linux-btrfs.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Qu Wenruo <quwenruo@cn.fujitsu.com>
To: linux-btrfs@vger.kernel.org
Cc: Wang Xiaoguang <wangxg.fnst@cn.fujitsu.com>
Subject: [PATCH v9 16/19] btrfs: dedupe: Add support for on-disk hash search
Date: Wed, 30 Mar 2016 15:56:11 +0800	[thread overview]
Message-ID: <1459324574-28063-17-git-send-email-quwenruo@cn.fujitsu.com> (raw)
In-Reply-To: <1459324574-28063-1-git-send-email-quwenruo@cn.fujitsu.com>

Now on-disk backend should be able to search hash now.

Signed-off-by: Wang Xiaoguang <wangxg.fnst@cn.fujitsu.com>
Signed-off-by: Qu Wenruo <quwenruo@cn.fujitsu.com>
---
 fs/btrfs/dedupe.c | 134 +++++++++++++++++++++++++++++++++++++++++++++++-------
 fs/btrfs/dedupe.h |   1 +
 2 files changed, 118 insertions(+), 17 deletions(-)

diff --git a/fs/btrfs/dedupe.c b/fs/btrfs/dedupe.c
index a274c1c..f2c2dde 100644
--- a/fs/btrfs/dedupe.c
+++ b/fs/btrfs/dedupe.c
@@ -651,6 +651,79 @@ int btrfs_dedupe_disable(struct btrfs_fs_info *fs_info)
 	return 0;
 }
 
+ /*
+ * Return 0 for not found
+ * Return >0 for found and set bytenr_ret
+ * Return <0 for error
+ */
+static int ondisk_search_hash(struct btrfs_dedupe_info *dedupe_info, u8 *hash,
+			      u64 *bytenr_ret, u32 *num_bytes_ret)
+{
+	struct btrfs_path *path;
+	struct btrfs_key key;
+	struct btrfs_root *dedupe_root = dedupe_info->dedupe_root;
+	u8 *buf = NULL;
+	u64 hash_key;
+	int hash_len = btrfs_dedupe_sizes[dedupe_info->hash_type];
+	int ret;
+
+	path = btrfs_alloc_path();
+	if (!path)
+		return -ENOMEM;
+
+	buf = kmalloc(hash_len, GFP_NOFS);
+	if (!buf) {
+		ret = -ENOMEM;
+		goto out;
+	}
+
+	memcpy(&hash_key, hash + hash_len - 8, 8);
+	key.objectid = hash_key;
+	key.type = BTRFS_DEDUPE_HASH_ITEM_KEY;
+	key.offset = (u64)-1;
+
+	ret = btrfs_search_slot(NULL, dedupe_root, &key, path, 0, 0);
+	if (ret < 0)
+		goto out;
+	WARN_ON(ret == 0);
+	while (1) {
+		struct extent_buffer *node;
+		struct btrfs_dedupe_hash_item *hash_item;
+		int slot;
+
+		ret = btrfs_previous_item(dedupe_root, path, hash_key,
+					  BTRFS_DEDUPE_HASH_ITEM_KEY);
+		if (ret < 0)
+			goto out;
+		if (ret > 0) {
+			ret = 0;
+			goto out;
+		}
+
+		node = path->nodes[0];
+		slot = path->slots[0];
+		btrfs_item_key_to_cpu(node, &key, slot);
+
+		if (key.type != BTRFS_DEDUPE_HASH_ITEM_KEY ||
+		    memcmp(&key.objectid, hash + hash_len - 8, 8))
+			break;
+		hash_item = btrfs_item_ptr(node, slot,
+				struct btrfs_dedupe_hash_item);
+		read_extent_buffer(node, buf, (unsigned long)(hash_item + 1),
+				   hash_len);
+		if (!memcmp(buf, hash, hash_len)) {
+			ret = 1;
+			*bytenr_ret = key.offset;
+			*num_bytes_ret = btrfs_dedupe_hash_len(node, hash_item);
+			break;
+		}
+	}
+out:
+	kfree(buf);
+	btrfs_free_path(path);
+	return ret;
+}
+
 /*
  * Caller must ensure the corresponding ref head is not being run.
  */
@@ -681,9 +754,36 @@ inmem_search_hash(struct btrfs_dedupe_info *dedupe_info, u8 *hash)
 	return NULL;
 }
 
