Supervisor setups break easily when start/stop/status functions are not
default.
Applications that write multiple PIDs to a pidfile (eg HAProxy as
described in bug 601540), can also benefit from being able to call the
default start/stop/status with modified environment variables.
Expose the default start/stop/status functions as
default_start/stop/status, and use them for the defaults
start/stop/status.
Trivial usage example:
```
stop()
{
t=$(mktemp)
for pid in $(cat $pidfile) ; do
echo $pid >$t
pidfile=$t default_stop
done
rm -f $t
}
```
X-Gentoo-Bug: 601540
X-Gentoo-Bug-URL: https://bugs.gentoo.org/601540
Signed-off-by: Robin H. Johnson <robbat2@gentoo.org>
Busybox does not support the 'scope' argument on 'ip address add' or 'ip
route add', this is documented in BUSYBOX.md, but is no longer actually
needed, as the kernel does get it right without manual specification,
and the ifconfig variant already relies on the kernel to get it right.
This is part of #103.
X-Gentoo-Bug: 487208
X-Gentoo-Bug-URL: https://bugs.gentoo.org/show_bug.cgi?id=487208
If /tmp or / are read-only, the clean_run function can fail in some very
bad ways.
1. dir=$(mktemp -d) returns an EMPTY string on error.
2. "mount -o bind / $dir", and don't check the result of that,
3. "rm -rf $dir/run/*", which removes the REAL /run contents
4. box gets very weird from this point forward
Signed-Off-By: Robin H. Johnson <robbat2@gentoo.org>
Signed-Off-By: Chip Parker <infowolfe@gmail.com>
Reported-by: Chip Parker <infowolfe@gmail.com>
Tested-by: Chip Parker <infowolfe@gmail.com>
As the author of our tmpfiles.sh script, I hereby license it under
2-clause BSD, like the rest of openrc.
Signed-off-by: Robin H. Johnson <robbat2@gentoo.org>
In the 3.10 kernel, EFI variables are now provided by a dedicated
filesystem that needs to be mounted.
Signed-off-by: Robin H. Johnson <robbat2@gentoo.org>
In the 3.10 kernel, EFI variables are now provided by a dedicated
filesystem that needs to be mounted.
Signed-off-by: Robin H. Johnson <robbat2@gentoo.org>
miimon & mode must be set before other parameters, and then not changed
again. Prior commit f671e0a28 per bug #421757 introduced a small logic
error. Fixed & refactored to prevent it happening as easily.
X-Gentoo-Bug: 447790
X-Gentoo-Bug-URL: https://bugs.gentoo.org/show_bug.cgi?id=447790
Signed-off-by: Robin H. Johnson <robbat2@gentoo.org>
In a pathname expansion, specifically single-character match, the pure
POSIX specification uses '!' as the Negation character where a regular
expression would normally be '^'.
Regular expression: "a[^a]a"
Pathname expansion pattern: "a[!a]a"
Reference:
IEEE Std 1003.1, 2004 Edition
2. Shell Command Language
2.13 Pattern Matching Notation
2.13.1 Patterns Matching a Single Character
> The description of basic regular expression bracket expressions in the
> Base Definitions volume of IEEE Std 1003.1-2001, Section 9.3.5, RE
> Bracket Expression shall also apply to the pattern bracket expression,
> except that the exclamation mark character ( '!' ) shall replace the
> circumflex character ( '^' ) in its role in a "non-matching list" in
> the regular expression notation. A bracket expression starting with an
> unquoted circumflex character produces unspecified results.
Signed-off-by: Robin H. Johnson <robbat2@gentoo.org>
If sysfs is not available, you might still be able to create bond
interfaces, but only in very specific configurations, and you must have
the ifenslave binary instead to call ioctls.
Signed-off-by: Robin H. Johnson <robbat2@gentoo.org>
If the kernel is NFS-booting, it is critical that we don't down the
slave interfaces when we activate the bond. To do so will break the root
filesystem when networking is temporarily lost.
