* make CHLBufferReference not a SP anymore
* copy over release and acquire points in CHLBufferReference
* use CHLBufferReference in screencopy and toplevel export
TODO: use CHLBufferReference in direct scanout properly
the only problem is the scanout buffer release timing,
specifically the onBackendRelease mechanism
* cleanup SSurfaceState and surface pending commit tracking
* move surface code from DRMSyncobj, and move acquire to SSurfaceState
* use queue for comitted pending surface states like proto says
"The content update is placed in a queue until it becomes active." - wl_surface::commit
* drop, not release, prev buffer if 2nd buffer wl_surface.attach is sent
"A wl_buffer that has been attached and then replaced by another attach instead of committed will not receive a release event, and is not used by the compositor." - wl_surface::attach
* core: Don't damage the entire surface every frame
* core: Damage buffer on dims or transform change
* core: Use guards for scale and tr equality checks
reset shm buffers early to mitigate stuttering animations, also reuse
the monitors eglSync and store the eglsync per monitor. this however
reintroduces flickering in dbeaver nonsyncobj application.
* syncobj: cleanup and use uniqueptrs
cleanup a bit missing removals if resource not good, erasing from
containers etc. make use of unique ptrs instead. and add default
destructors.
* syncobj: rework syncobj entirerly
remove early buffer release that was breaking explicit sync, the buffer
needs to exist until the surface commit event has been emitted and draw
calls added egl sync points, move to eventfd signaling instead of
stalling sync point checks, and recommit pending commits if waiting on a
signal. add a CDRMSyncPointState helper class. move a few weak pointers
to shared pointers so they dont destruct before we need to use them.
* syncobj: queue pending states for eventfd
eventfd requires us to queue pending stats until ready and then apply to
current, and also when no ready state exist commit the client commit on
the current existing buffer, if there is one.
* syncobj: clear current buffer damage
clear current buffer damage on current buffer commits.
* syncobj: cleanup code and fix hyprlock
remove unused code, and ensure we dont commit a empty texture causing
locksession protocol and gtk4-layer-shell misbehaving.
* syncobj: ensure buffers are cleaned up
ensure the containers having the various buffers actually gets cleaned
up from their containers, incase the CSignal isnt signaled because of
expired smart pointers or just wrong order destruction because mishaps.
also move the acquire/point setting to buffer attaching. instead of on
precommit.
* syncobj: remove unused code, optimize
remove unused code and merge sync fds if fence is valid, remove manual
directscanout buffer dropping that signals release point on pageflip, it
can cause us to signal the release point while still keeping the current
buffer and rendering it yet again causing wrong things.
* syncobj: delay buffer release on non syncobj
delay buffer releases on non syncobj surfaces until next commit, and
check on async buffers if syncobj and drop and signal the release point
on backend buffer release.
* syncobj: ensure we follow protocol
ensure we follow protocol by replacing acquire/release points if they
arrive late and replace already existing ones. also remove unneded
brackets, and dont try to manual lock/release buffers when it comes to
explicit protocol. it doesnt care about buffer releases only about
acquire and release points and signaling them.
* syncobj: lets not complicate things
set points in precommit, before checking protocol errors and we catch
any pending acquire/release points arriving late.
* syncobj: move SSurfaceState to types
remove destructor resource destroying, let resources destroys them on
their events, and move SSurfaceStates to types/SurfaceState.hpp
* syncobj: actually store the merged fd
have to actually store the mergedfd to use it.
* syncobj: cleanup a bit around fences
ensure the current asynchronous buffer is actually released on pageflip
not the previous. cleanup a bit FD handling in
commitPendingAndDoExplicitSync, and reuse the in fence when syncing
surfaces.
* syncobjs: ensure fence FD doesnt leak
calling resetexplicitfence without properly ensuring the FD is closed
before will leak it, store it per monitor and let it close itself with
the CFileDescriptor class.
* syncobj: ensure buffers are actually released
buffers were never being sent released properly.
* types: Defer buffer sync releaser until unlock
* syncobj: store directscanout fence in monitor
ensure the infence fd survives the scope of attemptdirectscanout so it
doesnt close before it should have.
* syncobj: check if if acquire is expired
we might hit a race to finish on exit where the timeline just has
destructed but the buffer waiter is still pending. and such we
removeAllWaiters null dereferences.
* syncobj: code style changes
remove quack comment, change to m_foo and use a std::vector and
weakpointer in the waiter for removal instead of a std::list.
* syncobj: remove unused async buffer drop
remove unused async buffer drop, only related to directscanout and is
handled elsewhere.
---------
Co-authored-by: Lee Bousfield <ljbousfield@gmail.com>
this fixes direct scanout glitches by ensuring that attemptDirectScanout doesn't try to recommit the same buffer to AQ
which would cause a pageflip event and the backendRelease to release the same buffer too early
* config: make fd use CFileDescriptor
make use of the new hyprutils CFileDescriptor instead of manual FD
handling.
