From patchwork Tue Jul 19 08:50:42 2016 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Qu Wenruo X-Patchwork-Id: 9236451 Return-Path: Received: from mail.wl.linuxfoundation.org (pdx-wl-mail.web.codeaurora.org [172.30.200.125]) by pdx-korg-patchwork.web.codeaurora.org (Postfix) with ESMTP id 61FC360574 for ; Tue, 19 Jul 2016 08:51:42 +0000 (UTC) Received: from mail.wl.linuxfoundation.org (localhost [127.0.0.1]) by mail.wl.linuxfoundation.org (Postfix) with ESMTP id 50BEB200F5 for ; Tue, 19 Jul 2016 08:51:42 +0000 (UTC) Received: by mail.wl.linuxfoundation.org (Postfix, from userid 486) id 458A626253; Tue, 19 Jul 2016 08:51:42 +0000 (UTC) X-Spam-Checker-Version: SpamAssassin 3.3.1 (2010-03-16) on pdx-wl-mail.web.codeaurora.org X-Spam-Level: X-Spam-Status: No, score=-6.9 required=2.0 tests=BAYES_00,RCVD_IN_DNSWL_HI autolearn=ham version=3.3.1 Received: from vger.kernel.org (vger.kernel.org [209.132.180.67]) by mail.wl.linuxfoundation.org (Postfix) with ESMTP id 8B37C200F5 for ; Tue, 19 Jul 2016 08:51:41 +0000 (UTC) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1752876AbcGSIvg (ORCPT ); Tue, 19 Jul 2016 04:51:36 -0400 Received: from cn.fujitsu.com ([222.73.24.84]:28573 "EHLO song.cn.fujitsu.com" rhost-flags-OK-FAIL-OK-OK) by vger.kernel.org with ESMTP id S1752791AbcGSIva (ORCPT ); Tue, 19 Jul 2016 04:51:30 -0400 X-IronPort-AV: E=Sophos;i="5.20,367,1444665600"; d="scan'208";a="675664" Received: from unknown (HELO cn.fujitsu.com) ([10.167.250.3]) by song.cn.fujitsu.com with ESMTP; 19 Jul 2016 16:50:56 +0800 Received: from adam-work.localdomain (unknown [10.167.226.34]) by cn.fujitsu.com (Postfix) with ESMTP id C97A5405640D; Tue, 19 Jul 2016 16:50:52 +0800 (CST) From: Qu Wenruo To: linux-btrfs@vger.kernel.org Cc: Wang Xiaoguang Subject: [PATCH v12.2 07/15] btrfs: dedupe: Introduce function to search for an existing hash Date: Tue, 19 Jul 2016 16:50:42 +0800 Message-Id: <20160719085050.30564-8-quwenruo@cn.fujitsu.com> X-Mailer: git-send-email 2.9.0 In-Reply-To: <20160719085050.30564-1-quwenruo@cn.fujitsu.com> References: <20160719085050.30564-1-quwenruo@cn.fujitsu.com> MIME-Version: 1.0 X-yoursite-MailScanner-ID: C97A5405640D.AC26E X-yoursite-MailScanner: Found to be clean X-yoursite-MailScanner-From: quwenruo@cn.fujitsu.com Sender: linux-btrfs-owner@vger.kernel.org Precedence: bulk List-ID: X-Mailing-List: linux-btrfs@vger.kernel.org X-Virus-Scanned: ClamAV using ClamSMTP From: Wang Xiaoguang Introduce static function inmem_search() to handle the job for in-memory hash tree. The trick is, we must ensure the delayed ref head is not being run at the time we search the for the hash. With inmem_search(), we can implement the btrfs_dedupe_search() interface. Signed-off-by: Qu Wenruo Signed-off-by: Wang Xiaoguang Reviewed-by: Josef Bacik --- fs/btrfs/dedupe.c | 185 ++++++++++++++++++++++++++++++++++++++++++++++++++++++ 1 file changed, 185 insertions(+) diff --git a/fs/btrfs/dedupe.c b/fs/btrfs/dedupe.c index 14c57fa..eb1b110 100644 --- a/fs/btrfs/dedupe.c +++ b/fs/btrfs/dedupe.c @@ -20,6 +20,7 @@ #include "btrfs_inode.h" #include "transaction.h" #include "delayed-ref.h" +#include "qgroup.h" struct inmem_hash { struct rb_node hash_node; @@ -454,3 +455,187 @@ int btrfs_dedupe_disable(struct btrfs_fs_info *fs_info) kfree(dedupe_info); return 0; } + +/* + * Caller must ensure the corresponding ref head is not being run. + */ +static struct inmem_hash * +inmem_search_hash(struct btrfs_dedupe_info *dedupe_info, u8 *hash) +{ + struct rb_node **p = &dedupe_info->hash_root.rb_node; + struct rb_node *parent = NULL; + struct inmem_hash *entry = NULL; + u16 hash_algo = dedupe_info->hash_algo; + int hash_len = btrfs_hash_sizes[hash_algo]; + + while (*p) { + parent = *p; + entry = rb_entry(parent, struct inmem_hash, hash_node); + + if (memcmp(hash, entry->hash, hash_len) < 