Introducing a default limit on number of threads broke a lot of software which
regularly needs more, such as MySQL and RabbitMQ, or services that spawn off an
indefinite number of subtasks that are not in a scope, like LXC or cron.
15% is way too much for most "simple" services, and it's too little for others
such as the ones mentioned above. There is also no particular rationale about
any particular global limit, so even if we'd bump it higher we'd just make the
limit even less useful while still breaking software.
It is both much safer and also much more effective in terms of guarding against
berserk programs/bugs/unintended fork bombs etc. to set limits in units
individually. Once someone looks at one, this is then a great time to also flip
on the other resource and privilege limitations that systemd offers.
Martin Pitt [Sat, 27 Feb 2016 11:27:06 +0000 (12:27 +0100)]
Revert "core: set RLIMIT_CORE to unlimited by default"
Partially revert commit 15a900327ab as this completely breaks core dumps
without systemd-coredump. It's also contradicting core(8), and it's not
systemd's place to redefine the kernel definitions of core files.
Commit bdfd7b2c now honours the process' RLIMIT_CORE for systemd-coredump. This
isn't what RLIMIT_CORE is supposed to do (it limits the size of the core
*file*, but the kernel deliberately ignores it for piping), so set a static
2^63 core size limit for systemd-coredump to go back to the previous behaviour
(otherwise the change above would break systemd-coredump).
Bug-Debian: https://bugs.debian.org/815020
Gbp-Pq: Topic debian
Gbp-Pq: Name Revert-core-set-RLIMIT_CORE-to-unlimited-by-default.patch
Martin Pitt [Mon, 27 Apr 2015 13:29:13 +0000 (15:29 +0200)]
Revert "core: one step back again, for nspawn we actually can't wait for cgroups running empty since systemd will get exactly zero notifications about it"
Didier Roche [Fri, 22 May 2015 11:04:38 +0000 (13:04 +0200)]
fsckd daemon for inter-fsckd communication
Global logic:
Add systemd-fsckd multiplexer which accepts multiple (via systemd-fsck's
/run/systemd/fsck.progress socket) fsck instances to connect to it and sends
progress report. systemd-fsckd then computes and writes to /dev/console the
number of devices currently being checked and the minimum fsck progress.
Plymouth and user interaction:
Forward the progress to plymouth and support canellation of in progress fsck.
Try to connect and send to plymouth (if running) some checked report progress,
using direct plymouth protocole.
Update message is the following:
fsckd:<num_devices>:<progress>:<string>
* num_devices corresponds to the current number of devices being checked (int)
* progress corresponds to the current minimum percentage of all devices being
checked (float, from 0 to 100)
* string is a translated message ready to be displayed by the plymouth theme
displaying the information above. It can be overridden by plymouth themes
supporting i18n.
Grab in fsckd plymouth watch key Control+C, and propagate this cancel request
to systemd-fsck which will terminate fsck.
Send a message to signal to user what key we are grabbing for fsck cancel.
Message is: fsckd-cancel-msg:<string>
Where string is a translated string ready to be displayed by the plymouth theme
indicating that Control+C can be used to cancel current checks. It can be
overridden (matching only fsckd-cancel-msg prefix) for themes supporting i18n.
Misc:
systemd-fsckd stops on idle when no fsck is connected.
Add man page explaining the plymouth theme protocol, usage of the daemon
as well as the socket activation part. Adapt existing fsck man page.
Note that fsckd had lived in the upstream tree for a while, but was removed.
More information at
http://lists.freedesktop.org/archives/systemd-devel/2015-April/030175.html
-
Gbp-Pq: Topic debian
Gbp-Pq: Name fsckd-daemon-for-inter-fsckd-communication.patch
Martin Pitt [Mon, 9 Feb 2015 09:53:43 +0000 (10:53 +0100)]
Only start logind if dbus is installed
logind fails to start in environments without dbus, such as LXC containers or
servers. Add a startup condition to avoid the very noisy startup failure.
Part of #772700
Gbp-Pq: Topic debian
Gbp-Pq: Name Only-start-logind-if-dbus-is-installed.patch
for details. Once we grow a journal.conf.d/ directory, sysloggers can be moved
to pulling from the journal one by one and disable forwarding again in such a
conf.d snippet.
Gbp-Pq: Topic debian
Gbp-Pq: Name Re-enable-journal-forwarding-to-syslog.patch
systemd does not support non-mainline kernel features so upstream rejected this
patch.
It is however required for systemd integration by tuxonice-userui package.
We need to keep supporting systems with 75-persistent-net-generator.rules
generated names for a while after switching to net.ifnames. Re-apply this old
hack to make the renaming less likely to fail.
