---

yaml
---
r: 276185
b: refs/heads/master
c: 99b1bb6
h: refs/heads/master
i:
  276183: 264f313
v: v3

[refs]

@@ -1,2 +1,2 @@
 ---
-refs/heads/master: 93862d5e1ab875664c6cc95254fc365028a48bb1
+refs/heads/master: 99b1bb61b225c3eb4d3b196d4f1d041695b19a7e

trunk/fs/ocfs2/aops.c

@@ -290,7 +290,15 @@ static int ocfs2_readpage(struct file *file, struct page *page)
 	}
 
 	if (down_read_trylock(&oi->ip_alloc_sem) == 0) {
+		/*
+		 * Unlock the page and cycle ip_alloc_sem so that we don't
+		 * busyloop waiting for ip_alloc_sem to unlock
+		 */
 		ret = AOP_TRUNCATED_PAGE;
+		unlock_page(page);
+		unlock = 0;
+		down_read(&oi->ip_alloc_sem);
+		up_read(&oi->ip_alloc_sem);
 		goto out_inode_unlock;
 	}
 
@@ -563,13 +571,23 @@ static void ocfs2_dio_end_io(struct kiocb *iocb,
 {
 	struct inode *inode = iocb->ki_filp->f_path.dentry->d_inode;
 	int level;
+	wait_queue_head_t *wq = ocfs2_ioend_wq(inode);
 
 	/* this io's submitter should not have unlocked this before we could */
 	BUG_ON(!ocfs2_iocb_is_rw_locked(iocb));
 
 	if (ocfs2_iocb_is_sem_locked(iocb))
 		ocfs2_iocb_clear_sem_locked(iocb);
 
+	if (ocfs2_iocb_is_unaligned_aio(iocb)) {
+		ocfs2_iocb_clear_unaligned_aio(iocb);
+
+		if (atomic_dec_and_test(&OCFS2_I(inode)->ip_unaligned_aio) &&
+		    waitqueue_active(wq)) {
+			wake_up_all(wq);
+		}
+	}
+
 	ocfs2_iocb_clear_rw_locked(iocb);
 
 	level = ocfs2_iocb_rw_locked_level(iocb);

trunk/fs/ocfs2/aops.h

@@ -78,6 +78,7 @@ enum ocfs2_iocb_lock_bits {
 	OCFS2_IOCB_RW_LOCK = 0,
 	OCFS2_IOCB_RW_LOCK_LEVEL,
 	OCFS2_IOCB_SEM,
+	OCFS2_IOCB_UNALIGNED_IO,
 	OCFS2_IOCB_NUM_LOCKS
 };
 
@@ -91,4 +92,17 @@ enum ocfs2_iocb_lock_bits {
 	clear_bit(OCFS2_IOCB_SEM, (unsigned long *)&iocb->private)
 #define ocfs2_iocb_is_sem_locked(iocb) \
 	test_bit(OCFS2_IOCB_SEM, (unsigned long *)&iocb->private)
+
+#define ocfs2_iocb_set_unaligned_aio(iocb) \
+	set_bit(OCFS2_IOCB_UNALIGNED_IO, (unsigned long *)&iocb->private)
+#define ocfs2_iocb_clear_unaligned_aio(iocb) \
+	clear_bit(OCFS2_IOCB_UNALIGNED_IO, (unsigned long *)&iocb->private)
+#define ocfs2_iocb_is_unaligned_aio(iocb) \
+	test_bit(OCFS2_IOCB_UNALIGNED_IO, (unsigned long *)&iocb->private)
+
+#define OCFS2_IOEND_WQ_HASH_SZ	37
+#define ocfs2_ioend_wq(v)   (&ocfs2__ioend_wq[((unsigned long)(v)) %\
+					    OCFS2_IOEND_WQ_HASH_SZ])
+extern wait_queue_head_t ocfs2__ioend_wq[OCFS2_IOEND_WQ_HASH_SZ];
+
 #endif /* OCFS2_FILE_H */

trunk/fs/ocfs2/dlmglue.c

@@ -1692,7 +1692,7 @@ int ocfs2_open_lock(struct inode *inode)
 	mlog(0, "inode %llu take PRMODE open lock\n",
 	     (unsigned long long)OCFS2_I(inode)->ip_blkno);
 
-	if (ocfs2_mount_local(osb))
+	if (ocfs2_is_hard_readonly(osb) || ocfs2_mount_local(osb))
 		goto out;
 
 	lockres = &OCFS2_I(inode)->ip_open_lockres;
@@ -1718,6 +1718,12 @@ int ocfs2_try_open_lock(struct inode *inode, int write)
 	     (unsigned long long)OCFS2_I(inode)->ip_blkno,
 	     write ? "EXMODE" : "PRMODE");
 
+	if (ocfs2_is_hard_readonly(osb)) {
+		if (write)
+			status = -EROFS;
+		goto out;
+	}
+
 	if (ocfs2_mount_local(osb))
 		goto out;
 
