On Thu, Jan 28, 2016 at 05:33:29AM -0800, Christoph Hellwig wrote:

Now that we have this in 4.5-rc1 let's get this into man-pages as well.

quoted

--- /dev/null
+++ b/man2/ioctl_ficlone.2

think this should be a link to the clone range man page, and the
both ioctls should be documented together.  Basically just document
FICLONERANGE and then mentiond that FICLONE does a whole file clone.

Ok, that's an easy switch to make.

quoted

+.B #include <sys/ioctl.h>

Don't we need uapi/linux/fs.h as well?

Err... yes. :)

quoted

Not all filesystems support
+overlapping reflink ranges in the same file.

Which one doesn't?

At the moment, neither btrfs nor XFS support it.

I don't think it'd be difficult for XFS (make a copy of the extent mappings in
the source area, increase the refcounts of those blocks, do the usual
punch-and-remap, then delete & decrement that copy of the mappings) but that
can come later.

Also there should be a sentence that the clones are atomic vs
concurrent writes, so no locks need to be taken for a consistent
snapshot.

<nod>

I've updated the manpages and will repush the patchset ... optimistically,
when XFS reflink RFCv5 comes out, but sooner if anyone requests it. :)

Thank you for the review!

--D

--
To unsubscribe from this list: send the line "unsubscribe linux-api" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

--
To unsubscribe from this list: send the line "unsubscribe linux-man" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html