On 07/13/2009 12:00 PM, Nick Piggin wrote:

On Mon, Jul 13, 2009 at 11:54:15AM +0300, Boaz Harrosh wrote:

quoted

On 07/13/2009 09:59 AM, Nick Piggin wrote:

quoted

On Sun, Jul 12, 2009 at 10:47:18AM -0400, Christoph Hellwig wrote:

quoted

On Sun, Jul 12, 2009 at 11:55:51AM +0300, Boaz Harrosh wrote:

quoted

I wish you would split it.

one - helper to be called by converted file systems
      (Which just ignores the ATTR_SIZE)
second - to be set into .setattr which does the simple_setsize + above.

More clear for FS users like me (and that ugly unmask of ATTR_SIZE)

or it's just me?

Yeah, that seems be a lot cleaner.  But let's wait until we got
rid of ->truncate for all filesystems to have the bigger picture.

Agreed, if it is a common sequence / requirement for filesystems
then of course I will not object to a helper to make things clearer
or share code.

I would like to see inode_setattr renamed into simple_setattr, and
then also .setattr made mandatory, so I don't like to cut code out
of inode_setattr which makes it unable to be the simple_setattr
after the old truncate code is removed.

I thought you meant inode_setattr will go away. There will
only be simple_setattr() and inode_setattr_nosize()

For the time been simple_setattr() will also take care
of old ->truncate FSs. And in the absence of .setattr
simple_setattr() is called. Have I miss-understood?

again please tell me when all this is in effect I want
to do the conversion in exofs.

AFAIKS inode_setattr basically is simple_setattr, so I think at some
point it should just get renamed to simple_setattr. Adding
simple_setattr_nosize or similar helper would be fine too. I don't
care much about the exact details... But anyway these things are not
so important to this truncate patchset at the moment.

I see. So what is the schedule? when are we to convert all FSs?

quoted

[BTW these changes are a life saver for me in regard to
the kind of things I need to do for pNFS-exports]

You mean the truncate patches? Well that's nice to know. I
guess it has always been possible just to redefine your own
setattr, but now it should be a bit nicer with the truncate
helpers...

OK, yes redefine .setattr, do the right thing in write_begin/end
and the helpers do help a lot, to the point that it was not safe to
open-code all this work. The situation is much better after your
patchset.

Thanks,
Nick

Thanks
Boaz

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>