Merge commit 'v2.6.36-rc1' into kbuild/rc-fixes

5 files changed, 87 insertions, 131 deletions

diff --git a/fs/exofs/exofs.h b/fs/exofs/exofs.h
index 22721b2..2dc925f 100644
--- a/fs/exofs/exofs.h
+++ b/fs/exofs/exofs.h
@@ -256,7 +256,6 @@ static inline int exofs_oi_read(struct exofs_i_info *oi,
 }
 
 /* inode.c               */
-void exofs_truncate(struct inode *inode);
 int exofs_setattr(struct dentry *, struct iattr *);
 int exofs_write_begin(struct file *file, struct address_space *mapping,
 		loff_t pos, unsigned len, unsigned flags,
@@ -264,7 +263,7 @@ int exofs_write_begin(struct file *file, struct address_space *mapping,
 extern struct inode *exofs_iget(struct super_block *, unsigned long);
 struct inode *exofs_new_inode(struct inode *, int);
 extern int exofs_write_inode(struct inode *, struct writeback_control *wbc);
-extern void exofs_delete_inode(struct inode *);
+extern void exofs_evict_inode(struct inode *);
 
 /* dir.c:                */
 int exofs_add_link(struct dentry *, struct inode *);
diff --git a/fs/exofs/file.c b/fs/exofs/file.c
index fef6899..68cb23e 100644
--- a/fs/exofs/file.c
+++ b/fs/exofs/file.c
@@ -30,9 +30,6 @@
  * along with exofs; if not, write to the Free Software
  * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
  */
-
-#include <linux/buffer_head.h>
-
 #include "exofs.h"
 
 static int exofs_release_file(struct inode *inode, struct file *filp)
@@ -40,19 +37,27 @@ static int exofs_release_file(struct inode *inode, struct file *filp)
 	return 0;
 }
 
+/* exofs_file_fsync - flush the inode to disk
+ *
+ *   Note, in exofs all metadata is written as part of inode, regardless.
+ *   The writeout is synchronous
+ */
 static int exofs_file_fsync(struct file *filp, int datasync)
 {
 	int ret;
-	struct address_space *mapping = filp->f_mapping;
-	struct inode *inode = mapping->host;
+	struct inode *inode = filp->f_mapping->host;
+	struct writeback_control wbc = {
+		.sync_mode = WB_SYNC_ALL,
+		.nr_to_write = 0, /* metadata-only; caller takes care of data */
+	};
 	struct super_block *sb;
 
-	ret = filemap_write_and_wait(mapping);
-	if (ret)
-		return ret;
+	if (!(inode->i_state & I_DIRTY))
+		return 0;
+	if (datasync && !(inode->i_state & I_DIRTY_DATASYNC))
+		return 0;
 
-	/* sync the inode attributes */
-	ret = write_inode_now(inode, 1);
+	ret = sync_inode(inode, &wbc);
 
 	/* This is a good place to write the sb */
 	/* TODO: Sechedule an sb-sync on create */
@@ -65,9 +70,9 @@ static int exofs_file_fsync(struct file *filp, int datasync)
 
 static int exofs_flush(struct file *file, fl_owner_t id)
 {
-	exofs_file_fsync(file, 1);
+	int ret = vfs_fsync(file, 0);
 	/* TODO: Flush the OSD target */
-	return 0;
+	return ret;
 }
 
 const struct file_operations exofs_file_operations = {
@@ -86,6 +91,5 @@ const struct file_operations exofs_file_operations = {
 };
 
 const struct inode_operations exofs_file_inode_operations = {
-	.truncate	= exofs_truncate,
 	.setattr	= exofs_setattr,
 };
diff --git a/fs/exofs/inode.c b/fs/exofs/inode.c
index 4bb6ef8..eb7368e 100644
--- a/fs/exofs/inode.c
+++ b/fs/exofs/inode.c
@@ -32,9 +32,6 @@
  */
 
 #include <linux/slab.h>
-#include <linux/writeback.h>
-#include <linux/buffer_head.h>
-#include <scsi/scsi_device.h>
 
 #include "exofs.h"
 
@@ -697,6 +694,13 @@ static int exofs_writepage(struct page *page, struct writeback_control *wbc)
 	return write_exec(&pcol);
 }
 
