| Age | Commit message (Collapse) | Author |
|---|
|
Lots of small fixes
|
|
core: unit deps port to extract_first_word
|
|
core: still make progress when throttling the manager loop
|
|
|
|
Let's make sure "LimitCPU=30min" can be parsed properly, following the
usual logic how we parse time values. Similar for LimitRTTIME=.
While we are at it, extend a bit on the man page section about resource
limits.
Fixes: #1772
|
|
Let's not convert RLIM_INFINITY to "unsigned long long" and then back to
rlim_t, but let's leave it in the right type right-away.
Parse resource limits as 64 bit in all cases, as according to the man
page that's what libc does anyway.
Make sure setting a resource limit to (uint64_t) -1 results in a parsing
error, and isn't implicitly converted to RLIM_INFINITY.
|
|
Let's generate a simple error, and that's it. Let's not try to be smart
and record the last word that failed.
Also, let's make sure we don't compare numeric values with 0 by relying
on C's downgrade-to-bool feature, as suggested in CODING_STYLE.
|
|
The macro is generically useful for putting together search paths, hence
let's make it truly generic, by dropping the implicit ".d" appending it
does, and leave that to the caller. Also rename it from
CONF_DIRS_NULSTR() to CONF_PATHS_NULSTR(), since it's not strictly about
dirs that way, but any kind of file system path.
Also, mark CONF_DIR_SPLIT_USR() as internal macro by renaming it to
_CONF_PATHS_SPLIT_USR() so that the leading underscore indicates that
it's internal.
|
|
[v2] treewide: treatment of errno and other cleanups
|
|
doc: typo and ortho fixes
|
|
core/timer: in containers OnBootSec should count from container startup
|
|
core: support IEC suffixes for RLIMIT stuff
|
|
In Fedora dnf-makecache.timer specifies OnBootSec=10min, but it fired
during boot in systemd-nspawn, because the clock ticks since the host's
boot.
Fix it by treating OnBootSec the same as OnStartupSec when running in
a container.
|
|
unused since 7b2313f5
|
|
|
|
Let's make things more user-friendly and support for example
LimitAS=16G
rather than force users to always use LimitAS=16106127360.
The change is relevant for options:
[Default]Limit{FSIZE,DATA,STACK,CORE,RSS,AS,MEMLOCK,MSGQUEUE}
The patch introduces config_parse_bytes_limit(), it's the same as
config_parse_limit() but uses parse_size() tu support the suffixes.
Addresses: https://github.com/systemd/systemd/issues/1772
|
|
Our functions return negative error codes.
Do not rely on errno being set after calling our own functions.
|
|
Don't simply continue after sleeping, it potentially puts us in a state
of spinning doing nothing slowly, if the ratelimit_test() keeps
detecting the need for limiting.
Observed in vms after the host had been suspended for a while, on resume
systemd entered a loop of making zero progress spamming the console
with:
[431942.850090] systemd[1]: Looping too fast. Throttling execution a
little.
I see no reason to have a continue here, the intention should be to
throttle execution, not circumvent it altogether.
|
|
core: parse socket port to extract_first_word
|
|
|
|
* don't hide ENOMEM
* log r instead of 0
|
|
|
|
debug-generator: respect kernel parameters for default unit setting
|
|
|
|
core: drop check for /etc/mtab
|
|
For each service expose how many file descriptors there are currently in
the fd store.
(Also, fix the exporting of the fdstore limit, given that the field is
just an "unsigned" but we exported it as "uint32_t". Not that there way
any effective difference, but let's make this clean...)
|
|
Only that way it actually has an effect on all our sockets, including
$NOTIFY_SOCKET.
|
|
systemd-run can launch units with RuntimeDirectory
|
|
util-linux 2.27.1 now entirely stops looking at /etc/mtab, so we don't need to
verify /etc/mtab during early boot any more. Later on, tmpfiles.d/etc.conf will
fix /etc/mtab anyway, so there's not even a point in warning about it.
Drop test_mtab() and bump the util-linux dependency to >= 2.17.1.
Fixes #1495
|
|
Port capabiliy bounding set parsing to extract_first_word
|
|
|
|
The directive is called FileDescriptorStoreMax=, not FDStoreMax=.
|
|
Let's make sure we don't start blocking on sd_notify() earlier than
necessary, let's bump the socket buffer sizes to 8M.
We already do something similar for our logging socket buffers, hence
apply a similar bump here.
|
|
Let's make sure we don't even try to create the audit socket
|
|
|
|
|
|
|
|
|
|
bug: CapabilityBoundingSet= doesn't reset all caps
|
|
Currently, we dispatch NOTIFY messages in a tight loop. Regardless how
much data is incoming, we always dispatch everything that is queued.
This, however, completely breaks priority event-handling of sd-event.
When dispatching one NOTIFY event, another completely different event
might fire, or might be queued by the NOTIFY handling. However, this
event will not get dispatched until all other further NOTIFY messages are
handled. Those might even arrive _after_ the other event fired, and as
such completely break priority ordering of sd-event (which several code
paths rely on).
Break this by never dispatching multiple messages. Just return after each
message that was read and let sd-event handle everything else.
(The patch looks scarier that it is. It basically just drops the for(;;)
loop and re-indents the loop-content.)
|
|
fixup for previous pr
|
|
|
|
When coldplugging the unit state, make sure to follow the same basic
logic for all unit types: always verify whether the control PID is still
a waitable process before proceeding.
|
|
Since 5fd9b2c5467b0a42ccdabc7eb8e516d512609a8e passing a pid of 0 to
pid_is_unwaited() and pid_is_live() is considered as a request on the
current process, similar how the other calls in process-util.c handle a
PID of 0. This broke service.c, which passes a 0 PID and expects it to
be considered an unwaited process.
This fix make sure we can boot again.
|
|
core: fix CapabilityBoundingSet merging
|
|
|
|
|
|
|
|
In order to match the other get_process_xyz() calls.
|
|
|
