[PATCH 0/2] ext4: support encryption with blocksize != PAGE_SIZE

[PATCH 0/2] ext4: support encryption with blocksize != PAGE_SIZE

[Eric Biggers]

Hello,

This patchset makes ext4 support encryption on filesystems where the
filesystem block size is not equal to PAGE_SIZE.  This allows e.g.
PowerPC systems to use ext4 encryption.

Most of the work for this was already done in prior kernel releases; now
the only part missing is decryption support in block_read_full_page().
Chandan Rajendra has proposed a patchset "Consolidate FS read I/O
callbacks code" [1] to address this and do various other things like
make ext4 use mpage_readpages() again, and make ext4 and f2fs share more
code.  But it doesn't seem to be going anywhere.

Therefore, I propose we simply add decryption support to
block_read_full_page() for now.  This is a fairly small change, and it
gets ext4 encryption with subpage-sized blocks working.

Note: to keep things simple I'm just allocating the work object from the
bi_end_io function with GFP_ATOMIC.  But if people think it's necessary,
it could be changed to use preallocation like the page-based read path.

Tested with 'gce-xfstests -c ext4/encrypt_1k -g auto', using the new
"encrypt_1k" config I created.  All tests pass except for those that
already fail or are excluded with the encrypt or 1k configs, and 2 tests
that try to create 1023-byte symlinks which fails since encrypted
symlinks are limited to blocksize-3 bytes.  Also ran the dedicated
encryption tests using 'kvm-xfstests -c ext4/1k -g encrypt'; all pass,
including the on-disk ciphertext verification tests.

[1] https://lkml.kernel.org/linux-fsdevel/20190910155115.28550-1-chandan@linux.ibm.com/T/#u

Chandan Rajendra (1):
  ext4: Enable encryption for subpage-sized blocks

Eric Biggers (1):
  fs/buffer.c: support fscrypt in block_read_full_page()

 Documentation/filesystems/fscrypt.rst |  4 +--
 fs/buffer.c                           | 47 ++++++++++++++++++++++++---
 fs/ext4/super.c                       |  7 ----
 3 files changed, 44 insertions(+), 14 deletions(-)

-- 
2.23.0

[PATCH 1/2] fs/buffer.c: support fscrypt in block_read_full_page()

[Eric Biggers]

From: Eric Biggers <ebiggers@google.com>

After each filesystem block (as represented by a buffer_head) has been
read from disk by block_read_full_page(), decrypt it if needed.  The
decryption is done on the fscrypt_read_workqueue.

This is the final change needed to support ext4 encryption with
blocksize != PAGE_SIZE, and it's a fairly small change now that
CONFIG_FS_ENCRYPTION is a bool and fs/crypto/ exposes functions to
decrypt individual blocks and to enqueue work on the fscrypt workqueue.

Don't try to add fs-verity support yet, as the fs/verity/ support layer
isn't ready for sub-page blocks yet.  Just add fscrypt support for now.

Almost all the new code is compiled away when CONFIG_FS_ENCRYPTION=n.

Cc: Chandan Rajendra <chandan@linux.ibm.com>
Signed-off-by: Eric Biggers <ebiggers@google.com>
---
 fs/buffer.c | 47 ++++++++++++++++++++++++++++++++++++++++++-----
 1 file changed, 42 insertions(+), 5 deletions(-)

diff --git a/fs/buffer.c b/fs/buffer.c
index 86a38b979323..f26f64655246 100644
--- a/fs/buffer.c
+++ b/fs/buffer.c
@@ -47,6 +47,7 @@
 #include <linux/pagevec.h>
 #include <linux/sched/mm.h>
 #include <trace/events/block.h>
+#include <linux/fscrypt.h>
 
 static int fsync_buffers_list(spinlock_t *lock, struct list_head *list);
 static int submit_bh_wbc(int op, int op_flags, struct buffer_head *bh,
@@ -246,10 +247,6 @@ __find_get_block_slow(struct block_device *bdev, sector_t block)
 	return ret;
 }
 