+/* i_mutex held using inode->i_size directly */
+static void _write_failed(struct inode *inode, loff_t to)
+{
+	if (to > inode->i_size)
+		truncate_pagecache(inode, to, inode->i_size);
+}
+
 int exofs_write_begin(struct file *file, struct address_space *mapping,
 		loff_t pos, unsigned len, unsigned flags,
 		struct page **pagep, void **fsdata)
@@ -710,7 +714,7 @@ int exofs_write_begin(struct file *file, struct address_space *mapping,
 					 fsdata);
 		if (ret) {
 			EXOFS_DBGMSG("simple_write_begin faild\n");
-			return ret;
+			goto out;
 		}
 
 		page = *pagep;
@@ -725,6 +729,9 @@ int exofs_write_begin(struct file *file, struct address_space *mapping,
 			EXOFS_DBGMSG("__readpage_filler faild\n");
 		}
 	}
+out:
+	if (unlikely(ret))
+		_write_failed(mapping->host, pos + len);
 
 	return ret;
 }
@@ -750,6 +757,10 @@ static int exofs_write_end(struct file *file, struct address_space *mapping,
 	int ret;
 
 	ret = simple_write_end(file, mapping,pos, len, copied, page, fsdata);
+	if (unlikely(ret))
+		_write_failed(inode, pos + len);
+
+	/* TODO: once simple_write_end marks inode dirty remove */
 	if (i_size != inode->i_size)
 		mark_inode_dirty(inode);
 	return ret;
@@ -759,15 +770,13 @@ static int exofs_releasepage(struct page *page, gfp_t gfp)
 {
 	EXOFS_DBGMSG("page 0x%lx\n", page->index);
 	WARN_ON(1);
-	return try_to_free_buffers(page);
+	return 0;
 }
 
 static void exofs_invalidatepage(struct page *page, unsigned long offset)
 {
-	EXOFS_DBGMSG("page_has_buffers=>%d\n", page_has_buffers(page));
+	EXOFS_DBGMSG("page 0x%lx offset 0x%lx\n", page->index, offset);
 	WARN_ON(1);
-
-	block_invalidatepage(page, offset);
 }
 
 const struct address_space_operations exofs_aops = {
@@ -808,87 +817,55 @@ static inline int exofs_inode_is_fast_symlink(struct inode *inode)
 	return S_ISLNK(inode->i_mode) && (oi->i_data[0] != 0);
 }
 
-/*
- * get_block_t - Fill in a buffer_head
- * An OSD takes care of block allocation so we just fake an allocation by
- * putting in the inode's sector_t in the buffer_head.
- * TODO: What about the case of create==0 and @iblock does not exist in the
- * object?
- */
-static int exofs_get_block(struct inode *inode, sector_t iblock,
-		    struct buffer_head *bh_result, int create)
-{
-	map_bh(bh_result, inode->i_sb, iblock);
-	return 0;
-}
-
 const struct osd_attr g_attr_logical_length = ATTR_DEF(
 	OSD_APAGE_OBJECT_INFORMATION, OSD_ATTR_OI_LOGICAL_LENGTH, 8);
 
-static int _do_truncate(struct inode *inode)
+static int _do_truncate(struct inode *inode, loff_t newsize)
 {
 	struct exofs_i_info *oi = exofs_i(inode);
-	loff_t isize = i_size_read(inode);
 	int ret;
 
 	inode->i_mtime = inode->i_ctime = CURRENT_TIME;
 
-	nobh_truncate_page(inode->i_mapping, isize, exofs_get_block);
+	ret = exofs_oi_truncate(oi, (u64)newsize);
+	if (likely(!ret))
+		truncate_setsize(inode, newsize);
 
-	ret = exofs_oi_truncate(oi, (u64)isize);
-	EXOFS_DBGMSG("(0x%lx) size=0x%llx\n", inode->i_ino, isize);
+	EXOFS_DBGMSG("(0x%lx) size=0x%llx ret=>%d\n",
+		     inode->i_ino, newsize, ret);
 	return ret;
 }
 
 /*
- * Truncate a file to the specified size - all we have to do is set the size
- * attribute.  We make sure the object exists first.
- */
-void exofs_truncate(struct inode *inode)
-{
-	struct exofs_i_info *oi = exofs_i(inode);
-	int ret;
-
-	if (!(S_ISREG(inode->i_mode) || S_ISDIR(inode->i_mode)
-	     || S_ISLNK(inode->i_mode)))
-		return;
-	if (exofs_inode_is_fast_symlink(inode))
-		return;
-	if (IS_APPEND(inode) || IS_IMMUTABLE(inode))
-		return;
-
-	/* if we are about to truncate an object, and it hasn't been
-	 * created yet, wait
-	 */
-	if (unlikely(wait_obj_created(oi)))
-		goto fail;
-
-	ret = _do_truncate(inode);
-	if (ret)
-		goto fail;
-
-out:
-	mark_inode_dirty(inode);
-	return;
-fail:
-	make_bad_inode(inode);
-	goto out;
-}
-
-/*
- * Set inode attributes - just call generic functions.
+ * Set inode attributes - update size attribute on OSD if needed,
+ *                        otherwise just call generic functions.
  */
 int exofs_setattr(struct dentry *dentry, struct iattr *iattr)
 {
 	struct inode *inode = dentry->d_inode;
 	int error;
 
