WaLyong Cho [Thu, 15 May 2014 15:09:34 +0000 (00:09 +0900)]
core: add startup resource control option
Similar to CPUShares= and BlockIOWeight= respectively. However only
assign the specified weight during startup. Each control group
attribute is re-assigned as weight by CPUShares=weight and
BlockIOWeight=weight after startup. If not CPUShares= or
BlockIOWeight= be specified, then the attribute is re-assigned to each
default attribute value. (default cpu.shares=1024, blkio.weight=1000)
If only CPUShares=weight or BlockIOWeight=weight be specified, then
that implies StartupCPUShares=weight and StartupBlockIOWeight=weight.
This introduces SYSTEM_UID_MAX defined to the maximum UID of system
users. This value is determined compile-time, either as configure switch
or from /etc/login.defs. (We don't read that file at runtime, since this
is really a choice for a system builder, not the end user.)
While we are at it we then also update journald to use SYSTEM_UID_MAX
when we decide whether to split out log data for a specific client.
Tom Gundersen [Tue, 20 May 2014 11:07:19 +0000 (13:07 +0200)]
sd-dhcp: refactor parse_options
Similar to the previous patch, exchange a length and a pointer with only one offset variable.
Also fix the type of the options to be uint8_t[], rather than uint8_t*.
Tom Gundersen [Tue, 20 May 2014 09:04:50 +0000 (11:04 +0200)]
sd-dhcp: refactor dhcp_option_append
Store a pointer to the options in the DHCPMessage struct, and pass
this together with an offset around, rather than a uint8_t**.
This avoids us having to (re)compute the pointer; and changes
dhcp_option_append from adjusting both the pointer to the next
option and the remaining size of the options, to just adjusting
the current offset.
This makes the code a bit simpler to follow IMHO, but there should
be no functional change.
timesyncd: save clock to disk everytime we get an NTP fix, and bump clock at boot using this
This is useful to make sure the system clock stays monotonic even on
systems that lack an RTC.
Also, why we are at it, also use the systemd release time for bumping
the clock, since it's a slightly less bad than starting with jan 1st,
1970.
This also moves timesyncd into the early bootphase, in order to make
sure this initial bump is guaranteed to have finished by the time we
start real daemons which might write to the file systemd and thus
shouldn't leave 1970's timestamps all over the place...
Harald Hoyer [Tue, 20 May 2014 10:25:16 +0000 (12:25 +0200)]
udevadm-settle: fixed return code for empty queue
If the udev queue is empty and "/run/udev/queue" does not exist,
"udevadm settle" would return with EXIT_FAILURE, because the inotify on
"/run/udev/queue" would fail with ENOENT.
This patch lets "udevadm settle" exit with EXIT_SUCCESS in this case.
When we dropped support for creating a per-user to the "main" X11
display we stopped returning useful data in the "Display" user property.
With this change this is fixed and we again expose an appropriate
(graphical session) in the property that is useful as the "main" one, if
one is needed.
Will Woods [Fri, 25 Apr 2014 22:26:34 +0000 (18:26 -0400)]
core: let selinux_setup() load policy more than once
When you switch-root into a new root that has SELinux policy, you're
supposed to to run selinux_init_load_policy() to set up SELinux and load
policy. Normally this gets handled by selinux_setup().
But if SELinux was already initialized, selinux_setup() skips loading
policy and returns 0. So if you load policy normally, and then you
switch-root to a new root that has new policy, selinux_setup() never
loads the new policy. What gives?
As far as I can tell, this check is an artifact of how selinux_setup()
worked when it was first written (see commit c4dcdb9 / systemd v12):
* when systemd starts, run selinux_setup()
* if selinux_setup() loads policy OK, restart systemd
So the "if policy already loaded, skip load and return 0" check was
there to prevent an infinite re-exec loop.
Modern systemd only calls selinux_setup() on initial load and after
switch-root, and selinux_setup() no longer restarts systemd, so we don't
need that check to guard against the infinite loop anymore.
So: this patch removes the "return 0", thus allowing selinux_setup() to
actually perform SELinux setup after switch-root.
We still want to check to see if SELinux is initialized, because if
selinux_init_load_policy() fails *but* SELinux is initialized that means
we still have (old) policy active. So we don't need to halt if
enforce=1.
With proprietary graphics drivers, there won't be any 'drm' devices in
sysfs, so logind will never suspend the system upon closing the lid,
even if only one (internal) display is connected. This has been reported
by multiple users so far.
IMHO, it's better to suspend the system in this case for safety reasons,
to avoid having nvidia blob users' laptops overheat, for the same reason
that sleep inhibitors are overridden (LidSwitchIgnoreInhibited=yes).
Tanu Kaskinen [Sat, 3 May 2014 08:52:12 +0000 (11:52 +0300)]
path-util: add path_make_relative()
In user_dirs() in path-lookup.c, I want to replace this:
symlink("../../../.config/systemd/user", data_home);
with
symlink(config_home, data_home);
to avoid hardcoding .config when XDG_CONFIG_HOME is set.
The problem is that config_home is an absolute path, and it's better
to make the symlink relative. path_make_relative() is an utility
function that converts an absolute path into a relative one.
fsck: Allow to specify the fsck repair option in the cmdline
Some unattended systems do not have a console attached and entering
the default rescue mode will not be too helpful. Allow to specify
the "-y" option to attempt to fix all filesystem errors.
Manually verified by downloading an image.gz of e2fsprogs, using
losetup and running systemd-fsck on the loop device and varying
the fsck.repair=preen|yes|no option.
projects / thirdparty / systemd.git / log
