linux-erofs.lists.ozlabs.org archive mirror
 help / color / mirror / Atom feed
* [PATCH v3 1/2] erofs: decouple basic mount options from fs_context
@ 2021-10-06 19:44 Gao Xiang
  2021-10-06 19:44 ` [PATCH v3 2/2] erofs: add multiple device support Gao Xiang
  2021-10-07  6:10 ` [PATCH v3 1/2] erofs: decouple basic mount options from fs_context Liu Bo
  0 siblings, 2 replies; 5+ messages in thread
From: Gao Xiang @ 2021-10-06 19:44 UTC (permalink / raw)
  To: linux-erofs, Chao Yu, Liu Bo
  Cc: Yan Song, LKML, Peng Tao, Joseph Qi, Changwei Ge, Gao Xiang, Liu Jiang

Previously, EROFS mount options are all in the basic types, so
erofs_fs_context can be directly copied with assignment. However,
when the multiple device feature is introduced, it's hard to handle
multiple device information like the other basic mount options.

There is no need to allocate the whole sb info in advance, instead,
let's separate the basic mount options from fs_context, thus
multiple device information can be handled gracefully then.

No logic changes.

Cc: Liu Bo <bo.liu@linux.alibaba.com>
Cc: Chao Yu <chao@kernel.org>
Signed-off-by: Gao Xiang <hsiangkao@linux.alibaba.com>
---
changes since RFC v2:
 - add a new patch to decouple basic mount options from fs_context
   so that dev_context can be passed safely.

 fs/erofs/inode.c    |  2 +-
 fs/erofs/internal.h | 16 ++++++++-----
 fs/erofs/super.c    | 58 ++++++++++++++++++++++-----------------------
 fs/erofs/xattr.c    |  4 ++--
 fs/erofs/zdata.c    |  8 +++----
 5 files changed, 45 insertions(+), 43 deletions(-)

diff --git a/fs/erofs/inode.c b/fs/erofs/inode.c
index a552399e211d..2345f1de438e 100644
--- a/fs/erofs/inode.c
+++ b/fs/erofs/inode.c
@@ -192,7 +192,7 @@ static struct page *erofs_read_inode(struct inode *inode,
 	inode->i_atime.tv_nsec = inode->i_ctime.tv_nsec;
 
 	inode->i_flags &= ~S_DAX;
-	if (test_opt(&sbi->ctx, DAX_ALWAYS) && S_ISREG(inode->i_mode) &&
+	if (test_opt(&sbi->opt, DAX_ALWAYS) && S_ISREG(inode->i_mode) &&
 	    vi->datalayout == EROFS_INODE_FLAT_PLAIN)
 		inode->i_flags |= S_DAX;
 	if (!nblks)
diff --git a/fs/erofs/internal.h b/fs/erofs/internal.h
index 9524e155b38f..b1b9d1b5cb66 100644
--- a/fs/erofs/internal.h
+++ b/fs/erofs/internal.h
@@ -47,7 +47,7 @@ typedef u64 erofs_off_t;
 /* data type for filesystem-wide blocks number */
 typedef u32 erofs_blk_t;
 
-struct erofs_fs_context {
+struct erofs_mount_opts {
 #ifdef CONFIG_EROFS_FS_ZIP
 	/* current strategy of how to use managed cache */
 	unsigned char cache_strategy;
@@ -60,6 +60,10 @@ struct erofs_fs_context {
 	unsigned int mount_opt;
 };
 
+struct erofs_fs_context {
+	struct erofs_mount_opts opt;
+};
+
 /* all filesystem-wide lz4 configurations */
 struct erofs_sb_lz4_info {
 	/* # of pages needed for EROFS lz4 rolling decompression */
@@ -69,6 +73,8 @@ struct erofs_sb_lz4_info {
 };
 
 struct erofs_sb_info {
+	struct erofs_mount_opts opt;	/* options */
+
 #ifdef CONFIG_EROFS_FS_ZIP
 	/* list for all registered superblocks, mainly for shrinker */
 	struct list_head list;
@@ -108,8 +114,6 @@ struct erofs_sb_info {
 	u8 volume_name[16];             /* volume name */
 	u32 feature_compat;
 	u32 feature_incompat;
-
-	struct erofs_fs_context ctx;	/* options */
 };
 
 #define EROFS_SB(sb) ((struct erofs_sb_info *)(sb)->s_fs_info)
@@ -121,9 +125,9 @@ struct erofs_sb_info {
 #define EROFS_MOUNT_DAX_ALWAYS		0x00000040
 #define EROFS_MOUNT_DAX_NEVER		0x00000080
 
-#define clear_opt(ctx, option)	((ctx)->mount_opt &= ~EROFS_MOUNT_##option)
-#define set_opt(ctx, option)	((ctx)->mount_opt |= EROFS_MOUNT_##option)
-#define test_opt(ctx, option)	((ctx)->mount_opt & EROFS_MOUNT_##option)
+#define clear_opt(opt, option)	((opt)->mount_opt &= ~EROFS_MOUNT_##option)
+#define set_opt(opt, option)	((opt)->mount_opt |= EROFS_MOUNT_##option)
+#define test_opt(opt, option)	((opt)->mount_opt & EROFS_MOUNT_##option)
 
 enum {
 	EROFS_ZIP_CACHE_DISABLED,
diff --git a/fs/erofs/super.c b/fs/erofs/super.c
index 11b88559f8bf..25f6b8b37f28 100644
--- a/fs/erofs/super.c
+++ b/fs/erofs/super.c
@@ -340,15 +340,15 @@ static int erofs_read_superblock(struct super_block *sb)
 static void erofs_default_options(struct erofs_fs_context *ctx)
 {
 #ifdef CONFIG_EROFS_FS_ZIP
-	ctx->cache_strategy = EROFS_ZIP_CACHE_READAROUND;
-	ctx->max_sync_decompress_pages = 3;
-	ctx->readahead_sync_decompress = false;
+	ctx->opt.cache_strategy = EROFS_ZIP_CACHE_READAROUND;
+	ctx->opt.max_sync_decompress_pages = 3;
+	ctx->opt.readahead_sync_decompress = false;
 #endif
 #ifdef CONFIG_EROFS_FS_XATTR
-	set_opt(ctx, XATTR_USER);
+	set_opt(&ctx->opt, XATTR_USER);
 #endif
 #ifdef CONFIG_EROFS_FS_POSIX_ACL
-	set_opt(ctx, POSIX_ACL);
+	set_opt(&ctx->opt, POSIX_ACL);
 #endif
 }
 
@@ -392,12 +392,12 @@ static bool erofs_fc_set_dax_mode(struct fs_context *fc, unsigned int mode)
 	switch (mode) {
 	case EROFS_MOUNT_DAX_ALWAYS:
 		warnfc(fc, "DAX enabled. Warning: EXPERIMENTAL, use at your own risk");
-		set_opt(ctx, DAX_ALWAYS);
-		clear_opt(ctx, DAX_NEVER);
+		set_opt(&ctx->opt, DAX_ALWAYS);
+		clear_opt(&ctx->opt, DAX_NEVER);
 		return true;
 	case EROFS_MOUNT_DAX_NEVER:
-		set_opt(ctx, DAX_NEVER);
-		clear_opt(ctx, DAX_ALWAYS);
+		set_opt(&ctx->opt, DAX_NEVER);
+		clear_opt(&ctx->opt, DAX_ALWAYS);
 		return true;
 	default:
 		DBG_BUGON(1);
@@ -424,9 +424,9 @@ static int erofs_fc_parse_param(struct fs_context *fc,
 	case Opt_user_xattr:
 #ifdef CONFIG_EROFS_FS_XATTR
 		if (result.boolean)
-			set_opt(ctx, XATTR_USER);
+			set_opt(&ctx->opt, XATTR_USER);
 		else
-			clear_opt(ctx, XATTR_USER);
+			clear_opt(&ctx->opt, XATTR_USER);
 #else
 		errorfc(fc, "{,no}user_xattr options not supported");
 #endif
@@ -434,16 +434,16 @@ static int erofs_fc_parse_param(struct fs_context *fc,
 	case Opt_acl:
 #ifdef CONFIG_EROFS_FS_POSIX_ACL
 		if (result.boolean)
-			set_opt(ctx, POSIX_ACL);
+			set_opt(&ctx->opt, POSIX_ACL);
 		else
-			clear_opt(ctx, POSIX_ACL);
+			clear_opt(&ctx->opt, POSIX_ACL);
 #else
 		errorfc(fc, "{,no}acl options not supported");
 #endif
 		break;
 	case Opt_cache_strategy:
 #ifdef CONFIG_EROFS_FS_ZIP
-		ctx->cache_strategy = result.uint_32;
+		ctx->opt.cache_strategy = result.uint_32;
 #else
 		errorfc(fc, "compression not supported, cache_strategy ignored");
 #endif
@@ -540,15 +540,16 @@ static int erofs_fc_fill_super(struct super_block *sb, struct fs_context *fc)
 		return -ENOMEM;
 
 	sb->s_fs_info = sbi;
+	sbi->opt = ctx->opt;
 	sbi->dax_dev = fs_dax_get_by_bdev(sb->s_bdev);
 	err = erofs_read_superblock(sb);
 	if (err)
 		return err;
 
-	if (test_opt(ctx, DAX_ALWAYS) &&
+	if (test_opt(&sbi->opt, DAX_ALWAYS) &&
 	    !dax_supported(sbi->dax_dev, sb->s_bdev, EROFS_BLKSIZ, 0, bdev_nr_sectors(sb->s_bdev))) {
 		errorfc(fc, "DAX unsupported by block device. Turning off DAX.");
-		clear_opt(ctx, DAX_ALWAYS);
+		clear_opt(&sbi->opt, DAX_ALWAYS);
 	}
 	sb->s_flags |= SB_RDONLY | SB_NOATIME;
 	sb->s_maxbytes = MAX_LFS_FILESIZE;
@@ -557,13 +558,11 @@ static int erofs_fc_fill_super(struct super_block *sb, struct fs_context *fc)
 	sb->s_op = &erofs_sops;
 	sb->s_xattr = erofs_xattr_handlers;
 
-	if (test_opt(ctx, POSIX_ACL))
+	if (test_opt(&sbi->opt, POSIX_ACL))
 		sb->s_flags |= SB_POSIXACL;
 	else
 		sb->s_flags &= ~SB_POSIXACL;
 
-	sbi->ctx = *ctx;
-
 #ifdef CONFIG_EROFS_FS_ZIP
 	xa_init(&sbi->managed_pslots);
 #endif
@@ -607,12 +606,12 @@ static int erofs_fc_reconfigure(struct fs_context *fc)
 
 	DBG_BUGON(!sb_rdonly(sb));
 
-	if (test_opt(ctx, POSIX_ACL))
+	if (test_opt(&ctx->opt, POSIX_ACL))
 		fc->sb_flags |= SB_POSIXACL;
 	else
 		fc->sb_flags &= ~SB_POSIXACL;
 
-	sbi->ctx = *ctx;
+	sbi->opt = ctx->opt;
 
 	fc->sb_flags |= SB_RDONLY;
 	return 0;
@@ -640,7 +639,6 @@ static int erofs_init_fs_context(struct fs_context *fc)
 	erofs_default_options(fc->fs_private);
 
 	fc->ops = &erofs_context_ops;
-
 	return 0;
 }
 
@@ -763,31 +761,31 @@ static int erofs_statfs(struct dentry *dentry, struct kstatfs *buf)
 static int erofs_show_options(struct seq_file *seq, struct dentry *root)
 {
 	struct erofs_sb_info *sbi = EROFS_SB(root->d_sb);
-	struct erofs_fs_context *ctx = &sbi->ctx;
+	struct erofs_mount_opts *opt = &sbi->opt;
 
