Merge branch 'for-linus' of git://git.open-osd.org/linux-open-osd

* 'for-linus' of git://git.open-osd.org/linux-open-osd: (21 commits)
  ore: Enable RAID5 mounts
  exofs: Support for RAID5 read-4-write interface.
  ore: RAID5 Write
  ore: RAID5 read
  fs/Makefile: Always inspect exofs/
  ore: Make ore_calc_stripe_info EXPORT_SYMBOL
  ore/exofs: Change ore_check_io API
  ore/exofs: Define new ore_verify_layout
  ore: Support for partial component table
  ore: Support for short read/writes
  exofs: Support for short read/writes
  ore: Remove check for ios->kern_buff in _prepare_for_striping to later
  ore: cleanup: Embed an ore_striping_info inside ore_io_state
  ore: Only IO one group at a time (API change)
  ore/exofs: Change the type of the devices array (API change)
  ore: Make ore_striping_info and ore_calc_stripe_info public
  exofs: Remove unused data_map member from exofs_sb_info
  exofs: Rename struct ore_components comps => oc
  exofs/super.c: local functions should be static
  exofs/ore.c: local functions should be static
  ...

9 files changed, 1506 insertions, 367 deletions

diff --git a/fs/Makefile b/fs/Makefile
index afc109691a9..5c30a13341e 100644
--- a/fs/Makefile
+++ b/fs/Makefile
@@ -120,6 +120,6 @@ obj-$(CONFIG_DEBUG_FS)		+= debugfs/
 obj-$(CONFIG_OCFS2_FS)		+= ocfs2/
 obj-$(CONFIG_BTRFS_FS)		+= btrfs/
 obj-$(CONFIG_GFS2_FS)           += gfs2/
-obj-$(CONFIG_EXOFS_FS)          += exofs/
+obj-$(y)          		+= exofs/ # Multiple mods, used by nfs/objlayout
 obj-$(CONFIG_CEPH_FS)		+= ceph/
 obj-$(CONFIG_PSTORE)		+= pstore/
diff --git a/fs/exofs/Kbuild b/fs/exofs/Kbuild
index c5a5855a6c4..352ba149d23 100644
--- a/fs/exofs/Kbuild
+++ b/fs/exofs/Kbuild
@@ -13,7 +13,8 @@
 #
 
 # ore module library
-obj-$(CONFIG_ORE) += ore.o
+libore-y := ore.o ore_raid.o
+obj-$(CONFIG_ORE) += libore.o
 
 exofs-y := inode.o file.o symlink.o namei.o dir.o super.o
 obj-$(CONFIG_EXOFS_FS) += exofs.o
diff --git a/fs/exofs/Kconfig b/fs/exofs/Kconfig
index 70bae414929..fa9a286c877 100644
--- a/fs/exofs/Kconfig
+++ b/fs/exofs/Kconfig
@@ -1,10 +1,17 @@
+# Note ORE needs to "select ASYNC_XOR". So Not to force multiple selects
+# for every ORE user we do it like this. Any user should add itself here
+# at the "depends on EXOFS_FS || ..." with an ||. The dependencies are
+# selected here, and we default to "ON". So in effect it is like been
+# selected by any of the users.
 config ORE
 	tristate
+	depends on EXOFS_FS
+	select ASYNC_XOR
+	default SCSI_OSD_ULD
 
 config EXOFS_FS
 	tristate "exofs: OSD based file system support"
 	depends on SCSI_OSD_ULD
-	select ORE
 	help
 	  EXOFS is a file system that uses an OSD storage device,
 	  as its backing storage.
diff --git a/fs/exofs/exofs.h b/fs/exofs/exofs.h
index f4e442ec744..51f4b4c40f0 100644
--- a/fs/exofs/exofs.h
+++ b/fs/exofs/exofs.h
@@ -53,6 +53,10 @@
 /* u64 has problems with printk this will cast it to unsigned long long */
 #define _LLU(x) (unsigned long long)(x)
 
+struct exofs_dev {
+	struct ore_dev ored;
+	unsigned did;
+};
 /*
  * our extension to the in-memory superblock
  */
@@ -66,13 +70,9 @@ struct exofs_sb_info {
 	u32		s_next_generation;	/* next gen # to use          */
 	atomic_t	s_curr_pending;		/* number of pending commands */
 
-	struct pnfs_osd_data_map data_map;	/* Default raid to use
-						 * FIXME: Needed ?
-						 */
 	struct ore_layout	layout;		/* Default files layout       */
 	struct ore_comp one_comp;		/* id & cred of partition id=0*/
-	struct ore_components comps;		/* comps for the partition    */
-	struct osd_dev	*_min_one_dev[1];	/* Place holder for one dev   */
+	struct ore_components oc;		/* comps for the partition    */
 };
 
