This does all the boring stuff for an input subsystem:
- Input device discovery using udev
- Input device hotplug (add/remove)
- Reading the evdev data from the devices
- Publishing input events to a callback through the eloop
- Suspending and resuming the input capture
Currently the delivered input events are basically just evdev's
"struct input_event" (type, code, value). More work is required to
have it pass out useful data.
Signed-off-by: Ran Benita <ran234@gmail.com>
Signed-off-by: David Herrmann <dh.herrmann@googlemail.com>
vte: This adds a dummy VTE subsystem. It will be used to emulate a vt100
terminal based on our console subsystem.
terminal: This ties together several subsystems including the output-, the
console- and the vte-subsystem.
The test_terminal test application can be used to test the terminal. It is
supposed to provide a full vt100 implementation which is displayed on all
available screens.
Most of the functionality here is not implemented yet and only a dummy function.
These will be added in later commits.
Signed-off-by: David Herrmann <dh.herrmann@googlemail.com>
It doesn't make sense to print the size of new buffer objects as the resize
function does the same. We also currently access uninitialized memory here so it
is useless anyway.
Also move the message of the unref function to be consistent with other objects.
Signed-off-by: David Herrmann <dh.herrmann@googlemail.com>
There is no reason to prohibit a compositor from activating its GL context while
being asleep. It is even recommended to activate the GL context when being
asleep as there is no other way to initialize other subsystems without an GL
context.
Signed-off-by: David Herrmann <dh.herrmann@googlemail.com>
If wakeup fails we should drop master priviledges again, otherwise we might
break other applications.
Also remove the SetMaster call on compositor initialization. It is redundant and
not needed by us.
Signed-off-by: David Herrmann <dh.herrmann@googlemail.com>
Currently, we set up our VT with SIGUSR1/2 signals and after that the user may
connect the signal handlers to the eloop. However, if we receive a signal in
between, the signal gets lost.
Therefore, this simply merges the kmscon_vt_connect_eloop function into
kmscon_vt_open as there is no obvious reason to split them.
Signed-off-by: David Herrmann <dh.herrmann@googlemail.com>
If we have only an output but not the related compositor we currently cannot
figure out whether this output is awake. This adds such a helper function.
Signed-off-by: David Herrmann <dh.herrmann@googlemail.com>
We must not destroy the compositor object if there is still a console using the
GL context. Otherwise we get a SEGFAULT when calling any gl* function.
In future we may need propoper dependencies here. We use reference counts so
this should be easy to implement.
Signed-off-by: David Herrmann <dh.herrmann@googlemail.com>
kmscon_console_newline() can be used to produce a newline. Writing \n doesn't
work as this would write \n as character into the cell and not produce a
newline.
The console does not perform any parsing so we provide a separate function.
Signed-off-by: David Herrmann <dh.herrmann@googlemail.com>
Remove the *_set_res() function entirely. Instead, the *_resize() function now
accepts a height argument which is used as quality hint. Internally, we compute
the texture width by the selected font plus the given height. This simplifies
the API and the internal handling and improves output on non-standard
resolutions or with non-standard fonts.
Until now it was only optimized for my 16x9 output but 5x4 output was quite
garbled.
Signed-off-by: David Herrmann <dh.herrmann@googlemail.com>
We shouldn't do anything if buffer_resize is called with the same size. This
also suppresses useless "buffer resized to XxY" debug messages.
Signed-off-by: David Herrmann <dh.herrmann@googlemail.com>
When creating a new font we automatically measure the average width of a
character. This allows us to draw fonts even on weird resolutions.
Signed-off-by: David Herrmann <dh.herrmann@googlemail.com>
A kmscon_char has always a valid buffer so we can always return a valid pointer.
The user shouldn't use it if length is 0 but it's not our job to prohibit that.
Signed-off-by: David Herrmann <dh.herrmann@googlemail.com>
Forward resizing requests to the buffer but correctly update our size cache and
the cursor position.
Signed-off-by: David Herrmann <dh.herrmann@googlemail.com>
Add kmscon_console_write() to write a character to the current position of the
cursor. The cursor is automatically moved to the next cell.
Also extend the test_console example to use this new function.
Signed-off-by: David Herrmann <dh.herrmann@googlemail.com>
When resizing lines we shouldn't drop all information. Instead we only
initialize new cells.
Signed-off-by: David Herrmann <dh.herrmann@googlemail.com>
We now use a proper cache for the current screen and a linked list for the
scrollback buffer. This allows fast rotations and fast access.
Signed-off-by: David Herrmann <dh.herrmann@googlemail.com>
There is no reason to make all tests/* apps need cairo dependencies. We pass
this parameter internally so we can safely use void*.
Signed-off-by: David Herrmann <dh.herrmann@googlemail.com>
Our scrollback buffer needs to adjust the current position when being resized or
when a new line is pushed.
Signed-off-by: David Herrmann <dh.herrmann@googlemail.com>
Unless we benchmark the buffer implementation we shouldn't implement
speedup-hacks. Therefore, remove the buffer-cache and look how it works out.
