Re: adding proper O_SYNC/O_DSYNC, was Re: O_DIRECT and barriers

[Ulrich Drepper <drepper@redhat.com>]

On 08/28/2009 09:41 AM, Christoph Hellwig wrote:
> Yeah.  The implementation really is trivial in 2.6.32 - we basically
> just need to change one function to check the new O_REALLY_SYNC flag
> and pass down a 0 instead of a 1 to another routine in the generic
> fs code, plus doing the same in a few filesystems opencoding it instead
> of using the generic helpers.

I don't think you have to change anything.  As I wrote before, the 
kernel ignores unknown O_* flags.  It's usually a problem.  Here it is a 
positive thing.


> So the logistics of doing the flags really is the biggest work here.
> And I'm not entirely sure how to do it correctly.  Can we just switch
> the current O_SYNC defintion in the kernel headers to O_DSYNC while
> adding the new O_SYNC and everything will continue to work?

No, that's not a good idea.  This would mean a program compiled with 
newer headers is using O_SYNC which isn't known to old kernels and 
ignored.  Such programs will then not even get the current O_DSYNC benefits.


> That includes a write from another process?  So O_RSYNC basically means
> doing an range-fdatasync before the actual read request?

Yes.  You can easily see how this can be useful.


> Again, we could implement this easily if we care enough.

I think it can be useful at times.

-- 
➧ Ulrich Drepper ➧ Red Hat, Inc. ➧ 444 Castro St ➧ Mountain View, CA ❖