| Age | Commit message (Collapse) | Author |
|---|
|
Previously, an IPC connection error could shut down the entire process
without giving a hint as to what's wrong. Now, we report that error to
the debug console.
|
|
Previously we would shut down an ipc connection regardless of if there
were still bytes that have been read and not been handed over to
processing, causing WindowServer not to receive
WindowServer::SetFlashFlush messages sent by `wsctl -f` except the first
one.
This patch fixes that behavior by still shutting the connection down due
to having reached EOF while also processing remaining bytes.
Resolves #12954
See also #8912 which fixes the same issue that this patch fixes but also
seems to have initially broken SettingsWindow cancel not actually
closing the window unless the cursor got moved as described in #12003.
Pull request #12547 fixing the SettingsWindow behavior broke `wsctl`
again by always shutting down.
|
|
|
|
For the reasoning, see the earlier commit about Core::Stream::read().
|
|
|
|
Seems like a useful thing to have.
|
|
|
|
|
|
https://isocpp.github.io/CppCoreGuidelines/CppCoreGuidelines#cother-other-default-operation-rules
"The compiler is more likely to get the default semantics right and
you cannot implement these functions better than the compiler."
|
|
This was done with CLion's automatic rename feature.
|
|
This was done with CLion's automatic rename feature and with:
find . -name ClientConnection.h
| rename 's/ClientConnection\.h/ConnectionFromClient.h/'
find . -name ClientConnection.cpp
| rename 's/ClientConnection\.cpp/ConnectionFromClient.cpp/'
|
|
|
|
This functionality is required by Core::LocalServer and LibIPC depends
on LibCore.
take_over_accepted_socket_from_system_server has also been renamed to
take_over_socket_from_system_server as the socket need not be accepted
before taking it over. :^)
|
|
Apologies for the enormous commit, but I don't see a way to split this
up nicely. In the vast majority of cases it's a simple change. A few
extra places can use TRY instead of manual error checking though. :^)
|
|
This change unfortunately cannot be atomically made without a single
commit changing everything.
Most of the important changes are in LibIPC/Connection.cpp,
LibIPC/ServerConnection.cpp and LibCore/LocalServer.cpp.
The notable changes are:
- IPCCompiler now generates the decode and decode_message functions such
that they take a Core::Stream::LocalSocket instead of the socket fd.
- IPC::Decoder now uses the receive_fd method of LocalSocket instead of
doing system calls directly on the fd.
- IPC::ConnectionBase and related classes now use the Stream API
functions.
- IPC::ServerConnection no longer constructs the socket itself; instead,
a convenience macro, IPC_CLIENT_CONNECTION, is used in place of
C_OBJECT and will generate a static try_create factory function for
the ServerConnection subclass. The subclass is now responsible for
passing the socket constructed in this function to its
ServerConnection base; the socket is passed as the first argument to
the constructor (as a NonnullOwnPtr<Core::Stream::LocalServer>) before
any other arguments.
- The functionality regarding taking over sockets from SystemServer has
been moved to LibIPC/SystemServerTakeover.cpp. The Core::LocalSocket
implementation of this functionality hasn't been deleted due to my
intention of removing this class in the near future and to reduce
noise on this (already quite noisy) PR.
|
|
|
|
This is an encapsulation of the common work done by all of our
single-client IPC servers on startup:
1. Create a Core::LocalSocket, taking over an accepted fd.
2. Create an application-specific ClientConnection object,
wrapping the socket.
It's not a huge change in terms of lines saved, but I do feel that it
improves expressiveness. :^)
|
|
This encapsulates what our multi-client IPC servers typically do on
startup:
1. Create a Core::LocalServer
2. Take over a listening socket file descriptor from SystemServer
3. Set up an accept handler for incoming connections
IPC::MultiServer does all this for you! All you have to do is provide
the relevant client connection type as a template argument.
|
|
|
|
|
|
We can and should do more cleanups of this kind, but this is a quick fix
to unbreak the 32-bit HackStudio build.
|
|
This allows us to use TRY() in decoding helpers, leading to a nice
reduction in line count.
|
|
|
|
This dodges a heap allocation when sending 0 or 1 fd across the IPC
boundary (which covers every message.)
|
|
|
|
|
|
Before this, we only had ClientConnection::did_misbehave() to report an
error and shut the connection down. But it's not fair to say that *all*
errors are the client misbehaving! A typical non-misbehavior is resource
allocation failure on the server side.
|
|
|
|
This reverts commit 3bed7d5a5ee5870de4805dd2bf47e0523e387e76.
As discovered by tomuta, this caused a large increase in WindowServer
CPU usage.
|
|
This lets us avoid using Core::deferred_invoke() which is not usable
during application teardown (as there is no event loop to push the
deferred invocation onto.)
(Not that there is an event loop to fire the processing timer during
teardown *either*, but at least we can exit gracefully with pending
timers, unlike deferred invocations, which hang the process. This is an
area where more improvements are definitely needed!)
|
|
This simplifies some of the code, since it's no longer necessary for the
templated code to pass LocalEndpoint::static_magic() everywhere.
|
|
|
|
This patch moves the templated message parsing code to a virtual
try_parse_messages() helper. By doing that, we can move the rest of the
socket draining code up to ConnectionBase and keep it out of line.
|
|
This patch splits IPC::Connection into Connection and ConnectionBase.
ConnectionBase moves into Connection.cpp so we don't have to inline it
for every single templated subclass.
|
|
|
|
Default implementations allow for more optimizations.
See: https://pvs-studio.com/en/docs/warnings/v832/
|
|
I'm still wondering why nobody did this yet :^) Also changes the use of
unions for the more cleaner / less undefined AK::bit_cast.
|
|
|
|
|
|
Only one place used this argument and it was to hold on to a strong ref
for the object. Since we already do that now, there's no need to keep
this argument around since this can be easily captured.
This commit contains no changes.
|
|
Fixes #9015.
|
|
|
|
This fixes not processing any messages read up until a connection
close is detected. We were returning from the function despite having
read some messages.
|
|
This will close the socket when the recv() returns 0 indicating that
the peer has shutdown, and when there are no pending bytes to be
processed.
|
|
This will allow clients to see if the IPC socket is still open.
|
|
This fixes a Clang warning.
|
|
While structs being forward declared as classes is not strictly an
issue, Clang complains as this is not portable code, since some ABIs
treat classes declared as `class` and `struct` differently.
It's easier to fix these than to reason about explicitly disabling
another warning.
|
|
If an enum has a supported underlying type we can provide encoding and
decoding for the enum as well.
|
|
Clang produced an error on these pieces of code without the `typename`
keyword.
|
|
Caught by userspace UBSAN. :^)
|
