[illumos-Developer] Now with zfs-ops-in-an-ioctl: iSCSI UNMAP (1st draft)

[Dan McDonald]

On Wed, Feb 16, 2011 at 03:58:31PM -0500, Matthew Ahrens wrote:
> I don't understand the volume subsystems very well so take this with a
> grain of salt.

You understand them better than I.  I'm just trying to make sure this
contribution is clean and Illumos-safe.

> Why is this an ioctl on /dev/zfs, rather than an operation on the zvol
> itself?  I'd expect this to either come in from a call in sbd_zvol.c,
> or from a callback in struct cb_ops zfs_cb_ops.

The original code:

	http://10.1.1.13/~danmcd/webrevs/unmap-post-rich/

Had direct zvol manipulations to which Eric Schrock objected strongly.

> If you have to do it from a zfsdev ioctl, you will at least need to
> make sure the objset is a zvol.  And I'd prefer to put the actual
> free_long_range in zvol code (ie, have zfs_ioctl.c marshall the
> arguments and call zvol_trim() in zvol.c)

Does Eric agree with you?  If so, I'd like to hear both of you say this.  If
not, please debate this PUBLICALLY so we can see your thought processes.

Thanks,
Dan