From patchwork Tue Jun 7 15:07:52 2011 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: "Amir G." X-Patchwork-Id: 99247 X-Patchwork-Delegate: tytso@mit.edu Return-Path: X-Original-To: patchwork-incoming@ozlabs.org Delivered-To: patchwork-incoming@ozlabs.org Received: from vger.kernel.org (vger.kernel.org [209.132.180.67]) by ozlabs.org (Postfix) with ESMTP id 3987DB6F91 for ; Wed, 8 Jun 2011 01:10:10 +1000 (EST) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1756262Ab1FGPKI (ORCPT ); Tue, 7 Jun 2011 11:10:08 -0400 Received: from mail-ww0-f44.google.com ([74.125.82.44]:41767 "EHLO mail-ww0-f44.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1753053Ab1FGPKG (ORCPT ); Tue, 7 Jun 2011 11:10:06 -0400 Received: by mail-ww0-f44.google.com with SMTP id 36so5035787wwa.1 for ; Tue, 07 Jun 2011 08:10:05 -0700 (PDT) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=gamma; h=domainkey-signature:sender:from:to:cc:subject:date:message-id :x-mailer:in-reply-to:references; bh=OHlsfc8fEERSomQiC+8+IxOs6sEWa+F4ih1fzJ3HB8Q=; b=vwgT1AFvelbL0x25HSAoS1UzbSwOYxYqfpmVHJ8IlWUjZHD0FIhC4//YUtnFiU35PU RBwynbuvDIrQboSGy549n2IMZ9HLvE5Z4TvzLlRF98qHXwyspE1eEERN0sHxJN20tHsA 3cMBnWZ9wezx0IRlvJFkwicq/zPDJKswn0DUs= DomainKey-Signature: a=rsa-sha1; c=nofws; d=gmail.com; s=gamma; h=sender:from:to:cc:subject:date:message-id:x-mailer:in-reply-to :references; b=DntXtGtKhKe8HVRV+dRS9eVgRmIEHyiguv+m4szUvkXLLmeXVipRsdFS0SmEcVd3+g ekBGPxCodJ/dJjvXOaOMUtCYYp/pGh+w6+MYCe9D+ft017gtExCUVAkroI+0TiXw6sMk GbjHz+w9eIzxYrSde/meODYrtLy/+s0RFrRUs= Received: by 10.227.100.219 with SMTP id z27mr6452682wbn.45.1307459405575; Tue, 07 Jun 2011 08:10:05 -0700 (PDT) Received: from localhost.localdomain (bzq-218-153-66.cablep.bezeqint.net [81.218.153.66]) by mx.google.com with ESMTPS id en1sm3622645wbb.52.2011.06.07.08.10.02 (version=TLSv1/SSLv3 cipher=OTHER); Tue, 07 Jun 2011 08:10:05 -0700 (PDT) From: amir73il@users.sourceforge.net To: linux-ext4@vger.kernel.org Cc: tytso@mit.edu, lczerner@redhat.com, Amir Goldstein , Yongqiang Yang Subject: [PATCH v1 25/36] ext4: snapshot journaled - cache last COW tid in journal_head Date: Tue, 7 Jun 2011 18:07:52 +0300 Message-Id: <1307459283-22130-26-git-send-email-amir73il@users.sourceforge.net> X-Mailer: git-send-email 1.7.4.1 In-Reply-To: <1307459283-22130-1-git-send-email-amir73il@users.sourceforge.net> References: <1307459283-22130-1-git-send-email-amir73il@users.sourceforge.net> Sender: linux-ext4-owner@vger.kernel.org Precedence: bulk List-ID: X-Mailing-List: linux-ext4@vger.kernel.org From: Amir Goldstein Cache last COW transaction id in buffer's journal_head. The cache suppresses COW tests until the transaction in committed. By default, the running transaction is committed every 5 seconds which implies an average COW cache expiry of 2.5 seconds. Before taking a new snapshot, the journal is flushed to disk and the current transaction in committed, so the COW cache is invalidated (as it should be). Signed-off-by: Amir Goldstein Signed-off-by: Yongqiang Yang --- fs/ext4/snapshot.c | 93 ++++++++++++++++++++++++++++++++++++++++++++++ fs/ext4/snapshot.h | 2 + fs/ext4/snapshot_debug.c | 7 +++ 3 files changed, 102 insertions(+), 0 deletions(-) diff --git a/fs/ext4/snapshot.c b/fs/ext4/snapshot.c index e86dc42..2724381 100644 --- a/fs/ext4/snapshot.c +++ b/fs/ext4/snapshot.c @@ -400,6 +400,90 @@ __ext4_snapshot_trace_cow(const char *where, handle_t *handle, #define ext4_snapshot_trace_cow(where, handle, sb, inode, bh, blk, cnt, cmd) #endif /* + * The last transaction ID during which the buffer has been COWed is stored in + * the b_cow_tid field of the journal_head struct. If we know that the buffer + * was COWed during the current transaction, we don't need to COW it again. + * [jbd_lock_bh_state()] + */ + +void init_ext4_snapshot_cow_cache(void) +{ +#ifdef CONFIG_EXT4_DEBUG + cow_cache_enabled = 1; +#endif +} + +#ifdef CONFIG_EXT4_DEBUG +#define cow_cache_enabled() (cow_cache_enabled) +#else +#define cow_cache_enabled() (1) +#endif + +#define