Enhance mdmfs(8) to work with tmpfs(5).

Existing scripts and associated config such as rc.initdiskless, rc.d/var,
and others, use mdmfs to create memory filesystems. That program accepts a
size argument which allows SI suffixes and treats an unsuffixed number as a
count of 512 byte sectors. That makes it difficult to convert existing
scripts to use tmpfs instead of mdmfs, because tmpfs treats unsuffixed
numbers as a count of bytes. The script logic to deal with existing user
config that might include suffixed and unsuffixed numbers is... unpleasant.

Also, there is no g'tee that tmpfs will be available. It is sometimes
configured out of small-resource embedded systems to save memory and flash
storage space.

These changes enhance mdmfs(8) so that it accepts two new values for the
'md-device' arg: 'tmpfs' and 'auto'. With tmpfs, the program always uses
tmpfs(5) (and fails if it's not available). With 'auto' the program prefers
tmpfs, but falls back to using md(4) if tmpfs isn't available. It also
handles the -s <size> argument so that the mdconfig interpetation of
unsuffixed numbers applies when tmpfs is used as well, so that existing user
config keeps working after a switch to tmpfs.

A new rc setting, mfs_type, is added to etc/defaults/rc.conf to let users
force the use of tmpfs or md; the default value is "auto".

Differential Revision:	https://reviews.freebsd.org/D12301

1 files changed, 53 insertions, 12 deletions

diff --git a/sbin/mdmfs/mdmfs.8 b/sbin/mdmfs/mdmfs.8
index 30a81fa49cf6..7505ff214ca8 100644
--- a/sbin/mdmfs/mdmfs.8
+++ b/sbin/mdmfs/mdmfs.8
@@ -25,7 +25,7 @@
 .\"
 .\" $FreeBSD$
 .\"
-.Dd September 9, 2016
+.Dd September 9, 2017
 .Dt MDMFS 8
 .Os
 .Sh NAME
@@ -33,7 +33,9 @@
 .Nm mount_mfs
 .Nd configure and mount an in-memory file system using the
 .Xr md 4
-driver
+driver or the
+.Xr tmpfs 5
+filesystem
 .Sh SYNOPSIS
 .Nm
 .Op Fl DLlMNnPStTUX
@@ -63,9 +65,13 @@ utility is designed to be a work-alike and look-alike of the deprecated
 .Xr mount_mfs 8 .
 The end result is essentially the same,
 but is accomplished in a completely different way.
-The
+Based on
+.Ar md-device ,
+the
 .Nm
-utility configures an
+utility either creates a 
+.Xr tmpfs 5
+filesystem, or it configures an
 .Xr md 4
 disk using
 .Xr mdconfig 8 ,
@@ -81,6 +87,44 @@ compressed disk images, as long as the kernel supports this GEOM class.
 All the command line options are passed to the appropriate program
 at the appropriate stage in order to achieve the desired effect.
 .Pp
+When
+.Ar md-device
+is `auto',
+.Nm
+uses
+.Xr tmpfs 5
+if it is present in the kernel or can be loaded as a module,
+otherwise it falls back to using
+.Xr md 4
+auto-unit as if `md' had been specified.
+.Pp
+When
+.Ar md-device
+is `tmpfs',
+.Nm
+mounts a
+.Xr tmpfs 5
+filesystem, translating the
+.Fl s
+size option, if present, into a `-o size=' mount option.
+Any
+.Fl o
+options on the command line are passed through to the
+.Xr tmpfs 5
+mount.
+Options specific to
+.Xr mdconfig 8
+or
+.Xr newfs 8
+are ignored.
+.Pp
+When
+.Ar md-device
+does not result in
+.Xr tmpfs 5
+being used, then an
+.Xr md 4
+device is configured instead.
 By default,
 .Nm
 creates a swap-based
@@ -219,14 +263,10 @@ is
 .Em not
 specified.
 That is,
-this will work for the default swap-backed
-.Pq Dv MD_SWAP
-disks,
-and the optional
-.Pq Fl M
-.Xr malloc 9
-backed disks
-.Pq Dv MD_MALLOC .
+this will work when the backing storage is some form of
+memory, as opposed to a fixed-size file.
+The size may include the usual SI suffixes (k, m, g, t, p).
+A number without a suffix is interpreted as a count of 512-byte sectors.
 .It Fl t
 Turn on the TRIM enable flag for
 .Xr newfs 8 .
@@ -392,6 +432,7 @@ was given on the command line.
 .Sh SEE ALSO
 .Xr md 4 ,
 .Xr fstab 5 ,
+.Xr tmpfs 5 ,
 .Xr mdconfig 8 ,
 .Xr mount 8 ,
 .Xr newfs 8