What kind of failures should we expect with this missing?

how to test this in chrome, Is there any test case ?

per reveman@ Chrome is landing changes to use this and it should use it automatically if the extension exists. Also the simple ozone test that does buffer flipping should use it when available.

Support for EGL_ANDROID_native_fence_sync in amdgpu is already present along with libdrm/mesa changes. Was not sure how to test it or if any particular test is present in chrome for it.

Have the necessary changes to use this landed in Chrome? Is there a way to verify that the path that uses this extension is getting exercised?

Not yet.

RFC CL here:
https://chromium-review.googlesource.com/c/chromium/src/+/857499

that probably doesn't work yet but it's on the right track so I'm expecting it to land in some form soon.

Thanks for the update. In the meantime, is there another way that EGL_ANDROID_native_fence_sync can be tested? Are there any other tests or code paths that would exercise this?

The rects wayland client test uses this extension when available: https://cs.chromium.org/chromium/src/components/exo/wayland/clients/rects.cc

We'll update that code to pass the fence fd to chrome eventually but maybe updating it to use libsync to wait on the fence fd is a good first step.

ddavenport said:
> Thanks for the update. In the meantime, is there another way that EGL_ANDROID_native_fence_sync can be tested? Are there any other tests or code paths that would exercise this?

If you're willing to dive into ARC++, you could "test" EGL_ANDROID_native_fence_sync by running any app and strategically inserting logcat messages into Mesa.

This bug has an owner, thus, it's been triaged. Changing status to "assigned".