-/*
- * I/O completion handler for block_read_full_page() - pages
- * which come unlocked at the end of I/O.
- */
 static void end_buffer_async_read(struct buffer_head *bh, int uptodate)
 {
 	unsigned long flags;
@@ -307,6 +304,46 @@ static void end_buffer_async_read(struct buffer_head *bh, int uptodate)
 	return;
 }
 
+struct decrypt_bh_ctx {
+	struct work_struct work;
+	struct buffer_head *bh;
+};
+
+static void decrypt_bh(struct work_struct *work)
+{
+	struct decrypt_bh_ctx *ctx =
+		container_of(work, struct decrypt_bh_ctx, work);
+	struct buffer_head *bh = ctx->bh;
+	int err;
+
+	err = fscrypt_decrypt_pagecache_blocks(bh->b_page, bh->b_size,
+					       bh_offset(bh));
+	end_buffer_async_read(bh, err == 0);
+	kfree(ctx);
+}
+
+/*
+ * I/O completion handler for block_read_full_page() - pages
+ * which come unlocked at the end of I/O.
+ */
+static void end_buffer_async_read_io(struct buffer_head *bh, int uptodate)
+{
+	/* Decrypt if needed */
+	if (uptodate && IS_ENABLED(CONFIG_FS_ENCRYPTION) &&
+	    IS_ENCRYPTED(bh->b_page->mapping->host)) {
+		struct decrypt_bh_ctx *ctx = kmalloc(sizeof(*ctx), GFP_ATOMIC);
+
+		if (ctx) {
+			INIT_WORK(&ctx->work, decrypt_bh);
+			ctx->bh = bh;
+			fscrypt_enqueue_decrypt_work(&ctx->work);
+			return;
+		}
+		uptodate = 0;
+	}
+	end_buffer_async_read(bh, uptodate);
+}
+
 /*
  * Completion handler for block_write_full_page() - pages which are unlocked
  * during I/O, and which have PageWriteback cleared upon I/O completion.
@@ -379,7 +416,7 @@ EXPORT_SYMBOL(end_buffer_async_write);
  */
 static void mark_buffer_async_read(struct buffer_head *bh)
 {
-	bh->b_end_io = end_buffer_async_read;
+	bh->b_end_io = end_buffer_async_read_io;
 	set_buffer_async_read(bh);
 }
 
-- 
2.23.0

Re: [PATCH 1/2] fs/buffer.c: support fscrypt in block_read_full_page()

[Eric Biggers]

On Wed, Oct 16, 2019 at 03:11:41PM -0700, Eric Biggers wrote:
> +static void end_buffer_async_read_io(struct buffer_head *bh, int uptodate)
> +{
> +	/* Decrypt if needed */
> +	if (uptodate && IS_ENABLED(CONFIG_FS_ENCRYPTION) &&
> +	    IS_ENCRYPTED(bh->b_page->mapping->host)) {
> +		struct decrypt_bh_ctx *ctx = kmalloc(sizeof(*ctx), GFP_ATOMIC);

Technically this should check S_ISREG() too (though it happens not to make a
difference currently).   I'll fix it in the next version of this patchset.

We probably should add a helper function fscrypt_needs_contents_encryption()
that returns IS_ENABLED(CONFIG_FS_ENCRYPTION) && IS_ENCRYPTED() && S_ISREG().

- Eric

[PATCH 2/2] ext4: Enable encryption for subpage-sized blocks

[Eric Biggers]

From: Chandan Rajendra <chandan@linux.ibm.com>

Now that we have the code to support encryption for subpage-sized
blocks, this commit removes the conditional check in filesystem mount
code.

The commit also changes the support statement in
Documentation/filesystems/fscrypt.rst to reflect the fact that
encryption on filesystems with blocksize less than page size now works.

[EB: Tested with 'gce-xfstests -c ext4/encrypt_1k -g auto', using the
new "encrypt_1k" config I created.  All tests pass except for those that
already fail or are excluded with the encrypt or 1k configs, and 2 tests
that try to create 1023-byte symlinks which fails since encrypted
symlinks are limited to blocksize-3 bytes.  Also ran the dedicated
encryption tests using 'kvm-xfstests -c ext4/1k -g encrypt'; all pass,
including the on-disk ciphertext verification tests.]