 /*
@@ -86,7 +86,7 @@ struct exofs_i_info {
 	uint32_t       i_dir_start_lookup; /* which page to start lookup      */
 	uint64_t       i_commit_size;      /* the object's written length     */
 	struct ore_comp one_comp;	   /* same component for all devices  */
-	struct ore_components comps;	   /* inode view of the device table  */
+	struct ore_components oc;	   /* inode view of the device table  */
 };
 
 static inline osd_id exofs_oi_objno(struct exofs_i_info *oi)
@@ -207,7 +207,7 @@ extern const struct inode_operations exofs_fast_symlink_inode_operations;
  * bigger and that the device table repeats twice.
  * See: exofs_read_lookup_dev_table()
  */
-static inline void exofs_init_comps(struct ore_components *comps,
+static inline void exofs_init_comps(struct ore_components *oc,
 				    struct ore_comp *one_comp,
 				    struct exofs_sb_info *sbi, osd_id oid)
 {
@@ -217,13 +217,15 @@ static inline void exofs_init_comps(struct ore_components *comps,
 	one_comp->obj.id = oid;
 	exofs_make_credential(one_comp->cred, &one_comp->obj);
 
-	comps->numdevs = sbi->comps.numdevs;
-	comps->single_comp = EC_SINGLE_COMP;
-	comps->comps = one_comp;
+	oc->first_dev = 0;
+	oc->numdevs = sbi->layout.group_width * sbi->layout.mirrors_p1 *
+							sbi->layout.group_count;
+	oc->single_comp = EC_SINGLE_COMP;
+	oc->comps = one_comp;
 
 	/* Round robin device view of the table */
-	first_dev = (dev_mod * sbi->layout.mirrors_p1) % sbi->comps.numdevs;
-	comps->ods = sbi->comps.ods + first_dev;
+	first_dev = (dev_mod * sbi->layout.mirrors_p1) % sbi->oc.numdevs;
+	oc->ods = &sbi->oc.ods[first_dev];
 }
 
 #endif
diff --git a/fs/exofs/inode.c b/fs/exofs/inode.c
index f39a38fc234..3e5f3a6be90 100644
--- a/fs/exofs/inode.c
+++ b/fs/exofs/inode.c
@@ -37,11 +37,7 @@
 
 #define EXOFS_DBGMSG2(M...) do {} while (0)
 
-enum { BIO_MAX_PAGES_KMALLOC =
-		(PAGE_SIZE - sizeof(struct bio)) / sizeof(struct bio_vec),
-	MAX_PAGES_KMALLOC =
-		PAGE_SIZE / sizeof(struct page *),
-};
+enum {MAX_PAGES_KMALLOC = PAGE_SIZE / sizeof(struct page *), };
 
 unsigned exofs_max_io_pages(struct ore_layout *layout,
 			    unsigned expected_pages)
@@ -49,8 +45,7 @@ unsigned exofs_max_io_pages(struct ore_layout *layout,
 	unsigned pages = min_t(unsigned, expected_pages, MAX_PAGES_KMALLOC);
 
 	/* TODO: easily support bio chaining */
-	pages =  min_t(unsigned, pages,
-		       layout->group_width * BIO_MAX_PAGES_KMALLOC);
+	pages =  min_t(unsigned, pages, layout->max_io_length / PAGE_SIZE);
 	return pages;
 }
 
@@ -68,6 +63,7 @@ struct page_collect {
 	bool read_4_write; /* This means two things: that the read is sync
 			    * And the pages should not be unlocked.
 			    */
+	struct page *that_locked_page;
 };
 
 static void _pcol_init(struct page_collect *pcol, unsigned expected_pages,
@@ -86,6 +82,7 @@ static void _pcol_init(struct page_collect *pcol, unsigned expected_pages,
 	pcol->length = 0;
 	pcol->pg_first = -1;
 	pcol->read_4_write = false;
+	pcol->that_locked_page = NULL;
 }
 
 static void _pcol_reset(struct page_collect *pcol)
@@ -98,6 +95,7 @@ static void _pcol_reset(struct page_collect *pcol)
 	pcol->length = 0;
 	pcol->pg_first = -1;
 	pcol->ios = NULL;
+	pcol->that_locked_page = NULL;
 
 	/* this is probably the end of the loop but in writes
 	 * it might not end here. don't be left with nothing
@@ -149,14 +147,17 @@ static int pcol_add_page(struct page_collect *pcol, struct page *page,
 	return 0;
 }
 
+enum {PAGE_WAS_NOT_IN_IO = 17};
 static int update_read_page(struct page *page, int ret)
 {
-	if (ret == 0) {
+	switch (ret) {
+	case 0:
 		/* Everything is OK */
 		SetPageUptodate(page);
 		if (PageError(page))
 			ClearPageError(page);
-	} else if (ret == -EFAULT) {
+		break;
+	case -EFAULT:
 		/* In this case we were trying to read something that wasn't on
 		 * disk yet - return a page full of zeroes.  This should be OK,
 		 * because the object should be empty (if there was a write
@@ -167,16 +168,22 @@ static int update_read_page(struct page *page, int ret)
 		SetPageUptodate(page);
 		if (PageError(page))
 			ClearPageError(page);
-		ret = 0; /* recovered error */
 		EXOFS_DBGMSG("recovered read error\n");
-	} else /* Error */
+		/* fall through */
+	case PAGE_WAS_NOT_IN_IO:
+		ret = 0; /* recovered error */
+		break;
+	default:
 		SetPageError(page);
-
+	}
 	return ret;
 }
 
 static void update_write_page(struct page *page, int ret)
 {
+	if (unlikely(ret == PAGE_WAS_NOT_IN_IO))
+		return; /* don't pass start don't collect $200 */
+
 	if (ret) {
 		mapping_set_error(page->mapping, ret);
 		SetPageError(page);
@@ -190,15 +197,16 @@ static void update_write_page(struct page *page, int ret)
 static int __readpages_done(struct page_collect *pcol)
 {
 	int i;
-	u64 resid;
 	u64 good_bytes;
 	u64 length = 0;
-	int ret = ore_check_io(pcol->ios, &resid);
+	int ret = ore_check_io(pcol->ios, NULL);
 
