The glBlitFramebuffer man page says: When the color buffer is transferred, values are taken from the read buffer of the read framebuffer and written to each of the draw buffers of the draw framebuffer. Commit a6022e54 modified one path through meta to use gl_FragColor to implement this behavior for multisample-to-singlesample "broadcast" blits where the destination has multiple enabled color buffers. This fixed the GLES3 confromance test framebuffer_blit_functionality_multisampled_to_singlesampled_blit. Based on inspection of the code (and the commit message for a6022e54), broadcast blits involving integer (or unsigned integer) color buffers or multisample-to-multisample blits are still broken. I am currently working on patches that will happen to fix the integer cases. However, we need test cases. There is one related test case on the GLES3 CTS, and there are zero related test cases in piglit. We need to verify the full cross product of (MS-to-MS, MS-to-SS) x (normalized, integer, unsigned integer, depth component).
It's actually a 3-dimensional matrix: (MS-to-MS, MS-to-SS) × (normalized, integer, unsigned integer, depth component) × (1:1, scaled).
Remove depth component. There can only be one depth buffer. It's only color buffers that can be broadcast. Duh.
-- GitLab Migration Automatic Message -- This bug has been migrated to freedesktop.org's GitLab instance and has been closed from further activity. You can subscribe and participate further through the new bug through this link to our GitLab instance: https://gitlab.freedesktop.org/mesa/mesa/issues/1513.
Use of freedesktop.org services, including Bugzilla, is subject to our Code of Conduct. How we collect and use information is described in our Privacy Policy.