ocfs2: _really_ sync the right range

[pandora-kernel.git] / fs / ocfs2 / file.c

diff --git a/fs/ocfs2/file.c b/fs/ocfs2/file.c
index de4ea1a..6a7a3d9 100644 (file)
--- a/fs/ocfs2/file.c
+++ b/fs/ocfs2/file.c
@@ -1950,6 +1950,9 @@ static int __ocfs2_change_file_space(struct file *file, struct inode *inode,
        if (ret < 0)
                mlog_errno(ret);
 
+       if (file && (file->f_flags & O_SYNC))
+               handle->h_sync = 1;
+
        ocfs2_commit_trans(osb, handle);
 
 out_inode_unlock:
@@ -2052,6 +2055,23 @@ out:
        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 +2250,7 @@ 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 +2318,10 @@ relock:
                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 +2336,18 @@ relock:
                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.
@@ -2350,10 +2387,14 @@ out_dio:
        /* buffered aio wouldn't have proper lock coverage today */
        BUG_ON(ret == -EIOCBQUEUED && !(file->f_flags & O_DIRECT));
 
+       if (unlikely(written <= 0))
+               goto no_sync;
+
        if (((file->f_flags & O_DSYNC) && !direct_io) || IS_SYNC(inode) ||
            ((file->f_flags & O_DIRECT) && !direct_io)) {
-               ret = filemap_fdatawrite_range(file->f_mapping, pos,
-                                              pos + count - 1);
+               ret = filemap_fdatawrite_range(file->f_mapping,
+                                              iocb->ki_pos - written,
+                                              iocb->ki_pos - 1);
                if (ret < 0)
                        written = ret;
 
@@ -2366,10 +2407,12 @@ out_dio:
                }
 
                if (!ret)
-                       ret = filemap_fdatawait_range(file->f_mapping, pos,
-                                                     pos + count - 1);
+                       ret = filemap_fdatawait_range(file->f_mapping,
+                                                     iocb->ki_pos - written,
+                                                     iocb->ki_pos - 1);
        }
 
+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
@@ -2382,6 +2425,12 @@ out_dio:
        if ((ret == -EIOCBQUEUED) || (!ocfs2_iocb_is_rw_locked(iocb))) {
                rw_level = -1;
                have_alloc_sem = 0;
+               unaligned_dio = 0;
+       }
+
+       if (unaligned_dio) {
+               ocfs2_iocb_clear_unaligned_aio(iocb);
+               atomic_dec(&OCFS2_I(inode)->ip_unaligned_aio);
        }
 
 out:
@@ -2425,9 +2474,7 @@ static ssize_t ocfs2_file_splice_write(struct pipe_inode_info *pipe,
        struct address_space *mapping = out->f_mapping;
        struct inode *inode = mapping->host;
        struct splice_desc sd = {
-               .total_len = len,
                .flags = flags,
-               .pos = *ppos,
                .u.file = out,
        };
 
@@ -2437,6 +2484,12 @@ static ssize_t ocfs2_file_splice_write(struct pipe_inode_info *pipe,
                        out->f_path.dentry->d_name.len,
                        out->f_path.dentry->d_name.name, len);
 
+       ret = generic_write_checks(out, ppos, &len, 0);
+       if (ret)
+               return ret;
+       sd.total_len = len;
+       sd.pos = *ppos;
+
        if (pipe->inode)
                mutex_lock_nested(&pipe->inode->i_mutex, I_MUTEX_PARENT);
 
@@ -2591,6 +2644,57 @@ bail:
        return ret;
 }
 
+/* Refer generic_file_llseek_unlocked() */
+static loff_t ocfs2_file_llseek(struct file *file, loff_t offset, int origin)
+{
+       struct inode *inode = file->f_mapping->host;
+       int ret = 0;
+
+       mutex_lock(&inode->i_mutex);
+
+       switch (origin) {
+       case SEEK_SET:
+               break;
+       case SEEK_END:
+               offset += inode->i_size;
+               break;
+       case SEEK_CUR:
+               if (offset == 0) {
+                       offset = file->f_pos;
+                       goto out;
+               }
+               offset += file->f_pos;
+               break;
+       case SEEK_DATA:
+       case SEEK_HOLE:
+               ret = ocfs2_seek_data_hole_offset(file, &offset, origin);
+               if (ret)
+                       goto out;
+               break;
+       default:
+               ret = -EINVAL;
+               goto out;
+       }
+
+       if (offset < 0 && !(file->f_mode & FMODE_UNSIGNED_OFFSET))
+               ret = -EINVAL;
+       if (!ret && offset > inode->i_sb->s_maxbytes)
+               ret = -EINVAL;
+       if (ret)
+               goto out;
+
+       if (offset != file->f_pos) {
+               file->f_pos = offset;
+               file->f_version = 0;
+       }
+
+out:
+       mutex_unlock(&inode->i_mutex);
+       if (ret)
+               return ret;
+       return offset;
+}
+
 const struct inode_operations ocfs2_file_iops = {
        .setattr        = ocfs2_setattr,
        .getattr        = ocfs2_getattr,
@@ -2615,7 +2719,7 @@ const struct inode_operations ocfs2_special_file_iops = {
  * ocfs2_fops_no_plocks and ocfs2_dops_no_plocks!
  */
 const struct file_operations ocfs2_fops = {
-       .llseek         = generic_file_llseek,
+       .llseek         = ocfs2_file_llseek,
        .read           = do_sync_read,
        .write          = do_sync_write,
        .mmap           = ocfs2_mmap,
@@ -2663,7 +2767,7 @@ const struct file_operations ocfs2_dops = {
  * the cluster.
  */
 const struct file_operations ocfs2_fops_no_plocks = {
-       .llseek         = generic_file_llseek,
+       .llseek         = ocfs2_file_llseek,
        .read           = do_sync_read,
        .write          = do_sync_write,
        .mmap           = ocfs2_mmap,