gfs2: use iomap for buffered I/O in ordered and writeback mode

Switch to using the iomap readpage and writepage helpers for all I/O in
the ordered and writeback modes, and thus eliminate using buffer_heads
for I/O in these cases.  The journaled data mode is left untouched.

(Andreas Gruenbacher: In gfs2_unstuffer_page, switch from mark_buffer_dirty
to set_page_dirty instead of accidentally leaving the page / buffer clean.)
Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Andreas Gruenbacher <agruenba@redhat.com>

fs/gfs2/aops.c

@@ -91,22 +91,13 @@ static int gfs2_writepage(struct page *page, struct writeback_control *wbc)
 	struct inode *inode = page->mapping->host;
 	struct gfs2_inode *ip = GFS2_I(inode);
 	struct gfs2_sbd *sdp = GFS2_SB(inode);
-	loff_t i_size = i_size_read(inode);
-	pgoff_t end_index = i_size >> PAGE_SHIFT;
-	unsigned offset;
+	struct iomap_writepage_ctx wpc = { };
 
 	if (gfs2_assert_withdraw(sdp, gfs2_glock_is_held_excl(ip->i_gl)))
 		goto out;
 	if (current->journal_info)
 		goto redirty;
-	/* Is the page fully outside i_size? (truncate in progress) */
-	offset = i_size & (PAGE_SIZE-1);
-	if (page->index > end_index || (page->index == end_index && !offset)) {
-		page->mapping->a_ops->invalidatepage(page, 0, PAGE_SIZE);
-		goto out;
-	}
-
-	return nobh_writepage(page, gfs2_get_block_noalloc, wbc);
+	return iomap_writepage(page, wbc, &wpc, &gfs2_writeback_ops);
 
 redirty:
 	redirty_page_for_writepage(wbc, page);
@@ -208,7 +199,8 @@ static int gfs2_writepages(struct address_space *mapping,
 			   struct writeback_control *wbc)
 {
 	struct gfs2_sbd *sdp = gfs2_mapping2sbd(mapping);
-	int ret = mpage_writepages(mapping, wbc, gfs2_get_block_noalloc);
+	struct iomap_writepage_ctx wpc = { };
+	int ret;
 
 	/*
 	 * Even if we didn't write any pages here, we might still be holding
@@ -216,9 +208,9 @@ static int gfs2_writepages(struct address_space *mapping,
 	 * want balance_dirty_pages() to loop indefinitely trying to write out
 	 * pages held in the ail that it can't find.
 	 */
+	ret = iomap_writepages(mapping, wbc, &wpc, &gfs2_writeback_ops);
 	if (ret == 0)
 		set_bit(SDF_FORCE_AIL_FLUSH, &sdp->sd_flags);
-
 	return ret;
 }
 
@@ -470,12 +462,13 @@ static int stuffed_readpage(struct gfs2_inode *ip, struct page *page)
 
 static int __gfs2_readpage(void *file, struct page *page)
 {
-	struct gfs2_inode *ip = GFS2_I(page->mapping->host);
-	struct gfs2_sbd *sdp = GFS2_SB(page->mapping->host);
+	struct inode *inode = page->mapping->host;
+	struct gfs2_inode *ip = GFS2_I(inode);
+	struct gfs2_sbd *sdp = GFS2_SB(inode);
 	int error;
 
-	if (i_blocksize(page->mapping->host) == PAGE_SIZE &&
-	    !page_has_buffers(page)) {
+	if (!gfs2_is_jdata(ip) ||
+	    (i_blocksize(inode) == PAGE_SIZE && !page_has_buffers(page))) {
 		error = iomap_readpage(page, &gfs2_iomap_ops);
 	} else if (gfs2_is_stuffed(ip)) {
 		error = stuffed_readpage(ip, page);
@@ -563,8 +556,12 @@ static void gfs2_readahead(struct readahead_control *rac)
 	struct inode *inode = rac->mapping->host;
 	struct gfs2_inode *ip = GFS2_I(inode);
 
-	if (!gfs2_is_stuffed(ip))
+	if (gfs2_is_stuffed(ip))
+		;
+	else if (gfs2_is_jdata(ip))
 		mpage_readahead(rac, gfs2_block_map);
+	else
+		iomap_readahead(rac, &gfs2_iomap_ops);
 }
 
