))
> +		ret = -EINVAL;
> +	else
> +		ret = 0;
> +	return ret;
> +}
> +EXPORT_SYMBOL(generic_ci_d_hash);

I thought this was discussed before, but the 'dentry' passed to ->d_hash() is
the parent dentry, not the one being hashed.

Therefore checking DCACHE_ENCRYPTED_NAME on 'dentry' is wrong here.  Instead we
need to use !fscrypt_has_encryption_key() here.  (IOW, while checking
DCACHE_ENCRYPTED_NAME is better *when possible*, it's not possible here.)

Note that the whole point of ->d_hash() is to hash the filename so that the VFS
can find the dentry.  If the VFS already had the dentry, there would be no need
for ->d_hash().

Also, did you consider my suggestion to not handle encrypt+casefold in this
patch?  I'd like to get this series in as a refactoring for 5.9.  The encryption
handling (which is new) might better belong in a later patch series.

- Eric