md: Replace BIO_DELETE emulation with vn_deallocate(9)

[khng]

Summary

Both zero-filling and/or deallocation can be done with vn_deallocate(9).

Submitted by: Ka Ho Ng <[e-mail]>
Sponsored by: The FreeBSD Foundation

[khng]

Rebased.

[khng]

Rebased

[khng]

Rebased

[khng]

Updated commit message.

[khng]

vn_deallocate KPI changes.

[khng]

vn_deallocate() changed.

[markj]

sys/dev/md/md.c

919

Looks like there is no way to honour MD_SYNC? For writes it is translated to IO_SYNC, but here it looks like it would not be plumbed through VOP_DEALLOCATE.

[markj]

sys/dev/md/md.c

888

Harmless, but there's a missing break here.

[khng]

sys/dev/md/md.c

919

For the fallback part, do you think it is reasonable if we add a ioflg in VOP_DEALLOCATE that can accept IO_SYNC? Or do we just make it a vn_deallocate + VOP_FSYNC? On the other hand, posix_fallocate is another call that currently bypass IO_SYNC.

[markj]

sys/dev/md/md.c

919

I think that adding an additional parameter would be reasonable. Calling VOP_FSYNC() should also work but that seems like a hack, and it can in principle result in unnecessary I/O. I do not insist on fixing this for this review, but I would note it with an XXX comment or so.

So posix_fallocate() does not respect O_SYNC, effectively, which is a bug IMO.

[khng]

• Gotchas on the current behavior
• missing break

[markj]

sys/dev/md/md.c

921

I believe we need to update bio->bp_resid as well.

[khng]

Update bp->bio_resid.

[markj]

A test case in sbin/mdconfig/tests would be nice. I believe trim(8) can be used to easily submit DIOCGDELETE requests against a md device.

[khng]

In D28899#709851, @markj wrote:

A test case in sbin/mdconfig/tests would be nice. I believe trim(8) can be used to easily submit DIOCGDELETE requests against a md device.

I will make that a separate differential for review.

[khng]

• Use IO_SYNC for non-MD_ASYNC case.