+	/* if we are about to modify an object, and it hasn't been
+	 * created yet, wait
+	 */
+	error = wait_obj_created(exofs_i(inode));
+	if (unlikely(error))
+		return error;
+
 	error = inode_change_ok(inode, iattr);
-	if (error)
+	if (unlikely(error))
 		return error;
 
-	error = inode_setattr(inode, iattr);
-	return error;
+	if ((iattr->ia_valid & ATTR_SIZE) &&
+	    iattr->ia_size != i_size_read(inode)) {
+		error = _do_truncate(inode, iattr->ia_size);
+		if (unlikely(error))
+			return error;
+	}
+
+	setattr_copy(inode, iattr);
+	mark_inode_dirty(inode);
+	return 0;
 }
 
 static const struct osd_attr g_attr_inode_file_layout = ATTR_DEF(
@@ -1325,7 +1302,7 @@ static void delete_done(struct exofs_io_state *ios, void *p)
  * from the OSD here.  We make sure the object was created before we try and
  * delete it.
  */
-void exofs_delete_inode(struct inode *inode)
+void exofs_evict_inode(struct inode *inode)
 {
 	struct exofs_i_info *oi = exofs_i(inode);
 	struct super_block *sb = inode->i_sb;
@@ -1335,30 +1312,27 @@ void exofs_delete_inode(struct inode *inode)
 
 	truncate_inode_pages(&inode->i_data, 0);
 
-	if (is_bad_inode(inode))
+	/* TODO: should do better here */
+	if (inode->i_nlink || is_bad_inode(inode))
 		goto no_delete;
 
-	mark_inode_dirty(inode);
-	exofs_update_inode(inode, inode_needs_sync(inode));
-
 	inode->i_size = 0;
-	if (inode->i_blocks)
-		exofs_truncate(inode);
+	end_writeback(inode);
 
-	clear_inode(inode);
+	/* if we are deleting an obj that hasn't been created yet, wait */
+	if (!obj_created(oi)) {
+		BUG_ON(!obj_2bcreated(oi));
+		wait_event(oi->i_wq, obj_created(oi));
+		/* ignore the error attempt a remove anyway */
+	}
 
+	/* Now Remove the OSD objects */
 	ret = exofs_get_io_state(&sbi->layout, &ios);
 	if (unlikely(ret)) {
 		EXOFS_ERR("%s: exofs_get_io_state failed\n", __func__);
 		return;
 	}
 
-	/* if we are deleting an obj that hasn't been created yet, wait */
-	if (!obj_created(oi)) {
-		BUG_ON(!obj_2bcreated(oi));
-		wait_event(oi->i_wq, obj_created(oi));
-	}
-
 	ios->obj.id = exofs_oi_objno(oi);
 	ios->done = delete_done;
 	ios->private = sbi;
@@ -1374,5 +1348,5 @@ void exofs_delete_inode(struct inode *inode)
 	return;
 
 no_delete:
-	clear_inode(inode);
+	end_writeback(inode);
 }
diff --git a/fs/exofs/ios.c b/fs/exofs/ios.c
index 4337cad..6550bf7 100644
--- a/fs/exofs/ios.c
+++ b/fs/exofs/ios.c
@@ -305,8 +305,6 @@ int exofs_check_io(struct exofs_io_state *ios, u64 *resid)
 struct _striping_info {
 	u64 obj_offset;
 	u64 group_length;
-	u64 total_group_length;
-	u64 Major;
 	unsigned dev;
 	unsigned unit_off;
 };
@@ -343,8 +341,6 @@ static void _calc_stripe_info(struct exofs_io_state *ios, u64 file_offset,
 				  (M * group_depth * stripe_unit);
 
 	si->group_length = T - H;
-	si->total_group_length = T;
-	si->Major = M;
 }
 
 static int _add_stripe_unit(struct exofs_io_state *ios,  unsigned *cur_pg,
@@ -392,20 +388,19 @@ static int _add_stripe_unit(struct exofs_io_state *ios,  unsigned *cur_pg,
 }
 