-	if (likely(!ret))
+	if (likely(!ret)) {
 		good_bytes = pcol->length;
-	else
-		good_bytes = pcol->length - resid;
+		ret = PAGE_WAS_NOT_IN_IO;
+	} else {
+		good_bytes = 0;
+	}
 
 	EXOFS_DBGMSG2("readpages_done(0x%lx) good_bytes=0x%llx"
 		     " length=0x%lx nr_pages=%u\n",
@@ -259,6 +267,46 @@ static void _unlock_pcol_pages(struct page_collect *pcol, int ret, int rw)
 	}
 }
 
+static int _maybe_not_all_in_one_io(struct ore_io_state *ios,
+	struct page_collect *pcol_src, struct page_collect *pcol)
+{
+	/* length was wrong or offset was not page aligned */
+	BUG_ON(pcol_src->nr_pages < ios->nr_pages);
+
+	if (pcol_src->nr_pages > ios->nr_pages) {
+		struct page **src_page;
+		unsigned pages_less = pcol_src->nr_pages - ios->nr_pages;
+		unsigned long len_less = pcol_src->length - ios->length;
+		unsigned i;
+		int ret;
+
+		/* This IO was trimmed */
+		pcol_src->nr_pages = ios->nr_pages;
+		pcol_src->length = ios->length;
+
+		/* Left over pages are passed to the next io */
+		pcol->expected_pages += pages_less;
+		pcol->nr_pages = pages_less;
+		pcol->length = len_less;
+		src_page = pcol_src->pages + pcol_src->nr_pages;
+		pcol->pg_first = (*src_page)->index;
+
+		ret = pcol_try_alloc(pcol);
+		if (unlikely(ret))
+			return ret;
+
+		for (i = 0; i < pages_less; ++i)
+			pcol->pages[i] = *src_page++;
+
+		EXOFS_DBGMSG("Length was adjusted nr_pages=0x%x "
+			"pages_less=0x%x expected_pages=0x%x "
+			"next_offset=0x%llx next_len=0x%lx\n",
+			pcol_src->nr_pages, pages_less, pcol->expected_pages,
+			pcol->pg_first * PAGE_SIZE, pcol->length);
+	}
+	return 0;
+}
+
 static int read_exec(struct page_collect *pcol)
 {
 	struct exofs_i_info *oi = exofs_i(pcol->inode);
@@ -270,7 +318,7 @@ static int read_exec(struct page_collect *pcol)
 		return 0;
 
 	if (!pcol->ios) {
-		int ret = ore_get_rw_state(&pcol->sbi->layout, &oi->comps, true,
+		int ret = ore_get_rw_state(&pcol->sbi->layout, &oi->oc, true,
 					     pcol->pg_first << PAGE_CACHE_SHIFT,
 					     pcol->length, &pcol->ios);
 
@@ -280,7 +328,6 @@ static int read_exec(struct page_collect *pcol)
 
 	ios = pcol->ios;
 	ios->pages = pcol->pages;
-	ios->nr_pages = pcol->nr_pages;
 
 	if (pcol->read_4_write) {
 		ore_read(pcol->ios);
@@ -296,17 +343,23 @@ static int read_exec(struct page_collect *pcol)
 	*pcol_copy = *pcol;
 	ios->done = readpages_done;
 	ios->private = pcol_copy;
+
+	/* pages ownership was passed to pcol_copy */
+	_pcol_reset(pcol);
+
+	ret = _maybe_not_all_in_one_io(ios, pcol_copy, pcol);
+	if (unlikely(ret))
+		goto err;
+
+	EXOFS_DBGMSG2("read_exec(0x%lx) offset=0x%llx length=0x%llx\n",
+		pcol->inode->i_ino, _LLU(ios->offset), _LLU(ios->length));
+
 	ret = ore_read(ios);
 	if (unlikely(ret))
 		goto err;
 
 	atomic_inc(&pcol->sbi->s_curr_pending);
 
-	EXOFS_DBGMSG2("read_exec obj=0x%llx start=0x%llx length=0x%lx\n",
-		  oi->one_comp.obj.id, _LLU(ios->offset), pcol->length);
-
-	/* pages ownership was passed to pcol_copy */
-	_pcol_reset(pcol);
 	return 0;
 
 err:
@@ -341,6 +394,8 @@ static int readpage_strip(void *data, struct page *page)
 		EXOFS_ERR("PageUptodate(0x%lx, 0x%lx)\n", pcol->inode->i_ino,
 			  page->index);
 
+	pcol->that_locked_page = page;
+
 	if (page->index < end_index)
 		len = PAGE_CACHE_SIZE;
 	else if (page->index == end_index)
@@ -429,6 +484,10 @@ static int exofs_readpages(struct file *file, struct address_space *mapping,
 		return ret;
 	}
 
+	ret = read_exec(&pcol);
+	if (unlikely(ret))
+		return ret;
+
 	return read_exec(&pcol);
 }
 
