WIFEXITED() returns a bool, so previously we were setting
exited_successfully to true when the service was terminated by a signal,
and false if it exited, regardless of the exit status. To test the exit
status, we have to use WEXITSTATUS() instead.
This causes us to correctly use the "3 tries then give up" logic for
services that crash, instead of infinitely attempting to respawn them.
This commit un-deprecates DeprecatedString, and repurposes it as a byte
string.
As the null state has already been removed, there are no other
particularly hairy blockers in repurposing this type as a byte string
(what it _really_ is).
This commit is auto-generated:
$ xs=$(ack -l \bDeprecatedString\b\|deprecated_string AK Userland \
Meta Ports Ladybird Tests Kernel)
$ perl -pie 's/\bDeprecatedString\b/ByteString/g;
s/deprecated_string/byte_string/g' $xs
$ clang-format --style=file -i \
$(git diff --name-only | grep \.cpp\|\.h)
$ gn format $(git ls-files '*.gn' '*.gni')
This commit removes DeprecatedString's "null" state, and replaces all
its users with one of the following:
- A normal, empty DeprecatedString
- Optional<DeprecatedString>
Note that null states of DeprecatedFlyString/StringView/etc are *not*
affected by this commit. However, DeprecatedString::empty() is now
considered equal to a null StringView.
Userspace initially didn't have any sort of mechanism to handle
device hotplug (either removing or inserting a device).
This meant that after a short term of scanning all known devices, by
fetching device events (DeviceEvent packets) from /dev/devctl, we
basically never try to read it again after SystemServer initialization
code.
To accommodate hotplug needs, we change SystemServer by ensuring it will
generate a known set of device nodes at their location during the its
main initialization code. This includes devices like /dev/mem, /dev/zero
and /dev/full, etc.
The actual responsible userspace program to handle hotplug events is a
new userspace program called DeviceMapper, with following key points:
- Its current task is to to constantly read the /dev/devctl device node.
Because we already created generic devices, we only handle devices
that are dynamically-generated in nature, like storage devices, audio
channels, etc.
- Since dynamically-generated device nodes could have an infinite minor
numbers, but major numbers are decoded to a device type, we create an
internal registry based on two structures - DeviceNodeFamily, and
RegisteredDeviceNode. DeviceNodeFamily objects are attached in the
main logic code, when handling a DeviceEvent device insertion packet.
A DeviceNodeFamily object has an internal HashTable to hold objects of
RegisteredDeviceNode class.
- Because some device nodes could still share the same major number (TTY
and serial TTY devices), we have two modes of allocation - limited
allocation (so a range is defined for a major number), or infinite
range. Therefore, two (or more) separate DeviceNodeFamily objects can
can exist albeit sharing the same major number, but they are required
to allocate from a different minor numbers' range to ensure there are
no collisions.
- As for KCOV, we handle this device differently. In case the user
compiled the kernel with such support - this happens to be a singular
device node that we usually don't need, so it's dynamically-generated
too, and because it has only one instance, we don't register it in our
internal registry to not make it complicated needlessly.
The Kernel code is modified to allow proper blocking in case of no
events in the DeviceControlDevice class, because otherwise we will need
to poll periodically the device to check if a new event is available,
which would waste CPU time for no good reason.
This change ensures that code in the Service class doesn't try to check
the g_system_mode variable, but instead is asked on whether it supports
a given system mode string value.
Also, don't assume that we should create sockets for any new Service
instance, but instead do that only if the Service should run in the
current system mode.
We used to call `did_exit()` directly with the status returned from
`waitpid` but the function expected an exit code. We now use several
of `wait`-related macros to deduce the correct information.
If we try to launch a lazily-spawned service and the SystemServer as a
(running --user) session leader is running with root permissions, then
if it is instructed to drop the root permissions for a the new service
then it will make sense to abort the entire spawn procedure if dropping
of privileges failed.
For other users, trying to change UID/GID to something else doesn't make
sense (and will always actually fail) as we are already running in non
root permissions, hence we don't attempt to do this anymore.
It should be noted that if an explicit User configuration was actually
specified for a Service to be used with, we would still try to login
with the requested User option value, which would fail when running as
non-root user.
This is useful for example when trying to run the pro utility with pls
to elevate to root permissions, but the session leader is still the same
so trying to "drop" privileges to UID 0 doesn't make sense.
It occurred to me that when trying to running "pls pro SOME_URL" with a
subsequent failure (which will be fixed in a future patch), that a small
error message was printed to the debug log about "Failed to drop
privileges (GID=0, UID=0)".
To actually understand where it failed, I added the actual errno to
printed message which helped me with further debugging, but this could
easily help others in similar scenarios so let's print the actual error.
Not a single client of this API actually used the event mask feature to
listen for readability AND writability.
