On Wed, Apr 13, 2016 at 11:17:45AM -0500, Eric Sandeen wrote:
> On 4/13/16 10:08 AM, Jan Tulak wrote:
>
> ...
>
> > All right, I will keep the O_TRUNC there. However, should it
> > truncate the file every time, or should we offer a way how to avoid
> > the file truncating? Until now, mkfs behaved differently based on
> > whether -d file was given, or not. Your explanation suggests that we
> > should truncate every time, right?
>
> There are probably valid reasons to keep size as well as to truncate;
> it's not immediately clear to me how we should handle it.
That's easy: don't specify -dfile or -dsize, and instead let mkfs
detect that it is file automatically. Because it's not given a size,
or that it's supposed to be a file, it should not attempt to change
the "device" in any way.
When block devices grow fallocate support, we can unify the discard
behaviour to be a hole punch, and then it will work identically on
devices and files treated this way.
> Honestly, at this point, in the interest of getting the other fixes in,
> I think I might rather see the truncating behavior unchanged from what
> we have today; we can tackle that as a separate problem at a later date.
Yup, don't change it in this patch series.
Looking on this patch again, I think that it can be removed entirely. After reverting the O_CREATE part, the only thing it does is differentiating error messages between files and block devices, which is not worth of a patch.
Cheers,
Jan
_______________________________________________ xfs mailing list xfs@xxxxxxxxxxx http://oss.sgi.com/mailman/listinfo/xfs