Gbp-Pq: Topic debian
Gbp-Pq: Name Revert-udev-network-device-renaming-immediately-give.patch
Michael Biebl [Thu, 4 Sep 2014 23:15:16 +0000 (01:15 +0200)]
Make /run/lock tmpfs an API fs
The /run/lock directory is world-writable in Debian due to historic
reasons. To avoid user processes filling up /run, we mount a separate
tmpfs for /run/lock. As this directory needs to be available during
early boot, we make it an API fs.
Drop it from tmpfiles.d/legacy.conf to not clobber the permissions.
Closes: #751392
Gbp-Pq: Topic debian
Gbp-Pq: Name Make-run-lock-tmpfs-an-API-fs.patch
Michael Biebl [Thu, 10 Jan 2019 11:58:27 +0000 (12:58 +0100)]
meson: stop setting -fPIE globally
Setting -fPIE globally can lead to miscompilations on certain
architectures.
This is caused by both -fPIE and -fPIC options being added to various
compilation commands. Only -fPIC is being recorded in the LTO options
section of the object. The gcc-8 LTO plugin merges -fPIC + -fPIE to
nothing. So, the compilations done by the plugin are not
position-independent and fail to link with -pie.
The simplest solution is to stop setting -fPIE globally and instead
using meson's b_pie=true option. This requires meson 0.49 or later.
Since we don't set this option in meson.build but leave it up to the
distro maintainer to set this option, do not bump the meson version
requirement.
Yu Watanabe [Fri, 11 Jan 2019 20:24:54 +0000 (05:24 +0900)]
sd-device-monitor: fix ordering of setting buffer size
By b1c097af8df58a94cba031a347061b7cb9b62d9b (#10239), the receive buffer
size for uevents was set by SO_RCVBUF at first, and fallback to
use SO_RCVBUFFORCE. So, as SO_RCVBUF limits to the buffer size
net.core.rmem_max, which is usually much smaller than 128MB udevd requests,
uevents buffer size was not sufficient.
This fixes the ordering of the request: SO_RCVBUFFORCE first, and
fallback to SO_RCVBUF. Then, udevd's uevent buffer size can be set to
128MB.
udev: open control and netlink sockets before daemonization
c4b69e990f962128cc6975e36e91e9ad838fa2c4 effectively moved the initalization of socket.
Before that commit:
run → listen_fds → udev_ctrl_new → udev_ctrl_new_from_fd → socket()
After:
run → main_loop → manager_new → udev_ctrl_new_from_fd → socket()
The problem is that main_loop was called after daemonization. Move manager_new
out of main_loop and before daemonization.
Fixes #11314 (hopefully ;)).
v2: Yu Watanabe
sd_event is initialized in main_loop().
journal-remote: set a limit on the number of fields in a message
Existing use of E2BIG is replaced with ENOBUFS (entry too long), and E2BIG is
reused for the new error condition (too many fields).
This matches the change done for systemd-journald, hence forming the second
part of the fix for CVE-2018-16865
(https://bugzilla.redhat.com/show_bug.cgi?id=1653861).
Calling mhd_respond(), which ulimately calls MHD_queue_response() is
ineffective at point, becuase MHD_queue_response() immediately returns
MHD_NO signifying an error, because the connection is in state
MHD_CONNECTION_CONTINUE_SENT.
As Christian Grothoff kindly explained:
> You are likely calling MHD_queue_repsonse() too late: once you are
> receiving upload_data, HTTP forces you to process it all. At this time,
> MHD has already sent "100 continue" and cannot take it back (hence you
> get MHD_NO!).
>
> In your request handler, the first time when you are called for a
> connection (and when hence *upload_data_size == 0 and upload_data ==
> NULL) you must check the content-length header and react (with
> MHD_queue_response) based on this (to prevent MHD from automatically
> generating 100 continue).
If we ever encounter this kind of error, print a warning and immediately
abort the connection. (The alternative would be to keep reading the data,
but ignore it, and return an error after we get to the end of data.
That is possible, but of course puts additional load on both the
sender and reciever, and doesn't seem important enough just to return
a good error message.)
Note that sending of the error does not work (the connection is always aborted
when MHD_queue_response is used with MHD_RESPMEM_MUST_FREE, as in this case)
with libµhttpd 0.59, but works with 0.61:
https://src.fedoraproject.org/rpms/libmicrohttpd/pull-request/1
journald: lower the maximum entry size limit to ½ for non-sealed fds
We immediately read the whole contents into memory, making thigs much more
expensive. Sealed fds should be used instead since they are more efficient
on our side.
journald: when processing a native message, bail more quickly on overbig messages
We'd first parse all or most of the message, and only then consider if it
is not too large. Also, when encountering a single field over the limit,
we'd still process the preceding part of the message. Let's be stricter,
and check size limits early, and let's refuse the whole message if it fails
any of the size limits.
journald: set a limit on the number of fields (1k)
We allocate a iovec entry for each field, so with many short entries,
our memory usage and processing time can be large, even with a relatively
small message size. Let's refuse overly long entries.