Let's simplify the API and have only one hook: on_activation.
This reverts the SystemServer exec() logic to how it was before
81bd91c, but now with some extra TRY()s. This allows the HOME var
to always be propagated from LoginServer which prevents needing
to unveil() /etc/passwd everywhere.
This patch also includes some changes in the way that the environment
and arguments are passed to `exec`. It was needed to fit the signature
of `Core::System::exec`. That's beneficial though, as we are now doing
`String` manipulation in a fallible environment, so we can propagate
more errors.
Simplify a lot of uses of ElapsedTimer by converting the callers to
elapsed_time from elapsed, as the AK::Time returned is better for unit
conversions and comparisons against constants.
This will make it easier to support both string types at the same time
while we convert code, and tracking down remaining uses.
One big exception is Value::to_string() in LibJS, where the name is
dictated by the ToString AO.
We have a new, improved string type coming up in AK (OOM aware, no null
state), and while it's going to use UTF-8, the name UTF8String is a
mouthful - so let's free up the String name by renaming the existing
class.
Making the old one have an annoying name will hopefully also help with
quick adoption :^)
In theory our peer process could die between the call to getsockopt()
and Core::system::stat() and another process could end up with the same
PID which would result in us incorrectly launching the service as
another user (e.g. root).
This commit does three things atomically:
- switch over Core::Account+SystemServer+LoginServer to sid based socket
names.
- change socket names with %uid to %sid.
- add/update necessary pledges and unveils.
Userland: Switch over servers to sid based sockets
Userland: Properly pledge and unveil for sid based sockets
This patch allows to insert "%uid" in `IPC_CLIENT_CONNECTION`
declaration and in SystemServer's ini files. This pattern is replaced
then replaced by the UID of the owner of the service. It opens a path
for seamlessly managed, per-user portal.
Various changes are needed to support this:
- The directory is created by Core::Account on login (and located in
/tmp).
- Service's sockets are now deleted on exit (to allow re-creation)
- SystemServer needs to handle SIGTERM to correctly destroy services.
Each of these strings would previously rely on StringView's char const*
constructor overload, which would call __builtin_strlen on the string.
Since we now have operator ""sv, we can replace these with much simpler
versions. This opens the door to being able to remove
StringView(char const*).
No functional changes.
SystemServer now invokes services with the same uid as the process that
made the request.
This allows the superuser to have a normal GUI workflow. For example,
read and write its own files in TextEditor.
'bootmode' now only controls which set of services are started by
SystemServer, so it is more appropriate to rename it to system_mode, and
no longer validate it in the Kernel.
I use the `configure-components` functionality locally during
development. There are a few services (SpiceAgent) which aren't marked
as required components, and hence aren't built in all configurations,
but we still try to launch them in all configurations.
Instead of letting the forked SystemServer process crash, lets
gracefully handle the situation of a missing binary and provide a
message to the user.
SPDX License Identifiers are a more compact / standardized
way of representing file license information.
See: https://spdx.dev/resources/use/#identifiers
This was done with the `ambr` search and replace tool.
ambr --no-parent-ignore --key-from-file --rep-from-file key.txt rep.txt *
SystemServer only allowed a single socket to be created for a service
before this. Now, SystemServer will allow any amount of sockets. The
sockets can be defined like so:
[SomeService]
Socket=/tmp/portal/socket1,/tmp/portal/socket2,/tmp/portal/socket3
SocketPermissions=660,600
The last item in SocketPermissions is applied to the remainder of the
sockets in the Socket= line, so multiple sockets can have the same
permissions without having to repeat them.
Defining multiple sockets is not allowed for socket-activated services
at the moment, and wouldn't make much sense anyway.
This patch also makes socket takeovers more robust by removing the
assumption that the socket will always be passed in fd 3. Now, the
SOCKET_TAKEOVER environment variable carries information about which
endpoint corresponds to which socket, like so:
SOCKET_TAKEOVER=/tmp/portal/socket1:3 /tmp/portal/socket2:4
and LocalServer/LocalService will parse this automatically and select
the correct one. The old behavior of getting the default socket is
preserved so long as the service only requests a single socket in
SystemServer.ini.
According to the Single UNIX Specification, Version 2 that's where
those macros should be defined. This fixes the libiconv port.
This also fixes some (but not all) build errors for the diffutils and nano ports.
(...and ASSERT_NOT_REACHED => VERIFY_NOT_REACHED)
Since all of these checks are done in release builds as well,
let's rename them to VERIFY to prevent confusion, as everyone is
used to assertions being compiled out in release.
We can introduce a new ASSERT macro that is specifically for debug
checks, but I'm doing this wholesale conversion first since we've
accumulated thousands of these already, and it's not immediately
obvious which ones are suitable for ASSERT.