Reported-by: Walter <walter@pratyeka.org>
X-Gentoo-Bug: 428604
X-Gentoo-Bug-URL: https://bugs.gentoo.org/show_bug.cgi?id=428604
Signed-off-by: Robin H. Johnson <robbat2@gentoo.org>
Now that the tmpfiles.d code is more tested, actually call it from
init.d. It assumes that /run is already available when it runs.
Please note it runs TWICE.
- During sysinit, ideally just after /dev/shm is created, but before
udev has started. After udev is also acceptable, but not ideal.
- During boot, ideally just after localmount has completed.
Signed-off-by: Robin H. Johnson <robbat2@gentoo.org>
Using the new dev-mount virtual, with udev-mount included until new udev
version is rolled out, we run devfs earlier now, before udev/mdev.
It only needs (u)dev-mount before it, so that /dev is mounted.
This opens the way for tmpfiles.d, which needs to be sandwiched in the
middle.
Signed-off-by: Robin H. Johnson <robbat2@gentoo.org>
net.* and module loading require sysfs now, and if udev is not in use,
it is not always loaded early enough, esp for net.lo. Force it to come
up during sysinit instead.
Signed-off-by: Robin H. Johnson <robbat2@gentoo.org>
sysfs needs to be available before we can check interfaces. On udev
systems this was not a problem, but with mdev/static-dev, there is no
other need for sysfs until later in the init.
X-Gentoo-Bug: 410701
X-Gentoo-Bug-URL: https://bugs.gentoo.org/show_bug.cgi?id=410701
Signed-off-by: Robin H. Johnson <robbat2@gentoo.org>
In most cases, changing the MAC on a bond manually is wrong. The bonding
module will do it as needed to failover between interfaces, or to get
multiple interfaces to correctly have the same MAC.
We cannot however enforce it, as there are some corner cases where it is
actually valid (hardware that requires specific MAC configuration, like
some quad-port NICs).
Suggested-by: Martin Mokrejs <mmokrejs@fold.natur.cuni.cz>
Signed-off-by: Robin H. Johnson <robbat2@gentoo.org>
The newer bonding code using sysfs does not clearly show the most common
bond parameter: mode
Also include a path to the referenced kernel documentation for the other
settings.
X-Gentoo-Bug: 408333
X-Gentoo-Bug-URL: https://bugs.gentoo.org/show_bug.cgi?id=408333
Signed-off-by: Robin H. Johnson <robbat2@gentoo.org>
Upstream has clarified via IRC:
- hardcoding /usr/lib/ is an explicit choice. It should NOT consider
$libdir at all.
- The z/Z relabel types should call restorecon, not chcon.
- Whitespace is not allowed in tmpfiles.d/*.conf path entries,
but is allowed in globs results. Fixed quoting of path arguments for
this.
Signed-off-by: Robin H. Johnson <robbat2@gentoo.org>
This is the baseline support for tmpfiles.d.
Still missing:
- SELinux relabel, pending upstream clarification
- LIBDIR vs multilib systems, pending upstream clarification
- Whitespace in paths?
- Clean support not implemented
- "x" exclude type not implemented
X-Gentoo-Bug: 396003
X-Gentoo-Bug-URL: https://bugs.gentoo.org/show_bug.cgi?id=396003
Signed-off-by: Robin H. Johnson <robbat2@gentoo.org>
The location of the ifconfig binary has changed in
net-tools-1.60_p20120127084908, and if we do not check both locations
for it, the user will get errors like:
_is_wireless: command not found
_exists: command not found
X-Gentoo-Bug: 407757
X-Gentoo-Bug-URL: https://bugs.gentoo.org/show_bug.cgi?id=407757
During early boot, the keywords were not being checked for
consolefont/termencoding and they were running anyway when they should
not be.
X-Gentoo-Bug: 400549
X-Gentoo-Bug-URL: https://bugs.gentoo.org/show_bug.cgi?id=400549
Signed-off-by: Robin H. Johnson <robbat2@gentoo.org>
Our sysctl script has read from /etc/sysctl.d/ but the directory was not
created by default, and we didn't document it.
