ocfs2: Read support for inline data
authorMark Fasheh <mark.fasheh@oracle.com>
Fri, 7 Sep 2007 21:05:51 +0000 (14:05 -0700)
committerMark Fasheh <mark.fasheh@oracle.com>
Fri, 12 Oct 2007 18:54:39 +0000 (11:54 -0700)
This hooks up ocfs2_readpage() to populate a page with data from an inode
block. Direct IO reads from inline data are modified to fall back to
buffered I/O. Appropriate checks are also placed in the extent map code to
avoid reading an extent list when inline data might be stored.

Signed-off-by: Mark Fasheh <mark.fasheh@oracle.com>
Reviewed-by: Joel Becker <joel.becker@oracle.com>
fs/ocfs2/aops.c
fs/ocfs2/extent_map.c

index 8416e38..fef0186 100644 (file)
@@ -206,9 +206,70 @@ bail:
        return err;
 }
 
+static int ocfs2_read_inline_data(struct inode *inode, struct page *page,
+                                 struct buffer_head *di_bh)
+{
+       void *kaddr;
+       unsigned int size;
+       struct ocfs2_dinode *di = (struct ocfs2_dinode *)di_bh->b_data;
+
+       if (!(le16_to_cpu(di->i_dyn_features) & OCFS2_INLINE_DATA_FL)) {
+               ocfs2_error(inode->i_sb, "Inode %llu lost inline data flag",
+                           (unsigned long long)OCFS2_I(inode)->ip_blkno);
+               return -EROFS;
+       }
+
+       size = i_size_read(inode);
+
+       if (size > PAGE_CACHE_SIZE ||
+           size > ocfs2_max_inline_data(inode->i_sb)) {
+               ocfs2_error(inode->i_sb,
+                           "Inode %llu has with inline data has bad size: %u",
+                           (unsigned long long)OCFS2_I(inode)->ip_blkno, size);
+               return -EROFS;
+       }
+
+       kaddr = kmap_atomic(page, KM_USER0);
+       if (size)
+               memcpy(kaddr, di->id2.i_data.id_data, size);
+       /* Clear the remaining part of the page */
+       memset(kaddr + size, 0, PAGE_CACHE_SIZE - size);
+       flush_dcache_page(page);
+       kunmap_atomic(kaddr, KM_USER0);
+
+       SetPageUptodate(page);
+
+       return 0;
+}
+
+static int ocfs2_readpage_inline(struct inode *inode, struct page *page)
+{
+       int ret;
+       struct buffer_head *di_bh = NULL;
+       struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
+
+       BUG_ON(!PageLocked(page));
+       BUG_ON(!OCFS2_I(inode)->ip_dyn_features & OCFS2_INLINE_DATA_FL);
+
+       ret = ocfs2_read_block(osb, OCFS2_I(inode)->ip_blkno, &di_bh,
+                              OCFS2_BH_CACHED, inode);
+       if (ret) {
+               mlog_errno(ret);
+               goto out;
+       }
+
+       ret = ocfs2_read_inline_data(inode, page, di_bh);
+out:
+       unlock_page(page);
+
+       brelse(di_bh);
+       return ret;
+}
+
 static int ocfs2_readpage(struct file *file, struct page *page)
 {
        struct inode *inode = page->mapping->host;
+       struct ocfs2_inode_info *oi = OCFS2_I(inode);
        loff_t start = (loff_t)page->index << PAGE_CACHE_SHIFT;
        int ret, unlock = 1;
 
@@ -222,7 +283,7 @@ static int ocfs2_readpage(struct file *file, struct page *page)
                goto out;
        }
 
-       if (down_read_trylock(&OCFS2_I(inode)->ip_alloc_sem) == 0) {
+       if (down_read_trylock(&oi->ip_alloc_sem) == 0) {
                ret = AOP_TRUNCATED_PAGE;
                goto out_meta_unlock;
        }
@@ -252,7 +313,10 @@ static int ocfs2_readpage(struct file *file, struct page *page)
                goto out_alloc;
        }
 
-       ret = block_read_full_page(page, ocfs2_get_block);
+       if (oi->ip_dyn_features & OCFS2_INLINE_DATA_FL)
+               ret = ocfs2_readpage_inline(inode, page);
+       else
+               ret = block_read_full_page(page, ocfs2_get_block);
        unlock = 0;
 
        ocfs2_data_unlock(inode, 0);
@@ -397,7 +461,9 @@ static sector_t ocfs2_bmap(struct address_space *mapping, sector_t block)
                down_read(&OCFS2_I(inode)->ip_alloc_sem);
        }
 
-       err = ocfs2_extent_map_get_blocks(inode, block, &p_blkno, NULL, NULL);
+       if (!(OCFS2_I(inode)->ip_dyn_features & OCFS2_INLINE_DATA_FL))
+               err = ocfs2_extent_map_get_blocks(inode, block, &p_blkno, NULL,
+                                                 NULL);
 
        if (!INODE_JOURNAL(inode)) {
                up_read(&OCFS2_I(inode)->ip_alloc_sem);
@@ -411,7 +477,6 @@ static sector_t ocfs2_bmap(struct address_space *mapping, sector_t block)
                goto bail;
        }
 
-
 bail:
        status = err ? 0 : p_blkno;
 
@@ -566,6 +631,13 @@ static ssize_t ocfs2_direct_IO(int rw,
 
        mlog_entry_void();
 
+       /*
+        * Fallback to buffered I/O if we see an inode without
+        * extents.
+        */
+       if (OCFS2_I(inode)->ip_dyn_features & OCFS2_INLINE_DATA_FL)
+               return 0;
+
        if (!ocfs2_sparse_alloc(OCFS2_SB(inode->i_sb))) {
                /*
                 * We get PR data locks even for O_DIRECT.  This
index 03c1d36..c58668a 100644 (file)
@@ -387,6 +387,12 @@ int ocfs2_get_clusters(struct inode *inode, u32 v_cluster,
        struct ocfs2_extent_rec *rec;
        u32 coff;
 
+       if (OCFS2_I(inode)->ip_dyn_features & OCFS2_INLINE_DATA_FL) {
+               ret = -ERANGE;
+               mlog_errno(ret);
+               goto out;
+       }
+
        ret = ocfs2_extent_map_lookup(inode, v_cluster, p_cluster,
                                      num_clusters, extent_flags);
        if (ret == 0)