Compilation fixes after rebase

• A leftover manpage fix

• Manpage clarifications and fixes
• Code cosmetic fixes
• Removed a stray #include "sys/fcntl.h"

manpage fixes.

manpage fixes.

manpage fixes.

Wording.

Wording around page fault behavior.

Updated. Document the behavior around page faults as well.

Updates.

.Dd updates and add document the direction from old EOF to new EOF as well.

bcr@'s suggestion: rewording document description

In D29408#663654, @bcr wrote:

OK, I only have one textual suggestion. Did you check the man page with "mandoc -Tlint" and textproc/igor?

Style.

@afedorov I would like to have some comments on this again.

Refactor the prototype to match vq_ring_enqueue_segments in sys/dev/virtio/virtqueue.c.

• Remove unused flags from pci_vt9p_notify

• Fixed loopy check fallout out when num of descs are equal to maximum number
• No need to use memset in vq_getchain

In D29413#658895, @philip wrote:

Approved by: philip (mentor)

But just wondering: how does this software behave when it is built on a system where sndstat.h is available but subsequently run on a system where it is not? i.e. what happens if we build a binary package on a new package builder and install the package on an older (but still supported) version of FreeBSD?

Xr VOP_SETATTR 9 -> .Xr VOP_SETATTR 9

Fix copyright notices

Manual: fix missing includes

Updated commit message.

Updated commit message.

Moved the following into this revision:

	/* Translate to posix_fallocate(2) return value as needed. */
	if (error == ENOMEM)
		error = ENOSPC;

Updated commit message.

• Move them to D28833.
• shm_deallocate: kill int error;. Just directly return (0).

Updated commit message.

In D28347#657998, @khng wrote:

For Linux's fallocate(), we currently do not have an equivalent of linux's fallocate(2)
either. In particular I do not quite like it putting FALLOC_FL_PUNCH_HOLE as a flag
despite it is actually one of the major function.

In D28347#657992, @asomers wrote:

Sorry I'm late to the party, but I have a few questions:

1. Why a new syscall? There are already at least three APIs for hole-punching. Instead of adding a fourth, can we just reuse one of these existing APIs?
   • Solaris uses fcntl() with F_FREESP. AFAIK it was the first major operating system to support hole punching.
   • Linux uses fallocate() with FALLOC_FL_PUNCH_HOLE.
   • GEOM uses ioctl() with DIOCGDELETE.

Nit

In D26209#656995, @c.koehne_beckhoff.com wrote:

I'm going to rebase this patch soon.

Rebased

Rebased

Rebased.