@@ -2298,7 +2304,7 @@ int ocfs2_inode_lock_full_nested(struct inode *inode,
 	if (ocfs2_is_hard_readonly(osb)) {
 		if (ex)
 			status = -EROFS;
-		goto bail;
+		goto getbh;
 	}
 
 	if (ocfs2_mount_local(osb))
@@ -2356,7 +2362,7 @@ int ocfs2_inode_lock_full_nested(struct inode *inode,
 			mlog_errno(status);
 		goto bail;
 	}
-
+getbh:
 	if (ret_bh) {
 		status = ocfs2_assign_bh(inode, ret_bh, local_bh);
 		if (status < 0) {
@@ -2628,8 +2634,11 @@ int ocfs2_dentry_lock(struct dentry *dentry, int ex)
 
 	BUG_ON(!dl);
 
-	if (ocfs2_is_hard_readonly(osb))
-		return -EROFS;
+	if (ocfs2_is_hard_readonly(osb)) {
+		if (ex)
+			return -EROFS;
+		return 0;
+	}
 
 	if (ocfs2_mount_local(osb))
 		return 0;
@@ -2647,7 +2656,7 @@ void ocfs2_dentry_unlock(struct dentry *dentry, int ex)
 	struct ocfs2_dentry_lock *dl = dentry->d_fsdata;
 	struct ocfs2_super *osb = OCFS2_SB(dentry->d_sb);
 
-	if (!ocfs2_mount_local(osb))
+	if (!ocfs2_is_hard_readonly(osb) && !ocfs2_mount_local(osb))
 		ocfs2_cluster_unlock(osb, &dl->dl_lockres, level);
 }
 

trunk/fs/ocfs2/file.c

@@ -2052,6 +2052,23 @@ int ocfs2_check_range_for_refcount(struct inode *inode, loff_t pos,
 	return ret;
 }
 
+static void ocfs2_aiodio_wait(struct inode *inode)
+{
+	wait_queue_head_t *wq = ocfs2_ioend_wq(inode);
+
+	wait_event(*wq, (atomic_read(&OCFS2_I(inode)->ip_unaligned_aio) == 0));
+}
+
+static int ocfs2_is_io_unaligned(struct inode *inode, size_t count, loff_t pos)
+{
+	int blockmask = inode->i_sb->s_blocksize - 1;
+	loff_t final_size = pos + count;
+
+	if ((pos & blockmask) || (final_size & blockmask))
+		return 1;
+	return 0;
+}
+
 static int ocfs2_prepare_inode_for_refcount(struct inode *inode,
 					    struct file *file,
 					    loff_t pos, size_t count,
@@ -2230,6 +2247,7 @@ static ssize_t ocfs2_file_aio_write(struct kiocb *iocb,
 	struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
 	int full_coherency = !(osb->s_mount_opt &
 			       OCFS2_MOUNT_COHERENCY_BUFFERED);
+	int unaligned_dio = 0;
 
 	trace_ocfs2_file_aio_write(inode, file, file->f_path.dentry,
 		(unsigned long long)OCFS2_I(inode)->ip_blkno,
@@ -2297,6 +2315,10 @@ static ssize_t ocfs2_file_aio_write(struct kiocb *iocb,
 		goto out;
 	}
 
+	if (direct_io && !is_sync_kiocb(iocb))
+		unaligned_dio = ocfs2_is_io_unaligned(inode, iocb->ki_left,
+						      *ppos);
+
 	/*
 	 * We can't complete the direct I/O as requested, fall back to
 	 * buffered I/O.
@@ -2311,6 +2333,18 @@ static ssize_t ocfs2_file_aio_write(struct kiocb *iocb,
 		goto relock;
 	}
 
+	if (unaligned_dio) {
+		/*
+		 * Wait on previous unaligned aio to complete before
+		 * proceeding.
+		 */
+		ocfs2_aiodio_wait(inode);
+
+		/* Mark the iocb as needing a decrement in ocfs2_dio_end_io */
+		atomic_inc(&OCFS2_I(inode)->ip_unaligned_aio);
+		ocfs2_iocb_set_unaligned_aio(iocb);
+	}
+
 	/*
 	 * To later detect whether a journal commit for sync writes is
 	 * necessary, we sample i_size, and cluster count here.
@@ -2382,8 +2416,12 @@ static ssize_t ocfs2_file_aio_write(struct kiocb *iocb,
 	if ((ret == -EIOCBQUEUED) || (!ocfs2_iocb_is_rw_locked(iocb))) {
 		rw_level = -1;
 		have_alloc_sem = 0;
+		unaligned_dio = 0;
 	}
 
+	if (unaligned_dio)
+		atomic_dec(&OCFS2_I(inode)->ip_unaligned_aio);
+
 out:
 	if (rw_level != -1)
 		ocfs2_rw_unlock(inode, rw_level);

trunk/fs/ocfs2/inode.h

@@ -43,6 +43,9 @@ struct ocfs2_inode_info
 	/* protects extended attribute changes on this inode */
 	struct rw_semaphore		ip_xattr_sem;
 