* hyprctl: make fd use CFileDescriptor
make use of the new hyprutils CFileDescriptor instead of manual FD
handling.
* ikeyboard: make fd use CFileDescriptor
make use of the new CFileDescriptor instead of manual FD handling, also
in sendKeymap remove dead code, it already early returns if keyboard
isnt valid, and dont try to close the FD that ikeyboard owns.
* core: make SHMFile functions use CFileDescriptor
make SHMFile misc functions use CFileDescriptor and its associated usage
in dmabuf and keyboard.
* core: make explicit sync use CFileDescriptor
begin using CFileDescriptor in explicit sync and its timelines and
eglsync usage in opengl, there is still a bit left with manual handling
that requires future aquamarine change aswell.
* eventmgr: make fd and sockets use CFileDescriptor
make use of the hyprutils CFileDescriptor instead of manual FD and
socket handling and closing.
* eventloopmgr: make timerfd use CFileDescriptor
make the timerfd use CFileDescriptor instead of manual fd handling
* opengl: make gbm fd use CFileDescriptor
make the gbm rendernode fd use CFileDescriptor instead of manual fd
handling
* core: make selection source/offer use CFileDescriptor
make data selection source and offers use CFileDescriptor and its
associated use in xwm and protocols
* protocols: convert protocols fd to CFileDescriptor
make most fd handling use CFileDescriptor in protocols
* shm: make SHMPool use CfileDescriptor
make SHMPool use CFileDescriptor instead of manual fd handling.
* opengl: duplicate fd with CFileDescriptor
duplicate fenceFD with CFileDescriptor duplicate instead.
* xwayland: make sockets and fds use CFileDescriptor
instead of manual opening/closing make sockets and fds use
CFileDescriptor
* keybindmgr: make sockets and fds use CFileDescriptor
make sockets and fds use CFileDescriptor instead of manual handling.
* compositor: reduce amount of window box copies
mousemoveunified can call this very frequently, the cbox copying
actually shows up as an impact in such cases, move it down in the scope
and only do it when necessery.
* core: constify and reference frequent calls
profiling shows these as frequent called functions try to reduce the
amount of copies with references and const the variables.
* pointermgr: remove not used local copy, const ref
remove unneded local copies and const ref cursorsize.
* inputmgr: reduce amount of calls to vectortowindow
the amount of calls to g_pCompositor->vectorToWindowUnified fast ramps
up in cpu usage with enough windows existing and moving the mouse, move
the PWINDOWIDEAL up and reuse it if its already the same.
* protocol: compositor remove unused local copy
remove unused local copy of accumulateCurrentBufferDamage and const
previousBuffer.
* renderer: reduce scope of variables and refactor
move a few variables down in their scopes to reduce the amount of calls
and copies when not needed, also add one more for loop in
renderWorkspaceWindows and store the windows in a vector with
weakpointers that should be rendered, this adds a loop but reduces the
amount of repeated calls to shouldRenderWindow and also makes the rest
of the loops go over way smaller vector when many windows exist.
#### Describe your PR, what does it fix/add?
Fix lag spikes when pressing more than 6 keys at the same time.
#### Is there anything you want to mention? (unchecked code, possible bugs, found problems, breaking compatibility, etc.)
Debugging process:
<details>
This is triggered by typing some applications, like CopyQ or XWayland.
Typing in Firefox doesn't lead to lag, however it itself does lag
handling these events.
Profiling CopyQ shows that paths leading to
`QtWaylandClient::QWaylandInputDevice::Keyboard::keyboard` take over
80% of processing time of an otherwise idle program.
Looking at output of 'wev' even when it's not focused shows same events
received over and over again.
```
[14: wl_keyboard] repeat_info: rate: 25 keys/sec; delay: 300 ms
[14: wl_keyboard] keymap: format: 1 (xkb v1), size: 64754
```
Looking at what passes through CInputManager::onKeyboardKey() ->
CSeatManager::setKeyboard() shows Hyprland 'switching' between endpoints
of the same keyboard, one of them being named like the other but with
'-1' suffix.
</details>
Tested changing layouts in Fcitx5 and with following config.
```
input:kb_layout = us,cz
input:kb_variant = ,qwerty
input:kb_options = grp:alt_shift_toggle
```
Also tested changing 'input:repeat_delay' while running.
Curiously, now these events appear in the output of 'wev' only once.
Changing layouts still seems to work fine though.
#### Is it ready for merging, or does it need work?
Ready for merging.
* deco: reduce local temporars and function calls
profiling shows this is a high used function, reduce the amount of
function calls and local temporar copies and also check if we even need
to add extents at all in the loop.
* popup: optimize bfhelper in popups
pass nodes as const reference and reserve the size of the children node
vector help reduce the reallocations.
* procotol: make compositor bfhelper use const ref
use const ref for nodes and reserve second nodes vector size to reduce
amount of reallocation needed.
