Received-SPF: pass (google.com: domain of linux-ext4-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
From:   Harshad Shirwadkar <harshadshirwadkar@gmail.com>
To:     linux-ext4@vger.kernel.org
Cc:     tytso@mit.edu, Harshad Shirwadkar <harshadshirwadkar@gmail.com>
Subject: [RFC PATCH v4 3/8] ext4: use extent status tree in fast commit path
Date:   Thu, 21 Jul 2022 06:02:41 +0000
Message-Id: <20220721060246.1696852-4-harshadshirwadkar@gmail.com>
In-Reply-To: <20220721060246.1696852-1-harshadshirwadkar@gmail.com>
References: <20220721060246.1696852-1-harshadshirwadkar@gmail.com>
MIME-Version: 1.0
Content-Transfer-Encoding: 8bit
Precedence: bulk

This patch moves fc commit path to use extent status tree to lookup
logical to physical mappings. In order to preserve the uncommitted
entries in the es cache, this patch makes all the inodes on fast
commit list as shrinker ineligible. Making the uncommitted entries in
es cache to stick around is left as a future enhancement.

Signed-off-by: Harshad Shirwadkar <harshadshirwadkar@gmail.com>
---
 fs/ext4/extents_status.c | 3 ++-
 fs/ext4/fast_commit.c    | 3 ++-
 2 files changed, 4 insertions(+), 2 deletions(-)

diff --git a/fs/ext4/extents_status.c b/fs/ext4/extents_status.c
index 9a3a8996aacf..07fb86746534 100644
--- a/fs/ext4/extents_status.c
+++ b/fs/ext4/extents_status.c
@@ -1500,7 +1500,8 @@ static int __es_shrink(struct ext4_sb_info *sbi, int nr_to_scan,
 			continue;
 		}
 
-		if (ei == locked_ei || !write_trylock(&ei->i_es_lock)) {
+		if (!list_empty(&ei->i_fc_list) || ei == locked_ei ||
+			!write_trylock(&ei->i_es_lock)) {
 			nr_skipped++;
 			continue;
 		}
diff --git a/fs/ext4/fast_commit.c b/fs/ext4/fast_commit.c
index 4d2384adcbb0..916f62cfa7f7 100644
--- a/fs/ext4/fast_commit.c
+++ b/fs/ext4/fast_commit.c
@@ -951,7 +951,8 @@ static int ext4_fc_write_inode_data(struct inode *inode, u32 *crc)
 	while (cur_lblk_off <= new_blk_size) {
 		map.m_lblk = cur_lblk_off;
 		map.m_len = new_blk_size - cur_lblk_off + 1;
-		ret = ext4_map_blocks(NULL, inode, &map, 0);
+		ret = ext4_map_blocks(NULL, inode, &map,
+			EXT4_GET_BLOCKS_CACHED_NOWAIT);
 		if (ret < 0)
 			return -ECANCELED;
 
-- 
2.37.0.170.g444d1eabd0-goog