From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from smtp.kernel.org (aws-us-west-2-korg-mail-1.web.codeaurora.org [10.30.226.201]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by smtp.subspace.kernel.org (Postfix) with ESMTPS id 4FE186FB5 for ; Sun, 17 Sep 2023 20:41:55 +0000 (UTC) Received: by smtp.kernel.org (Postfix) with ESMTPSA id 69122C433C8; Sun, 17 Sep 2023 20:41:54 +0000 (UTC) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/simple; d=linuxfoundation.org; s=korg; t=1694983314; bh=XL25KCxH4gaJhUTRKfPOQZ70KU0vbjr73f+sLGvoJ78=; h=From:To:Cc:Subject:Date:In-Reply-To:References:From; b=UNsmmqgL33AqanmxttRtpoOGF7HLawXq4mBrbu49QQGe36cGIao050umC2HHKYxNR Xzx+cXz1sLErltdTe8Di3GQDnJBCTOVPdwZJ8ARbvNx/YoCGh5Aeyp42yZXxn5BvHm Ts2roDMyhx2Oh+vcJT0dvO/LengVvw8w+X3elIuc= From: Greg Kroah-Hartman To: stable@vger.kernel.org Cc: Greg Kroah-Hartman , patches@lists.linux.dev, stable@kernel.org, Zhang Yi , Jan Kara , Theodore Tso Subject: [PATCH 5.15 466/511] jbd2: fix checkpoint cleanup performance regression Date: Sun, 17 Sep 2023 21:14:53 +0200 Message-ID: <20230917191125.003083101@linuxfoundation.org> X-Mailer: git-send-email 2.42.0 In-Reply-To: <20230917191113.831992765@linuxfoundation.org> References: <20230917191113.831992765@linuxfoundation.org> User-Agent: quilt/0.67 X-stable: review X-Patchwork-Hint: ignore Precedence: bulk X-Mailing-List: patches@lists.linux.dev List-Id: List-Subscribe: List-Unsubscribe: MIME-Version: 1.0 Content-Transfer-Encoding: 8bit 5.15-stable review patch. If anyone has any objections, please let me know. ------------------ From: Zhang Yi commit 373ac521799d9e97061515aca6ec6621789036bb upstream. journal_clean_one_cp_list() has been merged into journal_shrink_one_cp_list(), but do chekpoint buffer cleanup from the committing process is just a best effort, it should stop scan once it meet a busy buffer, or else it will cause a lot of invalid buffer scan and checks. We catch a performance regression when doing fs_mark tests below. Test cmd: ./fs_mark -d scratch -s 1024 -n 10000 -t 1 -D 100 -N 100 Before merging checkpoint buffer cleanup: FSUse% Count Size Files/sec App Overhead 95 10000 1024 8304.9 49033 After merging checkpoint buffer cleanup: FSUse% Count Size Files/sec App Overhead 95 10000 1024 7649.0 50012 FSUse% Count Size Files/sec App Overhead 95 10000 1024 2107.1 50871 After merging checkpoint buffer cleanup, the total loop count in journal_shrink_one_cp_list() could be up to 6,261,600+ (50,000+ ~ 100,000+ in general), most of them are invalid. This patch fix it through passing 'shrink_type' into journal_shrink_one_cp_list() and add a new 'SHRINK_BUSY_STOP' to indicate it should stop once meet a busy buffer. After fix, the loop count descending back to 10,000+. After this fix: FSUse% Count Size Files/sec App Overhead 95 10000 1024 8558.4 49109 Cc: stable@kernel.org Fixes: b98dba273a0e ("jbd2: remove journal_clean_one_cp_list()") Signed-off-by: Zhang Yi Reviewed-by: Jan Kara Link: https://lore.kernel.org/r/20230714025528.564988-2-yi.zhang@huaweicloud.com Signed-off-by: Theodore Ts'o Signed-off-by: Greg Kroah-Hartman --- fs/jbd2/checkpoint.c | 20 ++++++++++++++------ 1 file changed, 14 insertions(+), 6 deletions(-) --- a/fs/jbd2/checkpoint.c +++ b/fs/jbd2/checkpoint.c @@ -349,6 +349,8 @@ int jbd2_cleanup_journal_tail(journal_t /* Checkpoint list management */ +enum shrink_type {SHRINK_DESTROY, SHRINK_BUSY_STOP, SHRINK_BUSY_SKIP}; + /* * journal_shrink_one_cp_list * @@ -360,7 +362,8 @@ int jbd2_cleanup_journal_tail(journal_t * Called with j_list_lock held. */ static unsigned long journal_shrink_one_cp_list(struct journal_head *jh, - bool destroy, bool *released) + enum shrink_type type, + bool *released) { struct journal_head *last_jh; struct journal_head *next_jh = jh; @@ -376,12 +379,15 @@ static unsigned long journal_shrink_one_ jh = next_jh; next_jh = jh->b_cpnext; - if (destroy) { + if (type == SHRINK_DESTROY) { ret = __jbd2_journal_remove_checkpoint(jh); } else { ret = jbd2_journal_try_remove_checkpoint(jh); - if (ret < 0) - continue; + if (ret < 0) { + if (type == SHRINK_BUSY_SKIP) + continue; + break; + } } nr_freed++; @@ -445,7 +451,7 @@ again: tid = transaction->t_tid; freed = journal_shrink_one_cp_list(transaction->t_checkpoint_list, - false, &released); + SHRINK_BUSY_SKIP, &released); nr_freed += freed; (*nr_to_scan) -= min(*nr_to_scan, freed); if (*nr_to_scan == 0) @@ -485,19 +491,21 @@ out: void __jbd2_journal_clean_checkpoint_list(journal_t *journal, bool destroy) { transaction_t *transaction, *last_transaction, *next_transaction; + enum shrink_type type; bool released; transaction = journal->j_checkpoint_transactions; if (!transaction) return; + type = destroy ? SHRINK_DESTROY : SHRINK_BUSY_STOP; last_transaction = transaction->t_cpprev; next_transaction = transaction; do { transaction = next_transaction; next_transaction = transaction->t_cpnext; journal_shrink_one_cp_list(transaction->t_checkpoint_list, - destroy, &released); + type, &released); /* * This function only frees up some memory if possible so we * dont have an obligation to finish processing. Bail out if