+	/* Number of outstanding AIO's which are not page aligned */
+	atomic_t			ip_unaligned_aio;
+
 	/* These fields are protected by ip_lock */
 	spinlock_t			ip_lock;
 	u32				ip_open_count;

trunk/fs/ocfs2/ioctl.c

@@ -381,7 +381,7 @@ int ocfs2_info_handle_freeinode(struct inode *inode,
 	if (!oifi) {
 		status = -ENOMEM;
 		mlog_errno(status);
-		goto bail;
+		goto out_err;
 	}
 
 	if (o2info_from_user(*oifi, req))
@@ -431,7 +431,7 @@ int ocfs2_info_handle_freeinode(struct inode *inode,
 		o2info_set_request_error(&oifi->ifi_req, req);
 
 	kfree(oifi);
-
+out_err:
 	return status;
 }
 
@@ -666,7 +666,7 @@ int ocfs2_info_handle_freefrag(struct inode *inode,
 	if (!oiff) {
 		status = -ENOMEM;
 		mlog_errno(status);
-		goto bail;
+		goto out_err;
 	}
 
 	if (o2info_from_user(*oiff, req))
@@ -716,7 +716,7 @@ int ocfs2_info_handle_freefrag(struct inode *inode,
 		o2info_set_request_error(&oiff->iff_req, req);
 
 	kfree(oiff);
-
+out_err:
 	return status;
 }
 

trunk/fs/ocfs2/move_extents.c

@@ -746,7 +746,7 @@ static int ocfs2_move_extent(struct ocfs2_move_extents_context *context,
 	 */
 	ocfs2_probe_alloc_group(inode, gd_bh, &goal_bit, len, move_max_hop,
 				new_phys_cpos);
-	if (!new_phys_cpos) {
+	if (!*new_phys_cpos) {
 		ret = -ENOSPC;
 		goto out_commit;
 	}

trunk/fs/ocfs2/ocfs2.h

@@ -836,13 +836,13 @@ static inline unsigned int ocfs2_clusters_to_megabytes(struct super_block *sb,
 
 static inline void _ocfs2_set_bit(unsigned int bit, unsigned long *bitmap)
 {
-	__test_and_set_bit_le(bit, bitmap);
+	__set_bit_le(bit, bitmap);
 }
 #define ocfs2_set_bit(bit, addr) _ocfs2_set_bit((bit), (unsigned long *)(addr))
 
 static inline void _ocfs2_clear_bit(unsigned int bit, unsigned long *bitmap)
 {
-	__test_and_clear_bit_le(bit, bitmap);
+	__clear_bit_le(bit, bitmap);
 }
 #define ocfs2_clear_bit(bit, addr) _ocfs2_clear_bit((bit), (unsigned long *)(addr))
 

trunk/fs/ocfs2/super.c

@@ -54,6 +54,7 @@
 #include "ocfs1_fs_compat.h"
 
 #include "alloc.h"
+#include "aops.h"
 #include "blockcheck.h"
 #include "dlmglue.h"
 #include "export.h"
@@ -1616,12 +1617,17 @@ static int ocfs2_show_options(struct seq_file *s, struct vfsmount *mnt)
 	return 0;
 }
 
+wait_queue_head_t ocfs2__ioend_wq[OCFS2_IOEND_WQ_HASH_SZ];
+
 static int __init ocfs2_init(void)
 {
-	int status;
+	int status, i;
 
 	ocfs2_print_version();
 
+	for (i = 0; i < OCFS2_IOEND_WQ_HASH_SZ; i++)
+		init_waitqueue_head(&ocfs2__ioend_wq[i]);
+
 	status = init_ocfs2_uptodate_cache();
 	if (status < 0) {
 		mlog_errno(status);
@@ -1760,7 +1766,7 @@ static void ocfs2_inode_init_once(void *data)
 	ocfs2_extent_map_init(&oi->vfs_inode);
 	INIT_LIST_HEAD(&oi->ip_io_markers);
 	oi->ip_dir_start_lookup = 0;
-
+	atomic_set(&oi->ip_unaligned_aio, 0);
 	init_rwsem(&oi->ip_alloc_sem);
 	init_rwsem(&oi->ip_xattr_sem);
 	mutex_init(&oi->ip_io_mutex);
@@ -1974,7 +1980,8 @@ static void ocfs2_dismount_volume(struct super_block *sb, int mnt_err)
 	 * If we failed before we got a uuid_str yet, we can't stop
 	 * heartbeat.  Otherwise, do it.
 	 */
-	if (!mnt_err && !ocfs2_mount_local(osb) && osb->uuid_str)
+	if (!mnt_err && !ocfs2_mount_local(osb) && osb->uuid_str &&
+	    !ocfs2_is_hard_readonly(osb))
 		hangup_needed = 1;
 
 	if (osb->cconn)