-static int inmem_search(struct btrfs_dedupe_info *dedupe_info,
-			struct inode *inode, u64 file_pos,
-			struct btrfs_dedupe_hash *hash)
+/* Wapper for different backends, caller needs to hold dedupe_info->lock */
+static inline int generic_search_hash(struct btrfs_dedupe_info *dedupe_info,
+				      u8 *hash, u64 *bytenr_ret,
+				      u32 *num_bytes_ret)
+{
+	if (dedupe_info->backend == BTRFS_DEDUPE_BACKEND_INMEMORY) {
+		struct inmem_hash *found_hash;
+		int ret;
+
+		found_hash = inmem_search_hash(dedupe_info, hash);
+		if (found_hash) {
+			ret = 1;
+			*bytenr_ret = found_hash->bytenr;
+			*num_bytes_ret = found_hash->num_bytes;
+		} else {
+			ret = 0;
+			*bytenr_ret = 0;
+			*num_bytes_ret = 0;
+		}
+		return ret;
+	} else if (dedupe_info->backend == BTRFS_DEDUPE_BACKEND_ONDISK) {
+		return ondisk_search_hash(dedupe_info, hash, bytenr_ret,
+					  num_bytes_ret);
+	}
+	return -EINVAL;
+}
+
+static int generic_search(struct btrfs_dedupe_info *dedupe_info,
+			  struct inode *inode, u64 file_pos,
+			  struct btrfs_dedupe_hash *hash)
 {
 	int ret;
 	struct btrfs_root *root = BTRFS_I(inode)->root;
@@ -693,9 +793,9 @@ static int inmem_search(struct btrfs_dedupe_info *dedupe_info,
 	struct btrfs_delayed_ref_head *insert_head;
 	struct btrfs_delayed_data_ref *insert_dref;
 	struct btrfs_qgroup_extent_record *insert_qrecord = NULL;
-	struct inmem_hash *found_hash;
 	int free_insert = 1;
 	u64 bytenr;
+	u64 tmp_bytenr;
 	u32 num_bytes;
 
 	insert_head = kmem_cache_alloc(btrfs_delayed_ref_head_cachep, GFP_NOFS);
@@ -727,14 +827,9 @@ static int inmem_search(struct btrfs_dedupe_info *dedupe_info,
 
 again:
 	mutex_lock(&dedupe_info->lock);
-	found_hash = inmem_search_hash(dedupe_info, hash->hash);
-	/* If we don't find a duplicated extent, just return. */
-	if (!found_hash) {
-		ret = 0;
+	ret = generic_search_hash(dedupe_info, hash->hash, &bytenr, &num_bytes);
+	if (ret <= 0)
 		goto out;
-	}
-	bytenr = found_hash->bytenr;
-	num_bytes = found_hash->num_bytes;
 
 	delayed_refs = &trans->transaction->delayed_refs;
 
@@ -773,13 +868,17 @@ again:
 
 	mutex_lock(&dedupe_info->lock);
 	/* Search again to ensure the hash is still here */
-	found_hash = inmem_search_hash(dedupe_info, hash->hash);
-	if (!found_hash) {
-		ret = 0;
+	ret = generic_search_hash(dedupe_info, hash->hash, &tmp_bytenr,
+				  &num_bytes);
+	if (ret <= 0) {
 		mutex_unlock(&head->mutex);
 		goto out;
 	}
-	ret = 1;
+	if (tmp_bytenr != bytenr) {
+		mutex_unlock(&head->mutex);
+		mutex_unlock(&dedupe_info->lock);
+		goto again;
+	}
 	hash->bytenr = bytenr;
 	hash->num_bytes = num_bytes;
 
@@ -824,8 +923,9 @@ int btrfs_dedupe_search(struct btrfs_fs_info *fs_info,
 	if (WARN_ON(btrfs_dedupe_hash_hit(hash)))
 		return -EINVAL;
 
-	if (dedupe_info->backend == BTRFS_DEDUPE_BACKEND_INMEMORY)
-		ret = inmem_search(dedupe_info, inode, file_pos, hash);
+	if (dedupe_info->backend == BTRFS_DEDUPE_BACKEND_INMEMORY ||
+	    dedupe_info->backend == BTRFS_DEDUPE_BACKEND_ONDISK)
+		ret = generic_search(dedupe_info, inode, file_pos, hash);
 
