Re: [PATCH] exofs: exofs_file_fsync correctness

[Boaz Harrosh <bharrosh@panasas.com>]

On 06/01/2010 06:34 PM, Christoph Hellwig wrote:
> On Tue, Jun 01, 2010 at 06:30:55PM +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
>> + */
> 
> You still want to check it.  It's really only need for checking the
> various dirty flags in the inode, which you're still missing.
> 

will do

>> +	struct writeback_control wbc = {
>> +		.sync_mode = WB_SYNC_ALL,
>> +		.nr_to_write = LONG_MAX,
> 
> By setting a nr_to_write you still write out data (at least in theory).
> 

So when do I also sync the data? is that done for me at the VFS layer?

> I'd recommend just copying the code from generic_file_fsync..
> 

I was actually mimicking the code from nfs/write.c which has similar
semantics as mine.

And if so then I'll need to not reuse the above in .flush

Boaz