If pidfile does not exist when we are stopping the daemon, assume it is
already stopped, and report success.
hostapd is an example of a daemon which removes its pidfile when it is
exiting. If this daemon terminates prematurely, that is, without s-s-d
involvement, then openrc fails to restart it, because s-s-d "stop"
command fails when pidfile is missing.
X-Gentoo-Bug: 646274
X-Gentoo-Bug-URL: https://bugs.gentoo.org/646274
This is our own version of asprintf(). This original code was written by
Mike Frysinger, and I was able to modify it to use our memory helper
functions.
We need a version of this code because it is not available on glibc at
least without defining _GNU_SOURCE, and I would rather not do that.
This is the first step in improving string handling in OpenRC for #207.
This is related to #195.
This is an attempt to shorten the window for the first two issues
discussed by using a file descriptor which does not follow symbolic
links and using the fchmod and fchown calls instead of chown and chmod.
with.
rc_deptree_update_needed would return early as soon as it found
any file newer than the existing dependency cache. Unfortunately,
the first file found may not be the newest one there; so the
clock skew workaround in rc-misc:_rc_deptree_load would be given
a timestamp that was still too old.
This fix forces a full scan of all relevant files, so as to
ensure that we return a timestamp that will allow the clock skew
fix to operate. The runtime cost is no worse than the case where
the cache is up to date (ie. we must check every possible file).
This fixes#161.
Ignore namespaces if there are errors reading either the pid namespace
for the current process or the process we aare testing.
This fixes https://github.com/openrc/openrc/issues/180.
- Harden against dying by handling all signals that would terminate the
program and adding --reexec support
- factor the supervisor into its own function
- fix test for whether we are already running
Prior to this change, we were logging unexpected terminations of daemons
we were supervising at the info level. This change moves the logs to
warnings.
The service binary was just a synonym for rc-service, so use rc-service
instead of service. If you want a "service" binary, it should be
something that can determine which service manager you are running and
run the appropriate service manager commands.
rc-selinux.c: In function ‘selinux_setup’:
rc-selinux.c:361:9: warning: assignment discards ‘const’ qualifier from pointer target type [-Wdiscarded-qualifiers]
curr_t = context_type_get(curr_con);
^
The --retry option for supervise-daemon defines how the supervisor will
attempt to stop the child process it is monitoring. It is defined when
the supervisor is started since stopping the supervisor just sends a
signal to the active supervisor.
This fixes#160.
How to reproduce 1-byte overflow:
```
$ FEATURES=-test CFLAGS="-fsanitize=address -O0 -ggdb3" emerge -1 openrc
=================================================================
==1==ERROR: AddressSanitizer: stack-buffer-overflow on address 0x7fff0efd8710
at pc 0x000000402076 bp 0x7fff0efd7d50 sp 0x7fff0efd7d40
WRITE of size 1 at 0x7fff0efd8710 thread T0
#0 0x402075 (/sbin/openrc-init+0x402075)
#1 0x3cf6e2070f in __libc_start_main (/lib64/libc.so.6+0x3cf6e2070f)
#2 0x4013b8 (/sbin/openrc-init+0x4013b8)
Address 0x7fff0efd8710 is located in stack of thread T0 at offset 2432 in frame
#0 0x401cfb (/sbin/openrc-init+0x401cfb)
This frame has 3 object(s):
[32, 160) 'signals'
[192, 344) 'sa'
[384, 2432) 'buf' <== Memory access at offset 2432 overflows this variable
HINT: this may be a false positive if your program uses some custom stack unwind mechanism or swapcontext
(longjmp and C++ exceptions *are* supported)
SUMMARY: AddressSanitizer: stack-buffer-overflow ??:0 ??
```
The problem here is in the code handling reads from 'init.ctl':
```
int main(int argc, char **argv) {
...
char buf[2048];
for (;;) {
/* This will block until a command is sent down the pipe... */
fifo = fopen(RC_INIT_FIFO, "r");
count = fread(buf, 1, 2048, fifo);
buf[count] = 0;
...