 #ifdef CONFIG_EROFS_FS_XATTR
-	if (test_opt(ctx, XATTR_USER))
+	if (test_opt(opt, XATTR_USER))
 		seq_puts(seq, ",user_xattr");
 	else
 		seq_puts(seq, ",nouser_xattr");
 #endif
 #ifdef CONFIG_EROFS_FS_POSIX_ACL
-	if (test_opt(ctx, POSIX_ACL))
+	if (test_opt(opt, POSIX_ACL))
 		seq_puts(seq, ",acl");
 	else
 		seq_puts(seq, ",noacl");
 #endif
 #ifdef CONFIG_EROFS_FS_ZIP
-	if (ctx->cache_strategy == EROFS_ZIP_CACHE_DISABLED)
+	if (opt->cache_strategy == EROFS_ZIP_CACHE_DISABLED)
 		seq_puts(seq, ",cache_strategy=disabled");
-	else if (ctx->cache_strategy == EROFS_ZIP_CACHE_READAHEAD)
+	else if (opt->cache_strategy == EROFS_ZIP_CACHE_READAHEAD)
 		seq_puts(seq, ",cache_strategy=readahead");
-	else if (ctx->cache_strategy == EROFS_ZIP_CACHE_READAROUND)
+	else if (opt->cache_strategy == EROFS_ZIP_CACHE_READAROUND)
 		seq_puts(seq, ",cache_strategy=readaround");
 #endif
-	if (test_opt(ctx, DAX_ALWAYS))
+	if (test_opt(opt, DAX_ALWAYS))
 		seq_puts(seq, ",dax=always");
-	if (test_opt(ctx, DAX_NEVER))
+	if (test_opt(opt, DAX_NEVER))
 		seq_puts(seq, ",dax=never");
 	return 0;
 }
diff --git a/fs/erofs/xattr.c b/fs/erofs/xattr.c
index 778f2c52295d..01c581e93c5f 100644
--- a/fs/erofs/xattr.c
+++ b/fs/erofs/xattr.c
@@ -429,7 +429,7 @@ static int shared_getxattr(struct inode *inode, struct getxattr_iter *it)
 
 static bool erofs_xattr_user_list(struct dentry *dentry)
 {
-	return test_opt(&EROFS_SB(dentry->d_sb)->ctx, XATTR_USER);
+	return test_opt(&EROFS_SB(dentry->d_sb)->opt, XATTR_USER);
 }
 
 static bool erofs_xattr_trusted_list(struct dentry *dentry)
@@ -476,7 +476,7 @@ static int erofs_xattr_generic_get(const struct xattr_handler *handler,
 
 	switch (handler->flags) {
 	case EROFS_XATTR_INDEX_USER:
-		if (!test_opt(&sbi->ctx, XATTR_USER))
+		if (!test_opt(&sbi->opt, XATTR_USER))
 			return -EOPNOTSUPP;
 		break;
 	case EROFS_XATTR_INDEX_TRUSTED:
diff --git a/fs/erofs/zdata.c b/fs/erofs/zdata.c
index 11c7a1aaebad..e59e22852c78 100644
--- a/fs/erofs/zdata.c
+++ b/fs/erofs/zdata.c
@@ -695,7 +695,7 @@ static int z_erofs_do_read_page(struct z_erofs_decompress_frontend *fe,
 		goto err_out;
 
 	/* preload all compressed pages (maybe downgrade role if necessary) */
-	if (should_alloc_managed_pages(fe, sbi->ctx.cache_strategy, map->m_la))
+	if (should_alloc_managed_pages(fe, sbi->opt.cache_strategy, map->m_la))
 		cache_strategy = TRYALLOC;
 	else
 		cache_strategy = DONTALLOC;
@@ -796,7 +796,7 @@ static void z_erofs_decompress_kickoff(struct z_erofs_decompressqueue *io,
 	/* Use workqueue and sync decompression for atomic contexts only */
 	if (in_atomic() || irqs_disabled()) {
 		queue_work(z_erofs_workqueue, &io->u.work);
-		sbi->ctx.readahead_sync_decompress = true;
+		sbi->opt.readahead_sync_decompress = true;
 		return;
 	}
 	z_erofs_decompressqueue_work(&io->u.work);
@@ -1411,8 +1411,8 @@ static void z_erofs_readahead(struct readahead_control *rac)
 	struct erofs_sb_info *const sbi = EROFS_I_SB(inode);
 
 	unsigned int nr_pages = readahead_count(rac);
-	bool sync = (sbi->ctx.readahead_sync_decompress &&
-			nr_pages <= sbi->ctx.max_sync_decompress_pages);
+	bool sync = (sbi->opt.readahead_sync_decompress &&
+			nr_pages <= sbi->opt.max_sync_decompress_pages);
 	struct z_erofs_decompress_frontend f = DECOMPRESS_FRONTEND_INIT(inode);
 	struct page *page, *head = NULL;
 	LIST_HEAD(pagepool);
-- 
2.24.4


^ permalink raw reply	[flat|nested] 5+ messages in thread

* [PATCH v3 2/2] erofs: add multiple device support
  2021-10-06 19:44 [PATCH v3 1/2] erofs: decouple basic mount options from fs_context Gao Xiang
@ 2021-10-06 19:44 ` Gao Xiang
  2021-10-07  6:42   ` Liu Bo
  2021-10-07  6:10 ` [PATCH v3 1/2] erofs: decouple basic mount options from fs_context Liu Bo
  1 sibling, 1 reply; 5+ messages in thread
From: Gao Xiang @ 2021-10-06 19:44 UTC (permalink / raw)
  To: linux-erofs, Chao Yu, Liu Bo
  Cc: Yan Song, LKML, Peng Tao, Joseph Qi, Changwei Ge, Gao Xiang, Liu Jiang

In order to support multi-layer container images, add multiple
device feature to EROFS. Two ways are available to use for now:

 - Devices can be mapped into 32-bit global block address space;
 - Device ID can be specified with the chunk indexes format.

Note that it assumes no extent would cross device boundary and mkfs
should take care of it seriously.

In the future, a dedicated device manager could be introduced then
so extra devices can be automatically scanned by uuid as well.

Cc: Liu Bo <bo.liu@linux.alibaba.com>
Cc: Chao Yu <chao@kernel.org>
Signed-off-by: Gao Xiang <hsiangkao@linux.alibaba.com>
---
 Documentation/filesystems/erofs.rst |  12 ++-
 fs/erofs/data.c                     |  73 ++++++++++---
 fs/erofs/erofs_fs.h                 |  20 +++-
 fs/erofs/internal.h                 |  35 ++++++-
 fs/erofs/super.c                    | 155 ++++++++++++++++++++++++++--
 fs/erofs/zdata.c                    |  20 +++-
 6 files changed, 279 insertions(+), 36 deletions(-)

diff --git a/Documentation/filesystems/erofs.rst b/Documentation/filesystems/erofs.rst
index b97579b7d8fb..01df283c7d04 100644
--- a/Documentation/filesystems/erofs.rst
+++ b/Documentation/filesystems/erofs.rst
@@ -19,9 +19,10 @@ It is designed as a better filesystem solution for the following scenarios:
    immutable and bit-for-bit identical to the official golden image for
    their releases due to security and other considerations and
 
- - hope to save some extra storage space with guaranteed end-to-end performance
-   by using reduced metadata and transparent file compression, especially
-   for those embedded devices with limited memory (ex, smartphone);
+ - hope to minimize extra storage space with guaranteed end-to-end performance
+   by using compact layout, transparent file compression and direct access,
+   especially for those embedded devices with limited memory and high-density
+   hosts with numerous containers;
 
 Here is the main features of EROFS:
 
@@ -51,7 +52,9 @@ Here is the main features of EROFS:
  - Support POSIX.1e ACLs by using xattrs;
 
  - Support transparent data compression as an option:
-   LZ4 algorithm with the fixed-sized output compression for high performance.
+   LZ4 algorithm with the fixed-sized output compression for high performance;
+
+ - Multiple device support for multi-layer container images.
 
 The following git tree provides the file system user-space tools under
 development (ex, formatting tool mkfs.erofs):
@@ -87,6 +90,7 @@ cache_strategy=%s      Select a strategy for cached decompression from now on:
 dax={always,never}     Use direct access (no page cache).  See
                        Documentation/filesystems/dax.rst.
 dax                    A legacy option which is an alias for ``dax=always``.
+device=%s              Specify a path to an extra device to be used together.
 ===================    =========================================================
 
 On-disk details
diff --git a/fs/erofs/data.c b/fs/erofs/data.c
index 9db829715652..808234d9190c 100644
--- a/fs/erofs/data.c
+++ b/fs/erofs/data.c
@@ -89,6 +89,7 @@ static int erofs_map_blocks(struct inode *inode,
 	erofs_off_t pos;
 	int err = 0;
 
+	map->m_deviceid = 0;
 	if (map->m_la >= inode->i_size) {
 		/* leave out-of-bound access unmapped */
 		map->m_flags = 0;
@@ -135,14 +136,8 @@ static int erofs_map_blocks(struct inode *inode,
 		map->m_flags = 0;
 		break;
 	default:
-		/* only one device is supported for now */
-		if (idx->device_id) {
-			erofs_err(sb, "invalid device id %u @ %llu for nid %llu",
-				  le16_to_cpu(idx->device_id),
-				  chunknr, vi->nid);
-			err = -EFSCORRUPTED;
-			goto out_unlock;
-		}
+		map->m_deviceid = le16_to_cpu(idx->device_id) &
+			EROFS_SB(sb)->device_id_mask;
 		map->m_pa = blknr_to_addr(le32_to_cpu(idx->blkaddr));
 		map->m_flags = EROFS_MAP_MAPPED;
 		break;
@@ -155,11 +150,55 @@ static int erofs_map_blocks(struct inode *inode,
 	return err;
 }
 
+int erofs_map_dev(struct super_block *sb, struct erofs_map_dev *map)
+{
+	struct erofs_dev_context *devs = EROFS_SB(sb)->devs;
+	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;
+
+	if (map->m_deviceid) {
+		down_read(&devs->rwsem);
+		dif = idr_find(&devs->tree, map->m_deviceid - 1);
+		if (!dif) {
+			up_read(&devs->rwsem);
+			return -ENODEV;
+		}
+		map->m_bdev = dif->bdev;
+		map->m_daxdev = dif->dax_dev;
+		up_read(&devs->rwsem);
+	} else if (devs->extra_devices) {
+		down_read(&devs->rwsem);
+		idr_for_each_entry(&devs->tree, dif, id) {
+			erofs_off_t startoff, length;
+
+			if (!dif->mapped_blkaddr)
+				continue;
+			startoff = blknr_to_addr(dif->mapped_blkaddr);
+			length = blknr_to_addr(dif->blocks);
+
+			if (map->m_pa >= startoff &&
+			    map->m_pa < startoff + length) {
+				map->m_pa -= startoff;
+				map->m_bdev = dif->bdev;
+				map->m_daxdev = dif->dax_dev;
+				break;
+			}
+		}
+		up_read(&devs->rwsem);
+	}
+	return 0;
+}
+
 static int erofs_iomap_begin(struct inode *inode, loff_t offset, loff_t length,
 		unsigned int flags, struct iomap *iomap, struct iomap *srcmap)
 {
 	int ret;
 	struct erofs_map_blocks map;
+	struct erofs_map_dev mdev;
 
 	map.m_la = offset;
 	map.m_llen = length;
@@ -168,8 +207,16 @@ static int erofs_iomap_begin(struct inode *inode, loff_t offset, loff_t length,
 	if (ret < 0)
 		return ret;
 
