depot/third_party/nixpkgs/nixos/modules/tasks/filesystems.nix
Default email c7cb07f092 Project import generated by Copybara.
GitOrigin-RevId: 1536926ef5621b09bba54035ae2bb6d806d72ac8
2024-02-29 21:09:43 +01:00

452 lines
16 KiB
Nix
Raw Blame History

This file contains ambiguous Unicode characters

This file contains Unicode characters that might be confused with other characters. If you think that this is intentional, you can safely ignore this warning. Use the Escape button to reveal them.

{ config, lib, pkgs, utils, ... }:
with lib;
with utils;
let
addCheckDesc = desc: elemType: check: types.addCheck elemType check
// { description = "${elemType.description} (with check: ${desc})"; };
isNonEmpty = s: (builtins.match "[ \t\n]*" s) == null;
nonEmptyStr = addCheckDesc "non-empty" types.str isNonEmpty;
fileSystems' = toposort fsBefore (attrValues config.fileSystems);
fileSystems = if fileSystems' ? result
then # use topologically sorted fileSystems everywhere
fileSystems'.result
else # the assertion below will catch this,
# but we fall back to the original order
# anyway so that other modules could check
# their assertions too
(attrValues config.fileSystems);
specialFSTypes = [ "proc" "sysfs" "tmpfs" "ramfs" "devtmpfs" "devpts" ];
nonEmptyWithoutTrailingSlash = addCheckDesc "non-empty without trailing slash" types.str
(s: isNonEmpty s && (builtins.match ".+/" s) == null);
coreFileSystemOpts = { name, config, ... }: {
options = {
mountPoint = mkOption {
example = "/mnt/usb";
type = nonEmptyWithoutTrailingSlash;
description = lib.mdDoc "Location of the mounted file system.";
};
stratis.poolUuid = lib.mkOption {
type = types.uniq (types.nullOr types.str);
description = lib.mdDoc ''
UUID of the stratis pool that the fs is located in
'';
example = "04c68063-90a5-4235-b9dd-6180098a20d9";
default = null;
};
device = mkOption {
default = null;
example = "/dev/sda";
type = types.nullOr nonEmptyStr;
description = lib.mdDoc "Location of the device.";
};
fsType = mkOption {
default = "auto";
example = "ext3";
type = nonEmptyStr;
description = lib.mdDoc "Type of the file system.";
};
options = mkOption {
default = [ "defaults" ];
example = [ "data=journal" ];
description = lib.mdDoc "Options used to mount the file system.";
type = types.nonEmptyListOf nonEmptyStr;
};
depends = mkOption {
default = [ ];
example = [ "/persist" ];
type = types.listOf nonEmptyWithoutTrailingSlash;
description = lib.mdDoc ''
List of paths that should be mounted before this one. This filesystem's
{option}`device` and {option}`mountPoint` are always
checked and do not need to be included explicitly. If a path is added
to this list, any other filesystem whose mount point is a parent of
the path will be mounted before this filesystem. The paths do not need
to actually be the {option}`mountPoint` of some other filesystem.
'';
};
};
config = {
mountPoint = mkDefault name;
device = mkIf (elem config.fsType specialFSTypes) (mkDefault config.fsType);
};
};
fileSystemOpts = { config, ... }: {
options = {
label = mkOption {
default = null;
example = "root-partition";
type = types.nullOr nonEmptyStr;
description = lib.mdDoc "Label of the device (if any).";
};
autoFormat = mkOption {
default = false;
type = types.bool;
description = lib.mdDoc ''
If the device does not currently contain a filesystem (as
determined by {command}`blkid`, then automatically
format it with the filesystem type specified in
{option}`fsType`. Use with caution.
'';
};
formatOptions = mkOption {
visible = false;
type = types.unspecified;
default = null;
};
autoResize = mkOption {
default = false;
type = types.bool;
description = lib.mdDoc ''
If set, the filesystem is grown to its maximum size before
being mounted. (This is typically the size of the containing
partition.) This is currently only supported for ext2/3/4
filesystems that are mounted during early boot.