}
```
`buf[count] = 0;` writes outside the buffer when `fread()` returns non-truncated read.
This fixes#138.
Only close the log if we successfully opened it.
Reported-by: Brian Evans <grknight@gentoo.org>
Tested-by: Brian Evans <grknight@gentoo.org>
Signed-off-by: Sergei Trofimovich <slyfox@gentoo.org>
This creates --respawn-delay, --respawn-max and --respawn-period. It was
suggested that it would be easier to follow if the options were
separated.
This is for #126.
Allow limiting the number of times supervise-daemon will attempt to respawn a
daemon once it has died to prevent infinite respawning. Also, set a
reasonable default limit (10 times in a 5 second period).
This is for issue #126.
openrc-init.c and openrc-shutdown.c are based on code which was written by
James Hammons <jlhamm@acm.org>, so I would like to publically
thank him for his work.
Since deptree2dot and the perl requirement are completely optional, we
can move this tool to the support folder. This gives the user the option
of using it if they have perl installed, and means we do not have an
optional runtime dependency on perl.
Documentation for this tool has also been added to the support folder.
X-Gentoo-Bug: 600742
X-Gentoo-Bug-URL: https://bugs.gentoo.org/show_bug.cgi?id=600742
This path should not be hard coded in the open call.
Linux prior to 2.4.19 did not have /proc/self/mounts, so for now I'm
making this value /proc/mounts everywhere, but that may change to
/proc/self/mounts on linux; I'm not sure we should care about <2.4.19.
X-Gentoo-Bug: 604646
X-Gentoo-Bug-URL: https://bugs.gentoo.org/show_bug.cgi?id=604646
This reverts commit 8b4fc05ff2.
The original commit did not explain why this feature was disabled, and I
now have a request to enable it.
This fixes#24.
The original auto detection of Docker containers assumed the presence of
a container environment variable. However, Docker-1.12 does not
implement this, and I'm not sure which versions of docker implemented
it.
The new test is for the presence of a file named .dockerenv in the
root directory.
These warnings were inserted in verbose only mode in OpenRC-0.13.A
Now, we are making them more visible in preparation for removing these
compatibility binaries in the future.
The whitelist of environment variables we pass to service scripts
included several unnecessary variables.
The default whitelist now includes EERROR_QUIET, EINFO_QUIET,
IN_BACKGROUND and IN_HOTPLUG.
X-Gentoo-Bug: 569542
X-Gentoo-Bug-URL: https://bugs.gentoo.org/show_bug.cgi?id=569542
In the past, OpenRC was a hybrid of a centralized and file-scope
license/copyright structure.
I followed the instructions from the Software Freedom Law Center [1] to
convert to a Centralized structure where possible, for easier future
maintenance.
[1] https://softwarefreedom.org/resources/2012/ManagingCopyrightInformation.html
These functions replace rc_sys so that we can detect containers and vms
separately.
Also, we copy file_regex() to rc-misc.c and open it to all operating
systems.
In src/rc/_usage.c, we were using bootlevel as the variable to hold the
return value of rc_sys.
This changes the variable name to systype because this function returns
a system type, not a runlevel.
These functions were never meant to be used outside of OpenRC, and they
were added when we thought we were going to do away with the automatic
detection of subsystems. Since the autodetection is not going away, we
can combine these functions into rc_sys.
The want dependency is similar to the use dependency. If a service
script, for example called service1, adds "want service2" to its depend
function, OpenRC will attempt to start service2, if it exists on the
system, when service1 is started.
However, service1 will start regardless of the status of
service2.
X-Gentoo-Bug: 406021
X-Gentoo-Bug-URL: https://bugs.gentoo.org/show_bug.cgi?id=406021
Rename the rc_conf_override function to describe its purpose better,
drop one conditional compile by making it available everywhere, and move
the call to it after the optional rc.conf.d directory is processed.