X-Gentoo-Bug: 398189
X-Gentoo-Bug-URL: https://bugs.gentoo.org/show_bug.cgi?id=398189
Signed-off-by: Robin H. Johnson <robbat2@gentoo.org>
The previous bonding change of ensuring interfaces were down to add
slave interfaces, but it clobbered the IFACE variable, because it was
being passed to a bash function rather than a command.
X-Gentoo-Bug: 400613
X-Gentoo-Bug-URL: https://bugs.gentoo.org/show_bug.cgi?id=400613
Signed-off-by: Robin H. Johnson <robbat2@gentoo.org>
In the case of a single line of configuration, where the address has
parameters, the parameters were being treated as seperate addresses.
Eg:
config_eth0="4321:0:1:2:3:4:567:89ab/64 nodad home preferred_lft 0"
Also document usage of parameters in the net example, and note that
multiple addresses on a single line cannot be mixed with parameters.
Newlines are required to seperate the addresses.
X-Gentoo-Bug: 398827
X-Gentoo-Bug-URL: https://bugs.gentoo.org/show_bug.cgi?id=398827
Signed-off-by: Robin H. Johnson <robbat2@gentoo.org>
This commit provides the checkpath applet with feature parity to
systemd's tmpfiles.c create_item function.
Very similarly to the systemd function, it does NOT do any of the
cleanup work in this function.
Signed-off-by: Robin H. Johnson <robbat2@gentoo.org>
The program function in depend blocks is now able to search paths by
itself. If passed multiple arguments or multiple calls, at least one of
the arguments passed must be a program or a shell builtin (eg ip built
into busybox). If a qualified path is specified, only that path will be
checked, otherwise it will be checked as a builtin, then $PATH will be
checked for the named binary (via type).
Signed-off-by: Robin H. Johnson <robbat2@gentoo.org>
Since commit 683a21b0a in Feburary 2011, iproute2 has been required for
new VLAN configuration. MACVLAN is also impossible to configure without
iproute2. However we did not check if iproute2 was actually in the
modules for a given interface, so it could end up hanging or giving
weird errors. Check for iproute2 before usage now.
X-Gentoo-Bug: 389437
X-Gentoo-Bug-URL: https://bugs.gentoo.org/show_bug.cgi?id=389437
Signed-off-by: Robin H. Johnson <robbat2@gentoo.org>
We need to create interfaces of custom names before we can sucessfully
start them. To do this, we have to add tunnels during prestart instead
of start. Split up the ip6to4 script to do this, saving the computed
variables for use in start with the new service data commands.
X-Gentoo-Bug: 372575
X-Gentoo-Bug-URL: https://bugs.gentoo.org/show_bug.cgi?id=372575
Signed-off-by: Robin H. Johnson <robbat2@gentoo.org>
Historically, we have tried to up interfaces before running preup, so
that the kernel setups up the device and makes things like ethtool work
(some hardware cannot be correct probed until then). However this ends
up breaking other hardware, so a variable has been introduced to allow
the up prior to preup to be disabled: up_before_preup_IFVAR=no
X-Gentoo-Bug: 389475
X-Gentoo-Bug-URL: https://bugs.gentoo.org/show_bug.cgi?id=389475
Signed-off-by: Robin H. Johnson <robbat2@gentoo.org>
Broadcast and IPv6 should not be used together. Do not try to set the
keyword for auto-generation of the broadcast address. If the user passes
a broadcast address for IPv6, throw an error.
Signed-off-by: Robin H. Johnson <robbat2@gentoo.org>
Several of the optional flags were not being handled correctly, they
were being passed as values only, without the keyword before them.
Affected keywords: anycast, label, scope, valid_lft, preferred_lft
Also change the handling of keywords to a common setup now, making
broadcast and peer strings the same as the above keywords.
Signed-off-by: Robin H. Johnson <robbat2@gentoo.org>
For creation of routing policy entries for IPv6, the family must be
explicitly specified to 'ip'.
