linux-erofs.lists.ozlabs.org archive mirror
 help / color / mirror / Atom feed
From: Jia Zhu <zhujia.zj@bytedance.com>
To: xiang@kernel.org, chao@kernel.org, gerry@linux.alibaba.com,
	linux-erofs@lists.ozlabs.org
Cc: huyue2@coolpad.com, linux-kernel@vger.kernel.org
Subject: [PATCH V2] erofs: support flattened block device for multi-blob images
Date: Wed,  1 Mar 2023 20:59:08 +0800	[thread overview]
Message-ID: <20230301125908.30879-1-zhujia.zj@bytedance.com> (raw)

In order to support mounting multi-blobs container image as a single
block device, add flattened block device feature for EROFS.

In this mode, all meta/data contents will be mapped into one block
address. User could compose a block device(by nbd/ublk/virtio-blk/
vhost-user-blk) from multiple sources and mount the block device by
EROFS directly. It can reduce the number of block devices used, and
it's also benefits in both VM file passthrough and distributed storage
scenarios.

You can test this using the method mentioned by:
https://github.com/dragonflyoss/image-service/pull/1111
1. Compose a (nbd)block device from multi-blobs.
2. Mount EROFS on mntdir/.
3. Compare the md5sum between source dir and mntdir/.

Later, we could also use it to refer original tar blobs.

Signed-off-by: Jia Zhu <zhujia.zj@bytedance.com>
Signed-off-by: Xin Yin <yinxin.x@bytedance.com>
---
v2:
1. Supplement commit message.
2. Add a bool field in erofs_dev_context to indicate flattened block
   device mode.
---
 fs/erofs/data.c     | 8 ++++++--
 fs/erofs/internal.h | 1 +
 fs/erofs/super.c    | 6 +++++-
 3 files changed, 12 insertions(+), 3 deletions(-)

diff --git a/fs/erofs/data.c b/fs/erofs/data.c
index e16545849ea7..818f78ce648c 100644
--- a/fs/erofs/data.c
+++ b/fs/erofs/data.c
@@ -197,7 +197,6 @@ int erofs_map_dev(struct super_block *sb, struct erofs_map_dev *map)
 	struct erofs_device_info *dif;
 	int id;
 
-	/* primary device by default */
 	map->m_bdev = sb->s_bdev;
 	map->m_daxdev = EROFS_SB(sb)->dax_dev;
 	map->m_dax_part_off = EROFS_SB(sb)->dax_part_off;
@@ -210,12 +209,17 @@ int erofs_map_dev(struct super_block *sb, struct erofs_map_dev *map)
 			up_read(&devs->rwsem);
 			return -ENODEV;
 		}
+		if (devs->flatdev) {
+			map->m_pa += blknr_to_addr(dif->mapped_blkaddr);
+			up_read(&devs->rwsem);
+			return 0;
+		}
 		map->m_bdev = dif->bdev;
 		map->m_daxdev = dif->dax_dev;
 		map->m_dax_part_off = dif->dax_part_off;
 		map->m_fscache = dif->fscache;
 		up_read(&devs->rwsem);
-	} else if (devs->extra_devices) {
+	} else if (devs->extra_devices && !devs->flatdev) {
 		down_read(&devs->rwsem);
 		idr_for_each_entry(&devs->tree, dif, id) {
 			erofs_off_t startoff, length;
diff --git a/fs/erofs/internal.h b/fs/erofs/internal.h
index 3f3561d37d1b..4fee380a98d9 100644
--- a/fs/erofs/internal.h
+++ b/fs/erofs/internal.h
@@ -81,6 +81,7 @@ struct erofs_dev_context {
 	struct rw_semaphore rwsem;
 
 	unsigned int extra_devices;
+	bool flatdev;
 };
 
 struct erofs_fs_context {
diff --git a/fs/erofs/super.c b/fs/erofs/super.c
index 19b1ae79cec4..307b3d2392cf 100644
--- a/fs/erofs/super.c
+++ b/fs/erofs/super.c
@@ -248,7 +248,7 @@ static int erofs_init_device(struct erofs_buf *buf, struct super_block *sb,
 		if (IS_ERR(fscache))
 			return PTR_ERR(fscache);
 		dif->fscache = fscache;
-	} else {
+	} else if (!sbi->devs->flatdev) {
 		bdev = blkdev_get_by_path(dif->path, FMODE_READ | FMODE_EXCL,
 					  sb->s_type);
 		if (IS_ERR(bdev))
@@ -281,6 +281,10 @@ static int erofs_scan_devices(struct super_block *sb,
 	else
 		ondisk_extradevs = le16_to_cpu(dsb->extra_devices);
 
+	if (!sbi->devs->extra_devices && ondisk_extradevs &&
+		!erofs_is_fscache_mode(sb))
+		sbi->devs->flatdev = true;
+
 	if (sbi->devs->extra_devices &&
 	    ondisk_extradevs != sbi->devs->extra_devices) {
 		erofs_err(sb, "extra devices don't match (ondisk %u, given %u)",
-- 
2.20.1


             reply	other threads:[~2023-03-01 12:59 UTC|newest]

Thread overview: 5+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2023-03-01 12:59 Jia Zhu [this message]
2023-03-02  6:39 ` [PATCH V2] erofs: support flattened block device for multi-blob images Jingbo Xu
2023-03-02  7:17   ` [PATCH V3] " Jia Zhu
2023-04-16 14:30     ` Chao Yu
2023-04-16 14:34     ` Gerry Liu

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20230301125908.30879-1-zhujia.zj@bytedance.com \
    --to=zhujia.zj@bytedance.com \
    --cc=chao@kernel.org \
    --cc=gerry@linux.alibaba.com \
    --cc=huyue2@coolpad.com \
    --cc=linux-erofs@lists.ozlabs.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=xiang@kernel.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).