This makes it possible to override settings in rc.conf by adding a
directory @SYSCONFDIR@/rc.conf.d and putting files in this directory.
The files will be processed in lexical order, and the last setting in
these files will be used.
On Linux, the --netdev and --nonetdev switches were not working. They
were both returning false. After this change, they operate based on the
presence or abscence of the _netdev option in mount options.
All of the dependency type lists had the types_ prefix in their names;
this has been changed to deptypes_ to make them more self documenting.
Along the same lines, the setup_types function was renamed
setup_deptypes.
If a service has the same name as the runlevel it is in, openrc will
crash on changing to such runlevel. It goes in a recursive madness and
eventually gets a SEGV while in snprintf (don't know why).
This fixes two errors:
1. ls_dir stats files not with full path -> stat always returns != 0
2. ls_dir adds files to list if stat failed
This fixes#53.
X-Gentoo-Bug: 537304
X-Gentoo-Bug-URL: https://bugs.gentoo.org/show_bug.cgi?id=537304
If selinux is disabled, then stub methods will be provided instead of
calling the real methods. This removes some warnings about unused
parameters which used to be covered up with #ifdef HAVE_SELINUX.
Signed-off-by: Jason Zaman <jason@perfinion.com>
The previous fix to --test (PR #34) prevented reading one too many
arguments when --exec -or --name was not specified, but created a
regression where the last argument would not print if either of those
arguments was specified. This corrects the issue.
Fixes#41.
This is another security fix. If you use chown() or chmod() on a
symbolic link, it affects the referenced file, not the symbolic link
itself.
X-Gentoo-Bug: 540006
X-Gentoo-Bug-URL: https://bugs.gentoo.org/show_bug.cgi?id=540006
Do not change permissions on the target if it is a file and has multiple
hard links. This is necessary because a hard link can be an attack
vector to gain privilege escalation.
X-Gentoo-Bug: 540006
X-Gentoo-Bug-URL: https://bugs.gentoo.org/show_bug.cgi?id=540006
runscript used to dlopen() runscript_selinux.so. This adds equivalent
functionality directly in to runscript instead. It authenticates with
either PAM or shadow and optionally has a dep on audit.
X-Gentoo-Bug: 517450
X-Gentoo-Bug-URL: https://bugs.gentoo.org/show_bug.cgi?id=517450
Check for __FreeBSD_kernel instead of __GLIBC__ in source files.
note from William Hubbs:
I was told this is a better check for GNU/kFreeBSD than checking the
C library the source is being compiled against.
GNU/kFreeBSD than checking which library we are using.
The stat structure was not being initialized correctly in do_check. This
was causing the owner adjustment to be skipped if the first path had the
correct owner.
Also, the "correcting owner" message should always be printed when the
owner is being changed.
X-Gentoo-Bug: 518042
X-Gentoo-Bug-URL: https://bugs.gentoo.org/show_bug.cgi?id=518042
Remove the recursive call in print_stacked_services which was causing an
infinite loop when using stacked runlevels.
I would like to thank Doug Freed and Jason Zaman for assisting with
tracking this down.
X-Gentoo-Bug: 514972
X-Gentoo-Bug-URL: https://bugs.gentoo.org/show_bug.cgi?id=514972
The -W option does not need an argument of its own; it can take the
first path after all other options are processed on the command line.
Also, move the processing for the -W option out of the switch so it will
be in the same loop as the other processing.
Before this commit, not specifying -d, -f, -p or -W in a checkpath
command meant the command exited successfully but actually did nothing.
This is an error condition, so report it as such.
Thanks to info and testing done by Daniel Robbins <drobbins@funtoo.org>,
there is now a fix for this. Below is his description of the steps
OpenRC needed to use.
1) See if /proc/<pid>/status exists
2) If it does, see if it has a "envID:" field
3) If it does, see if "envID:" is set to "0"
4) If so, then it's one of the host's processes and should be a
candidate for the list. Otherwise, it is one of the container's
processes and should be ignored.
