1 files changed, 46 insertions, 61 deletions

diff --git a/kernel/fs/ocfs2/file.c b/kernel/fs/ocfs2/file.c
index d8b670cbd..0e5b4515f 100644
--- a/kernel/fs/ocfs2/file.c
+++ b/kernel/fs/ocfs2/file.c
@@ -37,6 +37,7 @@
 #include <linux/falloc.h>
 #include <linux/quotaops.h>
 #include <linux/blkdev.h>
+#include <linux/backing-dev.h>
 
 #include <cluster/masklog.h>
 
@@ -104,8 +105,11 @@ static int ocfs2_file_open(struct inode *inode, struct file *file)
 			      file->f_path.dentry->d_name.len,
 			      file->f_path.dentry->d_name.name, mode);
 
-	if (file->f_mode & FMODE_WRITE)
-		dquot_initialize(inode);
+	if (file->f_mode & FMODE_WRITE) {
+		status = dquot_initialize(inode);
+		if (status)
+			goto leave;
+	}
 
 	spin_lock(&oi->ip_lock);
 
@@ -1126,6 +1130,7 @@ out:
 int ocfs2_setattr(struct dentry *dentry, struct iattr *attr)
 {
 	int status = 0, size_change;
+	int inode_locked = 0;
 	struct inode *inode = d_inode(dentry);
 	struct super_block *sb = inode->i_sb;
 	struct ocfs2_super *osb = OCFS2_SB(sb);
@@ -1154,8 +1159,11 @@ int ocfs2_setattr(struct dentry *dentry, struct iattr *attr)
 	if (status)
 		return status;
 
-	if (is_quota_modification(inode, attr))
-		dquot_initialize(inode);
+	if (is_quota_modification(inode, attr)) {
+		status = dquot_initialize(inode);
+		if (status)
+			return status;
+	}
 	size_change = S_ISREG(inode->i_mode) && attr->ia_valid & ATTR_SIZE;
 	if (size_change) {
 		status = ocfs2_rw_lock(inode, 1);
@@ -1171,6 +1179,7 @@ int ocfs2_setattr(struct dentry *dentry, struct iattr *attr)
 			mlog_errno(status);
 		goto bail_unlock_rw;
 	}
+	inode_locked = 1;
 
 	if (size_change) {
 		status = inode_newsize_ok(inode, attr->ia_size);
@@ -1208,8 +1217,8 @@ int ocfs2_setattr(struct dentry *dentry, struct iattr *attr)
 		    && OCFS2_HAS_RO_COMPAT_FEATURE(sb,
 		    OCFS2_FEATURE_RO_COMPAT_USRQUOTA)) {
 			transfer_to[USRQUOTA] = dqget(sb, make_kqid_uid(attr->ia_uid));
-			if (!transfer_to[USRQUOTA]) {
-				status = -ESRCH;
+			if (IS_ERR(transfer_to[USRQUOTA])) {
+				status = PTR_ERR(transfer_to[USRQUOTA]);
 				goto bail_unlock;
 			}
 		}
@@ -1217,8 +1226,8 @@ int ocfs2_setattr(struct dentry *dentry, struct iattr *attr)
 		    && OCFS2_HAS_RO_COMPAT_FEATURE(sb,
 		    OCFS2_FEATURE_RO_COMPAT_GRPQUOTA)) {
 			transfer_to[GRPQUOTA] = dqget(sb, make_kqid_gid(attr->ia_gid));
-			if (!transfer_to[GRPQUOTA]) {
-				status = -ESRCH;
+			if (IS_ERR(transfer_to[GRPQUOTA])) {
+				status = PTR_ERR(transfer_to[GRPQUOTA]);
 				goto bail_unlock;
 			}
 		}
@@ -1251,7 +1260,10 @@ int ocfs2_setattr(struct dentry *dentry, struct iattr *attr)
 bail_commit:
 	ocfs2_commit_trans(osb, handle);
 bail_unlock:
-	ocfs2_inode_unlock(inode, 1);
+	if (status) {
+		ocfs2_inode_unlock(inode, 1);
+		inode_locked = 0;
+	}
 bail_unlock_rw:
 	if (size_change)
 		ocfs2_rw_unlock(inode, 1);
@@ -1267,6 +1279,8 @@ bail:
 		if (status < 0)
 			mlog_errno(status);
 	}
+	if (inode_locked)
+		ocfs2_inode_unlock(inode, 1);
 
 	return status;
 }
@@ -2250,13 +2264,11 @@ out:
 static ssize_t ocfs2_file_write_iter(struct kiocb *iocb,
 				    struct iov_iter *from)
 {
-	int direct_io, appending, rw_level, have_alloc_sem  = 0;
+	int direct_io, appending, rw_level;
 	int can_do_direct, has_refcount = 0;
 	ssize_t written = 0;
 	ssize_t ret;
 	size_t count = iov_iter_count(from), orig_count;
-	loff_t old_size;
-	u32 old_clusters;
 	struct file *file = iocb->ki_filp;
 	struct inode *inode = file_inode(file);
 	struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
@@ -2264,6 +2276,8 @@ static ssize_t ocfs2_file_write_iter(struct kiocb *iocb,
 			       OCFS2_MOUNT_COHERENCY_BUFFERED);
 	int unaligned_dio = 0;
 	int dropped_dio = 0;
+	int append_write = ((iocb->ki_pos + count) >=
+			i_size_read(inode) ? 1 : 0);
 
 	trace_ocfs2_file_aio_write(inode, file, file->f_path.dentry,
 		(unsigned long long)OCFS2_I(inode)->ip_blkno,
@@ -2279,26 +2293,18 @@ static ssize_t ocfs2_file_write_iter(struct kiocb *iocb,
 
 	mutex_lock(&inode->i_mutex);
 