-	iomap->bdev = inode->i_sb->s_bdev;
-	iomap->dax_dev = EROFS_I_SB(inode)->dax_dev;
+	mdev = (struct erofs_map_dev) {
+		.m_deviceid = map.m_deviceid,
+		.m_pa = map.m_pa,
+	};
+	ret = erofs_map_dev(inode->i_sb, &mdev);
+	if (ret)
+		return ret;
+
+	iomap->bdev = mdev.m_bdev;
+	iomap->dax_dev = mdev.m_daxdev;
 	iomap->offset = map.m_la;
 	iomap->length = map.m_llen;
 	iomap->flags = 0;
@@ -188,15 +235,15 @@ static int erofs_iomap_begin(struct inode *inode, loff_t offset, loff_t length,
 
 		iomap->type = IOMAP_INLINE;
 		ipage = erofs_get_meta_page(inode->i_sb,
-					    erofs_blknr(map.m_pa));
+					    erofs_blknr(mdev.m_pa));
 		if (IS_ERR(ipage))
 			return PTR_ERR(ipage);
 		iomap->inline_data = page_address(ipage) +
-					erofs_blkoff(map.m_pa);
+					erofs_blkoff(mdev.m_pa);
 		iomap->private = ipage;
 	} else {
 		iomap->type = IOMAP_MAPPED;
-		iomap->addr = map.m_pa;
+		iomap->addr = mdev.m_pa;
 	}
 	return 0;
 }
diff --git a/fs/erofs/erofs_fs.h b/fs/erofs/erofs_fs.h
index b0b23f41abc3..c2f1c2453a55 100644
--- a/fs/erofs/erofs_fs.h
+++ b/fs/erofs/erofs_fs.h
@@ -21,14 +21,27 @@
 #define EROFS_FEATURE_INCOMPAT_COMPR_CFGS	0x00000002
 #define EROFS_FEATURE_INCOMPAT_BIG_PCLUSTER	0x00000002
 #define EROFS_FEATURE_INCOMPAT_CHUNKED_FILE	0x00000004
+#define EROFS_FEATURE_INCOMPAT_DEVICE_TABLE	0x00000008
 #define EROFS_ALL_FEATURE_INCOMPAT		\
 	(EROFS_FEATURE_INCOMPAT_LZ4_0PADDING | \
 	 EROFS_FEATURE_INCOMPAT_COMPR_CFGS | \
 	 EROFS_FEATURE_INCOMPAT_BIG_PCLUSTER | \
-	 EROFS_FEATURE_INCOMPAT_CHUNKED_FILE)
+	 EROFS_FEATURE_INCOMPAT_CHUNKED_FILE | \
+	 EROFS_FEATURE_INCOMPAT_DEVICE_TABLE)
 
 #define EROFS_SB_EXTSLOT_SIZE	16
 
+struct erofs_deviceslot {
+	union {
+		u8 uuid[16];		/* used for device manager later */
+		u8 userdata[64];	/* digest(sha256), etc. */
+	} u;
+	__le32 blocks;
+	__le32 mapped_blkaddr;
+	u8 reserved[56];
+};
+#define EROFS_DEVT_SLOT_SIZE	sizeof(struct erofs_deviceslot)
+
 /* erofs on-disk super block (currently 128 bytes) */
 struct erofs_super_block {
 	__le32 magic;           /* file system magic number */
@@ -54,7 +67,9 @@ struct erofs_super_block {
 		/* customized sliding window size instead of 64k by default */
 		__le16 lz4_max_distance;
 	} __packed u1;
-	__u8 reserved2[42];
+	__le16 extra_devices;	/* # of devices besides the primary device */
+	__le16 devt_slotoff;	/* startoff = devt_slotoff * devt_slotsize */
+	__u8 reserved2[38];
 };
 
 /*
@@ -384,6 +399,7 @@ static inline void erofs_check_ondisk_layout_definitions(void)
 	/* keep in sync between 2 index structures for better extendibility */
 	BUILD_BUG_ON(sizeof(struct erofs_inode_chunk_index) !=
 		     sizeof(struct z_erofs_vle_decompressed_index));
+	BUILD_BUG_ON(sizeof(struct erofs_deviceslot) != 128);
 
 	BUILD_BUG_ON(BIT(Z_EROFS_VLE_DI_CLUSTER_TYPE_BITS) <
 		     Z_EROFS_VLE_CLUSTER_TYPE_MAX - 1);
diff --git a/fs/erofs/internal.h b/fs/erofs/internal.h
index b1b9d1b5cb66..34f0933fd0c0 100644
--- a/fs/erofs/internal.h
+++ b/fs/erofs/internal.h
@@ -47,6 +47,15 @@ typedef u64 erofs_off_t;
 /* data type for filesystem-wide blocks number */
 typedef u32 erofs_blk_t;
 
+struct erofs_device_info {
+	char *path;
+	struct block_device *bdev;
+	struct dax_device *dax_dev;
+
+	u32 blocks;
+	u32 mapped_blkaddr;
+};
+
 struct erofs_mount_opts {
 #ifdef CONFIG_EROFS_FS_ZIP
 	/* current strategy of how to use managed cache */
@@ -60,8 +69,16 @@ struct erofs_mount_opts {
 	unsigned int mount_opt;
 };
 
+struct erofs_dev_context {
+	struct idr tree;
+	struct rw_semaphore rwsem;
+
+	unsigned int extra_devices;
+};
+
 struct erofs_fs_context {
 	struct erofs_mount_opts opt;
+	struct erofs_dev_context *devs;
 };
 
 /* all filesystem-wide lz4 configurations */
@@ -74,7 +91,6 @@ struct erofs_sb_lz4_info {
 
 struct erofs_sb_info {
 	struct erofs_mount_opts opt;	/* options */
-
 #ifdef CONFIG_EROFS_FS_ZIP
 	/* list for all registered superblocks, mainly for shrinker */
 	struct list_head list;
@@ -91,12 +107,16 @@ struct erofs_sb_info {
 
 	struct erofs_sb_lz4_info lz4;
 #endif	/* CONFIG_EROFS_FS_ZIP */
+	struct erofs_dev_context *devs;
 	struct dax_device *dax_dev;
-	u32 blocks;
+	u64 total_blocks;
+	u32 primarydevice_blocks;
+
 	u32 meta_blkaddr;
 #ifdef CONFIG_EROFS_FS_XATTR
 	u32 xattr_blkaddr;
 #endif
+	u16 device_id_mask;
 
 	/* inode slot unit size in bit shift */
 	unsigned char islotbits;
@@ -241,6 +261,7 @@ static inline bool erofs_sb_has_##name(struct erofs_sb_info *sbi) \
 EROFS_FEATURE_FUNCS(lz4_0padding, incompat, INCOMPAT_LZ4_0PADDING)
 EROFS_FEATURE_FUNCS(compr_cfgs, incompat, INCOMPAT_COMPR_CFGS)
 EROFS_FEATURE_FUNCS(big_pcluster, incompat, INCOMPAT_BIG_PCLUSTER)
+EROFS_FEATURE_FUNCS(device_table, incompat, INCOMPAT_DEVICE_TABLE)
 EROFS_FEATURE_FUNCS(sb_chksum, compat, COMPAT_SB_CHKSUM)
 
 /* atomic flag definitions */
@@ -359,6 +380,7 @@ struct erofs_map_blocks {
 	erofs_off_t m_pa, m_la;
 	u64 m_plen, m_llen;
 
+	unsigned int m_deviceid;
 	unsigned int m_flags;
 
 	struct page *mpage;
@@ -390,9 +412,18 @@ static inline int z_erofs_map_blocks_iter(struct inode *inode,
 }
 #endif	/* !CONFIG_EROFS_FS_ZIP */
 
+struct erofs_map_dev {
+	struct block_device *m_bdev;
+	struct dax_device *m_daxdev;
+
+	erofs_off_t m_pa;
+	unsigned int m_deviceid;
+};
+
 /* data.c */
 extern const struct file_operations erofs_file_fops;
 struct page *erofs_get_meta_page(struct super_block *sb, erofs_blk_t blkaddr);
+int erofs_map_dev(struct super_block *sb, struct erofs_map_dev *dev);
 int erofs_fiemap(struct inode *inode, struct fiemap_extent_info *fieinfo,
 		 u64 start, u64 len);
 
diff --git a/fs/erofs/super.c b/fs/erofs/super.c
index 25f6b8b37f28..f4dd20f76862 100644
--- a/fs/erofs/super.c
+++ b/fs/erofs/super.c
@@ -252,6 +252,79 @@ static int erofs_load_compr_cfgs(struct super_block *sb,
 }
 #endif
 
+static int erofs_init_devices(struct super_block *sb,
+			      struct erofs_super_block *dsb)
+{
+	struct erofs_sb_info *sbi = EROFS_SB(sb);
+	unsigned int ondisk_extradevs;
+	erofs_off_t pos;
+	struct page *page = NULL;
+	struct erofs_device_info *dif;
+	struct erofs_deviceslot *dis;
+	void *ptr;
+	int id, err = 0;
+
+	sbi->total_blocks = sbi->primarydevice_blocks;
+	if (!erofs_sb_has_device_table(sbi))
+		ondisk_extradevs = 0;
+	else
+		ondisk_extradevs = le16_to_cpu(dsb->extra_devices);
+
+	if (ondisk_extradevs != sbi->devs->extra_devices) {
+		erofs_err(sb, "extra devices don't match (ondisk %u, given %u)",
+			  ondisk_extradevs, sbi->devs->extra_devices);
+		return -EINVAL;
+	}
+	if (!ondisk_extradevs)
+		return 0;
+
+	sbi->device_id_mask = roundup_pow_of_two(ondisk_extradevs + 1) - 1;
+	pos = le16_to_cpu(dsb->devt_slotoff) * EROFS_DEVT_SLOT_SIZE;
+	down_read(&sbi->devs->rwsem);
+	idr_for_each_entry(&sbi->devs->tree, dif, id) {
+		erofs_blk_t blk = erofs_blknr(pos);
+		struct block_device *bdev;
+
+		if (!page || page->index != blk) {
+			if (page) {
+				kunmap(page);
+				unlock_page(page);
+				put_page(page);
+			}
+
+			page = erofs_get_meta_page(sb, blk);
+			if (IS_ERR(page)) {
+				up_read(&sbi->devs->rwsem);
+				return PTR_ERR(page);
+			}
+			ptr = kmap(page);
+		}
+		dis = ptr + erofs_blkoff(pos);
+
+		bdev = blkdev_get_by_path(dif->path,
+					  FMODE_READ | FMODE_EXCL,
+					  sb->s_type);
+		if (IS_ERR(bdev)) {
+			err = PTR_ERR(bdev);
+			goto err_out;
+		}
+		dif->bdev = bdev;
+		dif->dax_dev = fs_dax_get_by_bdev(bdev);
+		dif->blocks = le32_to_cpu(dis->blocks);
+		dif->mapped_blkaddr = le32_to_cpu(dis->mapped_blkaddr);
+		sbi->total_blocks += dif->blocks;
+		pos += sizeof(*dis);
+	}
+err_out:
+	up_read(&sbi->devs->rwsem);
+	if (page) {
+		kunmap(page);
+		unlock_page(page);
+		put_page(page);
+	}
+	return err;
+}
+
 static int erofs_read_superblock(struct super_block *sb)
 {
 	struct erofs_sb_info *sbi;
@@ -303,7 +376,7 @@ static int erofs_read_superblock(struct super_block *sb)
 			  sbi->sb_size);
 		goto out;
 	}
-	sbi->blocks = le32_to_cpu(dsb->blocks);
+	sbi->primarydevice_blocks = le32_to_cpu(dsb->blocks);
 	sbi->meta_blkaddr = le32_to_cpu(dsb->meta_blkaddr);
 #ifdef CONFIG_EROFS_FS_XATTR
 	sbi->xattr_blkaddr = le32_to_cpu(dsb->xattr_blkaddr);
@@ -330,6 +403,11 @@ static int erofs_read_superblock(struct super_block *sb)
 		ret = erofs_load_compr_cfgs(sb, dsb);
 	else
 		ret = z_erofs_load_lz4_config(sb, dsb, NULL, 0);
+	if (ret < 0)
+		goto out;
+
+	/* handle multiple devices */
+	ret = erofs_init_devices(sb, dsb);
 out:
 	kunmap(page);
 	put_page(page);
@@ -358,6 +436,7 @@ enum {
 	Opt_cache_strategy,
 	Opt_dax,
 	Opt_dax_enum,
+	Opt_device,
 	Opt_err
 };
 