@@ -462,17 +521,18 @@ static void writepages_done(struct ore_io_state *ios, void *p)
 {
 	struct page_collect *pcol = p;
 	int i;
-	u64 resid;
 	u64  good_bytes;
 	u64  length = 0;
-	int ret = ore_check_io(ios, &resid);
+	int ret = ore_check_io(ios, NULL);
 
 	atomic_dec(&pcol->sbi->s_curr_pending);
 
-	if (likely(!ret))
+	if (likely(!ret)) {
 		good_bytes = pcol->length;
-	else
-		good_bytes = pcol->length - resid;
+		ret = PAGE_WAS_NOT_IN_IO;
+	} else {
+		good_bytes = 0;
+	}
 
 	EXOFS_DBGMSG2("writepages_done(0x%lx) good_bytes=0x%llx"
 		     " length=0x%lx nr_pages=%u\n",
@@ -505,6 +565,56 @@ static void writepages_done(struct ore_io_state *ios, void *p)
 	EXOFS_DBGMSG2("writepages_done END\n");
 }
 
+static struct page *__r4w_get_page(void *priv, u64 offset, bool *uptodate)
+{
+	struct page_collect *pcol = priv;
+	pgoff_t index = offset / PAGE_SIZE;
+
+	if (!pcol->that_locked_page ||
+	    (pcol->that_locked_page->index != index)) {
+		struct page *page = find_get_page(pcol->inode->i_mapping, index);
+
+		if (!page) {
+			page = find_or_create_page(pcol->inode->i_mapping,
+						   index, GFP_NOFS);
+			if (unlikely(!page)) {
+				EXOFS_DBGMSG("grab_cache_page Failed "
+					"index=0x%llx\n", _LLU(index));
+				return NULL;
+			}
+			unlock_page(page);
+		}
+		if (PageDirty(page) || PageWriteback(page))
+			*uptodate = true;
+		else
+			*uptodate = PageUptodate(page);
+		EXOFS_DBGMSG("index=0x%lx uptodate=%d\n", index, *uptodate);
+		return page;
+	} else {
+		EXOFS_DBGMSG("YES that_locked_page index=0x%lx\n",
+			     pcol->that_locked_page->index);
+		*uptodate = true;
+		return pcol->that_locked_page;
+	}
+}
+
+static void __r4w_put_page(void *priv, struct page *page)
+{
+	struct page_collect *pcol = priv;
+
+	if (pcol->that_locked_page != page) {
+		EXOFS_DBGMSG("index=0x%lx\n", page->index);
+		page_cache_release(page);
+		return;
+	}
+	EXOFS_DBGMSG("that_locked_page index=0x%lx\n", page->index);
+}
+
+static const struct _ore_r4w_op _r4w_op = {
+	.get_page = &__r4w_get_page,
+	.put_page = &__r4w_put_page,
+};
+
 static int write_exec(struct page_collect *pcol)
 {
 	struct exofs_i_info *oi = exofs_i(pcol->inode);
@@ -516,10 +626,9 @@ static int write_exec(struct page_collect *pcol)
 		return 0;
 
 	BUG_ON(pcol->ios);
-	ret = ore_get_rw_state(&pcol->sbi->layout, &oi->comps, false,
+	ret = ore_get_rw_state(&pcol->sbi->layout, &oi->oc, false,
 				 pcol->pg_first << PAGE_CACHE_SHIFT,
 				 pcol->length, &pcol->ios);
-
 	if (unlikely(ret))
 		goto err;
 
@@ -534,10 +643,20 @@ static int write_exec(struct page_collect *pcol)
 
 	ios = pcol->ios;
 	ios->pages = pcol_copy->pages;
-	ios->nr_pages = pcol_copy->nr_pages;
 	ios->done = writepages_done;
+	ios->r4w = &_r4w_op;
 	ios->private = pcol_copy;
 
+	/* pages ownership was passed to pcol_copy */
+	_pcol_reset(pcol);
+
+	ret = _maybe_not_all_in_one_io(ios, pcol_copy, pcol);
+	if (unlikely(ret))
+		goto err;
+
+	EXOFS_DBGMSG2("write_exec(0x%lx) offset=0x%llx length=0x%llx\n",
+		pcol->inode->i_ino, _LLU(ios->offset), _LLU(ios->length));
+
 	ret = ore_write(ios);
 	if (unlikely(ret)) {
 		EXOFS_ERR("write_exec: ore_write() Failed\n");
@@ -545,11 +664,6 @@ static int write_exec(struct page_collect *pcol)
 	}
 
 	atomic_inc(&pcol->sbi->s_curr_pending);
-	EXOFS_DBGMSG2("write_exec(0x%lx, 0x%llx) start=0x%llx length=0x%lx\n",
-		  pcol->inode->i_ino, pcol->pg_first, _LLU(ios->offset),
-		  pcol->length);
-	/* pages ownership was passed to pcol_copy */
-	_pcol_reset(pcol);
 	return 0;
 
 err:
@@ -689,14 +803,33 @@ static int exofs_writepages(struct address_space *mapping,
 	_pcol_init(&pcol, expected_pages, mapping->host);
 
 	ret = write_cache_pages(mapping, wbc, writepage_strip, &pcol);
-	if (ret) {
+	if (unlikely(ret)) {
 		EXOFS_ERR("write_cache_pages => %d\n", ret);
 		return ret;
 	}
 
