From patchwork Tue Feb 2 03:05:51 2016 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Qu Wenruo X-Patchwork-Id: 8186401 Return-Path: X-Original-To: patchwork-linux-btrfs@patchwork.kernel.org Delivered-To: patchwork-parsemail@patchwork2.web.kernel.org Received: from mail.kernel.org (mail.kernel.org [198.145.29.136]) by patchwork2.web.kernel.org (Postfix) with ESMTP id 15924BEEE5 for ; Tue, 2 Feb 2016 03:09:02 +0000 (UTC) Received: from mail.kernel.org (localhost [127.0.0.1]) by mail.kernel.org (Postfix) with ESMTP id 2AC1020254 for ; Tue, 2 Feb 2016 03:09:01 +0000 (UTC) Received: from vger.kernel.org (vger.kernel.org [209.132.180.67]) by mail.kernel.org (Postfix) with ESMTP id 4241920218 for ; Tue, 2 Feb 2016 03:09:00 +0000 (UTC) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1752219AbcBBDID (ORCPT ); Mon, 1 Feb 2016 22:08:03 -0500 Received: from cn.fujitsu.com ([222.73.24.84]:7891 "EHLO song.cn.fujitsu.com" rhost-flags-OK-FAIL-OK-OK) by vger.kernel.org with ESMTP id S1751342AbcBBDIB (ORCPT ); Mon, 1 Feb 2016 22:08:01 -0500 X-IronPort-AV: E=Sophos;i="5.20,367,1444665600"; d="scan'208";a="310378" Received: from unknown (HELO cn.fujitsu.com) ([10.167.250.3]) by song.cn.fujitsu.com with ESMTP; 02 Feb 2016 11:07:55 +0800 Received: from localhost.localdomain (unknown [10.167.226.34]) by cn.fujitsu.com (Postfix) with ESMTP id DC5114091CED; Tue, 2 Feb 2016 11:07:52 +0800 (CST) From: Qu Wenruo To: linux-btrfs@vger.kernel.org Cc: Wang Xiaoguang Subject: [PATCH v5 19/19] btrfs: try more times to alloc metadata reserve space Date: Tue, 2 Feb 2016 11:05:51 +0800 Message-Id: <1454382351-31775-20-git-send-email-quwenruo@cn.fujitsu.com> X-Mailer: git-send-email 2.7.0 In-Reply-To: <1454382351-31775-1-git-send-email-quwenruo@cn.fujitsu.com> References: <1454382351-31775-1-git-send-email-quwenruo@cn.fujitsu.com> MIME-Version: 1.0 X-yoursite-MailScanner-ID: DC5114091CED.ADC04 X-yoursite-MailScanner: Found to be clean X-yoursite-MailScanner-From: quwenruo@cn.fujitsu.com X-Spam-Status: No, score=-7.4 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 From: Wang Xiaoguang In btrfs_delalloc_reserve_metadata(), the number of metadata bytes we try to reserve is calculated by the difference between outstanding_extents and reserved_extents. When reserve_metadata_bytes() fails to reserve desited metadata space, it has already done some reclaim work, such as write ordered extents. In that case, outstanding_extents and reserved_extents may already changed, and we may reserve enough metadata space then. So this patch will try to call reserve_metadata_bytes() at most 3 times to ensure we really run out of space. Such false ENOSPC is mainly caused by small file extents and time consuming delalloc functions, which mainly affects in-band de-duplication. (Compress should also be affected, but LZO/zlib is faster than SHA256, so still harder to trigger than dedup). Signed-off-by: Wang Xiaoguang --- fs/btrfs/extent-tree.c | 23 +++++++++++++++++++++-- 1 file changed, 21 insertions(+), 2 deletions(-) diff --git a/fs/btrfs/extent-tree.c b/fs/btrfs/extent-tree.c index f9fc25c..0a6d172 100644 --- a/fs/btrfs/extent-tree.c +++ b/fs/btrfs/extent-tree.c @@ -5671,6 +5671,7 @@ int btrfs_delalloc_reserve_metadata(struct inode *inode, u64 num_bytes) bool delalloc_lock = true; u64 to_free = 0; unsigned dropped; + int loops = 0; /* If we are a free space inode we need to not flush since we will be in * the middle of a transaction commit. We also don't need the delalloc @@ -5686,11 +5687,12 @@ int btrfs_delalloc_reserve_metadata(struct inode *inode, u64 num_bytes) btrfs_transaction_in_commit(root->fs_info)) schedule_timeout(1); + num_bytes = ALIGN(num_bytes, root->sectorsize); + +again: if (delalloc_lock) mutex_lock(&BTRFS_I(inode)->delalloc_mutex); - num_bytes = ALIGN(num_bytes, root->sectorsize); - spin_lock(&BTRFS_I(inode)->lock); nr_extents = (unsigned)div64_u64(num_bytes + BTRFS_MAX_EXTENT_SIZE - 1, @@ -5811,6 +5813,23 @@ out_fail: } if (delalloc_lock) mutex_unlock(&BTRFS_I(inode)->delalloc_mutex); + /* + * The number of metadata bytes is calculated by the difference + * between outstanding_extents and reserved_extents. Sometimes though + * reserve_metadata_bytes() fails to reserve the wanted metadata bytes, + * indeed it has already done some work to reclaim metadata space, hence + * both outstanding_extents and reserved_extents would have changed and + * the bytes we try to reserve would also has changed(may be smaller). + * So here we try to reserve again. This is much useful for online + * dedup, which will easily eat almost all meta space. + * + * XXX: Indeed here 3 is arbitrarily choosed, it's a good workaround for + * online dedup, later we should find a better method to avoid dedup + * enospc issue. + */ + if (unlikely(ret == -ENOSPC && loops++ < 3)) + goto again; + return ret; }