This should fix the bug and allow start-stop-daemon to work properly on
OpenVZ hosts.
X-Gentoo-Bug: 376817
X-Gentoo-Bug-URL: https://bugs.gentoo.org/show_bug.cgi?id=376817
The mount and remount options should always be processed. They were only
being processed if -q was not on the command line.
X-Gentoo-Bug: 498206
X-Gentoo-Bug-URL: https://bugs.gentoo.org/show_bug.cgi?id=498206
The -a option,which only applies to the del command, is used to remove a
service from all runlevels.
X-Gentoo-Bug: 497740
X-Gentoo-Bug-URL: https://bugs.gentoo.org/show_bug.cgi?id=497740
This was requested by Debian, because the minicom software, which is
available on Debian and other distros, has a binary named runscript. We
are keeping a backward compatibility symlink for now, but this allows
Debian or any other distro to safely remove the symlink.
X-Gentoo-Bug: 494220
X-Gentoo-Bug-URL: https://bugs.gentoo.org/show_bug.cgi?id=494220
The message about a service being scheduled to start was confusing to
some of our users; I was told this wording is more clear.
X-Gentoo-Bug: 493070
X-Gentoo-Bug-URL: https://bugs.gentoo.org/show_bug.cgi?id=493070
Debian requested this rename due to the "rc" binary conflicting with the
"rc" binary from the plan 9 shell.
We also add a deprecation warning to the binary when it is run as rc to
encourage users to switch to openrc instead.
X-Gentoo-Bug: 493958
X-Gentoo-Bug-URL: https://bugs.gentoo.org/show_bug.cgi?id=493958
The rc_runlevel_exists function was attempting to treat "." and ".." as
valid runlevels; however, this should not be allowed.
X-Gentoo-Bug: 488710
X-Gentoo-Bug-URL: https://bugs.gentoo.org/show_bug.cgi?id=488710
Convert these messages to warnings that are only displayed if
EINFO_VERBOSE is set to yes in the environment.
This is based on a suggestion from Patrick Lauer <patrick@gentoo.org>.
X-Gentoo-Bug: 487588
X-Gentoo-Bug-URL: https://bugs.gentoo.org/show_bug.cgi?id=487588
Several calls to do_stop were forcing the test parameter to be true,
which was causing extra output to the terminal, such as:
* Would send signal 0 to pid xxxxx
This should only happen if the --test command line option was used.
We should first check if we are within bounds and then read rather than
the opposite.
This makes valgrind happy.
Signed-off-by: Natanael Copa <ncopa@alpinelinux.org>
The eerror calls in this function make it too verbose, so change them to
ewarnv() calls instead. This means that they will only print if the
--verbose option is used or EINFO_VERBOSE=yes is set in the environment.
The libeinfo library has no consumers other than OpenRC, so there is no
reason for it to be maintained as a library. The einfo routines are now
an object that links with the rc binary.
Along the same lines as the quiet variable, this is controlled by an
environment variable for the einfo code, so we do not need a separate
boolean flag.
The suppression of output is controlled in the e* functions themselves,
so there is no need for a variable to test in start-stop-daemon.
X-Gentoo-Bug: 482396
X-Gentoo-Bug-URL: https://bugs.gentoo.org/show_bug.cgi?id=482396
Patch was provided by Max Hacking <max.gentoo.bugzilla@hacking.co.uk>
and slightly fixed by Alexander Vershilov <qnikst@gentoo.org> and
William Hubbs <williamh@gentoo.org>.
Fixes:
1). Rebase to newest OpenRC version.
2). Remove code style fixes. Port to currect code style.
3). Fix rc_runlevel_stack instead of introducing new function.
4). Make get_runlevel_chain a private function.
X-Gentoo-Bug: 467368
X-Gentoo-Bug-URL: https://bugs.gentoo.org/show_bug.cgi?id=467368