What from I can see, the problem is not from an alloca, despite what the CVE
description says, but from the attack multiplication that comes from creating
many very small iovecs: (void* + size_t) for each three bytes of input message.
coredump: fix message when we fail to save a journald coredump
If creation of the message failed, we'd write a bogus entry:
systemd-coredump[1400]: Cannot store coredump of 416 (systemd-journal): No space left on device
systemd-coredump[1400]: MESSAGE=Process 416 (systemd-journal) of user 0 dumped core.
systemd-coredump[1400]: Coredump diverted to
2MB might be hard for some clients to use meaningfully, but OTOH, it is
important to log the full commandline sometimes. For example, when the program
is crashing, the exact argument list is useful.
journald: do not store the iovec entry for process commandline on stack
This fixes a crash where we would read the commandline, whose length is under
control of the sending program, and then crash when trying to create a stack
allocation for it.
coredump: remove duplicate MESSAGE= prefix from message
systemd-coredump[9982]: MESSAGE=Process 771 (systemd-journal) of user 0 dumped core.
systemd-coredump[9982]: Coredump diverted to /var/lib/systemd/coredump/core...
log_dispatch() calls log_dispatch_internal() which calls write_to_journal()
which appends MESSAGE= on its own.
udev: rework how we handle the return value from spawned programs
When running PROGRAM="...", we would log
systemd-udevd[447]: Failed to wait spawned command '...': Input/output error
no matter why the program actually failed, at error level.
The code wouldn't distinguish between an internal failure and a failure in the
program being called and run sd_event_exit(..., -EIO) on any kind of error. EIO
is rather misleading here, becuase it suggests a serious error.
on_spawn_sigchld is updated to set the return code to distinguish failure to
spawn, including the program being killed by a signal (a negative return value),
and the program failing (positive return value).
The logging levels are adjusted, so that for PROGRAM= calls, which are
essentially "if" statements, we only log at debug level (unless we get a
timeout or segfault or another unexpected error).
Removing the environment variable causes problems, e.g. Xfce and Chromium and
... don't communicate with the running dbus instance. If they attempt to start their
own instance, things become even more confusing. Those packages could be fixed
one by one, but removing the variable right now is causing too many problems.
YunQiang Su [Tue, 25 Dec 2018 11:01:17 +0000 (19:01 +0800)]
Pass separate dev_t var to device_path_parse_major_minor
MIPS/O32's st_rdev member of struct stat is unsigned long, which
is 32bit, while dev_t is defined as 64bit, which make some problems
in device_path_parse_major_minor.
Don't pass st.st_rdev, st_mode to device_path_parse_major_minor,
while pass 2 seperate variables. The result of stat is alos copied
out into these 2 variables. Fixes: #11247
test-json: check absolute and relative difference in floating point test
The test fails under valgrind, so there was an exception for valgrind.
Unfortunately that check only works when valgrind-devel headers are
available during build. But it is possible to have just valgrind installed,
or simply install it after the build, and then "valgrind test-json" would
fail.
It also seems that even without valgrind, this fails on some arm32 CPUs.
Let's do the usual-style test for absolute and relative differences.
Michael Biebl [Sat, 12 Jan 2019 20:49:44 +0000 (20:49 +0000)]
systemd (240-4) unstable; urgency=medium
[ Benjamin Drung ]
* Fix shellcheck issues in initramfs-tools scripts
[ Michael Biebl ]
* Import patches from v240-stable branch (up to f02b5472c6)
- Fixes a problem in logind closing the controlling terminal when using
startx. (Closes: #918927)
- Fixes various journald vulnerabilities via attacker controlled alloca.
(CVE-2018-16864, CVE-2018-16865, Closes: #918841, Closes: #918848)
* sd-device-monitor: Fix ordering of setting buffer size.
Fixes an issue with uevents not being processed properly during coldplug
stage and some kernel modules not being loaded via "udevadm trigger".
(Closes: #917607)
* meson: Stop setting -fPIE globally.
Setting -fPIE globally can lead to miscompilations on certain
architectures. Instead use the b_pie=true build option, which was
introduced in meson 0.49. Bump the Build-Depends accordingly.
(Closes: #909396)
projects / systemd.git / log