From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: X-Spam-Checker-Version: SpamAssassin 3.4.0 (2014-02-07) on aws-us-west-2-korg-lkml-1.web.codeaurora.org Received: from vger.kernel.org (vger.kernel.org [23.128.96.18]) by smtp.lore.kernel.org (Postfix) with ESMTP id 8FA3CC4167B for ; Fri, 2 Dec 2022 09:37:20 +0000 (UTC) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S232681AbiLBJhS (ORCPT ); Fri, 2 Dec 2022 04:37:18 -0500 Received: from lindbergh.monkeyblade.net ([23.128.96.19]:38762 "EHLO lindbergh.monkeyblade.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S232591AbiLBJhR (ORCPT ); Fri, 2 Dec 2022 04:37:17 -0500 Received: from out199-11.us.a.mail.aliyun.com (out199-11.us.a.mail.aliyun.com [47.90.199.11]) by lindbergh.monkeyblade.net (Postfix) with ESMTPS id D1C0110B6; Fri, 2 Dec 2022 01:37:11 -0800 (PST) X-Alimail-AntiSpam: AC=PASS;BC=-1|-1;BR=01201311R971e4;CH=green;DM=||false|;DS=||;FP=0|-1|-1|-1|0|-1|-1|-1;HT=ay29a033018046051;MF=ziyangzhang@linux.alibaba.com;NM=1;PH=DS;RN=5;SR=0;TI=SMTPD_---0VWDD4WH_1669973823; Received: from localhost.localdomain(mailfrom:ZiyangZhang@linux.alibaba.com fp:SMTPD_---0VWDD4WH_1669973823) by smtp.aliyun-inc.com; Fri, 02 Dec 2022 17:37:08 +0800 From: Ziyang Zhang To: fstests@vger.kernel.org, linux-xfs@vger.kernel.org Cc: djwong@kernel.org, Ziyang Zhang , Gao Xiang Subject: [PATCH V2] common/populate: Ensure that S_IFDIR.FMT_BTREE is in btree format Date: Fri, 2 Dec 2022 17:35:34 +0800 Message-Id: <20221202093534.940907-1-ZiyangZhang@linux.alibaba.com> X-Mailer: git-send-email 2.18.4 Precedence: bulk List-ID: X-Mailing-List: fstests@vger.kernel.org Sometimes "$((128 * dblksz / 40))" dirents cannot make sure that S_IFDIR.FMT_BTREE could become btree format for its DATA fork. Actually we just observed it can fail after apply our inode extent-to-btree workaround. The root cause is that the kernel may be too good at allocating consecutive blocks so that the data fork is still in extents format. Therefore instead of using a fixed number, let's make sure the number of extents is large enough than (inode size - inode core size) / sizeof(xfs_bmbt_rec_t). Suggested-by: "Darrick J. Wong" Signed-off-by: Gao Xiang Signed-off-by: Ziyang Zhang --- V2: take Darrick's advice to cleanup code common/populate | 28 +++++++++++++++++++++++++++- common/xfs | 17 +++++++++++++++++ 2 files changed, 44 insertions(+), 1 deletion(-) diff --git a/common/populate b/common/populate index 6e004997..c6b879fa 100644 --- a/common/populate +++ b/common/populate @@ -71,6 +71,31 @@ __populate_create_dir() { done } +# Create a large directory and ensure that it's a btree format +__populate_xfs_create_btree_dir() { + local name="$1" + local isize="$2" + local icore_size=$(_xfs_inode_core_bytes) + # We need enough extents to guarantee that the data fork is in + # btree format. Cycling the mount to use xfs_db is too slow, so + # watch for when the extent count exceeds the space after the + # inode core. + local max_nextents="$(((isize - 176) / 16))" + + mkdir -p "${name}" + d=0 + while true; do + creat=mkdir + test "$((d % 20))" -eq 0 && creat=touch + $creat "${name}/$(printf "%.08d" "$d")" + if [ "$((d % 40))" -eq 0 ]; then + nextents=$(_xfs_get_fsxattr nextents $name) + [ $nextents -gt $max_nextents ] && break + fi + d=$((d+1)) + done +} + # Add a bunch of attrs to a file __populate_create_attr() { name="$1" @@ -176,6 +201,7 @@ _scratch_xfs_populate() { blksz="$(stat -f -c '%s' "${SCRATCH_MNT}")" dblksz="$(_xfs_get_dir_blocksize "$SCRATCH_MNT")" + isize="$(_xfs_inode_size "$SCRATCH_MNT")" crc="$(_xfs_has_feature "$SCRATCH_MNT" crc -v)" if [ $crc -eq 1 ]; then leaf_hdr_size=64 @@ -226,7 +252,7 @@ _scratch_xfs_populate() { # - BTREE echo "+ btree dir" - __populate_create_dir "${SCRATCH_MNT}/S_IFDIR.FMT_BTREE" "$((128 * dblksz / 40))" true + __populate_xfs_create_btree_dir "${SCRATCH_MNT}/S_IFDIR.FMT_BTREE" "$isize" # Symlinks # - FMT_LOCAL diff --git a/common/xfs b/common/xfs index 8ac1964e..0359e422 100644 --- a/common/xfs +++ b/common/xfs @@ -1486,3 +1486,20 @@ _require_xfsrestore_xflag() $XFSRESTORE_PROG -h 2>&1 | grep -q -e '-x' || \ _notrun 'xfsrestore does not support -x flag.' } + + +# Number of bytes reserved for a full inode record, which includes the +# immediate fork areas. +_xfs_inode_size() +{ + local mntpoint="$1" + + $XFS_INFO_PROG "$mntpoint" | grep 'meta-data=.*isize' | sed -e 's/^.*isize=\([0-9]*\).*$/\1/g' +} + +# Number of bytes reserved for only the inode record, excluding the +# immediate fork areas. +_xfs_inode_core_bytes() +{ + echo 176 +} -- 2.18.4