@@ -381,6 +460,7 @@ static const struct fs_parameter_spec erofs_fs_parameters[] = {
 		     erofs_param_cache_strategy),
 	fsparam_flag("dax",             Opt_dax),
 	fsparam_enum("dax",		Opt_dax_enum, erofs_dax_param_enums),
+	fsparam_string("device",	Opt_device),
 	{}
 };
 
@@ -412,9 +492,10 @@ static bool erofs_fc_set_dax_mode(struct fs_context *fc, unsigned int mode)
 static int erofs_fc_parse_param(struct fs_context *fc,
 				struct fs_parameter *param)
 {
-	struct erofs_fs_context *ctx __maybe_unused = fc->fs_private;
+	struct erofs_fs_context *ctx = fc->fs_private;
 	struct fs_parse_result result;
-	int opt;
+	struct erofs_device_info *dif;
+	int opt, ret;
 
 	opt = fs_parse(fc, erofs_fs_parameters, param, &result);
 	if (opt < 0)
@@ -456,6 +537,25 @@ static int erofs_fc_parse_param(struct fs_context *fc,
 		if (!erofs_fc_set_dax_mode(fc, result.uint_32))
 			return -EINVAL;
 		break;
+	case Opt_device:
+		dif = kzalloc(sizeof(*dif), GFP_KERNEL);
+		if (!dif)
+			return -ENOMEM;
+		dif->path = kstrdup(param->string, GFP_KERNEL);
+		if (!dif->path) {
+			kfree(dif);
+			return -ENOMEM;
+		}
+		down_write(&ctx->devs->rwsem);
+		ret = idr_alloc(&ctx->devs->tree, dif, 0, 0, GFP_KERNEL);
+		up_write(&ctx->devs->rwsem);
+		if (ret < 0) {
+			kfree(dif->path);
+			kfree(dif);
+			return ret;
+		}
+		++ctx->devs->extra_devices;
+		break;
 	default:
 		return -ENOPARAM;
 	}
@@ -542,6 +642,9 @@ static int erofs_fc_fill_super(struct super_block *sb, struct fs_context *fc)
 	sb->s_fs_info = sbi;
 	sbi->opt = ctx->opt;
 	sbi->dax_dev = fs_dax_get_by_bdev(sb->s_bdev);
+	sbi->devs = ctx->devs;
+	ctx->devs = NULL;
+
 	err = erofs_read_superblock(sb);
 	if (err)
 		return err;
@@ -617,9 +720,32 @@ static int erofs_fc_reconfigure(struct fs_context *fc)
 	return 0;
 }
 
+static int erofs_release_device_info(int id, void *ptr, void *data)
+{
+	struct erofs_device_info *dif = ptr;
+
+	fs_put_dax(dif->dax_dev);
+	if (dif->bdev)
+		blkdev_put(dif->bdev, FMODE_READ | FMODE_EXCL);
+	kfree(dif->path);
+	kfree(dif);
+	return 0;
+}
+
+static void erofs_free_dev_context(struct erofs_dev_context *devs)
+{
+	if (!devs)
+		return;
+	idr_for_each(&devs->tree, &erofs_release_device_info, NULL);
+	idr_destroy(&devs->tree);
+}
+
 static void erofs_fc_free(struct fs_context *fc)
 {
-	kfree(fc->fs_private);
+	struct erofs_fs_context *ctx = fc->fs_private;
+
+	erofs_free_dev_context(ctx->devs);
+	kfree(ctx);
 }
 
 static const struct fs_context_operations erofs_context_ops = {
@@ -631,13 +757,20 @@ static const struct fs_context_operations erofs_context_ops = {
 
 static int erofs_init_fs_context(struct fs_context *fc)
 {
-	fc->fs_private = kzalloc(sizeof(struct erofs_fs_context), GFP_KERNEL);
-	if (!fc->fs_private)
-		return -ENOMEM;
+	struct erofs_fs_context *ctx = kzalloc(sizeof(*ctx), GFP_KERNEL);
 
-	/* set default mount options */
-	erofs_default_options(fc->fs_private);
+	if (!ctx)
+		return -ENOMEM;
+	ctx->devs = kzalloc(sizeof(struct erofs_dev_context), GFP_KERNEL);
+	if (!ctx->devs) {
+		kfree(ctx);
+		return -ENOMEM;
+	}
+	fc->fs_private = ctx;
 
+	idr_init(&ctx->devs->tree);
+	init_rwsem(&ctx->devs->rwsem);
+	erofs_default_options(ctx);
 	fc->ops = &erofs_context_ops;
 	return 0;
 }
@@ -657,6 +790,8 @@ static void erofs_kill_sb(struct super_block *sb)
 	sbi = EROFS_SB(sb);
 	if (!sbi)
 		return;
+
+	erofs_free_dev_context(sbi->devs);
 	fs_put_dax(sbi->dax_dev);
 	kfree(sbi);
 	sb->s_fs_info = NULL;
@@ -746,7 +881,7 @@ static int erofs_statfs(struct dentry *dentry, struct kstatfs *buf)
 
 	buf->f_type = sb->s_magic;
 	buf->f_bsize = EROFS_BLKSIZ;
-	buf->f_blocks = sbi->blocks;
+	buf->f_blocks = sbi->total_blocks;
 	buf->f_bfree = buf->f_bavail = 0;
 
 	buf->f_files = ULLONG_MAX;
diff --git a/fs/erofs/zdata.c b/fs/erofs/zdata.c
index e59e22852c78..8c947ed49299 100644
--- a/fs/erofs/zdata.c
+++ b/fs/erofs/zdata.c
@@ -1266,8 +1266,9 @@ static void z_erofs_submit_queue(struct super_block *sb,
 	struct z_erofs_decompressqueue *q[NR_JOBQUEUES];
 	void *bi_private;
 	z_erofs_next_pcluster_t owned_head = f->clt.owned_head;
-	/* since bio will be NULL, no need to initialize last_index */
+	/* bio is NULL initially, so no need to initialize last_{index,bdev} */
 	pgoff_t last_index;
+	struct block_device *last_bdev;
 	unsigned int nr_bios = 0;
 	struct bio *bio = NULL;
 
@@ -1279,6 +1280,7 @@ static void z_erofs_submit_queue(struct super_block *sb,
 	q[JQ_SUBMIT]->head = owned_head;
 
 	do {
+		struct erofs_map_dev mdev;
 		struct z_erofs_pcluster *pcl;
 		pgoff_t cur, end;
 		unsigned int i = 0;
@@ -1290,7 +1292,13 @@ static void z_erofs_submit_queue(struct super_block *sb,
 
 		pcl = container_of(owned_head, struct z_erofs_pcluster, next);
 
-		cur = pcl->obj.index;
+		/* no device id here, thus it will always succeed */
+		mdev = (struct erofs_map_dev) {
+			.m_pa = blknr_to_addr(pcl->obj.index),
+		};
+		(void)erofs_map_dev(sb, &mdev);
+
+		cur = erofs_blknr(mdev.m_pa);
 		end = cur + pcl->pclusterpages;
 
 		/* close the main owned chain at first */
@@ -1306,7 +1314,8 @@ static void z_erofs_submit_queue(struct super_block *sb,
 			if (!page)
 				continue;
 
-			if (bio && cur != last_index + 1) {
+			if (bio && (cur != last_index + 1 ||
+				    last_bdev != mdev.m_bdev)) {
 submit_bio_retry:
 				submit_bio(bio);
 				bio = NULL;
@@ -1314,9 +1323,10 @@ static void z_erofs_submit_queue(struct super_block *sb,
 
 			if (!bio) {
 				bio = bio_alloc(GFP_NOIO, BIO_MAX_VECS);
-
 				bio->bi_end_io = z_erofs_decompressqueue_endio;
-				bio_set_dev(bio, sb->s_bdev);
+
+				bio_set_dev(bio, mdev.m_bdev);
+				last_bdev = mdev.m_bdev;
 				bio->bi_iter.bi_sector = (sector_t)cur <<
 					LOG_SECTORS_PER_BLOCK;
 				bio->bi_private = bi_private;
-- 
2.24.4


^ permalink raw reply	[flat|nested] 5+ messages in thread

* Re: [PATCH v3 1/2] erofs: decouple basic mount options from fs_context
  2021-10-06 19:44 [PATCH v3 1/2] erofs: decouple basic mount options from fs_context Gao Xiang
  2021-10-06 19:44 ` [PATCH v3 2/2] erofs: add multiple device support Gao Xiang
@ 2021-10-07  6:10 ` Liu Bo
  2021-10-07  6:29   ` Gao Xiang
  1 sibling, 1 reply; 5+ messages in thread
From: Liu Bo @ 2021-10-07  6:10 UTC (permalink / raw)
  To: Gao Xiang
  Cc: Yan Song, LKML, Peng Tao, Joseph Qi, Changwei Ge, Liu Jiang, linux-erofs

On Thu, Oct 07, 2021 at 03:44:52AM +0800, Gao Xiang wrote:
> Previously, EROFS mount options are all in the basic types, so
> erofs_fs_context can be directly copied with assignment. However,
> when the multiple device feature is introduced, it's hard to handle
> multiple device information like the other basic mount options.
> 
> There is no need to allocate the whole sb info in advance, instead,
> let's separate the basic mount options from fs_context, thus
> multiple device information can be handled gracefully then.

