public inbox for linux-btrfs@vger.kernel.org
 help / color / mirror / Atom feed
From: Josef Bacik <josef@toxicpanda.com>
To: Sweet Tea Dorminy <sweettea-kernel@dorminy.me>
Cc: "Theodore Y. Ts'o" <tytso@mit.edu>,
	Jaegeuk Kim <jaegeuk@kernel.org>,
	Eric Biggers <ebiggers@kernel.org>, Chris Mason <clm@fb.com>,
	David Sterba <dsterba@suse.com>,
	linux-fscrypt@vger.kernel.org, linux-btrfs@vger.kernel.org,
	kernel-team@meta.com
Subject: Re: [PATCH v2 15/17] btrfs: start tracking extent encryption context info
Date: Mon, 17 Jul 2023 14:11:23 -0400	[thread overview]
Message-ID: <20230717181123.GO691303@perftesting> (raw)
In-Reply-To: <dc1087224f6a330acd8ae80e7c81c70a0cd8ad30.1689564024.git.sweettea-kernel@dorminy.me>

On Sun, Jul 16, 2023 at 11:52:46PM -0400, Sweet Tea Dorminy wrote:
> This puts the long-preserved 1-byte encryption field to work, storing
> whether the extent is encrypted and the length of its fscrypt_context.
> Since there is no way for a btrfs inode to be encrypted right now, we
> set context length to be 0 in all locations. We then update all the
> locations which check the size of an extent item to expect the size to
> agree with the context length.
> 
> The 1-byte field packs the encryption type and context length together,
> to avoid a format change. Right now we don't anticipate very many
> encryption policies, so 2 bits should be ample; similarly right now we
> can't imagine a context larger than fscrypt's current contexts, which
> are 40 bytes, so 6 bits is ample to store the context's size; and
> therefore we can pack these together into the one-byte encryption field
> without touching other space reserved for future use.
> 

I don't like this design, I think it doesn't give us enough flexibility.

I think I'd rather use the whole u8 to have the encryption policy, and then the
first u32 of the extent context is the size.  So something like

struct btrfs_encryption_context {
	__le32 len;
	__u8 context[0];
};

struct btrfs_file_extent {
	struct btrfs_encryption_context fscrypt_context[0];
};

It takes up a little more space but is more flexible and enables us more
flexibility in the future.

> Signed-off-by: Sweet Tea Dorminy <sweettea-kernel@dorminy.me>
> ---
>  fs/btrfs/accessors.h    | 31 +++++++++++++++++++++++++++++++
>  fs/btrfs/file-item.c    |  8 ++++++++
>  fs/btrfs/fscrypt.h      | 24 ++++++++++++++++++++++++
>  fs/btrfs/inode.c        | 26 ++++++++++++++++++++++----
>  fs/btrfs/tree-checker.c | 37 +++++++++++++++++++++++++++++--------
>  fs/btrfs/tree-log.c     |  3 +++
>  6 files changed, 117 insertions(+), 12 deletions(-)
> 
> diff --git a/fs/btrfs/accessors.h b/fs/btrfs/accessors.h
> index ceadfc5d6c66..fbee45b9d9c2 100644
> --- a/fs/btrfs/accessors.h
> +++ b/fs/btrfs/accessors.h
> @@ -3,6 +3,8 @@
>  #ifndef BTRFS_ACCESSORS_H
>  #define BTRFS_ACCESSORS_H
>  
> +#include "fscrypt.h"
> +

I don't want to pull other dependencies into accessors.h, I want to make syncing
*something* into btrfs-progs easy.

>  struct btrfs_map_token {
>  	struct extent_buffer *eb;
>  	char *kaddr;
> @@ -936,6 +938,16 @@ BTRFS_SETGET_STACK_FUNCS(stack_file_extent_disk_num_bytes,
>  			 struct btrfs_file_extent_item, disk_num_bytes, 64);
>  BTRFS_SETGET_STACK_FUNCS(stack_file_extent_compression,
>  			 struct btrfs_file_extent_item, compression, 8);
> +BTRFS_SETGET_STACK_FUNCS(stack_file_extent_encryption,
> +			 struct btrfs_file_extent_item, encryption, 8);
> +static inline u8 btrfs_stack_file_extent_encryption_ctxsize(
> +	struct btrfs_file_extent_item *e)
> +{
> +	u8 ctxsize;
> +
> +	btrfs_unpack_encryption(e->encryption, NULL, &ctxsize);
> +	return ctxsize;
> +}

This helper can go into fscrypt with the other related helpers.

