From patchwork Tue Jun 7 15:07:30 2011 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: "Amir G." X-Patchwork-Id: 99225 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 EBFF8B6FC1 for ; Wed, 8 Jun 2011 01:09:06 +1000 (EST) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1752832Ab1FGPJF (ORCPT ); Tue, 7 Jun 2011 11:09:05 -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 S1754241Ab1FGPJD (ORCPT ); Tue, 7 Jun 2011 11:09:03 -0400 Received: by mail-ww0-f44.google.com with SMTP id 36so5035787wwa.1 for ; Tue, 07 Jun 2011 08:09:02 -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=RwKVvk2M+UXrJOyE9UBvV+ztbuwteMVnuC4ZC6nyWnQ=; b=sV5ytN9LGEa4y3pHm8jfz4K/VviRA++s9YUpBAPhDIm6E6xcnzilJhoIuW7+GA1lwL oTIcacq5dfAT4STiwRna+i1LiLAyRjRQMIFcvb4YA+0BvukTFB5Bw+wZ84LasuxYY0WI yUkYJZ56PGRxGFLwUPB25fYVxL32gezojB4IY= 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=WJKdl3r8J78h4NoSpOoW5WoHZ4QaYSwidPFT9pwmAzqa282PGlbbjbOpXm9aKtcuX1 ms1n7IU8Iq3SLZcy5p3qGCTxwGk/PQ0NbZtyZVI9/VqKqoX5WNsV9RpYuPKtHqwDZK5I CjGClZHHClX/tYOscmZZPjxVFOzvsQifApVrA= Received: by 10.227.12.18 with SMTP id v18mr6384243wbv.89.1307459342559; Tue, 07 Jun 2011 08:09:02 -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.52.2011.06.07.08.09.00 (version=TLSv1/SSLv3 cipher=OTHER); Tue, 07 Jun 2011 08:09:02 -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 03/36] ext4: snapshot hooks - inside JBD hooks Date: Tue, 7 Jun 2011 18:07:30 +0300 Message-Id: <1307459283-22130-4-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 Before every metadata buffer write, the journal API is called, namely, one of the ext4_journal_get_XXX_access() functions. We use these journal hooks to call the snapshot API, namely ext4_snapshot_get_XXX_access(), to COW the metadata buffer before it is modified for the first time. Signed-off-by: Amir Goldstein Signed-off-by: Yongqiang Yang --- fs/ext4/ext4_jbd2.c | 9 ++++++- fs/ext4/ext4_jbd2.h | 15 ++++++++++--- fs/ext4/extents.c | 3 +- fs/ext4/inode.c | 22 ++++++++++++++------ fs/ext4/move_extent.c | 3 +- fs/ext4/snapshot.h | 51 +++++++++++++++++++++++++++++++++++++++++++++++++ 6 files changed, 88 insertions(+), 15 deletions(-) diff --git a/fs/ext4/ext4_jbd2.c b/fs/ext4/ext4_jbd2.c index 560020d..833969b 100644 --- a/fs/ext4/ext4_jbd2.c +++ b/fs/ext4/ext4_jbd2.c @@ -23,13 +23,16 @@ int __ext4_journal_get_undo_access(const char *where, unsigned int line, return err; } -int __ext4_journal_get_write_access(const char *where, unsigned int line, - handle_t *handle, struct buffer_head *bh) +int __ext4_journal_get_write_access_inode(const char *where, unsigned int line, + handle_t *handle, struct inode *inode, + struct buffer_head *bh, int exclude) { int err = 0; if (ext4_handle_valid(handle)) { err = jbd2_journal_get_write_access(handle, bh); + if (!err && !exclude) + err = ext4_snapshot_get_write_access(handle, inode, bh); if (err) ext4_journal_abort_handle(where, line, __func__, bh, handle, err); @@ -111,6 +114,8 @@ int __ext4_journal_get_create_access(const char *where, unsigned int line, if (ext4_handle_valid(handle)) { err = jbd2_journal_get_create_access(handle, bh); + if (!err) + err = ext4_snapshot_get_create_access(handle, bh); if (err) ext4_journal_abort_handle(where, line, __func__, bh, handle, err); diff --git a/fs/ext4/ext4_jbd2.h b/fs/ext4/ext4_jbd2.h index 3da2092..ca6e135 100644 --- a/fs/ext4/ext4_jbd2.h +++ b/fs/ext4/ext4_jbd2.h @@ -132,9 +132,9 @@ void ext4_journal_abort_handle(const char *caller, unsigned int line, int __ext4_journal_get_undo_access(const char *where, unsigned int line, handle_t *handle, struct buffer_head *bh); -int __ext4_journal_get_write_access(const char *where, unsigned int line, - handle_t *handle, struct buffer_head *bh); - +int __ext4_journal_get_write_access_inode(const char *where, unsigned int line, + handle_t *handle, struct inode *inode, + struct buffer_head *bh, int exclude); int __ext4_forget(const char *where, unsigned int line, handle_t *handle, int is_metadata, struct inode *inode, struct buffer_head *bh, ext4_fsblk_t blocknr); @@ -151,8 +151,15 @@ int __ext4_handle_dirty_super(const char *where, unsigned int line, #define ext4_journal_get_undo_access(handle, bh) \ __ext4_journal_get_undo_access(__func__, __LINE__, (handle), (bh)) +#define ext4_journal_get_write_access_exclude(handle, bh) \ + __ext4_journal_get_write_access_inode(__func__, __LINE__, \ + (handle), NULL, (bh), 1) #define ext4_journal_get_write_access(handle, bh) \ - __ext4_journal_get_write_access(__func__, __LINE__, (handle), (bh)) + __ext4_journal_get_write_access_inode(__func__, __LINE__, \ + (handle), NULL, (bh), 0) +#define ext4_journal_get_write_access_inode(handle, inode, bh) \ + __ext4_journal_get_write_access_inode(__func__, __LINE__, \ + (handle), (inode), (bh), 0) #define ext4_forget(handle, is_metadata, inode, bh, block_nr) \ __ext4_forget(__func__, __LINE__, (handle), (is_metadata), (inode), \ (bh), (block_nr)) diff --git a/fs/ext4/extents.c b/fs/ext4/extents.c index 7598224..6f0a711 100644 --- a/fs/ext4/extents.c +++ b/fs/ext4/extents.c @@ -79,7 +79,8 @@ static int ext4_ext_get_access(handle_t *handle, struct inode *inode, { if (path->p_bh) { /* path points to block */ - return ext4_journal_get_write_access(handle, path->p_bh); + return ext4_journal_get_write_access_inode(handle, + inode, path->p_bh); } /* path points to leaf/index in inode body */ /* we use in-core data, no need to protect them */ diff --git a/fs/ext4/inode.c b/fs/ext4/inode.c index 9dbd806..80e3393 100644 --- a/fs/ext4/inode.c +++ b/fs/ext4/inode.c @@ -874,7 +874,8 @@ static int ext4_splice_branch(handle_t *handle, struct inode *inode, */ if (where->bh) { BUFFER_TRACE(where->bh, "get_write_access"); - err = ext4_journal_get_write_access(handle, where->bh); + err = ext4_journal_get_write_access_inode(handle, inode, + where->bh); if (err) goto err_out; } @@ -4149,7 +4150,8 @@ static int ext4_clear_blocks(handle_t *handle, struct inode *inode, goto out_err; if (bh) { BUFFER_TRACE(bh, "retaking write access"); - err = ext4_journal_get_write_access(handle, bh); + err = ext4_journal_get_write_access_inode(handle, + inode, bh); if (unlikely(err)) goto out_err; } @@ -4200,7 +4202,8 @@ static void ext4_free_data(handle_t *handle, struct inode *inode, if (this_bh) { /* For indirect block */ BUFFER_TRACE(this_bh, "get_write_access"); - err = ext4_journal_get_write_access(handle, this_bh); + err = ext4_journal_get_write_access_inode(handle, inode, + this_bh); /* Important: if we can't update the indirect pointers * to the blocks, we can't free them. */ if (err) @@ -4363,8 +4366,8 @@ static void ext4_free_branches(handle_t *handle, struct inode *inode, * pointed to by an indirect block: journal it */ BUFFER_TRACE(parent_bh, "get_write_access"); - if (!ext4_journal_get_write_access(handle, - parent_bh)){ + if (!ext4_journal_get_write_access_inode( + handle, inode, parent_bh)){ *p = 0; BUFFER_TRACE(parent_bh, "call ext4_handle_dirty_metadata"); @@ -4741,9 +4744,14 @@ has_buffer: int ext4_get_inode_loc(struct inode *inode, struct ext4_iloc *iloc) { - /* We have all inode data except xattrs in memory here. */ - return __ext4_get_inode_loc(inode, iloc, + int in_mem = (!EXT4_SNAPSHOTS(inode->i_sb) && !ext4_test_inode_state(inode, EXT4_STATE_XATTR)); + + /* + * We have all inode's data except xattrs in memory here, + * but we must always read-in the entire inode block for COW. + */ + return __ext4_get_inode_loc(inode, iloc, in_mem); } void ext4_set_inode_flags(struct inode *inode) diff --git a/fs/ext4/move_extent.c b/fs/ext4/move_extent.c index b9f3e78..ad5409a 100644 --- a/fs/ext4/move_extent.c +++ b/fs/ext4/move_extent.c @@ -421,7 +421,8 @@ mext_insert_extents(handle_t *handle, struct inode *orig_inode, if (depth) { /* Register to journal */ - ret = ext4_journal_get_write_access(handle, orig_path->p_bh); + ret = ext4_journal_get_write_access_inode(handle, + orig_inode, orig_path->p_bh); if (ret) return ret; } diff --git a/fs/ext4/snapshot.h b/fs/ext4/snapshot.h index d0c985b..54241b9 100644 --- a/fs/ext4/snapshot.h +++ b/fs/ext4/snapshot.h @@ -145,6 +145,57 @@ static inline int EXT4_SNAPSHOTS(struct super_block *sb) * Block access functions */ +/* + * get_write_access() is called before writing to a metadata block + * if @inode is not NULL, then this is an inode's indirect block + * otherwise, this is a file system global metadata block + * + * Return values: + * = 0 - block was COWed or doesn't need to be COWed + * < 0 - error + */ +static inline int ext4_snapshot_get_write_access(handle_t *handle, + struct inode *inode, struct buffer_head *bh) +{ + struct super_block *sb; + + sb = handle->h_transaction->t_journal->j_private; + if (!EXT4_SNAPSHOTS(sb)) + return 0; + + return ext4_snapshot_cow(handle, inode, bh->b_blocknr, bh, 1); +} + +/* + * get_create_access() is called after allocating a new metadata block + * + * Return values: + * = 0 - block was COWed or doesn't need to be COWed + * < 0 - error + */ +static inline int ext4_snapshot_get_create_access(handle_t *handle, + struct buffer_head *bh) +{ + struct super_block *sb; + int err; + + sb = handle->h_transaction->t_journal->j_private; + if (!EXT4_SNAPSHOTS(sb)) + return 0; + + /* Should block be COWed? */ + err = ext4_snapshot_cow(handle, NULL, bh->b_blocknr, bh, 0); + /* + * A new block shouldn't need to be COWed if get_delete_access() was + * called for all deleted blocks. However, it may need to be COWed + * if fsck was run and if it had freed some blocks without moving them + * to snapshot. In the latter case, -EIO will be returned. + */ + if (err > 0) + err = -EIO; + return err; +} + /* snapshot_ctl.c */