This simplifies buffer handling alot.
Signed-off-by: David Herrmann <dh.herrmann@googlemail.com>
Resetting a char should never fail so we provide a separate function that sets a
char to its default value.
Signed-off-by: David Herrmann <dh.herrmann@googlemail.com>
Add two functions to enter/leave our VT object. This allows to
implement to expected behavior of automatically switching to the kmscon
when it is running on a new tty, and switching back to the tty we came
from when the program finishes. Presumably this behavior will be
controlled by a config variable or command line argument later on (like
Xorg's -novtswitch).
There's a bit of a subtlety in this because of VT_PROCESS. We need
permission from ourselves to switch in/out of out VT; this is done when
processing SIGUSR[12] in the eloop. We therefore must dispatch the loop
at least once after switching out. The usual case is demonstrated in
test_vt.c.
Signed-off-by: Ran Benita <ran234@gmail.com>
Signed-off-by: David Herrmann <dh.herrmann@googlemail.com>
The buffer object manages the cells and scrollback buffer. It is optimized for
speed: fast rotations, fast resize, etc.
Signed-off-by: David Herrmann <dh.herrmann@googlemail.com>
We add a new helper to initialize a new character directly to avoid two
allocations. This also refactors the character code and simplifies it alot.
Signed-off-by: David Herrmann <dh.herrmann@googlemail.com>
If we switch VT, we must release DRM master to allow other applications to
access the DRM. When waking up the compositor we try to acquire the master again
to access the DRM.
Signed-off-by: David Herrmann <dh.herrmann@googlemail.com>
As long as we are run in a VT we need to correctly handle VT-switches to avoid
blocking the crtc/input.
This is copied from wayland-compositor demos and modified to fit to our needs.
Signed-off-by: David Herrmann <dh.herrmann@googlemail.com>
Idle events are dispatched everytime kmscon_eloop_dispatch() is called. To allow
the callbacks to add/remove/modify all current idle events (including themself),
we need to keep a pointer to the currently dispatched event.
Signed-off-by: David Herrmann <dh.herrmann@googlemail.com>
We currently use the first mode in the list if no mode is explicitely specified
by the caller. However, this is most often the lowest available resolution
reported by the kernel.
This patch adds a new pointer to every output which remembers a sane default.
This is currently the first mode reported by the kernel but may later be
optimized to choose a better default mode.
Signed-off-by: David Herrmann <dh.herrmann@googlemail.com>
We simply forward all log messages to stderr. We use syslog/prink-like severity
prefixes. Use systemd or alike to forward stderr to syslog, kernel-log or
similar.
Signed-off-by: David Herrmann <dh.herrmann@googlemail.com>
Further commits will need a Signed-off-by line, otherwise I cannot accept them.
MIT license is compatible with GPL so we still can switch to GPL in the future
if we want.
Signed-off-by: David Herrmann <dh.herrmann@googlemail.com>
We didn't calculate the correct cell number when drawing a cell. Fix that.
Reported-by: Ran Benita <ran234@gmail.com>
Signed-off-by: David Herrmann <dh.herrmann@googlemail.com>
The crtc is saved from kmscon_output_activate, and restored from
kmscon_output_deactivate.
This means that when the program exits, the screen is back to what it
was - we don't need to switch back VTs if we started the program from an
fbcon, etc.
Signed-off-by: David Herrmann <dh.herrmann@googlemail.com>
This commit adds basic autoconf + automake files to build the project.
It also adds a main.c stub in order to simulate the main binary.
The configure script uses pkg-config to find the libraries. The usual
stuff should work. The only additional option right now is:
./configure --enable-debug [To enable debugging symbols]
The Makefile should also support the standard stuff:
make [To build the kmscon binary]
make check [To build the test_* binaries]
make dist [To create a tarball]
make clean
make install
etc.
To start from a clean tree (e.g. git clean -dfx), do something like the
following:
./autogen.sh
./configure --enable-debug CFLAGS=-O0
make
It all should work well enough for now.
Signed-off-by: Ran Benita <ran234@gmail.com>
Signed-off-by: David Herrmann <dh.herrmann@googlemail.com>
This function can be used to change the content of the current cell. It
automatically moves the cursor to the next cell.
Signed-off-by: David Herrmann <dh.herrmann@googlemail.com>
To avoid redrawing the whole layout every time, we now support caching a single
glyph-string if it consists of only one run.
This should speed up drawing considerably. We still support drawing the whole
layout as fall-back method.
Signed-off-by: David Herrmann <dh.herrmann@googlemail.com>
We set the font size to the absolute size we have per cell instead of using a
fixed font-size.
Signed-off-by: David Herrmann <dh.herrmann@googlemail.com>
This sets the console resolution to 800x600 on startup. We no longer have a
console with an invalid resolution after startup so we now can start calculating
font sizes properly.
Signed-off-by: David Herrmann <dh.herrmann@googlemail.com>