test_cow_tid(jh, handle) \ + ((jh)->b_cow_tid == (handle)->h_transaction->t_tid) +#define set_cow_tid(jh, handle) \ + ((jh)->b_cow_tid = (handle)->h_transaction->t_tid) + +/* + * Journal COW cache functions. + * a block can only be COWed once per snapshot, + * so a block can only be COWed once per transaction, + * so a buffer that was COWed in the current transaction, + * doesn't need to be COWed. + * + * Return values: + * 1 - block was COWed in current transaction + * 0 - block wasn't COWed in current transaction + */ +static int +ext4_snapshot_test_cowed(handle_t *handle, struct buffer_head *bh) +{ + struct journal_head *jh; + + if (!cow_cache_enabled()) + return 0; + + /* check the COW tid in the journal head */ + if (bh && buffer_jbd(bh)) { + jbd_lock_bh_state(bh); + jh = bh2jh(bh); + if (jh && !test_cow_tid(jh, handle)) + jh = NULL; + jbd_unlock_bh_state(bh); + if (jh) + /* + * Block was already COWed in the running transaction, + * so we don't need to COW it again. + */ + return 1; + } + return 0; +} + +static void +ext4_snapshot_mark_cowed(handle_t *handle, struct buffer_head *bh) +{ + struct journal_head *jh; + + if (!cow_cache_enabled()) + return; + + if (bh && buffer_jbd(bh)) { + jbd_lock_bh_state(bh); + jh = bh2jh(bh); + if (jh && !test_cow_tid(jh, handle)) + /* + * this is the first time this block was COWed + * in the running transaction. + * update the COW tid in the journal head + * to mark that this block doesn't need to be COWed. + */ + set_cow_tid(jh, handle); + jbd_unlock_bh_state(bh); + } +} + +/* * Begin COW or move operation. * No locks needed here, because @handle is a per-task struct. */ @@ -479,6 +563,13 @@ int ext4_snapshot_test_and_cow(const char *where, handle_t *handle, snapshot_debug_hl(4, "active snapshot access denied!\n"); return -EPERM; } + /* check if the buffer was COWed in the current transaction */ + if (ext4_snapshot_test_cowed(handle, bh)) { + snapshot_debug_hl(4, "buffer found in COW cache - " + "skip block cow!\n"); + trace_cow_inc(handle, ok_jh); + return 0; + } /* BEGIN COWing */ ext4_snapshot_cow_begin(handle); @@ -572,6 +663,8 @@ int ext4_snapshot_test_and_cow(const char *where, handle_t *handle, test_pending_cow: cowed: + /* mark the buffer COWed in the current transaction */ + ext4_snapshot_mark_cowed(handle, bh); out: brelse(sbh); /* END COWing */ diff --git a/fs/ext4/snapshot.h b/fs/ext4/snapshot.h index 007fec0..44bac96 100644 --- a/fs/ext4/snapshot.h +++ b/fs/ext4/snapshot.h @@ -342,6 +342,7 @@ static inline int ext4_snapshot_get_delete_access(handle_t *handle, return ext4_snapshot_move(handle, inode, block, pcount, 1); } +extern void init_ext4_snapshot_cow_cache(void); /* snapshot_ctl.c */ @@ -364,6 +365,7 @@ extern void ext4_snapshot_destroy(struct super_block *sb); static inline int init_ext4_snapshot(void) { + init_ext4_snapshot_cow_cache(); return 0; } diff --git a/fs/ext4/snapshot_debug.c b/fs/ext4/snapshot_debug.c index 35f552a..265ee2c 100644 --- a/fs/ext4/snapshot_debug.c +++ b/fs/ext4/snapshot_debug.c @@ -47,10 +47,12 @@ static const char *snapshot_test_names[SNAPSHOT_TESTS_NUM] = { u16 snapshot_enable_test[SNAPSHOT_TESTS_NUM] __read_mostly = {0}; u8 snapshot_enable_debug __read_mostly = 1; +u8 cow_cache_enabled __read_mostly = 1; static struct dentry *snapshot_debug; static struct dentry *snapshot_version; static struct dentry *snapshot_test[SNAPSHOT_TESTS_NUM]; +static struct dentry *cow_cache; static char snapshot_version_str[] = EXT4_SNAPSHOT_VERSION; static struct debugfs_blob_wrapper snapshot_version_blob = { @@ -79,6 +81,9 @@ void ext4_snapshot_create_debugfs_entry(struct dentry *debugfs_dir) S_IRUGO|S_IWUSR, debugfs_dir, &snapshot_enable_test[i]); + cow_cache = debugfs_create_u8("cow-cache", S_IRUGO|S_IWUSR, + debugfs_dir, + &cow_cache_enabled); } /* @@ -89,6 +94,8 @@ void ext4_snapshot_remove_debugfs_entry(void) { int i; + if (cow_cache) + debugfs_remove(cow_cache); for (i = 0; i < SNAPSHOT_TESTS_NUM && i < SNAPSHOT_TEST_NAMES; i++) if (snapshot_test[i]) debugfs_remove(snapshot_test[i]);