This is a bit confusing, IIRC, currently erofs does not allocate the whole
sb info in advance, does it?

thanks,
liubo

> 
> No logic changes.
> 
> Cc: Liu Bo <bo.liu@linux.alibaba.com>
> Cc: Chao Yu <chao@kernel.org>
> Signed-off-by: Gao Xiang <hsiangkao@linux.alibaba.com>
> ---
> changes since RFC v2:
>  - add a new patch to decouple basic mount options from fs_context
>    so that dev_context can be passed safely.
> 
>  fs/erofs/inode.c    |  2 +-
>  fs/erofs/internal.h | 16 ++++++++-----
>  fs/erofs/super.c    | 58 ++++++++++++++++++++++-----------------------
>  fs/erofs/xattr.c    |  4 ++--
>  fs/erofs/zdata.c    |  8 +++----
>  5 files changed, 45 insertions(+), 43 deletions(-)
> 
> diff --git a/fs/erofs/inode.c b/fs/erofs/inode.c
> index a552399e211d..2345f1de438e 100644
> --- a/fs/erofs/inode.c
> +++ b/fs/erofs/inode.c
> @@ -192,7 +192,7 @@ static struct page *erofs_read_inode(struct inode *inode,
>  	inode->i_atime.tv_nsec = inode->i_ctime.tv_nsec;
>  
>  	inode->i_flags &= ~S_DAX;
> -	if (test_opt(&sbi->ctx, DAX_ALWAYS) && S_ISREG(inode->i_mode) &&
> +	if (test_opt(&sbi->opt, DAX_ALWAYS) && S_ISREG(inode->i_mode) &&
>  	    vi->datalayout == EROFS_INODE_FLAT_PLAIN)
>  		inode->i_flags |= S_DAX;
>  	if (!nblks)
> diff --git a/fs/erofs/internal.h b/fs/erofs/internal.h
> index 9524e155b38f..b1b9d1b5cb66 100644
> --- a/fs/erofs/internal.h
> +++ b/fs/erofs/internal.h
> @@ -47,7 +47,7 @@ typedef u64 erofs_off_t;
>  /* data type for filesystem-wide blocks number */
>  typedef u32 erofs_blk_t;
>  
> -struct erofs_fs_context {
> +struct erofs_mount_opts {
>  #ifdef CONFIG_EROFS_FS_ZIP
>  	/* current strategy of how to use managed cache */
>  	unsigned char cache_strategy;
> @@ -60,6 +60,10 @@ struct erofs_fs_context {
>  	unsigned int mount_opt;
>  };
>  
> +struct erofs_fs_context {
> +	struct erofs_mount_opts opt;
> +};
> +
>  /* all filesystem-wide lz4 configurations */
>  struct erofs_sb_lz4_info {
>  	/* # of pages needed for EROFS lz4 rolling decompression */
> @@ -69,6 +73,8 @@ struct erofs_sb_lz4_info {
>  };
>  
>  struct erofs_sb_info {
> +	struct erofs_mount_opts opt;	/* options */
> +
>  #ifdef CONFIG_EROFS_FS_ZIP
>  	/* list for all registered superblocks, mainly for shrinker */
>  	struct list_head list;
> @@ -108,8 +114,6 @@ struct erofs_sb_info {
>  	u8 volume_name[16];             /* volume name */
>  	u32 feature_compat;
>  	u32 feature_incompat;
> -
> -	struct erofs_fs_context ctx;	/* options */
>  };
>  
>  #define EROFS_SB(sb) ((struct erofs_sb_info *)(sb)->s_fs_info)
> @@ -121,9 +125,9 @@ struct erofs_sb_info {
>  #define EROFS_MOUNT_DAX_ALWAYS		0x00000040
>  #define EROFS_MOUNT_DAX_NEVER		0x00000080
>  
> -#define clear_opt(ctx, option)	((ctx)->mount_opt &= ~EROFS_MOUNT_##option)
> -#define set_opt(ctx, option)	((ctx)->mount_opt |= EROFS_MOUNT_##option)
> -#define test_opt(ctx, option)	((ctx)->mount_opt & EROFS_MOUNT_##option)
> +#define clear_opt(opt, option)	((opt)->mount_opt &= ~EROFS_MOUNT_##option)
> +#define set_opt(opt, option)	((opt)->mount_opt |= EROFS_MOUNT_##option)
> +#define test_opt(opt, option)	((opt)->mount_opt & EROFS_MOUNT_##option)
>  
>  enum {
>  	EROFS_ZIP_CACHE_DISABLED,
> diff --git a/fs/erofs/super.c b/fs/erofs/super.c
> index 11b88559f8bf..25f6b8b37f28 100644
> --- a/fs/erofs/super.c
> +++ b/fs/erofs/super.c
> @@ -340,15 +340,15 @@ static int erofs_read_superblock(struct super_block *sb)
>  static void erofs_default_options(struct erofs_fs_context *ctx)
>  {
>  #ifdef CONFIG_EROFS_FS_ZIP
> -	ctx->cache_strategy = EROFS_ZIP_CACHE_READAROUND;
> -	ctx->max_sync_decompress_pages = 3;
> -	ctx->readahead_sync_decompress = false;
> +	ctx->opt.cache_strategy = EROFS_ZIP_CACHE_READAROUND;
> +	ctx->opt.max_sync_decompress_pages = 3;
> +	ctx->opt.readahead_sync_decompress = false;
>  #endif
>  #ifdef CONFIG_EROFS_FS_XATTR
> -	set_opt(ctx, XATTR_USER);
> +	set_opt(&ctx->opt, XATTR_USER);
>  #endif
>  #ifdef CONFIG_EROFS_FS_POSIX_ACL
> -	set_opt(ctx, POSIX_ACL);
> +	set_opt(&ctx->opt, POSIX_ACL);
>  #endif
>  }
>  
> @@ -392,12 +392,12 @@ static bool erofs_fc_set_dax_mode(struct fs_context *fc, unsigned int mode)
>  	switch (mode) {
>  	case EROFS_MOUNT_DAX_ALWAYS:
>  		warnfc(fc, "DAX enabled. Warning: EXPERIMENTAL, use at your own risk");
> -		set_opt(ctx, DAX_ALWAYS);
> -		clear_opt(ctx, DAX_NEVER);
> +		set_opt(&ctx->opt, DAX_ALWAYS);
> +		clear_opt(&ctx->opt, DAX_NEVER);
>  		return true;
>  	case EROFS_MOUNT_DAX_NEVER:
> -		set_opt(ctx, DAX_NEVER);
> -		clear_opt(ctx, DAX_ALWAYS);
> +		set_opt(&ctx->opt, DAX_NEVER);
> +		clear_opt(&ctx->opt, DAX_ALWAYS);
>  		return true;
>  	default:
>  		DBG_BUGON(1);
> @@ -424,9 +424,9 @@ static int erofs_fc_parse_param(struct fs_context *fc,
>  	case Opt_user_xattr:
>  #ifdef CONFIG_EROFS_FS_XATTR
>  		if (result.boolean)
> -			set_opt(ctx, XATTR_USER);
> +			set_opt(&ctx->opt, XATTR_USER);
>  		else
> -			clear_opt(ctx, XATTR_USER);
> +			clear_opt(&ctx->opt, XATTR_USER);
>  #else
>  		errorfc(fc, "{,no}user_xattr options not supported");
>  #endif
> @@ -434,16 +434,16 @@ static int erofs_fc_parse_param(struct fs_context *fc,
>  	case Opt_acl:
>  #ifdef CONFIG_EROFS_FS_POSIX_ACL
>  		if (result.boolean)
> -			set_opt(ctx, POSIX_ACL);
> +			set_opt(&ctx->opt, POSIX_ACL);
>  		else
> -			clear_opt(ctx, POSIX_ACL);
> +			clear_opt(&ctx->opt, POSIX_ACL);
>  #else
>  		errorfc(fc, "{,no}acl options not supported");
>  #endif
>  		break;
>  	case Opt_cache_strategy:
>  #ifdef CONFIG_EROFS_FS_ZIP
> -		ctx->cache_strategy = result.uint_32;
> +		ctx->opt.cache_strategy = result.uint_32;
>  #else
>  		errorfc(fc, "compression not supported, cache_strategy ignored");
>  #endif
> @@ -540,15 +540,16 @@ static int erofs_fc_fill_super(struct super_block *sb, struct fs_context *fc)
>  		return -ENOMEM;
>  
>  	sb->s_fs_info = sbi;
> +	sbi->opt = ctx->opt;
>  	sbi->dax_dev = fs_dax_get_by_bdev(sb->s_bdev);
>  	err = erofs_read_superblock(sb);
>  	if (err)
>  		return err;
>  
> -	if (test_opt(ctx, DAX_ALWAYS) &&
> +	if (test_opt(&sbi->opt, DAX_ALWAYS) &&
>  	    !dax_supported(sbi->dax_dev, sb->s_bdev, EROFS_BLKSIZ, 0, bdev_nr_sectors(sb->s_bdev))) {
>  		errorfc(fc, "DAX unsupported by block device. Turning off DAX.");
> -		clear_opt(ctx, DAX_ALWAYS);
> +		clear_opt(&sbi->opt, DAX_ALWAYS);
>  	}
>  	sb->s_flags |= SB_RDONLY | SB_NOATIME;
>  	sb->s_maxbytes = MAX_LFS_FILESIZE;
> @@ -557,13 +558,11 @@ static int erofs_fc_fill_super(struct super_block *sb, struct fs_context *fc)
>  	sb->s_op = &erofs_sops;
>  	sb->s_xattr = erofs_xattr_handlers;
>  
> -	if (test_opt(ctx, POSIX_ACL))
> +	if (test_opt(&sbi->opt, POSIX_ACL))
>  		sb->s_flags |= SB_POSIXACL;
>  	else
>  		sb->s_flags &= ~SB_POSIXACL;
>  
> -	sbi->ctx = *ctx;
> -
>  #ifdef CONFIG_EROFS_FS_ZIP
>  	xa_init(&sbi->managed_pslots);
>  #endif
> @@ -607,12 +606,12 @@ static int erofs_fc_reconfigure(struct fs_context *fc)
>  
>  	DBG_BUGON(!sb_rdonly(sb));
>  
> -	if (test_opt(ctx, POSIX_ACL))
> +	if (test_opt(&ctx->opt, POSIX_ACL))
>  		fc->sb_flags |= SB_POSIXACL;
>  	else
>  		fc->sb_flags &= ~SB_POSIXACL;
>  
> -	sbi->ctx = *ctx;
> +	sbi->opt = ctx->opt;
>  
>  	fc->sb_flags |= SB_RDONLY;
>  	return 0;
> @@ -640,7 +639,6 @@ static int erofs_init_fs_context(struct fs_context *fc)
>  	erofs_default_options(fc->fs_private);
>  
>  	fc->ops = &erofs_context_ops;
> -
>  	return 0;
>  }
>  
> @@ -763,31 +761,31 @@ static int erofs_statfs(struct dentry *dentry, struct kstatfs *buf)
>  static int erofs_show_options(struct seq_file *seq, struct dentry *root)
>  {
>  	struct erofs_sb_info *sbi = EROFS_SB(root->d_sb);
> -	struct erofs_fs_context *ctx = &sbi->ctx;
> +	struct erofs_mount_opts *opt = &sbi->opt;
>  
>  #ifdef CONFIG_EROFS_FS_XATTR
> -	if (test_opt(ctx, XATTR_USER))
> +	if (test_opt(opt, XATTR_USER))
>  		seq_puts(seq, ",user_xattr");
>  	else
>  		seq_puts(seq, ",nouser_xattr");
>  #endif
>  #ifdef CONFIG_EROFS_FS_POSIX_ACL
> -	if (test_opt(ctx, POSIX_ACL))
> +	if (test_opt(opt, POSIX_ACL))
>  		seq_puts(seq, ",acl");
>  	else
>  		seq_puts(seq, ",noacl");
>  #endif
>  #ifdef CONFIG_EROFS_FS_ZIP
> -	if (ctx->cache_strategy == EROFS_ZIP_CACHE_DISABLED)
> +	if (opt->cache_strategy == EROFS_ZIP_CACHE_DISABLED)
>  		seq_puts(seq, ",cache_strategy=disabled");
> -	else if (ctx->cache_strategy == EROFS_ZIP_CACHE_READAHEAD)
> +	else if (opt->cache_strategy == EROFS_ZIP_CACHE_READAHEAD)
>  		seq_puts(seq, ",cache_strategy=readahead");
> -	else if (ctx->cache_strategy == EROFS_ZIP_CACHE_READAROUND)
> +	else if (opt->cache_strategy == EROFS_ZIP_CACHE_READAROUND)
>  		seq_puts(seq, ",cache_strategy=readaround");
>  #endif
> -	if (test_opt(ctx, DAX_ALWAYS))
> +	if (test_opt(opt, DAX_ALWAYS))
>  		seq_puts(seq, ",dax=always");
> -	if (test_opt(ctx, DAX_NEVER))
> +	if (test_opt(opt, DAX_NEVER))
>  		seq_puts(seq, ",dax=never");
>  	return 0;
>  }
> diff --git a/fs/erofs/xattr.c b/fs/erofs/xattr.c
> index 778f2c52295d..01c581e93c5f 100644
> --- a/fs/erofs/xattr.c
> +++ b/fs/erofs/xattr.c
> @@ -429,7 +429,7 @@ static int shared_getxattr(struct inode *inode, struct getxattr_iter *it)
>  
>  static bool erofs_xattr_user_list(struct dentry *dentry)
>  {
> -	return test_opt(&EROFS_SB(dentry->d_sb)->ctx, XATTR_USER);
> +	return test_opt(&EROFS_SB(dentry->d_sb)->opt, XATTR_USER);
>  }
>  
>  static bool erofs_xattr_trusted_list(struct dentry *dentry)
> @@ -476,7 +476,7 @@ static int erofs_xattr_generic_get(const struct xattr_handler *handler,
>  
>  	switch (handler->flags) {
>  	case EROFS_XATTR_INDEX_USER:
> -		if (!test_opt(&sbi->ctx, XATTR_USER))
> +		if (!test_opt(&sbi->opt, XATTR_USER))
>  			return -EOPNOTSUPP;
>  		break;
>  	case EROFS_XATTR_INDEX_TRUSTED:
> diff --git a/fs/erofs/zdata.c b/fs/erofs/zdata.c
> index 11c7a1aaebad..e59e22852c78 100644
> --- a/fs/erofs/zdata.c
> +++ b/fs/erofs/zdata.c
> @@ -695,7 +695,7 @@ static int z_erofs_do_read_page(struct z_erofs_decompress_frontend *fe,
>  		goto err_out;
>  
>  	/* preload all compressed pages (maybe downgrade role if necessary) */
> -	if (should_alloc_managed_pages(fe, sbi->ctx.cache_strategy, map->m_la))
> +	if (should_alloc_managed_pages(fe, sbi->opt.cache_strategy, map->m_la))
>  		cache_strategy = TRYALLOC;
>  	else
>  		cache_strategy = DONTALLOC;
> @@ -796,7 +796,7 @@ static void z_erofs_decompress_kickoff(struct z_erofs_decompressqueue *io,
>  	/* Use workqueue and sync decompression for atomic contexts only */
>  	if (in_atomic() || irqs_disabled()) {
>  		queue_work(z_erofs_workqueue, &io->u.work);
> -		sbi->ctx.readahead_sync_decompress = true;
> +		sbi->opt.readahead_sync_decompress = true;
>  		return;
>  	}
>  	z_erofs_decompressqueue_work(&io->u.work);
> @@ -1411,8 +1411,8 @@ static void z_erofs_readahead(struct readahead_control *rac)
>  	struct erofs_sb_info *const sbi = EROFS_I_SB(inode);
>  
>  	unsigned int nr_pages = readahead_count(rac);
> -	bool sync = (sbi->ctx.readahead_sync_decompress &&
> -			nr_pages <= sbi->ctx.max_sync_decompress_pages);
> +	bool sync = (sbi->opt.readahead_sync_decompress &&
> +			nr_pages <= sbi->opt.max_sync_decompress_pages);
>  	struct z_erofs_decompress_frontend f = DECOMPRESS_FRONTEND_INIT(inode);
>  	struct page *page, *head = NULL;
>  	LIST_HEAD(pagepool);
> -- 
> 2.24.4