 static int _prepare_one_group(struct exofs_io_state *ios, u64 length,
-			      struct _striping_info *si, unsigned first_comp)
+			      struct _striping_info *si)
 {
 	unsigned stripe_unit = ios->layout->stripe_unit;
 	unsigned mirrors_p1 = ios->layout->mirrors_p1;
 	unsigned devs_in_group = ios->layout->group_width * mirrors_p1;
 	unsigned dev = si->dev;
 	unsigned first_dev = dev - (dev % devs_in_group);
-	unsigned comp = first_comp + (dev - first_dev);
 	unsigned max_comp = ios->numdevs ? ios->numdevs - mirrors_p1 : 0;
 	unsigned cur_pg = ios->pages_consumed;
 	int ret = 0;
 
 	while (length) {
-		struct exofs_per_dev_state *per_dev = &ios->per_dev[comp];
+		struct exofs_per_dev_state *per_dev = &ios->per_dev[dev];
 		unsigned cur_len, page_off = 0;
 
 		if (!per_dev->length) {
@@ -424,11 +419,8 @@ static int _prepare_one_group(struct exofs_io_state *ios, u64 length,
 				cur_len = stripe_unit;
 			}
 
-			if (max_comp < comp)
-				max_comp = comp;
-
-			dev += mirrors_p1;
-			dev = (dev % devs_in_group) + first_dev;
+			if (max_comp < dev)
+				max_comp = dev;
 		} else {
 			cur_len = stripe_unit;
 		}
@@ -440,8 +432,8 @@ static int _prepare_one_group(struct exofs_io_state *ios, u64 length,
 		if (unlikely(ret))
 			goto out;
 
-		comp += mirrors_p1;
-		comp = (comp % devs_in_group) + first_comp;
+		dev += mirrors_p1;
+		dev = (dev % devs_in_group) + first_dev;
 
 		length -= cur_len;
 	}
@@ -454,18 +446,15 @@ out:
 static int _prepare_for_striping(struct exofs_io_state *ios)
 {
 	u64 length = ios->length;
+	u64 offset = ios->offset;
 	struct _striping_info si;
-	unsigned devs_in_group = ios->layout->group_width *
-				 ios->layout->mirrors_p1;
-	unsigned first_comp = 0;
 	int ret = 0;
 
-	_calc_stripe_info(ios, ios->offset, &si);
-
 	if (!ios->pages) {
 		if (ios->kern_buff) {
 			struct exofs_per_dev_state *per_dev = &ios->per_dev[0];
 
+			_calc_stripe_info(ios, ios->offset, &si);
 			per_dev->offset = si.obj_offset;
 			per_dev->dev = si.dev;
 
@@ -479,26 +468,17 @@ static int _prepare_for_striping(struct exofs_io_state *ios)
 	}
 
 	while (length) {
+		_calc_stripe_info(ios, offset, &si);
+
 		if (length < si.group_length)
 			si.group_length = length;
 
-		ret = _prepare_one_group(ios, si.group_length, &si, first_comp);
+		ret = _prepare_one_group(ios, si.group_length, &si);
 		if (unlikely(ret))
 			goto out;
 
+		offset += si.group_length;
 		length -= si.group_length;
-
-		si.group_length = si.total_group_length;
-		si.unit_off = 0;
-		++si.Major;
-		si.obj_offset = si.Major * ios->layout->stripe_unit *
-						ios->layout->group_depth;
-
-		si.dev = (si.dev - (si.dev % devs_in_group)) + devs_in_group;
-		si.dev %= ios->layout->s_numdevs;
-
-		first_comp += devs_in_group;
-		first_comp %= ios->layout->s_numdevs;
 	}
 
 out:
@@ -599,7 +579,7 @@ static int _sbi_write_mirror(struct exofs_io_state *ios, int cur_comp)
 			} else {
 				bio = master_dev->bio;
 				/* FIXME: bio_set_dir() */
-				bio->bi_rw |= (1 << BIO_RW);
+				bio->bi_rw |= REQ_WRITE;
 			}
 
 			osd_req_write(or, &ios->obj, per_dev->offset, bio,
diff --git a/fs/exofs/super.c b/fs/exofs/super.c
index 03149b9..047e92f 100644
--- a/fs/exofs/super.c
+++ b/fs/exofs/super.c
@@ -31,7 +31,6 @@
  * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
  */
 
-#include <linux/smp_lock.h>
 #include <linux/string.h>
 #include <linux/parser.h>
 #include <linux/vfs.h>
@@ -743,7 +742,7 @@ static const struct super_operations exofs_sops = {
 	.alloc_inode    = exofs_alloc_inode,
 	.destroy_inode  = exofs_destroy_inode,
 	.write_inode    = exofs_write_inode,
-	.delete_inode   = exofs_delete_inode,
+	.evict_inode    = exofs_evict_inode,
 	.put_super      = exofs_put_super,
 	.write_super    = exofs_write_super,
 	.sync_fs	= exofs_sync_fs,