'';
};
noCheck = mkOption {
default = false;
type = types.bool;
description = lib.mdDoc "Disable running fsck on this filesystem.";
};
};
config.options = mkMerge [
(mkIf config.autoResize [ "x-systemd.growfs" ])
(mkIf config.autoFormat [ "x-systemd.makefs" ])
(mkIf (utils.fsNeededForBoot config) [ "x-initrd.mount" ])
];
};
# Makes sequence of `specialMount device mountPoint options fsType` commands.
# `systemMount` should be defined in the sourcing script.
makeSpecialMounts = mounts:
pkgs.writeText "mounts.sh" (concatMapStringsSep "\n" (mount: ''
specialMount "${mount.device}" "${mount.mountPoint}" "${concatStringsSep "," mount.options}" "${mount.fsType}"
'') mounts);
makeFstabEntries =
let
fsToSkipCheck = [
"none"
"auto"
"overlay"
"iso9660"
"bindfs"
"udf"
"btrfs"
"zfs"
"tmpfs"
"bcachefs"
"nfs"
"nfs4"
"nilfs2"
"vboxsf"
"squashfs"
"glusterfs"
"apfs"
"9p"
"cifs"
"prl_fs"
"vmhgfs"
] ++ lib.optionals (!config.boot.initrd.checkJournalingFS) [
"ext3"
"ext4"
"reiserfs"
"xfs"
"jfs"
"f2fs"
];
isBindMount = fs: builtins.elem "bind" fs.options;
skipCheck = fs: fs.noCheck || fs.device == "none" || builtins.elem fs.fsType fsToSkipCheck || isBindMount fs;
# https://wiki.archlinux.org/index.php/fstab#Filepath_spaces
escape = string: builtins.replaceStrings [ " " "\t" ] [ "\\040" "\\011" ] string;
in fstabFileSystems: { }: concatMapStrings (fs:
(if fs.device != null then escape fs.device
else if fs.label != null then "/dev/disk/by-label/${escape fs.label}"
else throw "No device specified for mount point ${fs.mountPoint}.")
+ " " + escape fs.mountPoint
+ " " + fs.fsType
+ " " + escape (builtins.concatStringsSep "," fs.options)
+ " 0 " + (if skipCheck fs then "0" else if fs.mountPoint == "/" then "1" else "2")
+ "\n"
) fstabFileSystems;
initrdFstab = pkgs.writeText "initrd-fstab" (makeFstabEntries (filter utils.fsNeededForBoot fileSystems) { });
in
{
###### interface
options = {
fileSystems = mkOption {
default = {};
example = literalExpression ''
{
"/".device = "/dev/hda1";
"/data" = {
device = "/dev/hda2";
fsType = "ext3";
options = [ "data=journal" ];
};
"/bigdisk".label = "bigdisk";
}
'';
type = types.attrsOf (types.submodule [coreFileSystemOpts fileSystemOpts]);
description = lib.mdDoc ''
The file systems to be mounted. It must include an entry for
the root directory (`mountPoint = "/"`). Each
entry in the list is an attribute set with the following fields:
`mountPoint`, `device`,
`fsType` (a file system type recognised by
{command}`mount`; defaults to
`"auto"`), and `options`
(the mount options passed to {command}`mount` using the
{option}`-o` flag; defaults to `[ "defaults" ]`).
Instead of specifying `device`, you can also
specify a volume label (`label`) for file
systems that support it, such as ext2/ext3 (see {command}`mke2fs -L`).
