From patchwork Wed Mar 30 07:56:11 2016 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Qu Wenruo X-Patchwork-Id: 8693311 Return-Path: X-Original-To: patchwork-linux-btrfs@patchwork.kernel.org Delivered-To: patchwork-parsemail@patchwork1.web.kernel.org Received: from mail.kernel.org (mail.kernel.org [198.145.29.136]) by patchwork1.web.kernel.org (Postfix) with ESMTP id 4B6DF9F39A for ; Wed, 30 Mar 2016 07:56:52 +0000 (UTC) Received: from mail.kernel.org (localhost [127.0.0.1]) by mail.kernel.org (Postfix) with ESMTP id AB36C2027D for ; Wed, 30 Mar 2016 07:56:47 +0000 (UTC) Received: from vger.kernel.org (vger.kernel.org [209.132.180.67]) by mail.kernel.org (Postfix) with ESMTP id 7F81A202EB for ; Wed, 30 Mar 2016 07:56:46 +0000 (UTC) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1753316AbcC3H4o (ORCPT ); Wed, 30 Mar 2016 03:56:44 -0400 Received: from cn.fujitsu.com ([222.73.24.84]:29930 "EHLO song.cn.fujitsu.com" rhost-flags-OK-FAIL-OK-OK) by vger.kernel.org with ESMTP id S1753035AbcC3H4n (ORCPT ); Wed, 30 Mar 2016 03:56:43 -0400 X-IronPort-AV: E=Sophos;i="5.20,367,1444665600"; d="scan'208";a="416531" Received: from unknown (HELO cn.fujitsu.com) ([10.167.250.3]) by song.cn.fujitsu.com with ESMTP; 30 Mar 2016 15:56:16 +0800 Received: from localhost.localdomain (unknown [10.167.226.34]) by cn.fujitsu.com (Postfix) with ESMTP id A82264056422; Wed, 30 Mar 2016 15:56:15 +0800 (CST) From: Qu Wenruo To: linux-btrfs@vger.kernel.org Cc: Wang Xiaoguang Subject: [PATCH v9 16/19] btrfs: dedupe: Add support for on-disk hash search Date: Wed, 30 Mar 2016 15:56:11 +0800 Message-Id: <1459324574-28063-17-git-send-email-quwenruo@cn.fujitsu.com> X-Mailer: git-send-email 2.7.4 In-Reply-To: <1459324574-28063-1-git-send-email-quwenruo@cn.fujitsu.com> References: <1459324574-28063-1-git-send-email-quwenruo@cn.fujitsu.com> MIME-Version: 1.0 X-yoursite-MailScanner-ID: A82264056422.AB795 X-yoursite-MailScanner: Found to be clean X-yoursite-MailScanner-From: quwenruo@cn.fujitsu.com X-Spam-Status: No, score=-7.9 required=5.0 tests=BAYES_00, RCVD_IN_DNSWL_HI, RP_MATCHES_RCVD, UNPARSEABLE_RELAY autolearn=ham version=3.3.1 Sender: linux-btrfs-owner@vger.kernel.org Precedence: bulk List-ID: X-Mailing-List: linux-btrfs@vger.kernel.org X-Spam-Checker-Version: SpamAssassin 3.3.1 (2010-03-16) on mail.kernel.org X-Virus-Scanned: ClamAV using ClamSMTP Now on-disk backend should be able to search hash now. Signed-off-by: Wang Xiaoguang Signed-off-by: Qu Wenruo --- 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,