xref: /freebsd/share/man/man4/md.4 (revision 123af6ec70016f5556da5972d4d63c7d175c06d3)
1.\" ----------------------------------------------------------------------------
2.\" "THE BEER-WARE LICENSE" (Revision 42):
3.\" <phk@FreeBSD.org> wrote this file.  As long as you retain this notice you
4.\" can do whatever you want with this stuff. If we meet some day, and you think
5.\" this stuff is worth it, you can buy me a beer in return.   Poul-Henning Kamp
6.\" ----------------------------------------------------------------------------
7.\"
8.\" $FreeBSD$
9.\"
10.Dd December 26, 2017
11.Dt MD 4
12.Os
13.Sh NAME
14.Nm md
15.Nd memory disk
16.Sh SYNOPSIS
17To compile this driver into the kernel,
18place the following lines in your
19kernel configuration file:
20.Bd -ragged -offset indent
21.Cd "device md"
22.Ed
23.Pp
24Alternatively, to load the driver as a
25module at boot time, place the following line in
26.Xr loader.conf 5 :
27.Bd -literal -offset indent
28geom_md_load="YES"
29.Ed
30.Sh DESCRIPTION
31The
32.Nm
33driver provides support for four kinds of memory backed virtual disks:
34.Bl -tag -width preload
35.It Cm malloc
36Backing store is allocated using
37.Xr malloc 9 .
38Only one malloc-bucket is used, which means that all
39.Nm
40devices with
41.Cm malloc
42backing must share the malloc-per-bucket-quota.
43The exact size of this quota varies, in particular with the amount
44of RAM in the
45system.
46The exact value can be determined with
47.Xr vmstat 8 .
48.It Cm preload
49A file loaded by
50.Xr loader 8
51with type
52.Sq md_image
53is used for backing store.
54For backwards compatibility the type
55.Sq mfs_root
56is also recognized.
57If the kernel is created with option
58.Dv MD_ROOT
59the first preloaded image found will become the root file system.
60.It Cm vnode
61A regular file is used as backing store.
62This allows for mounting ISO images without the tedious
63detour over actual physical media.
64.It Cm swap
65Backing store is allocated from buffer memory.
66Pages get pushed out to the swap when the system is under memory
67pressure, otherwise they stay in the operating memory.
68Using
69.Cm swap
70backing is generally preferable over
71.Cm malloc
72backing.
73.El
74.Pp
75For more information, please see
76.Xr mdconfig 8 .
77.Sh EXAMPLES
78To create a kernel with a ramdisk or MD file system, your kernel config
79needs the following options:
80.Bd -literal -offset indent
81options 	MD_ROOT			# MD is a potential root device
82options 	MD_ROOT_READONLY	# disallow mounting root writeable
83options 	MD_ROOT_SIZE=8192	# 8MB ram disk
84makeoptions	MFS_IMAGE=/h/foo/ARM-MD
85options 	ROOTDEVNAME=\\"ufs:md0\\"
86.Ed
87.Pp
88The image in
89.Pa /h/foo/ARM-MD
90will be loaded as the initial image each boot.
91To create the image to use, please follow the steps to create a file-backed
92disk found in the
93.Xr mdconfig 8
94man page.
95Other tools will also create these images, such as NanoBSD.
96.Sh ARM KERNEL OPTIONS
97On armv6 and armv7 architectures, an MD_ROOT image larger than
98approximately 55 MiB may require building a custom kernel using
99several tuning options related to kernel memory usage.
100.Bl -tag -width indent
101.It Cd options LOCORE_MAP_MB=<num>
102This configures how much memory is mapped for the kernel during
103the early initialization stages.
104The value must be at least as large as the kernel plus all preloaded
105modules, including the root image.
106There is no downside to setting this value too large, as long
107as it does not exceed the amount of physical memory.
108The default is 64 MiB.
109.It Cd options NKPT2PG=<num>
110This configures the number of kernel L2 page table pages to
111preallocate during kernel initialization.
112Each L2 page can map 4 MiB of kernel space.
113The value must be large enough to map the kernel plus all preloaded
114modules, including the root image.
115The default value is 32, which is sufficient to map 128 MiB.
116.It Cd options VM_KMEM_SIZE_SCALE=<num>
117This configures the amount of kernel virtual address (KVA) space to
118dedicate to the kmem_arena map.
119The scale value is the ratio of physical to virtual pages.
120The default value of 3 allocates a page of KVA for each 3 pages
121of physical ram in the system.
122
123The kernel and modules, including the root image, also consume KVA.
124The combination of a large root image and the default scaling
125may preallocate so much KVA that there is not enough
126remaining address space to allocate kernel stacks, IO buffers,
127and other resources that are not part of kmem_arena.
128Overallocating kmem_arena space is likely to manifest as failure to
129launch userland processes with "cannot allocate kernel stack" messages.
130
131Setting the scale value too high may result in kernel failure to allocate
132memory because kmem_arena is too small, and the failure may require
133significant runtime to manifest.
134Empirically, a value of 5 works well for a 200 MiB root image on
135a system with 2 GiB of physical ram.
136.El
137.Sh SEE ALSO
138.Xr gpart 8 ,
139.Xr loader 8 ,
140.Xr mdconfig 8 ,
141.Xr mdmfs 8 ,
142.Xr newfs 8 ,
143.Xr vmstat 8
144.Sh HISTORY
145The
146.Nm
147driver first appeared in
148.Fx 4.0
149as a cleaner replacement
150for the MFS functionality previously used in
151.Tn PicoBSD
152and in the
153.Fx
154installation process.
155.Pp
156The
157.Nm
158driver did a hostile takeover of the
159.Xr vn 4
160driver in
161.Fx 5.0 .
162.Sh AUTHORS
163The
164.Nm
165driver was written by
166.An Poul-Henning Kamp Aq Mt phk@FreeBSD.org .
167