'';
};
system.fsPackages = mkOption {
internal = true;
default = [ ];
description = lib.mdDoc "Packages supplying file system mounters and checkers.";
};
boot.supportedFilesystems = mkOption {
default = { };
example = lib.literalExpression ''
{
btrfs = true;
zfs = lib.mkForce false;
}
'';
type = types.coercedTo
(types.listOf types.str)
(enabled: lib.listToAttrs (map (fs: lib.nameValuePair fs true) enabled))
(types.attrsOf types.bool);
description = lib.mdDoc ''
Names of supported filesystem types, or an attribute set of file system types
and their state. The set form may be used together with `lib.mkForce` to
explicitly disable support for specific filesystems, e.g. to disable ZFS
with an unsupported kernel.
'';
};
boot.specialFileSystems = mkOption {
default = {};
type = types.attrsOf (types.submodule coreFileSystemOpts);
internal = true;
description = lib.mdDoc ''
Special filesystems that are mounted very early during boot.
'';
};
boot.devSize = mkOption {
default = "5%";
example = "32m";
type = types.str;
description = lib.mdDoc ''
Size limit for the /dev tmpfs. Look at mount(8), tmpfs size option,
for the accepted syntax.
'';
};
boot.devShmSize = mkOption {
default = "50%";
example = "256m";
type = types.str;
description = lib.mdDoc ''
Size limit for the /dev/shm tmpfs. Look at mount(8), tmpfs size option,
for the accepted syntax.
'';
};
boot.runSize = mkOption {
default = "25%";
example = "256m";
type = types.str;
description = lib.mdDoc ''
Size limit for the /run tmpfs. Look at mount(8), tmpfs size option,
for the accepted syntax.
'';
};
};
###### implementation
config = {
assertions = let
ls = sep: concatMapStringsSep sep (x: x.mountPoint);
resizableFSes = [
"ext3"
"ext4"
"btrfs"
"xfs"
];
notAutoResizable = fs: fs.autoResize && !(builtins.elem fs.fsType resizableFSes);
in [
{ assertion = ! (fileSystems' ? cycle);
message = "The fileSystems option can't be topologically sorted: mountpoint dependency path ${ls " -> " fileSystems'.cycle} loops to ${ls ", " fileSystems'.loops}";
}
{ assertion = ! (any notAutoResizable fileSystems);
message = let
fs = head (filter notAutoResizable fileSystems);
in ''
Mountpoint '${fs.mountPoint}': 'autoResize = true' is not supported for 'fsType = "${fs.fsType}"'
${optionalString (fs.fsType == "auto") "fsType has to be explicitly set and"}
only the following support it: ${lib.concatStringsSep ", " resizableFSes}.
'';
}
{
assertion = ! (any (fs: fs.formatOptions != null) fileSystems);
message = let
fs = head (filter (fs: fs.formatOptions != null) fileSystems);
in ''
'fileSystems.<name>.formatOptions' has been removed, since
systemd-makefs does not support any way to provide formatting
options.
'';
}
];
# Export for use in other modules
system.build.fileSystems = fileSystems;
system.build.earlyMountScript = makeSpecialMounts (toposort fsBefore (attrValues config.boot.specialFileSystems)).result;
boot.supportedFilesystems = map (fs: fs.fsType) fileSystems;
# Add the mount helpers to the system path so that `mount' can find them.
system.fsPackages = [ pkgs.dosfstools ];
environment.systemPackages = with pkgs; [ fuse3 fuse ] ++ config.system.fsPackages;
environment.etc.fstab.text =
let
swapOptions = sw: concatStringsSep "," (
sw.options
++ optional (sw.priority != null) "pri=${toString sw.priority}"
++ optional (sw.discardPolicy != null) "discard${optionalString (sw.discardPolicy != "both") "=${toString sw.discardPolicy}"}"
);
in ''