0) { + p = &(*p)->rb_left; + } else if (memcmp(hash, entry->hash, hash_len) > 0) { + p = &(*p)->rb_right; + } else { + /* Found, need to re-add it to LRU list head */ + list_del(&entry->lru_list); + list_add(&entry->lru_list, &dedupe_info->lru_list); + return entry; + } + } + return NULL; +} + +static int inmem_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; + struct btrfs_trans_handle *trans; + struct btrfs_delayed_ref_root *delayed_refs; + struct btrfs_delayed_ref_head *head; + 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; + u32 num_bytes; + + insert_head = kmem_cache_alloc(btrfs_delayed_ref_head_cachep, GFP_NOFS); + if (!insert_head) + return -ENOMEM; + insert_head->extent_op = NULL; + insert_dref = kmem_cache_alloc(btrfs_delayed_data_ref_cachep, GFP_NOFS); + if (!insert_dref) { + kmem_cache_free(btrfs_delayed_ref_head_cachep, insert_head); + return -ENOMEM; + } + if (root->fs_info->quota_enabled && + is_fstree(root->root_key.objectid)) { + insert_qrecord = kmalloc(sizeof(*insert_qrecord), GFP_NOFS); + if (!insert_qrecord) { + kmem_cache_free(btrfs_delayed_ref_head_cachep, + insert_head); + kmem_cache_free(btrfs_delayed_data_ref_cachep, + insert_dref); + return -ENOMEM; + } + } + + trans = btrfs_join_transaction(root); + if (IS_ERR(trans)) { + ret = PTR_ERR(trans); + goto free_mem; + } + +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; + goto out; + } + bytenr = found_hash->bytenr; + num_bytes = found_hash->num_bytes; + + delayed_refs = &trans->transaction->delayed_refs; + + spin_lock(&delayed_refs->lock); + head = btrfs_find_delayed_ref_head(trans, bytenr); + if (!head) { + /* + * We can safely insert a new delayed_ref as long as we + * hold delayed_refs->lock. + * Only need to use atomic inc_extent_ref() + */ + btrfs_add_delayed_data_ref_locked(root->fs_info, trans, + insert_dref, insert_head, insert_qrecord, + bytenr, num_bytes, 0, root->root_key.objectid, + btrfs_ino(inode), file_pos, 0, + BTRFS_ADD_DELAYED_REF); + spin_unlock(&delayed_refs->lock); + + /* add_delayed_data_ref_locked will free unused memory */ + free_insert = 0; + hash->bytenr = bytenr; + hash->num_bytes = num_bytes; + ret = 1; + goto out; + } + + /* + * We can't lock ref head with dedupe_info->lock hold or we will cause + * ABBA dead lock. + */ + mutex_unlock(&dedupe_info->lock); + ret = btrfs_delayed_ref_lock(trans, head); + spin_unlock(&delayed_refs->lock); + if (ret == -EAGAIN) + goto 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; + mutex_unlock(&head->mutex); + goto out; + } + ret = 1; + hash->bytenr = bytenr; + hash->num_bytes = num_bytes; + + /* + * Increase the extent ref right now, to avoid delayed ref run + * Or we may increase ref on non-exist extent. + */ + btrfs_inc_extent_ref(trans, root, bytenr, num_bytes, 0, + root->root_key.objectid, + btrfs_ino(inode), file_pos); + mutex_unlock(&head->mutex); +out: + mutex_unlock(&dedupe_info->lock); + btrfs_end_transaction(trans, root); + +free_mem: + if (free_insert) { + kmem_cache_free(btrfs_delayed_ref_head_cachep, insert_head); + kmem_cache_free(btrfs_delayed_data_ref_cachep, insert_dref); + kfree(insert_qrecord); + } + return ret; +} + +int btrfs_dedupe_search(struct btrfs_fs_info *fs_info, + struct inode *inode, u64 file_pos, + struct btrfs_dedupe_hash *hash) +{ + struct btrfs_dedupe_info *dedupe_info = fs_info->dedupe_info; + int ret = -EINVAL; + + if (!hash) + return 0; + + /* + * This function doesn't follow fs_info->dedupe_enabled as it will need + * to ensure any hashed extent to go through dedupe routine + */ + if (WARN_ON(dedupe_info == NULL)) + return -EINVAL; + + 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); + + /* It's possible hash->bytenr/num_bytenr already changed */ + if (ret == 0) { + hash->num_bytes = 0; + hash->bytenr = 0; + } + return ret; +}