^ permalink raw reply	[flat|nested] 5+ messages in thread

* Re: [PATCH v3 1/2] erofs: decouple basic mount options from fs_context
  2021-10-07  6:10 ` [PATCH v3 1/2] erofs: decouple basic mount options from fs_context Liu Bo
@ 2021-10-07  6:29   ` Gao Xiang
  0 siblings, 0 replies; 5+ messages in thread
From: Gao Xiang @ 2021-10-07  6:29 UTC (permalink / raw)
  To: Liu Bo
  Cc: Yan Song, LKML, Peng Tao, Joseph Qi, Changwei Ge, Liu Jiang, linux-erofs

Hi Bo,

On Thu, Oct 07, 2021 at 02:10:48PM +0800, Liu Bo wrote:
> On Thu, Oct 07, 2021 at 03:44:52AM +0800, Gao Xiang wrote:
> > Previously, EROFS mount options are all in the basic types, so
> > erofs_fs_context can be directly copied with assignment. However,
> > when the multiple device feature is introduced, it's hard to handle
> > multiple device information like the other basic mount options.
> > 
> > There is no need to allocate the whole sb info in advance, instead,
> > let's separate the basic mount options from fs_context, thus
> > multiple device information can be handled gracefully then.
> 
> This is a bit confusing, IIRC, currently erofs does not allocate the whole
> sb info in advance, does it?

No, it doesn't. I just mentioned there are 2 ways to pass fs private
contexts to fs_context:
 1) allocate sbi in advance by using fc->s_fs_info;
 2) allocate/pass a fs private fs_context by using fc->fs_private;

erofs uses the second way.... Actually I'm fine with getting rid of
such expression to avoid confusion for the next version...

Thanks,
Gao Xiang

> 
> thanks,
> liubo

^ permalink raw reply	[flat|nested] 5+ messages in thread

* Re: [PATCH v3 2/2] erofs: add multiple device support
  2021-10-06 19:44 ` [PATCH v3 2/2] erofs: add multiple device support Gao Xiang
@ 2021-10-07  6:42   ` Liu Bo
  0 siblings, 0 replies; 5+ messages in thread
From: Liu Bo @ 2021-10-07  6:42 UTC (permalink / raw)
  To: Gao Xiang
  Cc: Yan Song, LKML, Peng Tao, Joseph Qi, Changwei Ge, Liu Jiang, linux-erofs

On Thu, Oct 07, 2021 at 03:44:53AM +0800, Gao Xiang wrote:
> In order to support multi-layer container images, add multiple
> device feature to EROFS. Two ways are available to use for now:
> 
>  - Devices can be mapped into 32-bit global block address space;
>  - Device ID can be specified with the chunk indexes format.
> 
> Note that it assumes no extent would cross device boundary and mkfs
> should take care of it seriously.
> 
> In the future, a dedicated device manager could be introduced then
> so extra devices can be automatically scanned by uuid as well.
>

