Btrfs: RAID5 and RAID6

This builds on David Woodhouse's original Btrfs raid5/6 implementation.
The code has changed quite a bit, blame Chris Mason for any bugs.

Read/modify/write is done after the higher levels of the filesystem have
prepared a given bio.  This means the higher layers are not responsible
for building full stripes, and they don't need to query for the topology
of the extents that may get allocated during delayed allocation runs.
It also means different files can easily share the same stripe.

But, it does expose us to incorrect parity if we crash or lose power
while doing a read/modify/write cycle.  This will be addressed in a
later commit.

Scrub is unable to repair crc errors on raid5/6 chunks.

Discard does not work on raid5/6 (yet)

The stripe size is fixed at 64KiB per disk.  This will be tunable
in a later commit.

Signed-off-by: Chris Mason <chris.mason@fusionio.com>

1 files changed, 53 insertions, 9 deletions

diff --git a/fs/btrfs/disk-io.c b/fs/btrfs/disk-io.c
index 65f03670a952..e9fa7b4d18e3 100644
--- a/fs/btrfs/disk-io.c
+++ b/fs/btrfs/disk-io.c
@@ -46,6 +46,7 @@
 #include "check-integrity.h"
 #include "rcu-string.h"
 #include "dev-replace.h"
+#include "raid56.h"
 
 #ifdef CONFIG_X86
 #include <asm/cpufeature.h>
@@ -639,8 +640,15 @@ err:
 		btree_readahead_hook(root, eb, eb->start, ret);
 	}
 
-	if (ret)
+	if (ret) {
+		/*
+		 * our io error hook is going to dec the io pages
+		 * again, we have to make sure it has something
+		 * to decrement
+		 */
+		atomic_inc(&eb->io_pages);
 		clear_extent_buffer_uptodate(eb);
+	}
 	free_extent_buffer(eb);
 out:
 	return ret;
@@ -654,6 +662,7 @@ static int btree_io_failed_hook(struct page *page, int failed_mirror)
 	eb = (struct extent_buffer *)page->private;
 	set_bit(EXTENT_BUFFER_IOERR, &eb->bflags);
 	eb->read_mirror = failed_mirror;
+	atomic_dec(&eb->io_pages);
 	if (test_and_clear_bit(EXTENT_BUFFER_READAHEAD, &eb->bflags))
 		btree_readahead_hook(root, eb, eb->start, -EIO);
 	return -EIO;	/* we fixed nothing */
@@ -670,17 +679,23 @@ static void end_workqueue_bio(struct bio *bio, int err)
 	end_io_wq->work.flags = 0;
 
 	if (bio->bi_rw & REQ_WRITE) {
-		if (end_io_wq->metadata == 1)
+		if (end_io_wq->metadata == BTRFS_WQ_ENDIO_METADATA)
 			btrfs_queue_worker(&fs_info->endio_meta_write_workers,
 					   &end_io_wq->work);
-		else if (end_io_wq->metadata == 2)
+		else if (end_io_wq->metadata == BTRFS_WQ_ENDIO_FREE_SPACE)
 			btrfs_queue_worker(&fs_info->endio_freespace_worker,
 					   &end_io_wq->work);
+		else if (end_io_wq->metadata == BTRFS_WQ_ENDIO_RAID56)
+			btrfs_queue_worker(&fs_info->endio_raid56_workers,
+					   &end_io_wq->work);
 		else
 			btrfs_queue_worker(&fs_info->endio_write_workers,
 					   &end_io_wq->work);
 	} else {
-		if (end_io_wq->metadata)
+		if (end_io_wq->metadata == BTRFS_WQ_ENDIO_RAID56)
+			btrfs_queue_worker(&fs_info->endio_raid56_workers,
+					   &end_io_wq->work);
+		else if (end_io_wq->metadata)
 			btrfs_queue_worker(&fs_info->endio_meta_workers,
 					   &end_io_wq->work);
 		else
@@ -695,6 +710,7 @@ static void end_workqueue_bio(struct bio *bio, int err)
  * 0 - if data
  * 1 - if normal metadta
  * 2 - if writing to the free space cache area
+ * 3 - raid parity work
  */
 int btrfs_bio_wq_end_io(struct btrfs_fs_info *info, struct bio *bio,
 			int metadata)
@@ -2165,6 +2181,12 @@ int open_ctree(struct super_block *sb,
 	init_waitqueue_head(&fs_info->transaction_blocked_wait);
 	init_waitqueue_head(&fs_info->async_submit_wait);
 
+	ret = btrfs_alloc_stripe_hash_table(fs_info);
+	if (ret) {
+		err = -ENOMEM;
+		goto fail_alloc;
+	}
+
 	__setup_root(4096, 4096, 4096, 4096, tree_root,
 		     fs_info, BTRFS_ROOT_TREE_OBJECTID);
 
