Re: [PATCH ver2] exofs: exofs_file_fsync and exofs_file_flush correctness

[Boaz Harrosh <bharrosh@panasas.com>]

On 06/01/2010 07:17 PM, Christoph Hellwig wrote:
> On Tue, Jun 01, 2010 at 07:11:06PM +0300, Boaz Harrosh wrote:
>> +/* exofs_file_fsync - flush the inode to disk
>> + *
>> + *   @datasync is not used. All metadata is written in one place regardless.
>> + *   the writeout is synchronous
>> + */
> 
> It actually is used now :)
> 

Right ;-) fixing

>>  static int exofs_flush(struct file *file, fl_owner_t id)
>>  {
>> -	exofs_file_fsync(file, 1);
>> +	int ret = vfs_fsync_range(file, 0, LLONG_MAX, 0);
>>  	/* TODO: Flush the OSD target */
>> -	return 0;
>> +	return ret;
> 
> why not just:
> 
> 	return vfs_fsync(file, 0);
> 

Why, because I don't look passed my nose that's why.

Thank you Christoph. This is an area I was feeling bad about from the
get go. You can't imagine how happy I'm to have you help me on this.
(Things You Always Wanted to Know But Were Afraid to Ask)

Will post
Boaz