 	/* It's possible hash->bytenr/num_bytenr already changed */
 	if (ret == 0) {
diff --git a/fs/btrfs/dedupe.h b/fs/btrfs/dedupe.h
index 2038ab8..bfcacd7 100644
--- a/fs/btrfs/dedupe.h
+++ b/fs/btrfs/dedupe.h
@@ -163,6 +163,7 @@ int btrfs_dedupe_calc_hash(struct btrfs_fs_info *fs_info,
  * *INCREASED*, and hash->bytenr/num_bytes will record the existing
  * extent data.
  * Return 0 for a hash miss. Nothing is done
+ * Return < 0 for error
  */
 int btrfs_dedupe_search(struct btrfs_fs_info *fs_info,
 			struct inode *inode, u64 file_pos,
-- 
2.7.4




  parent reply	other threads:[~2016-03-30  7:56 UTC|newest]

Thread overview: 26+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2016-03-30  7:55 [PATCH v9 00/19] Btrfs dedupe framework Qu Wenruo
2016-03-30  7:55 ` [PATCH v9 01/19] btrfs: dedupe: Introduce dedupe framework and its header Qu Wenruo
2016-03-30  7:55 ` [PATCH v9 02/19] btrfs: dedupe: Introduce function to initialize dedupe info Qu Wenruo
2016-03-30 10:10   ` kbuild test robot
2016-03-30  7:55 ` [PATCH v9 03/19] btrfs: dedupe: Introduce function to add hash into in-memory tree Qu Wenruo
2016-03-30  7:55 ` [PATCH v9 04/19] btrfs: dedupe: Introduce function to remove hash from " Qu Wenruo
2016-03-30  7:56 ` [PATCH v9 05/19] btrfs: delayed-ref: Add support for increasing data ref under spinlock Qu Wenruo
2016-03-30  7:56 ` [PATCH v9 06/19] btrfs: dedupe: Introduce function to search for an existing hash Qu Wenruo
2016-03-30  7:56 ` [PATCH v9 07/19] btrfs: dedupe: Implement btrfs_dedupe_calc_hash interface Qu Wenruo
2016-03-30  7:56 ` [PATCH v9 08/19] btrfs: ordered-extent: Add support for dedupe Qu Wenruo
2016-03-30  7:56 ` [PATCH v9 09/19] btrfs: dedupe: Inband in-memory only de-duplication implement Qu Wenruo
2016-03-30  7:56 ` [PATCH v9 10/19] btrfs: dedupe: Add ioctl for inband dedupelication Qu Wenruo
2016-03-30  7:56 ` [PATCH v9 11/19] btrfs: dedupe: add an inode nodedupe flag Qu Wenruo
2016-03-30  7:56 ` [PATCH v9 12/19] btrfs: dedupe: add a property handler for online dedupe Qu Wenruo
2016-03-30  7:56 ` [PATCH v9 13/19] btrfs: dedupe: add per-file online dedupe control Qu Wenruo
2016-03-30  7:56 ` [PATCH v9 14/19] btrfs: dedupe: Add basic tree structure for on-disk dedupe method Qu Wenruo
2016-03-30  7:56 ` [PATCH v9 15/19] btrfs: dedupe: Introduce interfaces to resume and cleanup dedupe info Qu Wenruo
2016-03-30  9:45   ` kbuild test robot
2016-03-30  7:56 ` Qu Wenruo [this message]
2016-03-30  7:56 ` [PATCH v9 17/19] btrfs: dedupe: Add support to delete hash for on-disk backend Qu Wenruo
2016-03-30  7:56 ` [PATCH v9 18/19] btrfs: dedupe: Add support for adding " Qu Wenruo
2016-03-30  7:56 ` [PATCH v9 19/19] btrfs: dedupe: Preparation for compress-dedupe co-work Qu Wenruo
2016-03-31 16:12 ` [PATCH v9 00/19] Btrfs dedupe framework David Sterba
2016-04-01  0:26   ` Qu Wenruo
2016-04-01 16:41     ` David Sterba
2016-04-01  2:01   ` Qu Wenruo

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=1459324574-28063-17-git-send-email-quwenruo@cn.fujitsu.com \
    --to=quwenruo@cn.fujitsu.com \
    --cc=linux-btrfs@vger.kernel.org \
    --cc=wangxg.fnst@cn.fujitsu.com \
    /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;
as well as URLs for NNTP newsgroup(s).