erofs: support compressed inodes over fscache

Since fscache can utilize iov_iter to write dest buffers, bio_vec can
be used in this way too.

To simplify this, pseudo bios are prepared and bio_vec will be filled
with bio_add_page().  And a common .bi_end_io will be called directly
to handle I/O completions.

Signed-off-by: Jingbo Xu <jefflexu@linux.alibaba.com>
Reviewed-by: Gao Xiang <hsiangkao@linux.alibaba.com>
Link: https://lore.kernel.org/r/20240308094159.40547-2-jefflexu@linux.alibaba.com
Signed-off-by: Gao Xiang <hsiangkao@linux.alibaba.com>

diff --git a/fs/erofs/fscache.c b/fs/erofs/fscache.c
index b9fb4c4da39c154a2b201875c133f615eb8201c2..8aff1a724805a67ad3733fbce2127a39e563e01d 100644 (file)
--- a/fs/erofs/fscache.c
+++ b/fs/erofs/fscache.c
@@ -174,6 +174,53 @@ static int erofs_fscache_read_io_async(struct fscache_cookie *cookie,
        return 0;
 }
 
+struct erofs_fscache_bio {
+       struct erofs_fscache_io io;
+       struct bio bio;         /* w/o bdev to share bio_add_page/endio() */
+       struct bio_vec bvecs[BIO_MAX_VECS];
+};
+
+static void erofs_fscache_bio_endio(void *priv,
+               ssize_t transferred_or_error, bool was_async)
+{
+       struct erofs_fscache_bio *io = priv;
+
+       if (IS_ERR_VALUE(transferred_or_error))
+               io->bio.bi_status = errno_to_blk_status(transferred_or_error);
+       io->bio.bi_end_io(&io->bio);
+       BUILD_BUG_ON(offsetof(struct erofs_fscache_bio, io) != 0);
+       erofs_fscache_io_put(&io->io);
+}
+
+struct bio *erofs_fscache_bio_alloc(struct erofs_map_dev *mdev)
+{
+       struct erofs_fscache_bio *io;
+
+       io = kmalloc(sizeof(*io), GFP_KERNEL | __GFP_NOFAIL);
+       bio_init(&io->bio, NULL, io->bvecs, BIO_MAX_VECS, REQ_OP_READ);
+       io->io.private = mdev->m_fscache->cookie;
+       io->io.end_io = erofs_fscache_bio_endio;
+       refcount_set(&io->io.ref, 1);
+       return &io->bio;
+}
+
+void erofs_fscache_submit_bio(struct bio *bio)
+{
+       struct erofs_fscache_bio *io = container_of(bio,
+                       struct erofs_fscache_bio, bio);
+       int ret;
+
+       iov_iter_bvec(&io->io.iter, ITER_DEST, io->bvecs, bio->bi_vcnt,
+                     bio->bi_iter.bi_size);
+       ret = erofs_fscache_read_io_async(io->io.private,
+                               bio->bi_iter.bi_sector << 9, &io->io);
+       erofs_fscache_io_put(&io->io);
+       if (!ret)
+               return;
+       bio->bi_status = errno_to_blk_status(ret);
+       bio->bi_end_io(bio);
+}
+
 static int erofs_fscache_meta_read_folio(struct file *data, struct folio *folio)
 {
        struct erofs_fscache *ctx = folio->mapping->host->i_private;

diff --git a/fs/erofs/inode.c b/fs/erofs/inode.c
index 36e638e8b53a3d290fcb7ade23a40dc4805be9e6..0eb0e6f933c34b9b03eb69d3bc8d04e743596f43 100644 (file)
--- a/fs/erofs/inode.c
+++ b/fs/erofs/inode.c
@@ -259,14 +259,12 @@ static int erofs_fill_inode(struct inode *inode)
 
        if (erofs_inode_is_data_compressed(vi->datalayout)) {
 #ifdef CONFIG_EROFS_FS_ZIP
-               if (!erofs_is_fscache_mode(inode->i_sb)) {
-                       DO_ONCE_LITE_IF(inode->i_sb->s_blocksize != PAGE_SIZE,
-                                 erofs_info, inode->i_sb,
-                                 "EXPERIMENTAL EROFS subpage compressed block support in use. Use at your own risk!");
-                       inode->i_mapping->a_ops = &z_erofs_aops;
-                       err = 0;
-                       goto out_unlock;
-               }
+               DO_ONCE_LITE_IF(inode->i_blkbits != PAGE_SHIFT,
+                         erofs_info, inode->i_sb,
+                         "EXPERIMENTAL EROFS subpage compressed block support in use. Use at your own risk!");
+               inode->i_mapping->a_ops = &z_erofs_aops;
+               err = 0;
+               goto out_unlock;
 #endif
                err = -EOPNOTSUPP;
                goto out_unlock;