>  
>  
>  BTRFS_SETGET_FUNCS(file_extent_type, struct btrfs_file_extent_item, type, 8);
> @@ -958,6 +970,25 @@ BTRFS_SETGET_FUNCS(file_extent_encryption, struct btrfs_file_extent_item,
>  BTRFS_SETGET_FUNCS(file_extent_other_encoding, struct btrfs_file_extent_item,
>  		   other_encoding, 16);
>  
> +static inline u8
> +btrfs_file_extent_encryption_ctxsize(const struct extent_buffer *eb,
> +				     struct btrfs_file_extent_item *e)
> +{
> +	u8 ctxsize;
> +
> +	btrfs_unpack_encryption(btrfs_file_extent_encryption(eb, e),
> +				NULL, &ctxsize);
> +	return ctxsize;
> +}

Same for this one.  Thanks,

Josef

  reply	other threads:[~2023-07-17 18:11 UTC|newest]

Thread overview: 33+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2023-07-17  3:52 [PATCH v2 00/17] btrfs: add encryption feature Sweet Tea Dorminy
2023-07-17  3:52 ` [PATCH v2 01/17] btrfs: disable various operations on encrypted inodes Sweet Tea Dorminy
2023-07-17  3:52 ` [PATCH v2 02/17] btrfs: disable verity " Sweet Tea Dorminy
2023-07-17  3:52 ` [PATCH v2 03/17] fscrypt: expose fscrypt_nokey_name Sweet Tea Dorminy
2023-07-17  3:52 ` [PATCH v2 04/17] btrfs: start using fscrypt hooks Sweet Tea Dorminy
2023-07-17 15:34   ` Luís Henriques
2023-07-17 17:28     ` David Sterba
2023-07-18  8:36       ` Luís Henriques
2023-07-17  3:52 ` [PATCH v2 05/17] btrfs: add inode encryption contexts Sweet Tea Dorminy
2023-07-17 15:41   ` Josef Bacik
2023-07-17  3:52 ` [PATCH v2 06/17] btrfs: add new FEATURE_INCOMPAT_ENCRYPT flag Sweet Tea Dorminy
2023-07-17 15:42   ` Josef Bacik
2023-07-17  3:52 ` [PATCH v2 07/17] btrfs: adapt readdir for encrypted and nokey names Sweet Tea Dorminy
2023-07-17 15:34   ` Luís Henriques
2023-07-17 17:46   ` Josef Bacik
2023-07-17  3:52 ` [PATCH v2 08/17] btrfs: use correct name hash for " Sweet Tea Dorminy
2023-07-17  3:52 ` [PATCH v2 09/17] btrfs: implement fscrypt ioctls Sweet Tea Dorminy
2023-07-17  3:52 ` [PATCH v2 10/17] btrfs: add encryption to CONFIG_BTRFS_DEBUG Sweet Tea Dorminy
2023-07-17  3:52 ` [PATCH v2 11/17] btrfs: add get_devices hook for fscrypt Sweet Tea Dorminy
2023-07-17 17:51   ` Josef Bacik
2023-07-17  3:52 ` [PATCH v2 12/17] btrfs: turn on inlinecrypt mount option for encrypt Sweet Tea Dorminy
2023-07-17 15:34   ` Luís Henriques
2023-07-17 17:55   ` Josef Bacik
2023-07-17  3:52 ` [PATCH v2 13/17] btrfs: turn on the encryption ioctls Sweet Tea Dorminy
2023-07-17  3:52 ` [PATCH v2 14/17] btrfs: create and free extent fscrypt_infos Sweet Tea Dorminy
2023-07-17 17:58   ` Josef Bacik
2023-07-17  3:52 ` [PATCH v2 15/17] btrfs: start tracking extent encryption context info Sweet Tea Dorminy
2023-07-17 18:11   ` Josef Bacik [this message]
2023-07-17  3:52 ` [PATCH v2 16/17] btrfs: explicitly track file extent length and encryption Sweet Tea Dorminy
2023-07-17 15:30   ` Josef Bacik
2023-07-17 18:12   ` Josef Bacik
2023-07-17  3:52 ` [PATCH v2 17/17] btrfs: save and load fscrypt extent contexts Sweet Tea Dorminy
2023-07-17 18:15   ` Josef Bacik

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20230717181123.GO691303@perftesting \
    --to=josef@toxicpanda.com \
    --cc=clm@fb.com \
    --cc=dsterba@suse.com \
    --cc=ebiggers@kernel.org \
    --cc=jaegeuk@kernel.org \
    --cc=kernel-team@meta.com \
    --cc=linux-btrfs@vger.kernel.org \
    --cc=linux-fscrypt@vger.kernel.org \
    --cc=sweettea-kernel@dorminy.me \
    --cc=tytso@mit.edu \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox