How to add dummy bytes to make the line width multiple of 64/128 bytes?

[markszente]

Hi guys,

I'm using SoftwareRenderer in a no_std environment. ST's AN4861 speaks in detail about how we should

add a number of dummy bytes at the end of every line of pixels to match the closest frame line width
multiple of 64/128 bytes (for AHB/AXI)

for optimal performance.

My question is, how would I do it with Slint? I'm using 24bpp (RGB888) color, and the width of the screen is 800. That's 2,400 bytes if my math is correct, not a multiple of 64/128. I would need to add 32 bytes to the end of each line (and then set the pitch in LTDC, but that's beyond the point). Is it possible?

Thanks in advance.

[ogoffart]

If you are using the SoftwareRenderer::render, you can allocate a bigger buffer and change the stride. But unfortunately, the stride being in pixel and not byte, you can't add 32 bytes :-(.

What you can do is add 12 pixels (288 bytes) at the end of each lines. That's some wasted spaces, but maybe that helps.

In a future version of Slint we should change the software renderer to have a trait TargetBuffer instead of TargetPixel which would have a function to retrieve lines.

Another solution is to render line by line into the buffer using the render_by_line function