CloudNebulaProject/wayray
1
0
Fork 0
mirror of https://github.com/CloudNebulaProject/wayray.git synced 2026-04-10 13:10:41 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions

Fix stride handling and add render debug logging

Browse source 
Use pixman Image's actual stride instead of width*4 for pixel reads.
Add debug logging for element count and damage presence per frame.
This commit is contained in:
Till Wegmueller 2026-04-07 18:26:07 +02:00
parent 8bd8f2490c
commit 6968bcd69c
1 changed files with 11 additions and 10 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

21
crates/wrsrvd/src/backend/headless.rs
View file

@ -134,7 +134,9 @@ pub fn run(
state.space.map_output(&output, (0, 0));
// Initialize previous frame buffer (all zeros = black).
let frame_size = (output_size.w * output_size.h * 4) as usize;
// Use the pixman Image's stride for correct row alignment.
let frame_stride = render_buffer.stride();
let frame_size = frame_stride * output_size.h as usize;
let previous_frame = vec![0u8; frame_size];
let mut calloop_data = CalloopData {
@ -231,9 +233,7 @@ fn render_headless_frame(data: &mut CalloopData) {
};
let element_count = data.state.space.elements().count();
if element_count > 0 {
tracing::trace!(element_count, "rendering space with elements");
}
tracing::debug!(element_count, "render tick");
let custom_elements: &[TextureRenderElement<PixmanTexture>] = &[];
@ -251,16 +251,18 @@ fn render_headless_frame(data: &mut CalloopData) {
match render_result {
Ok(result) => {
let has_damage = result.damage.is_some();
let damage = result.damage.cloned();
tracing::debug!(has_damage, "render complete");
// Drop the render target to release the borrow on the buffer.
drop(target);
// Read pixels directly from the pixman Image's CPU memory.
// PixmanRenderer composites into the buffer in-place, so after
// rendering the data is already there — no GPU readback needed.
// Use the Image's stride (bytes per row) which may include padding.
let output_size = data.state.output.current_mode().unwrap().size;
let frame_bytes = (output_size.w * output_size.h * 4) as usize;
let stride = data.render_buffer.stride();
let frame_bytes = stride * output_size.h as usize;
let pixels = unsafe {
let ptr = data.render_buffer.data() as *const u8;
std::slice::from_raw_parts(ptr, frame_bytes)
@ -268,7 +270,7 @@ fn render_headless_frame(data: &mut CalloopData) {
// Send frame over network if a client is connected.
if data.client_connected {
send_frame_to_network(data, pixels, &damage, output_size.w, output_size.h);
send_frame_to_network(data, pixels, &damage, output_size.w, output_size.h, stride);
}
// Send frame callbacks to all mapped surfaces.
@ -293,9 +295,8 @@ fn send_frame_to_network(
damage: &Option<Vec<Rectangle<i32, smithay::utils::Physical>>>,
width: i32,
height: i32,
stride: usize,
) {
let stride = width as usize * 4;
// Compute XOR diff.
let diff = encoding::xor_diff(current_pixels, &data.previous_frame);