Reviewed-by: Liu Bo <bo.liu@linux.alibaba.com>

thanks,
liubo

> Cc: Liu Bo <bo.liu@linux.alibaba.com>
> Cc: Chao Yu <chao@kernel.org>
> Signed-off-by: Gao Xiang <hsiangkao@linux.alibaba.com>
> ---
>  Documentation/filesystems/erofs.rst |  12 ++-
>  fs/erofs/data.c                     |  73 ++++++++++---
>  fs/erofs/erofs_fs.h                 |  20 +++-
>  fs/erofs/internal.h                 |  35 ++++++-
>  fs/erofs/super.c                    | 155 ++++++++++++++++++++++++++--
>  fs/erofs/zdata.c                    |  20 +++-
>  6 files changed, 279 insertions(+), 36 deletions(-)
> 
> diff --git a/Documentation/filesystems/erofs.rst b/Documentation/filesystems/erofs.rst
> index b97579b7d8fb..01df283c7d04 100644
> --- a/Documentation/filesystems/erofs.rst
> +++ b/Documentation/filesystems/erofs.rst
> @@ -19,9 +19,10 @@ It is designed as a better filesystem solution for the following scenarios:
>     immutable and bit-for-bit identical to the official golden image for
>     their releases due to security and other considerations and
>  
> - - hope to save some extra storage space with guaranteed end-to-end performance
> -   by using reduced metadata and transparent file compression, especially
> -   for those embedded devices with limited memory (ex, smartphone);
> + - hope to minimize extra storage space with guaranteed end-to-end performance
> +   by using compact layout, transparent file compression and direct access,
> +   especially for those embedded devices with limited memory and high-density
> +   hosts with numerous containers;
>  
>  Here is the main features of EROFS:
>  
> @@ -51,7 +52,9 @@ Here is the main features of EROFS:
>   - Support POSIX.1e ACLs by using xattrs;
>  
>   - Support transparent data compression as an option:
> -   LZ4 algorithm with the fixed-sized output compression for high performance.
> +   LZ4 algorithm with the fixed-sized output compression for high performance;
> +
> + - Multiple device support for multi-layer container images.
>  
>  The following git tree provides the file system user-space tools under
>  development (ex, formatting tool mkfs.erofs):
> @@ -87,6 +90,7 @@ cache_strategy=%s      Select a strategy for cached decompression from now on:
>  dax={always,never}     Use direct access (no page cache).  See
>                         Documentation/filesystems/dax.rst.
>  dax                    A legacy option which is an alias for ``dax=always``.
> +device=%s              Specify a path to an extra device to be used together.
>  ===================    =========================================================
>  
>  On-disk details
> diff --git a/fs/erofs/data.c b/fs/erofs/data.c
> index 9db829715652..808234d9190c 100644
> --- a/fs/erofs/data.c
> +++ b/fs/erofs/data.c
> @@ -89,6 +89,7 @@ static int erofs_map_blocks(struct inode *inode,
>  	erofs_off_t pos;
>  	int err = 0;
>  
> +	map->m_deviceid = 0;
>  	if (map->m_la >= inode->i_size) {
>  		/* leave out-of-bound access unmapped */
>  		map->m_flags = 0;
> @@ -135,14 +136,8 @@ static int erofs_map_blocks(struct inode *inode,
>  		map->m_flags = 0;
>  		break;
>  	default:
> -		/* only one device is supported for now */
> -		if (idx->device_id) {
> -			erofs_err(sb, "invalid device id %u @ %llu for nid %llu",
> -				  le16_to_cpu(idx->device_id),
> -				  chunknr, vi->nid);
> -			err = -EFSCORRUPTED;
> -			goto out_unlock;
> -		}
> +		map->m_deviceid = le16_to_cpu(idx->device_id) &
> +			EROFS_SB(sb)->device_id_mask;
>  		map->m_pa = blknr_to_addr(le32_to_cpu(idx->blkaddr));
>  		map->m_flags = EROFS_MAP_MAPPED;
>  		break;
> @@ -155,11 +150,55 @@ static int erofs_map_blocks(struct inode *inode,
>  	return err;
>  }
>  
> +int erofs_map_dev(struct super_block *sb, struct erofs_map_dev *map)
> +{
> +	struct erofs_dev_context *devs = EROFS_SB(sb)->devs;
> +	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;
> +
> +	if (map->m_deviceid) {
> +		down_read(&devs->rwsem);
> +		dif = idr_find(&devs->tree, map->m_deviceid - 1);
> +		if (!dif) {
> +			up_read(&devs->rwsem);
> +			return -ENODEV;
> +		}
> +		map->m_bdev = dif->bdev;
> +		map->m_daxdev = dif->dax_dev;
> +		up_read(&devs->rwsem);
> +	} else if (devs->extra_devices) {
> +		down_read(&devs->rwsem);
> +		idr_for_each_entry(&devs->tree, dif, id) {
> +			erofs_off_t startoff, length;
> +
> +			if (!dif->mapped_blkaddr)
> +				continue;
> +			startoff = blknr_to_addr(dif->mapped_blkaddr);
> +			length = blknr_to_addr(dif->blocks);
> +
> +			if (map->m_pa >= startoff &&
> +			    map->m_pa < startoff + length) {
> +				map->m_pa -= startoff;
> +				map->m_bdev = dif->bdev;
> +				map->m_daxdev = dif->dax_dev;
> +				break;
> +			}
> +		}
> +		up_read(&devs->rwsem);
> +	}
> +	return 0;
> +}
> +
>  static int erofs_iomap_begin(struct inode *inode, loff_t offset, loff_t length,
>  		unsigned int flags, struct iomap *iomap, struct iomap *srcmap)
>  {
>  	int ret;
>  	struct erofs_map_blocks map;
> +	struct erofs_map_dev mdev;
>  
>  	map.m_la = offset;
>  	map.m_llen = length;
> @@ -168,8 +207,16 @@ static int erofs_iomap_begin(struct inode *inode, loff_t offset, loff_t length,
>  	if (ret < 0)
>  		return ret;
>  
> -	iomap->bdev = inode->i_sb->s_bdev;
> -	iomap->dax_dev = EROFS_I_SB(inode)->dax_dev;
> +	mdev = (struct erofs_map_dev) {
> +		.m_deviceid = map.m_deviceid,
> +		.m_pa = map.m_pa,
> +	};
> +	ret = erofs_map_dev(inode->i_sb, &mdev);
> +	if (ret)
> +		return ret;
> +
> +	iomap->bdev = mdev.m_bdev;
> +	iomap->dax_dev = mdev.m_daxdev;
>  	iomap->offset = map.m_la;
>  	iomap->length = map.m_llen;
>  	iomap->flags = 0;
> @@ -188,15 +235,15 @@ static int erofs_iomap_begin(struct inode *inode, loff_t offset, loff_t length,
>  
>  		iomap->type = IOMAP_INLINE;
>  		ipage = erofs_get_meta_page(inode->i_sb,
> -					    erofs_blknr(map.m_pa));
> +					    erofs_blknr(mdev.m_pa));
>  		if (IS_ERR(ipage))
>  			return PTR_ERR(ipage);
>  		iomap->inline_data = page_address(ipage) +
> -					erofs_blkoff(map.m_pa);
> +					erofs_blkoff(mdev.m_pa);
>  		iomap->private = ipage;
>  	} else {
>  		iomap->type = IOMAP_MAPPED;
> -		iomap->addr = map.m_pa;
> +		iomap->addr = mdev.m_pa;
>  	}
>  	return 0;
>  }
> diff --git a/fs/erofs/erofs_fs.h b/fs/erofs/erofs_fs.h
> index b0b23f41abc3..c2f1c2453a55 100644
> --- a/fs/erofs/erofs_fs.h
> +++ b/fs/erofs/erofs_fs.h
> @@ -21,14 +21,27 @@
>  #define EROFS_FEATURE_INCOMPAT_COMPR_CFGS	0x00000002
>  #define EROFS_FEATURE_INCOMPAT_BIG_PCLUSTER	0x00000002
>  #define EROFS_FEATURE_INCOMPAT_CHUNKED_FILE	0x00000004
> +#define EROFS_FEATURE_INCOMPAT_DEVICE_TABLE	0x00000008
>  #define EROFS_ALL_FEATURE_INCOMPAT		\
>  	(EROFS_FEATURE_INCOMPAT_LZ4_0PADDING | \
>  	 EROFS_FEATURE_INCOMPAT_COMPR_CFGS | \
>  	 EROFS_FEATURE_INCOMPAT_BIG_PCLUSTER | \
> -	 EROFS_FEATURE_INCOMPAT_CHUNKED_FILE)
> +	 EROFS_FEATURE_INCOMPAT_CHUNKED_FILE | \
> +	 EROFS_FEATURE_INCOMPAT_DEVICE_TABLE)
>  
>  #define EROFS_SB_EXTSLOT_SIZE	16
>  
> +struct erofs_deviceslot {
> +	union {
> +		u8 uuid[16];		/* used for device manager later */
> +		u8 userdata[64];	/* digest(sha256), etc. */
> +	} u;
> +	__le32 blocks;
> +	__le32 mapped_blkaddr;
> +	u8 reserved[56];
> +};
> +#define EROFS_DEVT_SLOT_SIZE	sizeof(struct erofs_deviceslot)
> +
>  /* erofs on-disk super block (currently 128 bytes) */
>  struct erofs_super_block {
>  	__le32 magic;           /* file system magic number */
> @@ -54,7 +67,9 @@ struct erofs_super_block {
>  		/* customized sliding window size instead of 64k by default */
>  		__le16 lz4_max_distance;
>  	} __packed u1;
> -	__u8 reserved2[42];
> +	__le16 extra_devices;	/* # of devices besides the primary device */
> +	__le16 devt_slotoff;	/* startoff = devt_slotoff * devt_slotsize */
> +	__u8 reserved2[38];
>  };
>  
>  /*
> @@ -384,6 +399,7 @@ static inline void erofs_check_ondisk_layout_definitions(void)
>  	/* keep in sync between 2 index structures for better extendibility */
>  	BUILD_BUG_ON(sizeof(struct erofs_inode_chunk_index) !=
>  		     sizeof(struct z_erofs_vle_decompressed_index));
> +	BUILD_BUG_ON(sizeof(struct erofs_deviceslot) != 128);
>  
>  	BUILD_BUG_ON(BIT(Z_EROFS_VLE_DI_CLUSTER_TYPE_BITS) <
>  		     Z_EROFS_VLE_CLUSTER_TYPE_MAX - 1);
> diff --git a/fs/erofs/internal.h b/fs/erofs/internal.h
> index b1b9d1b5cb66..34f0933fd0c0 100644
> --- a/fs/erofs/internal.h
> +++ b/fs/erofs/internal.h
> @@ -47,6 +47,15 @@ typedef u64 erofs_off_t;
>  /* data type for filesystem-wide blocks number */
>  typedef u32 erofs_blk_t;
>  
> +struct erofs_device_info {
> +	char *path;
> +	struct block_device *bdev;
> +	struct dax_device *dax_dev;
> +
> +	u32 blocks;
> +	u32 mapped_blkaddr;
> +};
> +
>  struct erofs_mount_opts {
>  #ifdef CONFIG_EROFS_FS_ZIP
>  	/* current strategy of how to use managed cache */
> @@ -60,8 +69,16 @@ struct erofs_mount_opts {
>  	unsigned int mount_opt;
>  };
>  
> +struct erofs_dev_context {
> +	struct idr tree;
> +	struct rw_semaphore rwsem;
> +
> +	unsigned int extra_devices;
> +};
> +
>  struct erofs_fs_context {
>  	struct erofs_mount_opts opt;
> +	struct erofs_dev_context *devs;
>  };
>  
>  /* all filesystem-wide lz4 configurations */
> @@ -74,7 +91,6 @@ struct erofs_sb_lz4_info {
>  
>  struct erofs_sb_info {
>  	struct erofs_mount_opts opt;	/* options */
> -
>  #ifdef CONFIG_EROFS_FS_ZIP
>  	/* list for all registered superblocks, mainly for shrinker */
>  	struct list_head list;
> @@ -91,12 +107,16 @@ struct erofs_sb_info {
>  
>  	struct erofs_sb_lz4_info lz4;
>  #endif	/* CONFIG_EROFS_FS_ZIP */
> +	struct erofs_dev_context *devs;
>  	struct dax_device *dax_dev;
> -	u32 blocks;
> +	u64 total_blocks;
> +	u32 primarydevice_blocks;
> +
>  	u32 meta_blkaddr;
>  #ifdef CONFIG_EROFS_FS_XATTR
>  	u32 xattr_blkaddr;
>  #endif
> +	u16 device_id_mask;
>  
>  	/* inode slot unit size in bit shift */
>  	unsigned char islotbits;
> @@ -241,6 +261,7 @@ static inline bool erofs_sb_has_##name(struct erofs_sb_info *sbi) \
>  EROFS_FEATURE_FUNCS(lz4_0padding, incompat, INCOMPAT_LZ4_0PADDING)
>  EROFS_FEATURE_FUNCS(compr_cfgs, incompat, INCOMPAT_COMPR_CFGS)
>  EROFS_FEATURE_FUNCS(big_pcluster, incompat, INCOMPAT_BIG_PCLUSTER)
> +EROFS_FEATURE_FUNCS(device_table, incompat, INCOMPAT_DEVICE_TABLE)
>  EROFS_FEATURE_FUNCS(sb_chksum, compat, COMPAT_SB_CHKSUM)
>  
>  /* atomic flag definitions */
> @@ -359,6 +380,7 @@ struct erofs_map_blocks {
>  	erofs_off_t m_pa, m_la;
>  	u64 m_plen, m_llen;
>  
> +	unsigned int m_deviceid;
>  	unsigned int m_flags;
>  
>  	struct page *mpage;
