<html> <head> <base href="https://bugs.freedesktop.org/" /> </head> <body> <div> <a class="bz_bug_link bz_status_NEW " title="NEW --- - Weston calls eglCreateImageKHR/eglDestroyImageKHR every frame" href="https://bugs.freedesktop.org/show_bug.cgi?id=84117#c2">Comment # 2</a> on <a class="bz_bug_link bz_status_NEW " title="NEW --- - Weston calls eglCreateImageKHR/eglDestroyImageKHR every frame" href="https://bugs.freedesktop.org/show_bug.cgi?id=84117">bug 84117</a> from <a class="email" href="mailto:sjwlaoda@163.com" title="Sunny <sjwlaoda@163.com>"> Sunny</a> <pre>(In reply to <a href="show_bug.cgi?id=84117#c1">comment #1</a>) > Yeah, I seem to recall noticing this while looking at the code in the past > too, and I agree it should be fixed. The DRM backend may have additional > similar issues when using overlays or direct scanout, which actually causes > re-doing the work every output frame, not just every input frame. > > I always forget what the EGL image extension specs say about content > guarantees, but obviously we should avoid unneeded work when possible. So, > we should verify if we can keep the EGLImage around while the client re-uses > it as long as the wl_buffer exists. If that is guaranteed to work, we > probably need some renderer-specific state attached to weston_buffer to > implement the fix. (Note: mind the renderer hand-over.) > > By "frame" you do mean every frame the client attach+commits that actually > gets into output repaint, right? Not every frame the compositor paints? thanks for your quick response. the "frame" I mentioned is attach+commits as you described.</pre> </div> <hr> You are receiving this mail because: <ul> <li>You are the assignee for the bug.</li> </ul> </body> </html>