# This is a generated file. Do not edit!
#
# To make changes, edit the fileSystems and swapDevices NixOS options
# in your /etc/nixos/configuration.nix file.
#
# <file system> <mount point> <type> <options> <dump> <pass>
# Filesystems.
${makeFstabEntries fileSystems {}}
# Swap devices.
${flip concatMapStrings config.swapDevices (sw:
"${sw.realDevice} none swap ${swapOptions sw}\n"
)}
'';
boot.initrd.systemd.storePaths = [initrdFstab];
boot.initrd.systemd.managerEnvironment.SYSTEMD_SYSROOT_FSTAB = initrdFstab;
boot.initrd.systemd.services.initrd-parse-etc.environment.SYSTEMD_SYSROOT_FSTAB = initrdFstab;
# Provide a target that pulls in all filesystems.
systemd.targets.fs =
{ description = "All File Systems";
wants = [ "local-fs.target" "remote-fs.target" ];
};
systemd.services = {
# Mount /sys/fs/pstore for evacuating panic logs and crashdumps from persistent storage onto the disk using systemd-pstore.
# This cannot be done with the other special filesystems because the pstore module (which creates the mount point) is not loaded then.
"mount-pstore" = {
serviceConfig = {
Type = "oneshot";
# skip on kernels without the pstore module
ExecCondition = "${pkgs.kmod}/bin/modprobe -b pstore";
ExecStart = pkgs.writeShellScript "mount-pstore.sh" ''
set -eu
# if the pstore module is builtin it will have mounted the persistent store automatically. it may also be already mounted for other reasons.
${pkgs.util-linux}/bin/mountpoint -q /sys/fs/pstore || ${pkgs.util-linux}/bin/mount -t pstore -o nosuid,noexec,nodev pstore /sys/fs/pstore
# wait up to 1.5 seconds for the backend to be registered and the files to appear. a systemd path unit cannot detect this happening; and succeeding after a restart would not start dependent units.
TRIES=15
while [ "$(cat /sys/module/pstore/parameters/backend)" = "(null)" ]; do
if (( $TRIES )); then
sleep 0.1
TRIES=$((TRIES-1))
else
echo "Persistent Storage backend was not registered in time." >&2
break
fi
done
'';
RemainAfterExit = true;
};
unitConfig = {
ConditionVirtualization = "!container";
DefaultDependencies = false; # needed to prevent a cycle
};
before = [ "systemd-pstore.service" "shutdown.target" ];
conflicts = [ "shutdown.target" ];
wantedBy = [ "systemd-pstore.service" ];
};
};
systemd.tmpfiles.rules = [
"d /run/keys 0750 root ${toString config.ids.gids.keys}"
"z /run/keys 0750 root ${toString config.ids.gids.keys}"
];
# Sync mount options with systemd's src/core/mount-setup.c: mount_table.
boot.specialFileSystems = {
"/proc" = { fsType = "proc"; options = [ "nosuid" "noexec" "nodev" ]; };
"/run" = { fsType = "tmpfs"; options = [ "nosuid" "nodev" "strictatime" "mode=755" "size=${config.boot.runSize}" ]; };
"/dev" = { fsType = "devtmpfs"; options = [ "nosuid" "strictatime" "mode=755" "size=${config.boot.devSize}" ]; };
"/dev/shm" = { fsType = "tmpfs"; options = [ "nosuid" "nodev" "strictatime" "mode=1777" "size=${config.boot.devShmSize}" ]; };
"/dev/pts" = { fsType = "devpts"; options = [ "nosuid" "noexec" "mode=620" "ptmxmode=0666" "gid=${toString config.ids.gids.tty}" ]; };
# To hold secrets that shouldn't be written to disk
"/run/keys" = { fsType = "ramfs"; options = [ "nosuid" "nodev" "mode=750" ]; };
} // optionalAttrs (!config.boot.isContainer) {
# systemd-nspawn populates /sys by itself, and remounting it causes all
# kinds of weird issues (most noticeably, waiting for host disk device
# nodes).
"/sys" = { fsType = "sysfs"; options = [ "nosuid" "noexec" "nodev" ]; };
};
};
}