 /**
@@ -784,12 +781,13 @@ static const struct address_space_operations gfs2_aops = {
 	.writepages = gfs2_writepages,
 	.readpage = gfs2_readpage,
 	.readahead = gfs2_readahead,
+	.set_page_dirty = iomap_set_page_dirty,
+	.releasepage = iomap_releasepage,
+	.invalidatepage = iomap_invalidatepage,
 	.bmap = gfs2_bmap,
-	.invalidatepage = gfs2_invalidatepage,
-	.releasepage = gfs2_releasepage,
 	.direct_IO = noop_direct_IO,
-	.migratepage = buffer_migrate_page,
-	.is_partially_uptodate = block_is_partially_uptodate,
+	.migratepage = iomap_migrate_page,
+	.is_partially_uptodate = iomap_is_partially_uptodate,
 	.error_remove_page = generic_error_remove_page,
 };
 

fs/gfs2/bmap.c

@@ -56,7 +56,6 @@ static int gfs2_unstuffer_page(struct gfs2_inode *ip, struct buffer_head *dibh,
 			       u64 block, struct page *page)
 {
 	struct inode *inode = &ip->i_inode;
-	struct buffer_head *bh;
 	int release = 0;
 
 	if (!page || page->index) {
@@ -80,20 +79,21 @@ static int gfs2_unstuffer_page(struct gfs2_inode *ip, struct buffer_head *dibh,
 		SetPageUptodate(page);
 	}
 
+	if (gfs2_is_jdata(ip)) {
+		struct buffer_head *bh;
+
 		if (!page_has_buffers(page))
 			create_empty_buffers(page, BIT(inode->i_blkbits),
 					     BIT(BH_Uptodate));
 
 		bh = page_buffers(page);
-
 		if (!buffer_mapped(bh))
 			map_bh(bh, inode->i_sb, block);
 
 		set_buffer_uptodate(bh);
-	if (gfs2_is_jdata(ip))
 		gfs2_trans_add_data(ip->i_gl, bh);
-	else {
-		mark_buffer_dirty(bh);
+	} else {
+		set_page_dirty(page);
 		gfs2_ordered_add_inode(ip);
 	}
 
@@ -1158,6 +1158,7 @@ static int gfs2_iomap_begin(struct inode *inode, loff_t pos, loff_t length,
 	struct metapath mp = { .mp_aheight = 1, };
 	int ret;
 
+	if (gfs2_is_jdata(ip))
 		iomap->flags |= IOMAP_F_BUFFER_HEAD;
 
 	trace_gfs2_iomap_start(ip, pos, length, flags);
@@ -2518,3 +2519,26 @@ int __gfs2_punch_hole(struct file *file, loff_t offset, loff_t length)
 		gfs2_trans_end(sdp);
 	return error;
 }
+
+static int gfs2_map_blocks(struct iomap_writepage_ctx *wpc, struct inode *inode,
+		loff_t offset)
+{
+	struct metapath mp = { .mp_aheight = 1, };
+	int ret;
+
+	if (WARN_ON_ONCE(gfs2_is_stuffed(GFS2_I(inode))))
+		return -EIO;
+
+	if (offset >= wpc->iomap.offset &&
+	    offset < wpc->iomap.offset + wpc->iomap.length)
+		return 0;
+
+	memset(&wpc->iomap, 0, sizeof(wpc->iomap));
+	ret = gfs2_iomap_get(inode, offset, INT_MAX, 0, &wpc->iomap, &mp);
+	release_metapath(&mp);
+	return ret;
+}
+
+const struct iomap_writeback_ops gfs2_writeback_ops = {
+	.map_blocks		= gfs2_map_blocks,
+};

fs/gfs2/bmap.h

@@ -44,6 +44,7 @@ static inline void gfs2_write_calc_reserv(const struct gfs2_inode *ip,
 }
 
 extern const struct iomap_ops gfs2_iomap_ops;
+extern const struct iomap_writeback_ops gfs2_writeback_ops;
 
 extern int gfs2_unstuff_dinode(struct gfs2_inode *ip, struct page *page);
 extern int gfs2_block_map(struct inode *inode, sector_t lblock,