-	return write_exec(&pcol);
+	ret = write_exec(&pcol);
+	if (unlikely(ret))
+		return ret;
+
+	if (wbc->sync_mode == WB_SYNC_ALL) {
+		return write_exec(&pcol); /* pump the last reminder */
+	} else if (pcol.nr_pages) {
+		/* not SYNC let the reminder join the next writeout */
+		unsigned i;
+
+		for (i = 0; i < pcol.nr_pages; i++) {
+			struct page *page = pcol.pages[i];
+
+			end_page_writeback(page);
+			set_page_dirty(page);
+			unlock_page(page);
+		}
+	}
+	return 0;
 }
 
+/*
 static int exofs_writepage(struct page *page, struct writeback_control *wbc)
 {
 	struct page_collect pcol;
@@ -712,7 +845,7 @@ 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)
 {
@@ -818,7 +951,7 @@ static void exofs_invalidatepage(struct page *page, unsigned long offset)
 const struct address_space_operations exofs_aops = {
 	.readpage	= exofs_readpage,
 	.readpages	= exofs_readpages,
-	.writepage	= exofs_writepage,
+	.writepage	= NULL,
 	.writepages	= exofs_writepages,
 	.write_begin	= exofs_write_begin_export,
 	.write_end	= exofs_write_end,
@@ -860,7 +993,7 @@ static int _do_truncate(struct inode *inode, loff_t newsize)
 
 	inode->i_mtime = inode->i_ctime = CURRENT_TIME;
 
-	ret = ore_truncate(&sbi->layout, &oi->comps, (u64)newsize);
+	ret = ore_truncate(&sbi->layout, &oi->oc, (u64)newsize);
 	if (likely(!ret))
 		truncate_setsize(inode, newsize);
 
@@ -927,14 +1060,14 @@ static int exofs_get_inode(struct super_block *sb, struct exofs_i_info *oi,
 	struct exofs_on_disk_inode_layout *layout;
 	int ret;
 
-	ret = ore_get_io_state(&sbi->layout, &oi->comps, &ios);
+	ret = ore_get_io_state(&sbi->layout, &oi->oc, &ios);
 	if (unlikely(ret)) {
 		EXOFS_ERR("%s: ore_get_io_state failed.\n", __func__);
 		return ret;
 	}
 
-	attrs[1].len = exofs_on_disk_inode_layout_size(sbi->comps.numdevs);
-	attrs[2].len = exofs_on_disk_inode_layout_size(sbi->comps.numdevs);
+	attrs[1].len = exofs_on_disk_inode_layout_size(sbi->oc.numdevs);
+	attrs[2].len = exofs_on_disk_inode_layout_size(sbi->oc.numdevs);
 
 	ios->in_attr = attrs;
 	ios->in_attr_len = ARRAY_SIZE(attrs);
@@ -1018,7 +1151,7 @@ struct inode *exofs_iget(struct super_block *sb, unsigned long ino)
 		return inode;
 	oi = exofs_i(inode);
 	__oi_init(oi);
-	exofs_init_comps(&oi->comps, &oi->one_comp, sb->s_fs_info,
+	exofs_init_comps(&oi->oc, &oi->one_comp, sb->s_fs_info,
 			 exofs_oi_objno(oi));
 
 	/* read the inode from the osd */
@@ -1172,13 +1305,13 @@ struct inode *exofs_new_inode(struct inode *dir, int mode)
 	spin_unlock(&sbi->s_next_gen_lock);
 	insert_inode_hash(inode);
 
-	exofs_init_comps(&oi->comps, &oi->one_comp, sb->s_fs_info,
+	exofs_init_comps(&oi->oc, &oi->one_comp, sb->s_fs_info,
 			 exofs_oi_objno(oi));
 	exofs_sbi_write_stats(sbi); /* Make sure new sbi->s_nextid is on disk */
 
 	mark_inode_dirty(inode);
 
-	ret = ore_get_io_state(&sbi->layout, &oi->comps, &ios);
+	ret = ore_get_io_state(&sbi->layout, &oi->oc, &ios);
 	if (unlikely(ret)) {
 		EXOFS_ERR("exofs_new_inode: ore_get_io_state failed\n");
 		return ERR_PTR(ret);
@@ -1267,7 +1400,7 @@ static int exofs_update_inode(struct inode *inode, int do_sync)
 	} else
 		memcpy(fcb->i_data, oi->i_data, sizeof(fcb->i_data));
 
-	ret = ore_get_io_state(&sbi->layout, &oi->comps, &ios);
+	ret = ore_get_io_state(&sbi->layout, &oi->oc, &ios);
 	if (unlikely(ret)) {
 		EXOFS_ERR("%s: ore_get_io_state failed.\n", __func__);
 		goto free_args;
@@ -1350,7 +1483,7 @@ void exofs_evict_inode(struct inode *inode)
 	/* ignore the error, attempt a remove anyway */
 
 	/* Now Remove the OSD objects */
-	ret = ore_get_io_state(&sbi->layout, &oi->comps, &ios);
+	ret = ore_get_io_state(&sbi->layout, &oi->oc, &ios);
 	if (unlikely(ret)) {
 		EXOFS_ERR("%s: ore_get_io_state failed\n", __func__);
 		return;
diff --git a/fs/exofs/ore.c b/fs/exofs/ore.c
index 25305af8819..fcfa86ae6fa 100644
--- a/fs/exofs/ore.c
+++ b/fs/exofs/ore.c
@@ -24,76 +24,287 @@
 