@@ -2332,6 +2354,12 @@ int open_ctree(struct super_block *sb,
 	btrfs_init_workers(&fs_info->endio_meta_write_workers,
 			   "endio-meta-write", fs_info->thread_pool_size,
 			   &fs_info->generic_worker);
+	btrfs_init_workers(&fs_info->endio_raid56_workers,
+			   "endio-raid56", fs_info->thread_pool_size,
+			   &fs_info->generic_worker);
+	btrfs_init_workers(&fs_info->rmw_workers,
+			   "rmw", fs_info->thread_pool_size,
+			   &fs_info->generic_worker);
 	btrfs_init_workers(&fs_info->endio_write_workers, "endio-write",
 			   fs_info->thread_pool_size,
 			   &fs_info->generic_worker);
@@ -2350,6 +2378,8 @@ int open_ctree(struct super_block *sb,
 	 */
 	fs_info->endio_workers.idle_thresh = 4;
 	fs_info->endio_meta_workers.idle_thresh = 4;
+	fs_info->endio_raid56_workers.idle_thresh = 4;
+	fs_info->rmw_workers.idle_thresh = 2;
 
 	fs_info->endio_write_workers.idle_thresh = 2;
 	fs_info->endio_meta_write_workers.idle_thresh = 2;
@@ -2366,6 +2396,8 @@ int open_ctree(struct super_block *sb,
 	ret |= btrfs_start_workers(&fs_info->fixup_workers);
 	ret |= btrfs_start_workers(&fs_info->endio_workers);
 	ret |= btrfs_start_workers(&fs_info->endio_meta_workers);
+	ret |= btrfs_start_workers(&fs_info->rmw_workers);
+	ret |= btrfs_start_workers(&fs_info->endio_raid56_workers);
 	ret |= btrfs_start_workers(&fs_info->endio_meta_write_workers);
 	ret |= btrfs_start_workers(&fs_info->endio_write_workers);
 	ret |= btrfs_start_workers(&fs_info->endio_freespace_worker);
@@ -2710,6 +2742,8 @@ fail_sb_buffer:
 	btrfs_stop_workers(&fs_info->workers);
 	btrfs_stop_workers(&fs_info->endio_workers);
 	btrfs_stop_workers(&fs_info->endio_meta_workers);
+	btrfs_stop_workers(&fs_info->endio_raid56_workers);
+	btrfs_stop_workers(&fs_info->rmw_workers);
 	btrfs_stop_workers(&fs_info->endio_meta_write_workers);
 	btrfs_stop_workers(&fs_info->endio_write_workers);
 	btrfs_stop_workers(&fs_info->endio_freespace_worker);
@@ -2728,6 +2762,7 @@ fail_bdi:
 fail_srcu:
 	cleanup_srcu_struct(&fs_info->subvol_srcu);
 fail:
+	btrfs_free_stripe_hash_table(fs_info);
 	btrfs_close_devices(fs_info->fs_devices);
 	return err;
 
@@ -3076,11 +3111,16 @@ int btrfs_calc_num_tolerated_disk_barrier_failures(
 				     ((flags & BTRFS_BLOCK_GROUP_PROFILE_MASK)
 				      == 0)))
 					num_tolerated_disk_barrier_failures = 0;
-				else if (num_tolerated_disk_barrier_failures > 1
-					 &&
-					 (flags & (BTRFS_BLOCK_GROUP_RAID1 |
-						   BTRFS_BLOCK_GROUP_RAID10)))
-					num_tolerated_disk_barrier_failures = 1;
+				else if (num_tolerated_disk_barrier_failures > 1) {
+					if (flags & (BTRFS_BLOCK_GROUP_RAID1 |
+					    BTRFS_BLOCK_GROUP_RAID5 |
+					    BTRFS_BLOCK_GROUP_RAID10)) {
+						num_tolerated_disk_barrier_failures = 1;
+					} else if (flags &
+						   BTRFS_BLOCK_GROUP_RAID5) {
+						num_tolerated_disk_barrier_failures = 2;
+					}
+				}
 			}
 		}
 		up_read(&sinfo->groups_sem);
@@ -3384,6 +3424,8 @@ int close_ctree(struct btrfs_root *root)
 	btrfs_stop_workers(&fs_info->workers);
 	btrfs_stop_workers(&fs_info->endio_workers);
 	btrfs_stop_workers(&fs_info->endio_meta_workers);
+	btrfs_stop_workers(&fs_info->endio_raid56_workers);
+	btrfs_stop_workers(&fs_info->rmw_workers);
 	btrfs_stop_workers(&fs_info->endio_meta_write_workers);
 	btrfs_stop_workers(&fs_info->endio_write_workers);
 	btrfs_stop_workers(&fs_info->endio_freespace_worker);
@@ -3404,6 +3446,8 @@ int close_ctree(struct btrfs_root *root)
 	bdi_destroy(&fs_info->bdi);
 	cleanup_srcu_struct(&fs_info->subvol_srcu);
 
+	btrfs_free_stripe_hash_table(fs_info);
+
 	return 0;
 }