15430 2024-03-07 15:09:51 +0000 qemuarmv5 has multiple display heads 2024-08-15 18:01:42 +0000 1 1 11 Runtime General Runtime General Runtime unspecified x86 Multiple RESOLVED OBSOLETE Medium+ normal 5.1 1 richard.purdie jon.mason pidge randy.macleod --- --- No (bug/feature does not impact docs) oldest_to_newest 98406 0 richard.purdie 2024-03-07 15:09:51 +0000 Loading something like core-image-sato and trying to capture a screenshot of the desktop results in a corrupt png due to the presence of multiple heads. The qemuarm image doesn't have this issue so it seems specific to the v5 machine. 98464 1 jon.mason 2024-03-13 19:33:05 +0000 I believe this is because this (virtual) board has PL110. During boot, I see (multiple times) in dmesg: [ 2.478044] clcd-pl11x 10120000.display: PL110 designer 41 rev0 at 0x10120000 [ 2.479426] clcd-pl11x: probe of 10120000.display failed with error -2 [ 2.709386] drm-clcd-pl111 10120000.display: set up callbacks for Versatile PL110 in bsp/arm-versatile-926ejs/arm-versatile-926ejs-gfx.cfg it is enabled via CONFIG_DRM_PL111=y CONFIG_FB_ARMCLCD=y Since the VirtIO GPU is also enabled via virtio-gpu-pci, there are 2 graphics chips on there. Disabling this via menuconfig removes the entires and probably resolves the issue. If you tell me how to test this, I can verify. Though (assuming this is the issue), is this something we want to remove? It is present on the actual hardware. 99544 2 jon.mason 2024-08-15 18:01:42 +0000 The framebuffer device driver for amba-clcd was removed in the v6.8 kernel. Bruce already removed it from the yocto kernel cache on Aug 5, 2024. So, I don't believe this is an issue going forward. If this is still an issue, than please re-open (as I think we'll have to remove the DRM_PL111 from the versatile config or have the virtio cfg disable all other drm drivers).