 #include <linux/slab.h>
 #include <asm/div64.h>
+#include <linux/lcm.h>
 
-#include <scsi/osd_ore.h>
+#include "ore_raid.h"
 
-#define ORE_ERR(fmt, a...) printk(KERN_ERR "ore: " fmt, ##a)
+MODULE_AUTHOR("Boaz Harrosh <bharrosh@panasas.com>");
+MODULE_DESCRIPTION("Objects Raid Engine ore.ko");
+MODULE_LICENSE("GPL");
+
+/* ore_verify_layout does a couple of things:
+ * 1. Given a minimum number of needed parameters fixes up the rest of the
+ *    members to be operatonals for the ore. The needed parameters are those
+ *    that are defined by the pnfs-objects layout STD.
+ * 2. Check to see if the current ore code actually supports these parameters
+ *    for example stripe_unit must be a multple of the system PAGE_SIZE,
+ *    and etc...
+ * 3. Cache some havily used calculations that will be needed by users.
+ */
+
+enum { BIO_MAX_PAGES_KMALLOC =
+		(PAGE_SIZE - sizeof(struct bio)) / sizeof(struct bio_vec),};
 
-#ifdef CONFIG_EXOFS_DEBUG
-#define ORE_DBGMSG(fmt, a...) \
-	printk(KERN_NOTICE "ore @%s:%d: " fmt, __func__, __LINE__, ##a)
-#else
-#define ORE_DBGMSG(fmt, a...) \
-	do { if (0) printk(fmt, ##a); } while (0)
-#endif
+int ore_verify_layout(unsigned total_comps, struct ore_layout *layout)
+{
+	u64 stripe_length;
+
+	switch (layout->raid_algorithm) {
+	case PNFS_OSD_RAID_0:
+		layout->parity = 0;
+		break;
+	case PNFS_OSD_RAID_5:
+		layout->parity = 1;
+		break;
+	case PNFS_OSD_RAID_PQ:
+	case PNFS_OSD_RAID_4:
+	default:
+		ORE_ERR("Only RAID_0/5 for now\n");
+		return -EINVAL;
+	}
+	if (0 != (layout->stripe_unit & ~PAGE_MASK)) {
+		ORE_ERR("Stripe Unit(0x%llx)"
+			  " must be Multples of PAGE_SIZE(0x%lx)\n",
+			  _LLU(layout->stripe_unit), PAGE_SIZE);
+		return -EINVAL;
+	}
+	if (layout->group_width) {
+		if (!layout->group_depth) {
+			ORE_ERR("group_depth == 0 && group_width != 0\n");
+			return -EINVAL;
+		}
+		if (total_comps < (layout->group_width * layout->mirrors_p1)) {
+			ORE_ERR("Data Map wrong, "
+				"numdevs=%d < group_width=%d * mirrors=%d\n",
+				total_comps, layout->group_width,
+				layout->mirrors_p1);
+			return -EINVAL;
+		}
+		layout->group_count = total_comps / layout->mirrors_p1 /
+						layout->group_width;
+	} else {
+		if (layout->group_depth) {
+			printk(KERN_NOTICE "Warning: group_depth ignored "
+				"group_width == 0 && group_depth == %lld\n",
+				_LLU(layout->group_depth));
+		}
+		layout->group_width = total_comps / layout->mirrors_p1;
+		layout->group_depth = -1;
+		layout->group_count = 1;
+	}
 
-/* u64 has problems with printk this will cast it to unsigned long long */
-#define _LLU(x) (unsigned long long)(x)
+	stripe_length = (u64)layout->group_width * layout->stripe_unit;
+	if (stripe_length >= (1ULL << 32)) {
+		ORE_ERR("Stripe_length(0x%llx) >= 32bit is not supported\n",
+			_LLU(stripe_length));
+		return -EINVAL;
+	}
 
-#define ORE_DBGMSG2(M...) do {} while (0)
-/* #define ORE_DBGMSG2 ORE_DBGMSG */
+	layout->max_io_length =
+		(BIO_MAX_PAGES_KMALLOC * PAGE_SIZE - layout->stripe_unit) *
+							layout->group_width;
+	if (layout->parity) {
+		unsigned stripe_length =
+				(layout->group_width - layout->parity) *
+				layout->stripe_unit;
 
-MODULE_AUTHOR("Boaz Harrosh <bharrosh@panasas.com>");
-MODULE_DESCRIPTION("Objects Raid Engine ore.ko");
-MODULE_LICENSE("GPL");
+		layout->max_io_length /= stripe_length;
+		layout->max_io_length *= stripe_length;
+	}
+	return 0;
+}
+EXPORT_SYMBOL(ore_verify_layout);
 
 static u8 *_ios_cred(struct ore_io_state *ios, unsigned index)
 {
-	return ios->comps->comps[index & ios->comps->single_comp].cred;
+	return ios->oc->comps[index & ios->oc->single_comp].cred;
 }
 
 static struct osd_obj_id *_ios_obj(struct ore_io_state *ios, unsigned index)
 {
-	return &ios->comps->comps[index & ios->comps->single_comp].obj;
+	return &ios->oc->comps[index & ios->oc->single_comp].obj;
 }
 
 static struct osd_dev *_ios_od(struct ore_io_state *ios, unsigned index)
 {
-	return ios->comps->ods[index];
+	ORE_DBGMSG2("oc->first_dev=%d oc->numdevs=%d i=%d oc->ods=%p\n",
+		    ios->oc->first_dev, ios->oc->numdevs, index,
+		    ios->oc->ods);
+
+	return ore_comp_dev(ios->oc, index);
 }
 
