Received-SPF: pass (google.com: domain of linux-ext4-owner@vger.kernel.org designates 23.128.96.18 as permitted sender) client-ip=23.128.96.18;
Sender: "saranyamohan via sendgmr" 
        <saranyamohan@saranyamohan.svl.corp.google.com>
Date:   Wed, 18 Nov 2020 07:39:31 -0800
In-Reply-To: <20201118153947.3394530-1-saranyamohan@google.com>
Message-Id: <20201118153947.3394530-46-saranyamohan@google.com>
Mime-Version: 1.0
References: <20201118153947.3394530-1-saranyamohan@google.com>
Subject: [RFC PATCH v3 45/61] e2fsck: make default smallest RA size to 1M
From:   Saranya Muruganandam <saranyamohan@google.com>
To:     linux-ext4@vger.kernel.org, tytso@mit.edu
Cc:     adilger.kernel@dilger.ca, Wang Shilong <wshilong@ddn.com>,
        Saranya Muruganandam <saranyamohan@google.com>
Content-Type: text/plain; charset="UTF-8"
Precedence: bulk

From: Wang Shilong <wshilong@ddn.com>

If we have a smaller inodes per group, default ra size could
be very small(etc 128KiB), this hurts performances.

Tune above 128K to 1M, i see pass1 time drop down from
677.12 seconds to 246 secons with 32 threads.

Signed-off-by: Wang Shilong <wshilong@ddn.com>
Signed-off-by: Saranya Muruganandam <saranyamohan@google.com>
---
 e2fsck/readahead.c | 4 ++++
 1 file changed, 4 insertions(+)

diff --git a/e2fsck/readahead.c b/e2fsck/readahead.c
index 38d4ec42..40b73664 100644
--- a/e2fsck/readahead.c
+++ b/e2fsck/readahead.c
@@ -234,6 +234,8 @@ int e2fsck_can_readahead(ext2_filsys fs)
 	return err != EXT2_ET_OP_NOT_SUPPORTED;
 }
 
+#define MIN_DEFAULT_RA	(1024 * 1024)
+
 unsigned long long e2fsck_guess_readahead(ext2_filsys fs)
 {
 	unsigned long long guess;
@@ -245,6 +247,8 @@ unsigned long long e2fsck_guess_readahead(ext2_filsys fs)
 	 * in e2fsck runtime.
 	 */
 	guess = 2ULL * fs->blocksize * fs->inode_blocks_per_group;
+	if (guess < MIN_DEFAULT_RA)
+		guess = MIN_DEFAULT_RA;
 
 	/* Disable RA if it'd use more 1/50th of RAM. */
 	if (get_memory_size() > (guess * 50))
-- 
2.29.2.299.gdc1121823c-goog