Signed-off-by: Chandan Rajendra <chandan@linux.ibm.com>
Signed-off-by: Eric Biggers <ebiggers@google.com>
---
 Documentation/filesystems/fscrypt.rst | 4 ++--
 fs/ext4/super.c                       | 7 -------
 2 files changed, 2 insertions(+), 9 deletions(-)

diff --git a/Documentation/filesystems/fscrypt.rst b/Documentation/filesystems/fscrypt.rst
index 8a0700af9596..b0d015a8cdc3 100644
--- a/Documentation/filesystems/fscrypt.rst
+++ b/Documentation/filesystems/fscrypt.rst
@@ -331,8 +331,8 @@ Contents encryption
 -------------------
 
 For file contents, each filesystem block is encrypted independently.
-Currently, only the case where the filesystem block size is equal to
-the system's page size (usually 4096 bytes) is supported.
+Starting from Linux kernel 5.5, encryption of filesystems with block
+size less than system's page size is supported.
 
 Each block's IV is set to the logical block number within the file as
 a little endian number, except that:
diff --git a/fs/ext4/super.c b/fs/ext4/super.c
index dd654e53ba3d..369f852bef20 100644
--- a/fs/ext4/super.c
+++ b/fs/ext4/super.c
@@ -4439,13 +4439,6 @@ static int ext4_fill_super(struct super_block *sb, void *data, int silent)
 		}
 	}
 
-	if ((DUMMY_ENCRYPTION_ENABLED(sbi) || ext4_has_feature_encrypt(sb)) &&
-	    (blocksize != PAGE_SIZE)) {
-		ext4_msg(sb, KERN_ERR,
-			 "Unsupported blocksize for fs encryption");
-		goto failed_mount_wq;
-	}
-
 	if (ext4_has_feature_verity(sb) && blocksize != PAGE_SIZE) {
 		ext4_msg(sb, KERN_ERR, "Unsupported blocksize for fs-verity");
 		goto failed_mount_wq;
-- 
2.23.0

Re: [PATCH 0/2] ext4: support encryption with blocksize != PAGE_SIZE

[Chandan Rajendra]

On Thursday, October 17, 2019 3:41 AM Eric Biggers wrote: 
> Hello,
> 
> This patchset makes ext4 support encryption on filesystems where the
> filesystem block size is not equal to PAGE_SIZE.  This allows e.g.
> PowerPC systems to use ext4 encryption.
> 
> Most of the work for this was already done in prior kernel releases; now
> the only part missing is decryption support in block_read_full_page().
> Chandan Rajendra has proposed a patchset "Consolidate FS read I/O
> callbacks code" [1] to address this and do various other things like
> make ext4 use mpage_readpages() again, and make ext4 and f2fs share more
> code.  But it doesn't seem to be going anywhere.
> 
> Therefore, I propose we simply add decryption support to
> block_read_full_page() for now.  This is a fairly small change, and it
> gets ext4 encryption with subpage-sized blocks working.
> 
> Note: to keep things simple I'm just allocating the work object from the
> bi_end_io function with GFP_ATOMIC.  But if people think it's necessary,
> it could be changed to use preallocation like the page-based read path.
> 
> Tested with 'gce-xfstests -c ext4/encrypt_1k -g auto', using the new
> "encrypt_1k" config I created.  All tests pass except for those that
> already fail or are excluded with the encrypt or 1k configs, and 2 tests
> that try to create 1023-byte symlinks which fails since encrypted
> symlinks are limited to blocksize-3 bytes.  Also ran the dedicated
> encryption tests using 'kvm-xfstests -c ext4/1k -g encrypt'; all pass,
> including the on-disk ciphertext verification tests.
> 
> [1] https://lkml.kernel.org/linux-fsdevel/20190910155115.28550-1-chandan@linux.ibm.com/T/#u
>

Hi Eric,

Thanks a lot for doing this.

The changes seem to be good. I have started test runs on my ppc64le guest and
I will reply with the test results once they complete.

-- 
chandan