-int  ore_get_rw_state(struct ore_layout *layout, struct ore_components *comps,
+int  _ore_get_io_state(struct ore_layout *layout,
+			struct ore_components *oc, unsigned numdevs,
+			unsigned sgs_per_dev, unsigned num_par_pages,
+			struct ore_io_state **pios)
+{
+	struct ore_io_state *ios;
+	struct page **pages;
+	struct osd_sg_entry *sgilist;
+	struct __alloc_all_io_state {
+		struct ore_io_state ios;
+		struct ore_per_dev_state per_dev[numdevs];
+		union {
+			struct osd_sg_entry sglist[sgs_per_dev * numdevs];
+			struct page *pages[num_par_pages];
+		};
+	} *_aios;
+
+	if (likely(sizeof(*_aios) <= PAGE_SIZE)) {
+		_aios = kzalloc(sizeof(*_aios), GFP_KERNEL);
+		if (unlikely(!_aios)) {
+			ORE_DBGMSG("Failed kzalloc bytes=%zd\n",
+				   sizeof(*_aios));
+			*pios = NULL;
+			return -ENOMEM;
+		}
+		pages = num_par_pages ? _aios->pages : NULL;
+		sgilist = sgs_per_dev ? _aios->sglist : NULL;
+		ios = &_aios->ios;
+	} else {
+		struct __alloc_small_io_state {
+			struct ore_io_state ios;
+			struct ore_per_dev_state per_dev[numdevs];
+		} *_aio_small;
+		union __extra_part {
+			struct osd_sg_entry sglist[sgs_per_dev * numdevs];
+			struct page *pages[num_par_pages];
+		} *extra_part;
+
+		_aio_small = kzalloc(sizeof(*_aio_small), GFP_KERNEL);
+		if (unlikely(!_aio_small)) {
+			ORE_DBGMSG("Failed alloc first part bytes=%zd\n",
+				   sizeof(*_aio_small));
+			*pios = NULL;
+			return -ENOMEM;
+		}
+		extra_part = kzalloc(sizeof(*extra_part), GFP_KERNEL);
+		if (unlikely(!extra_part)) {
+			ORE_DBGMSG("Failed alloc second part bytes=%zd\n",
+				   sizeof(*extra_part));
+			kfree(_aio_small);
+			*pios = NULL;
+			return -ENOMEM;
+		}
+
+		pages = num_par_pages ? extra_part->pages : NULL;
+		sgilist = sgs_per_dev ? extra_part->sglist : NULL;
+		/* In this case the per_dev[0].sgilist holds the pointer to
+		 * be freed
+		 */
+		ios = &_aio_small->ios;
+		ios->extra_part_alloc = true;
+	}
+
+	if (pages) {
+		ios->parity_pages = pages;
+		ios->max_par_pages = num_par_pages;
+	}
+	if (sgilist) {
+		unsigned d;
+
+		for (d = 0; d < numdevs; ++d) {
+			ios->per_dev[d].sglist = sgilist;
+			sgilist += sgs_per_dev;
+		}
+		ios->sgs_per_dev = sgs_per_dev;
+	}
+
+	ios->layout = layout;
+	ios->oc = oc;
+	*pios = ios;
+	return 0;
+}
+
+/* Allocate an io_state for only a single group of devices
+ *
+ * If a user needs to call ore_read/write() this version must be used becase it
+ * allocates extra stuff for striping and raid.
+ * The ore might decide to only IO less then @length bytes do to alignmets
+ * and constrains as follows:
+ * - The IO cannot cross group boundary.
+ * - In raid5/6 The end of the IO must align at end of a stripe eg.
+ *   (@offset + @length) % strip_size == 0. Or the complete range is within a
+ *   single stripe.
+ * - Memory condition only permitted a shorter IO. (A user can use @length=~0
+ *   And check the returned ios->length for max_io_size.)
+ *
+ * The caller must check returned ios->length (and/or ios->nr_pages) and
+ * re-issue these pages that fall outside of ios->length
+ */
+int  ore_get_rw_state(struct ore_layout *layout, struct ore_components *oc,
 		      bool is_reading, u64 offset, u64 length,
 		      struct ore_io_state **pios)
 {
 	struct ore_io_state *ios;
+	unsigned numdevs = layout->group_width * layout->mirrors_p1;
+	unsigned sgs_per_dev = 0, max_par_pages = 0;
+	int ret;
 