diff --git a/fs/erofs/internal.h b/fs/erofs/internal.h
index f8623113be6c16bbb59286a6579b4dc3256fad6f..92120bf4ab71a4c988cb6a6750457d86ab8d9e4a 100644 (file)
--- a/fs/erofs/internal.h
+++ b/fs/erofs/internal.h
@@ -512,6 +512,8 @@ void erofs_fscache_unregister_fs(struct super_block *sb);
 struct erofs_fscache *erofs_fscache_register_cookie(struct super_block *sb,
                                        char *name, unsigned int flags);
 void erofs_fscache_unregister_cookie(struct erofs_fscache *fscache);
+struct bio *erofs_fscache_bio_alloc(struct erofs_map_dev *mdev);
+void erofs_fscache_submit_bio(struct bio *bio);
 #else
 static inline int erofs_fscache_register_fs(struct super_block *sb)
 {
@@ -529,6 +531,8 @@ struct erofs_fscache *erofs_fscache_register_cookie(struct super_block *sb,
 static inline void erofs_fscache_unregister_cookie(struct erofs_fscache *fscache)
 {
 }
+static inline struct bio *erofs_fscache_bio_alloc(struct erofs_map_dev *mdev) { return NULL; }
+static inline void erofs_fscache_submit_bio(struct bio *bio) {}
 #endif
 
 #define EFSCORRUPTED    EUCLEAN         /* Filesystem is corrupted */

diff --git a/fs/erofs/zdata.c b/fs/erofs/zdata.c
index c1bd4d8392ebceb770eebaf3266b860b937e14bb..3216b920d369c84d22c425781af84fa74f89ad6f 100644 (file)
--- a/fs/erofs/zdata.c
+++ b/fs/erofs/zdata.c
@@ -1561,7 +1561,7 @@ static void move_to_bypass_jobqueue(struct z_erofs_pcluster *pcl,
        qtail[JQ_BYPASS] = &pcl->next;
 }
 
-static void z_erofs_submissionqueue_endio(struct bio *bio)
+static void z_erofs_endio(struct bio *bio)
 {
        struct z_erofs_decompressqueue *q = bio->bi_private;
        blk_status_t err = bio->bi_status;
@@ -1582,7 +1582,8 @@ static void z_erofs_submissionqueue_endio(struct bio *bio)
        if (err)
                q->eio = true;
        z_erofs_decompress_kickoff(q, -1);
-       bio_put(bio);
+       if (bio->bi_bdev)
+               bio_put(bio);
 }
 
 static void z_erofs_submit_queue(struct z_erofs_decompress_frontend *f,
@@ -1596,7 +1597,6 @@ static void z_erofs_submit_queue(struct z_erofs_decompress_frontend *f,
        z_erofs_next_pcluster_t owned_head = f->owned_head;
        /* bio is NULL initially, so no need to initialize last_{index,bdev} */
        erofs_off_t last_pa;
-       struct block_device *last_bdev;
        unsigned int nr_bios = 0;
        struct bio *bio = NULL;
        unsigned long pflags;
@@ -1643,9 +1643,13 @@ static void z_erofs_submit_queue(struct z_erofs_decompress_frontend *f,
                                continue;
 
                        if (bio && (cur != last_pa ||
-                                   last_bdev != mdev.m_bdev)) {
-submit_bio_retry:
-                               submit_bio(bio);
+                                   bio->bi_bdev != mdev.m_bdev)) {
+io_retry:
+                               if (!erofs_is_fscache_mode(sb))
+                                       submit_bio(bio);
+                               else
+                                       erofs_fscache_submit_bio(bio);
+
                                if (memstall) {
                                        psi_memstall_leave(&pflags);
                                        memstall = 0;
@@ -1660,15 +1664,16 @@ submit_bio_retry:
                        }
 
                        if (!bio) {
-                               bio = bio_alloc(mdev.m_bdev, BIO_MAX_VECS,
-                                               REQ_OP_READ, GFP_NOIO);
-                               bio->bi_end_io = z_erofs_submissionqueue_endio;
+                               bio = erofs_is_fscache_mode(sb) ?
+                                       erofs_fscache_bio_alloc(&mdev) :
+                                       bio_alloc(mdev.m_bdev, BIO_MAX_VECS,
+                                                 REQ_OP_READ, GFP_NOIO);
+                               bio->bi_end_io = z_erofs_endio;
                                bio->bi_iter.bi_sector = cur >> 9;
                                bio->bi_private = q[JQ_SUBMIT];
                                if (readahead)
                                        bio->bi_opf |= REQ_RAHEAD;
                                ++nr_bios;
-                               last_bdev = mdev.m_bdev;
                        }
 
                        if (cur + bvec.bv_len > end)
@@ -1676,7 +1681,7 @@ submit_bio_retry:
                        DBG_BUGON(bvec.bv_len < sb->s_blocksize);
                        if (!bio_add_page(bio, bvec.bv_page, bvec.bv_len,
                                          bvec.bv_offset))
-                               goto submit_bio_retry;
+                               goto io_retry;
 
                        last_pa = cur + bvec.bv_len;
                        bypass = false;
@@ -1689,7 +1694,10 @@ submit_bio_retry:
        } while (owned_head != Z_EROFS_PCLUSTER_TAIL);
 
        if (bio) {
-               submit_bio(bio);
+               if (!erofs_is_fscache_mode(sb))
+                       submit_bio(bio);
+               else
+                       erofs_fscache_submit_bio(bio);
                if (memstall)
                        psi_memstall_leave(&pflags);
        }