Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:18 as permitted sender) client-ip=2620:137:e000::1:18;
From:   Jeffle Xu <jefflexu@linux.alibaba.com>
To:     dhowells@redhat.com, linux-cachefs@redhat.com, xiang@kernel.org,
        chao@kernel.org, linux-erofs@lists.ozlabs.org
Cc:     torvalds@linux-foundation.org, gregkh@linuxfoundation.org,
        willy@infradead.org, linux-fsdevel@vger.kernel.org,
        joseph.qi@linux.alibaba.com, bo.liu@linux.alibaba.com,
        tao.peng@linux.alibaba.com, gerry@linux.alibaba.com,
        eguan@linux.alibaba.com, linux-kernel@vger.kernel.org
Subject: [PATCH v3 19/22] erofs: implement fscache-based data readahead for hole
Date:   Wed,  9 Feb 2022 14:01:05 +0800
Message-Id: <20220209060108.43051-20-jefflexu@linux.alibaba.com>
In-Reply-To: <20220209060108.43051-1-jefflexu@linux.alibaba.com>
References: <20220209060108.43051-1-jefflexu@linux.alibaba.com>
MIME-Version: 1.0
Content-Transfer-Encoding: 8bit
Precedence: bulk

Implement fscache-based data readahead. This patch only supports
readahead for hole, while the following patches will handle other cases.

Besides this patch also registers an individual bdi for each erofs
instance, so that readahead can be enabled.

Signed-off-by: Jeffle Xu <jefflexu@linux.alibaba.com>
---
 fs/erofs/fscache.c | 83 ++++++++++++++++++++++++++++++++++++++++++++++
 fs/erofs/super.c   |  4 +++
 2 files changed, 87 insertions(+)

diff --git a/fs/erofs/fscache.c b/fs/erofs/fscache.c
index c7762e154064..c8a0851230e5 100644
--- a/fs/erofs/fscache.c
+++ b/fs/erofs/fscache.c
@@ -195,12 +195,95 @@ static int erofs_fscache_readpage(struct file *file, struct page *page)
 	return ret;
 }
 
+enum erofs_fscache_readahead_type {
+	EROFS_FSCACHE_READAHEAD_TYPE_HOLE,
+};
+
+static int erofs_fscache_do_readahead(struct readahead_control *rac,
+				      struct erofs_fscache_map *fsmap,
+				      enum erofs_fscache_readahead_type type)
+{
+	size_t offset, length, done;
+	struct page *page;
+
+	/*
+	 * 1) For CHUNK_BASED (HOLE), the output map.m_la is rounded down to
+	 *    the nearest chunk boundary, and thus offset will be non-zero.
+	 */
+	offset = fsmap->o_la - fsmap->m_la;
+	length = fsmap->m_llen - offset;
+
+	for (done = 0; done < length; done += PAGE_SIZE) {
+		page = readahead_page(rac);
+		if (!page)
+			break;
+
+		switch (type) {
+		case EROFS_FSCACHE_READAHEAD_TYPE_HOLE:
+			zero_user(page, 0, PAGE_SIZE);
+			break;
+		default:
+			DBG_BUGON(1);
+			return -EINVAL;
+		}
+
+		SetPageUptodate(page);
+		unlock_page(page);
+	}
+
+	return done;
+}
+
+static void erofs_fscache_readahead(struct readahead_control *rac)
+{
+	struct inode *inode = rac->mapping->host;
+	struct erofs_inode *vi = EROFS_I(inode);
+	struct super_block *sb = inode->i_sb;
+	size_t length = readahead_length(rac);
+	struct erofs_map_blocks map;
+	struct erofs_fscache_map fsmap;
+	int ret;
+
+	if (erofs_inode_is_data_compressed(vi->datalayout)) {
+		erofs_info(sb, "compressed layout not supported yet");
+		return;
+	}
+
+	while (length) {
+		map.m_la = fsmap.o_la = readahead_pos(rac);
+
+		ret = erofs_map_blocks(inode, &map, EROFS_GET_BLOCKS_RAW);
+		if (ret)
+			return;
+
+		if (!(map.m_flags & EROFS_MAP_MAPPED)) {
+			/* Only CHUNK_BASED layout supports hole. */
+			fsmap.m_la   = map.m_la;
+			fsmap.m_llen = map.m_llen;
+			ret = erofs_fscache_do_readahead(rac, &fsmap,
+					EROFS_FSCACHE_READAHEAD_TYPE_HOLE);
+		} else {
+			switch (vi->datalayout) {
+			default:
+				DBG_BUGON(1);
+				return;
+			}
+		}
+
+		if (ret <= 0)
+			return;
+
+		length -= ret;
+	}
+}
+
 static const struct address_space_operations erofs_fscache_blob_aops = {
 	.readpage = erofs_fscache_readpage_blob,
 };
 
 const struct address_space_operations erofs_fscache_access_aops = {
 	.readpage = erofs_fscache_readpage,
+	.readahead = erofs_fscache_readahead,
 };
 
 struct page *erofs_fscache_read_cache_page(struct erofs_fscache_context *ctx,
diff --git a/fs/erofs/super.c b/fs/erofs/super.c
index f058a04a00c7..2942029a7049 100644
--- a/fs/erofs/super.c
+++ b/fs/erofs/super.c
@@ -616,6 +616,10 @@ static int erofs_fc_fill_super(struct super_block *sb, struct fs_context *fc)
 			return PTR_ERR(bootstrap);
 
 		sbi->bootstrap = bootstrap;
+
+		err = super_setup_bdi(sb);
+		if (err)
+			return err;
 	}
 
 	err = erofs_read_superblock(sb);
-- 
2.27.0