-	/*TODO: Maybe use kmem_cach per sbi of size
-	 * exofs_io_state_size(layout->s_numdevs)
-	 */
-	ios = kzalloc(ore_io_state_size(comps->numdevs), GFP_KERNEL);
-	if (unlikely(!ios)) {
-		ORE_DBGMSG("Failed kzalloc bytes=%d\n",
-			     ore_io_state_size(comps->numdevs));
-		*pios = NULL;
-		return -ENOMEM;
+	if (layout->parity && length) {
+		unsigned data_devs = layout->group_width - layout->parity;
+		unsigned stripe_size = layout->stripe_unit * data_devs;
+		unsigned pages_in_unit = layout->stripe_unit / PAGE_SIZE;
+		u32 remainder;
+		u64 num_stripes;
+		u64 num_raid_units;
+
+		num_stripes = div_u64_rem(length, stripe_size, &remainder);
+		if (remainder)
+			++num_stripes;
+
+		num_raid_units =  num_stripes * layout->parity;
+
+		if (is_reading) {
+			/* For reads add per_dev sglist array */
+			/* TODO: Raid 6 we need twice more. Actually:
+			*         num_stripes / LCMdP(W,P);
+			*         if (W%P != 0) num_stripes *= parity;
+			*/
+
+			/* first/last seg is split */
+			num_raid_units += layout->group_width;
+			sgs_per_dev = div_u64(num_raid_units, data_devs);
+		} else {
+			/* For Writes add parity pages array. */
+			max_par_pages = num_raid_units * pages_in_unit *
+						sizeof(struct page *);
+		}
 	}
 
-	ios->layout = layout;
-	ios->comps = comps;
-	ios->offset = offset;
-	ios->length = length;
+	ret = _ore_get_io_state(layout, oc, numdevs, sgs_per_dev, max_par_pages,
+				pios);
+	if (unlikely(ret))
+		return ret;
+
+	ios = *pios;
 	ios->reading = is_reading;
+	ios->offset = offset;
+
+	if (length) {
+		ore_calc_stripe_info(layout, offset, length, &ios->si);
+		ios->length = ios->si.length;
+		ios->nr_pages = (ios->length + PAGE_SIZE - 1) / PAGE_SIZE;
+		if (layout->parity)
+			_ore_post_alloc_raid_stuff(ios);
+	}
 
-	*pios = ios;
 	return 0;
 }
 EXPORT_SYMBOL(ore_get_rw_state);
 
-int  ore_get_io_state(struct ore_layout *layout, struct ore_components *comps,
-		      struct ore_io_state **ios)
+/* Allocate an io_state for all the devices in the comps array
+ *
+ * This version of io_state allocation is used mostly by create/remove
+ * and trunc where we currently need all the devices. The only wastful
+ * bit is the read/write_attributes with no IO. Those sites should
+ * be converted to use ore_get_rw_state() with length=0
+ */
+int  ore_get_io_state(struct ore_layout *layout, struct ore_components *oc,
+		      struct ore_io_state **pios)
 {
-	return ore_get_rw_state(layout, comps, true, 0, 0, ios);
+	return _ore_get_io_state(layout, oc, oc->numdevs, 0, 0, pios);
 }
 EXPORT_SYMBOL(ore_get_io_state);
 
@@ -111,6 +322,7 @@ void ore_put_io_state(struct ore_io_state *ios)
 				bio_put(per_dev->bio);
 		}
 
+		_ore_free_raid_stuff(ios);
 		kfree(ios);
 	}
 }
@@ -138,7 +350,7 @@ static void _done_io(struct osd_request *or, void *p)
 	kref_put(&ios->kref, _last_io);
 }
 
-static int ore_io_execute(struct ore_io_state *ios)
+int ore_io_execute(struct ore_io_state *ios)
 {
 	DECLARE_COMPLETION_ONSTACK(wait);
 	bool sync = (ios->done == NULL);
@@ -198,7 +410,7 @@ static void _clear_bio(struct bio *bio)
 	}
 }
 
-int ore_check_io(struct ore_io_state *ios, u64 *resid)
+int ore_check_io(struct ore_io_state *ios, ore_on_dev_error on_dev_error)
 {
 	enum osd_err_priority acumulated_osd_err = 0;
 	int acumulated_lin_err = 0;
@@ -206,7 +418,8 @@ int ore_check_io(struct ore_io_state *ios, u64 *resid)
 
 	for (i = 0; i < ios->numdevs; i++) {
 		struct osd_sense_info osi;
-		struct osd_request *or = ios->per_dev[i].or;
+		struct ore_per_dev_state *per_dev = &ios->per_dev[i];
+		struct osd_request *or = per_dev->or;
 		int ret;
 
 		if (unlikely(!or))
@@ -218,29 +431,31 @@ int ore_check_io(struct ore_io_state *ios, u64 *resid)
 
 		if (OSD_ERR_PRI_CLEAR_PAGES == osi.osd_err_pri) {
 			/* start read offset passed endof file */
-			_clear_bio(ios->per_dev[i].bio);
+			_clear_bio(per_dev->bio);
 			ORE_DBGMSG("start read offset passed end of file "
 				"offset=0x%llx, length=0x%llx\n",
-				_LLU(ios->per_dev[i].offset),
-				_LLU(ios->per_dev[i].length));
+				_LLU(per_dev->offset),
+				_LLU(per_dev->length));
 
 			continue; /* we recovered */
 		}
 
+		if (on_dev_error) {
+			u64 residual = ios->reading ?
+					or->in.residual : or->out.residual;
+			u64 offset = (ios->offset + ios->length) - residual;
+			struct ore_dev *od = ios->oc->ods[
+					per_dev->dev - ios->oc->first_dev];
+
+			on_dev_error(ios, od, per_dev->dev, osi.osd_err_pri,
+				     offse