> @@ -390,9 +412,18 @@ static inline int z_erofs_map_blocks_iter(struct inode *inode,
>  }
>  #endif	/* !CONFIG_EROFS_FS_ZIP */
>  
> +struct erofs_map_dev {
> +	struct block_device *m_bdev;
> +	struct dax_device *m_daxdev;
> +
> +	erofs_off_t m_pa;
> +	unsigned int m_deviceid;
> +};
> +
>  /* data.c */
>  extern const struct file_operations erofs_file_fops;
>  struct page *erofs_get_meta_page(struct super_block *sb, erofs_blk_t blkaddr);
> +int erofs_map_dev(struct super_block *sb, struct erofs_map_dev *dev);
>  int erofs_fiemap(struct inode *inode, struct fiemap_extent_info *fieinfo,
>  		 u64 start, u64 len);
>  
> diff --git a/fs/erofs/super.c b/fs/erofs/super.c
> index 25f6b8b37f28..f4dd20f76862 100644
> --- a/fs/erofs/super.c
> +++ b/fs/erofs/super.c
> @@ -252,6 +252,79 @@ static int erofs_load_compr_cfgs(struct super_block *sb,
>  }
>  #endif
>  
> +static int erofs_init_devices(struct super_block *sb,
> +			      struct erofs_super_block *dsb)
> +{
> +	struct erofs_sb_info *sbi = EROFS_SB(sb);
> +	unsigned int ondisk_extradevs;
> +	erofs_off_t pos;
> +	struct page *page = NULL;
> +	struct erofs_device_info *dif;
> +	struct erofs_deviceslot *dis;
> +	void *ptr;
> +	int id, err = 0;
> +
> +	sbi->total_blocks = sbi->primarydevice_blocks;
> +	if (!erofs_sb_has_device_table(sbi))
> +		ondisk_extradevs = 0;
> +	else
> +		ondisk_extradevs = le16_to_cpu(dsb->extra_devices);
> +
> +	if (ondisk_extradevs != sbi->devs->extra_devices) {
> +		erofs_err(sb, "extra devices don't match (ondisk %u, given %u)",
> +			  ondisk_extradevs, sbi->devs->extra_devices);
> +		return -EINVAL;
> +	}
> +	if (!ondisk_extradevs)
> +		return 0;
> +
> +	sbi->device_id_mask = roundup_pow_of_two(ondisk_extradevs + 1) - 1;
> +	pos = le16_to_cpu(dsb->devt_slotoff) * EROFS_DEVT_SLOT_SIZE;
> +	down_read(&sbi->devs->rwsem);
> +	idr_for_each_entry(&sbi->devs->tree, dif, id) {
> +		erofs_blk_t blk = erofs_blknr(pos);
> +		struct block_device *bdev;
> +
> +		if (!page || page->index != blk) {
> +			if (page) {
> +				kunmap(page);
> +				unlock_page(page);
> +				put_page(page);
> +			}
> +
> +			page = erofs_get_meta_page(sb, blk);
> +			if (IS_ERR(page)) {
> +				up_read(&sbi->devs->rwsem);
> +				return PTR_ERR(page);
> +			}
> +			ptr = kmap(page);
> +		}
> +		dis = ptr + erofs_blkoff(pos);
> +
> +		bdev = blkdev_get_by_path(dif->path,
> +					  FMODE_READ | FMODE_EXCL,
> +					  sb->s_type);
> +		if (IS_ERR(bdev)) {
> +			err = PTR_ERR(bdev);
> +			goto err_out;
> +		}
> +		dif->bdev = bdev;
> +		dif->dax_dev = fs_dax_get_by_bdev(bdev);
> +		dif->blocks = le32_to_cpu(dis->blocks);
> +		dif->mapped_blkaddr = le32_to_cpu(dis->mapped_blkaddr);
> +		sbi->total_blocks += dif->blocks;
> +		pos += sizeof(*dis);
> +	}
> +err_out:
> +	up_read(&sbi->devs->rwsem);
> +	if (page) {
> +		kunmap(page);
> +		unlock_page(page);
> +		put_page(page);
> +	}
> +	return err;
> +}
> +
>  static int erofs_read_superblock(struct super_block *sb)
>  {
>  	struct erofs_sb_info *sbi;
> @@ -303,7 +376,7 @@ static int erofs_read_superblock(struct super_block *sb)
>  			  sbi->sb_size);
>  		goto out;
>  	}
> -	sbi->blocks = le32_to_cpu(dsb->blocks);
> +	sbi->primarydevice_blocks = le32_to_cpu(dsb->blocks);
>  	sbi->meta_blkaddr = le32_to_cpu(dsb->meta_blkaddr);
>  #ifdef CONFIG_EROFS_FS_XATTR
>  	sbi->xattr_blkaddr = le32_to_cpu(dsb->xattr_blkaddr);
> @@ -330,6 +403,11 @@ static int erofs_read_superblock(struct super_block *sb)
>  		ret = erofs_load_compr_cfgs(sb, dsb);
>  	else
>  		ret = z_erofs_load_lz4_config(sb, dsb, NULL, 0);
> +	if (ret < 0)
> +		goto out;
> +
> +	/* handle multiple devices */
> +	ret = erofs_init_devices(sb, dsb);
>  out:
>  	kunmap(page);
>  	put_page(page);
> @@ -358,6 +436,7 @@ enum {
>  	Opt_cache_strategy,
>  	Opt_dax,
>  	Opt_dax_enum,
> +	Opt_device,
>  	Opt_err
>  };
>  
> @@ -381,6 +460,7 @@ static const struct fs_parameter_spec erofs_fs_parameters[] = {
>  		     erofs_param_cache_strategy),
>  	fsparam_flag("dax",             Opt_dax),
>  	fsparam_enum("dax",		Opt_dax_enum, erofs_dax_param_enums),
> +	fsparam_string("device",	Opt_device),
>  	{}
>  };
>  
> @@ -412,9 +492,10 @@ static bool erofs_fc_set_dax_mode(struct fs_context *fc, unsigned int mode)
>  static int erofs_fc_parse_param(struct fs_context *fc,
>  				struct fs_parameter *param)
>  {
> -	struct erofs_fs_context *ctx __maybe_unused = fc->fs_private;
> +	struct erofs_fs_context *ctx = fc->fs_private;
>  	struct fs_parse_result result;
> -	int opt;
> +	struct erofs_device_info *dif;
> +	int opt, ret;
>  
>  	opt = fs_parse(fc, erofs_fs_parameters, param, &result);
>  	if (opt < 0)
> @@ -456,6 +537,25 @@ static int erofs_fc_parse_param(struct fs_context *fc,
>  		if (!erofs_fc_set_dax_mode(fc, result.uint_32))
>  			return -EINVAL;
>  		break;
> +	case Opt_device:
> +		dif = kzalloc(sizeof(*dif), GFP_KERNEL);
> +		if (!dif)
> +			return -ENOMEM;
> +		dif->path = kstrdup(param->string, GFP_KERNEL);
> +		if (!dif->path) {
> +			kfree(dif);
> +			return -ENOMEM;
> +		}
> +		down_write(&ctx->devs->rwsem);
> +		ret = idr_alloc(&ctx->devs->tree, dif, 0, 0, GFP_KERNEL);
> +		up_write(&ctx->devs->rwsem);
> +		if (ret < 0) {
> +			kfree(dif->path);
> +			kfree(dif);
> +			return ret;
> +		}
> +		++ctx->devs->extra_devices;
> +		break;
>  	default:
>  		return -ENOPARAM;
>  	}
> @@ -542,6 +642,9 @@ static int erofs_fc_fill_super(struct super_block *sb, struct fs_context *fc)
>  	sb->s_fs_info = sbi;
>  	sbi->opt = ctx->opt;
>  	sbi->dax_dev = fs_dax_get_by_bdev(sb->s_bdev);
> +	sbi->devs = ctx->devs;
> +	ctx->devs = NULL;
> +
>  	err = erofs_read_superblock(sb);
>  	if (err)
>  		return err;
> @@ -617,9 +720,32 @@ static int erofs_fc_reconfigure(struct fs_context *fc)
>  	return 0;
>  }
>  
> +static int erofs_release_device_info(int id, void *ptr, void *data)
> +{
> +	struct erofs_device_info *dif = ptr;
> +
> +	fs_put_dax(dif->dax_dev);
> +	if (dif->bdev)
> +		blkdev_put(dif->bdev, FMODE_READ | FMODE_EXCL);
> +	kfree(dif->path);
> +	kfree(dif);
> +	return 0;
> +}
> +
> +static void erofs_free_dev_context(struct erofs_dev_context *devs)
> +{
> +	if (!devs)
> +		return;
> +	idr_for_each(&devs->tree, &erofs_release_device_info, NULL);
> +	idr_destroy(&devs->tree);
> +}
> +
>  static void erofs_fc_free(struct fs_context *fc)
>  {
> -	kfree(fc->fs_private);
> +	struct erofs_fs_context *ctx = fc->fs_private;
> +
> +	erofs_free_dev_context(ctx->devs);
> +	kfree(ctx);
>  }
>  
>  static const struct fs_context_operations erofs_context_ops = {
> @@ -631,13 +757,20 @@ static const struct fs_context_operations erofs_context_ops = {
>  
>  static int erofs_init_fs_context(struct fs_context *fc)
>  {
> -	fc->fs_private = kzalloc(sizeof(struct erofs_fs_context), GFP_KERNEL);
> -	if (!fc->fs_private)
> -		return -ENOMEM;
> +	struct erofs_fs_context *ctx = kzalloc(sizeof(*ctx), GFP_KERNEL);
>  
> -	/* set default mount options */
> -	erofs_default_options(fc->fs_private);
> +	if (!ctx)
> +		return -ENOMEM;
> +	ctx->devs = kzalloc(sizeof(struct erofs_dev_context), GFP_KERNEL);
> +	if (!ctx->devs) {
> +		kfree(ctx);
> +		return -ENOMEM;
> +	}
> +	fc->fs_private = ctx;
>  
> +	idr_init(&ctx->devs->tree);
> +	init_rwsem(&ctx->devs->rwsem);
> +	erofs_default_options(ctx);
>  	fc->ops = &erofs_context_ops;
>  	return 0;
>  }
> @@ -657,6 +790,8 @@ static void erofs_kill_sb(struct super_block *sb)
>  	sbi = EROFS_SB(sb);
>  	if (!sbi)
>  		return;
> +
> +	erofs_free_dev_context(sbi->devs);
>  	fs_put_dax(sbi->dax_dev);
>  	kfree(sbi);
>  	sb->s_fs_info = NULL;
> @@ -746,7 +881,7 @@ static int erofs_statfs(struct dentry *dentry, struct kstatfs *buf)
>  
>  	buf->f_type = sb->s_magic;
>  	buf->f_bsize = EROFS_BLKSIZ;
> -	buf->f_blocks = sbi->blocks;
> +	buf->f_blocks = sbi->total_blocks;
>  	buf->f_bfree = buf->f_bavail = 0;
>  
>  	buf->f_files = ULLONG_MAX;
> diff --git a/fs/erofs/zdata.c b/fs/erofs/zdata.c
> index e59e22852c78..8c947ed49299 100644
> --- a/fs/erofs/zdata.c
> +++ b/fs/erofs/zdata.c
> @@ -1266,8 +1266,9 @@ static void z_erofs_submit_queue(struct super_block *sb,
>  	struct z_erofs_decompressqueue *q[NR_JOBQUEUES];
>  	void *bi_private;
>  	z_erofs_next_pcluster_t owned_head = f->clt.owned_head;
> -	/* since bio will be NULL, no need to initialize last_index */
> +	/* bio is NULL initially, so no need to initialize last_{index,bdev} */
>  	pgoff_t last_index;
> +	struct block_device *last_bdev;
>  	unsigned int nr_bios = 0;
>  	struct bio *bio = NULL;
>  
> @@ -1279,6 +1280,7 @@ static void z_erofs_submit_queue(struct super_block *sb,
>  	q[JQ_SUBMIT]->head = owned_head;
>  
>  	do {
> +		struct erofs_map_dev mdev;
>  		struct z_erofs_pcluster *pcl;
>  		pgoff_t cur, end;
>  		unsigned int i = 0;
> @@ -1290,7 +1292,13 @@ static void z_erofs_submit_queue(struct super_block *sb,
>  
>  		pcl = container_of(owned_head, struct z_erofs_pcluster, next);
>  
> -		cur = pcl->obj.index;
> +		/* no device id here, thus it will always succeed */
> +		mdev = (struct erofs_map_dev) {
> +			.m_pa = blknr_to_addr(pcl->obj.index),
> +		};
> +		(void)erofs_map_dev(sb, &mdev);
> +
> +		cur = erofs_blknr(mdev.m_pa);
>  		end = cur + pcl->pclusterpages;
>  
>  		/* close the main owned chain at first */
> @@ -1306,7 +1314,8 @@ static void z_erofs_submit_queue(struct super_block *sb,
>  			if (!page)
>  				continue;
>  
> -			if (bio && cur != last_index + 1) {
> +			if (bio && (cur != last_index + 1 ||
> +				    last_bdev != mdev.m_bdev)) {
>  submit_bio_retry:
>  				submit_bio(bio);
>  				bio = NULL;
> @@ -1314,9 +1323,10 @@ static void z_erofs_submit_queue(struct super_block *sb,
>  
>  			if (!bio) {
>  				bio = bio_alloc(GFP_NOIO, BIO_MAX_VECS);
> -
>  				bio->bi_end_io = z_erofs_decompressqueue_endio;
> -				bio_set_dev(bio, sb->s_bdev);
> +
> +				bio_set_dev(bio, mdev.m_bdev);
> +				last_bdev = mdev.m_bdev;
>  				bio->bi_iter.bi_sector = (sector_t)cur <<
>  					LOG_SECTORS_PER_BLOCK;
>  				bio->bi_private = bi_private;
> -- 
> 2.24.4

^ permalink raw reply	[flat|nested] 5+ messages in thread

end of thread, other threads:[~2021-10-07  6:42 UTC | newest]

Thread overview: 5+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2021-10-06 19:44 [PATCH v3 1/2] erofs: decouple basic mount options from fs_context Gao Xiang
2021-10-06 19:44 ` [PATCH v3 2/2] erofs: add multiple device support Gao Xiang
2021-10-07  6:42   ` Liu Bo
2021-10-07  6:10 ` [PATCH v3 1/2] erofs: decouple basic mount options from fs_context Liu Bo
2021-10-07  6:29   ` Gao Xiang

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).