X-Gentoo-Bug: 385833
X-Gentoo-Bug-URL: https://bugs.gentoo.org/385833
Signed-off-by: Robin H. Johnson <robbat2@gentoo.org>
Provide consistent methods using iproute2/ifconfig to check operational
and administrative up/down state of interfaces. This is not the same as
ethtool's "Link detected" field, which is the state of the layer 2
medium.
TODO: How to check operational state in BSD?
Signed-off-by: Robin H. Johnson <robbat2@gentoo.org>
The old ifenslave binary forcibly set new slave interface to down before
adding into the kernel, as a way of show it's version to the kernel. We
need to duplicate this so that the kernel doesn't think it's an old
ifenslave version using the bonding setup interface.
X-Gentoo-Bug: 391881
X-Gentoo-Bug-URL: https://bugs.gentoo.org/391881
Signed-off-by: Robin H. Johnson <robbat2@gentoo.org>
Thanks-to: Yun Zheng Hu <hu@fox-it.com>
This was originally to fix the fact that our code did not handle certain
orders of arguments in conversion, but it was easier to rewrite the
entire argument handling to support more options at the same time.
Now supports all options documented in the ip manpage, including the
IPv6-specific options that must be passed after the interface argument.
Signed-off-by: Robin H. Johnson <robbat2@gentoo.org>
Reported-by: Tony Vroon <chainsaw@gentoo.org>
X-Gentoo-Bug: 366905
X-Gentoo-Bug-URL: https://bugs.gentoo.org/366905
The version iteration code missed certain combinations:
KV=1.2.3.4
skips: 1.2.3, 1
KV=1.2.3
skips: 1
Simplify the code to use a loop and build the list of versions directly
instead of unique variables per version component.
Signed-off-by: Robin H. Johnson <robbat2@gentoo.org>
- Add support for a global plug_timeout setting to match the
documentation.
- Update the documentation to also show per-interface configuration.
Signed-off-by: Robin H. Johnson <robbat2@gentoo.org>
This is primarily for users with set_name_type settings from previous
sysadmins can easily see what changes to make.
Signed-off-by: Robin H. Johnson <robbat2@gentoo.org>
Implement a consistent interface for changing ethtool parameters, as
suggested in bug 195479. All variable names are based on the long option
to ethtool to set each group of parameters. Multiple entries seperated
by newlines are permitted for variable values.
Signed-off-by: Robin H. Johnson <robbat2@gentoo.org>
X-Gentoo-Bug: 195479
Tunnel modes ipip6 and ip6ip6 require an explicit family selection for
the ip tunnel call.
Signed-off-by: Robin H. Johnson <robbat2@gentoo.org>
X-Gentoo-Bug: 347657
This replaces the vlan setup code that previously used the old vconfig
binary with a new implementation using the iproute2 interface.
vconfig does not handle many of the newer setups. No automatic migration
path is provided, as altering the configuration is non-trivial.
Signed-off-by: Robin H. Johnson <robbat2@gentoo.org>
Written-by: Guillaume Castagnino <casta@xwing.info>
X-Gentoo-Bug: 346365
Previously we checked if /proc was alive by reading /proc/uptime twice
with a 1 second sleep between calls, so that it had time to update.
This got a complaint of an entire 1 second delay, so we improve the
check to be much faster without sleep. We cannot continue to use
/proc/uptime as it only has a 10ms resolution.
X-Gentoo-Bug: 348416
X-Gentoo-Bug-URL: https://bugs.gentoo.org/show_bug.cgi?id=348416
Signed-off-by: Robin H. Johnson <robbat2@gentoo.org>
Refactor rc_conf_value into librc for use in library context.
Also requires moving:
- rc_conf internal static
- Defines: PROFILE_ENV, SYS_WHITELIST, USR_WHITELIST, RC_PATH_PREFIX
moved to rc.h with new RC_ prefix added.
- Defines: RC_CONF, RC_CONF_OLD moved to rc.h.
Signed-off-by: Robin H. Johnson <robbat2@gentoo.org>
In addition to detecting what multicall applet we want via argv[0],
provide an explicit override mode with a --applet initial argument.
Signed-off-by: Robin H. Johnson <robbat2@gentoo.org>
Robin H. Johnson