-	ocfs2_iocb_clear_sem_locked(iocb);
-
 relock:
-	/* to match setattr's i_mutex -> rw_lock ordering */
-	if (direct_io) {
-		have_alloc_sem = 1;
-		/* communicate with ocfs2_dio_end_io */
-		ocfs2_iocb_set_sem_locked(iocb);
-	}
-
 	/*
 	 * Concurrent O_DIRECT writes are allowed with
 	 * mount_option "coherency=buffered".
+	 * For append write, we must take rw EX.
 	 */
-	rw_level = (!direct_io || full_coherency);
+	rw_level = (!direct_io || full_coherency || append_write);
 
 	ret = ocfs2_rw_lock(inode, rw_level);
 	if (ret < 0) {
 		mlog_errno(ret);
-		goto out_sems;
+		goto out_mutex;
 	}
 
 	/*
@@ -2347,7 +2353,6 @@ relock:
 	if (direct_io && !can_do_direct) {
 		ocfs2_rw_unlock(inode, rw_level);
 
-		have_alloc_sem = 0;
 		rw_level = -1;
 
 		direct_io = 0;
@@ -2367,13 +2372,6 @@ relock:
 		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.
-	 */
-	old_size = i_size_read(inode);
-	old_clusters = OCFS2_I(inode)->ip_clusters;
-
 	/* communicate with ocfs2_dio_end_io */
 	ocfs2_iocb_set_rw_locked(iocb, rw_level);
 
@@ -2381,6 +2379,20 @@ relock:
 	/* buffered aio wouldn't have proper lock coverage today */
 	BUG_ON(written == -EIOCBQUEUED && !(iocb->ki_flags & IOCB_DIRECT));
 
+	/*
+	 * deep in g_f_a_w_n()->ocfs2_direct_IO we pass in a ocfs2_dio_end_io
+	 * function pointer which is called when o_direct io completes so that
+	 * it can unlock our rw lock.
+	 * Unfortunately there are error cases which call end_io and others
+	 * that don't.  so we don't have to unlock the rw_lock if either an
+	 * async dio is going to do it in the future or an end_io after an
+	 * error has already done it.
+	 */
+	if ((written == -EIOCBQUEUED) || (!ocfs2_iocb_is_rw_locked(iocb))) {
+		rw_level = -1;
+		unaligned_dio = 0;
+	}
+
 	if (unlikely(written <= 0))
 		goto no_sync;
 
@@ -2405,22 +2417,7 @@ relock:
 	}
 
 no_sync:
-	/*
-	 * deep in g_f_a_w_n()->ocfs2_direct_IO we pass in a ocfs2_dio_end_io
-	 * function pointer which is called when o_direct io completes so that
-	 * it can unlock our rw lock.
-	 * Unfortunately there are error cases which call end_io and others
-	 * that don't.  so we don't have to unlock the rw_lock if either an
-	 * async dio is going to do it in the future or an end_io after an
-	 * error has already done it.
-	 */
-	if ((ret == -EIOCBQUEUED) || (!ocfs2_iocb_is_rw_locked(iocb))) {
-		rw_level = -1;
-		have_alloc_sem = 0;
-		unaligned_dio = 0;
-	}
-
-	if (unaligned_dio) {
+	if (unaligned_dio && ocfs2_iocb_is_unaligned_aio(iocb)) {
 		ocfs2_iocb_clear_unaligned_aio(iocb);
 		mutex_unlock(&OCFS2_I(inode)->ip_unaligned_aio);
 	}
@@ -2429,10 +2426,7 @@ out:
 	if (rw_level != -1)
 		ocfs2_rw_unlock(inode, rw_level);
 
-out_sems:
-	if (have_alloc_sem)
-		ocfs2_iocb_clear_sem_locked(iocb);
-
+out_mutex:
 	mutex_unlock(&inode->i_mutex);
 
 	if (written)
@@ -2473,7 +2467,7 @@ bail:
 static ssize_t ocfs2_file_read_iter(struct kiocb *iocb,
 				   struct iov_iter *to)
 {
-	int ret = 0, rw_level = -1, have_alloc_sem = 0, lock_level = 0;
+	int ret = 0, rw_level = -1, lock_level = 0;
 	struct file *filp = iocb->ki_filp;
 	struct inode *inode = file_inode(filp);
 
@@ -2490,16 +2484,11 @@ static ssize_t ocfs2_file_read_iter(struct kiocb *iocb,
 		goto bail;
 	}
 
-	ocfs2_iocb_clear_sem_locked(iocb);
-
 	/*
 	 * buffered reads protect themselves in ->readpage().  O_DIRECT reads
 	 * need locks to protect pending reads from racing with truncate.
 	 */
 	if (iocb->ki_flags & IOCB_DIRECT) {
-		have_alloc_sem = 1;
-		ocfs2_iocb_set_sem_locked(iocb);
-
 		ret = ocfs2_rw_lock(inode, 0);
 		if (ret < 0) {
 			mlog_errno(ret);
@@ -2535,13 +2524,9 @@ static ssize_t ocfs2_file_read_iter(struct kiocb *iocb,
 	/* see ocfs2_file_write_iter */
 	if (ret == -EIOCBQUEUED || !ocfs2_iocb_is_rw_locked(iocb)) {
 		rw_level = -1;
-		have_alloc_sem = 0;
 	}
 
 bail:
-	if (have_alloc_sem)
-		ocfs2_iocb_clear_sem_locked(iocb);
-
 	if (rw_level != -1)
 		ocfs2_rw_unlock(inode, rw_level);