00:19 llyyr: mesa main fails to build with: src/vulkan/util/vk_enum_to_str.c:684:14: error: ‘VK_COMPRESSED_TRIANGLE_FORMAT_AMDX_MAX_ENUM’ undeclared (first use in this function); did you mean ‘VK_COMPRESSED_TRIANGLE_FORMAT_MAX_ENUM_AMDX’?
00:19 llyyr: am I doing something wrong? I can't imagine something would be merged if it didn't build on ci
03:09 mattst88: llyyr: that's very weird. my vk_enum_to_str.c:684 looks like this:
03:09 mattst88: case VK_COMPRESSED_TRIANGLE_FORMAT_MAX_ENUM_AMDX: return "VK_COMPRESSED_TRIANGLE_FORMAT_MAX_ENUM_AMDX";
03:09 llyyr: i fixed it and the fix got merged just now
03:10 llyyr: though im not sure how ci didn't catch that
03:10 mattst88: ahhh, nice
05:14 mareko: karolherbst: so it's the same as load_global(ACCESS_CAN_REORDER | ACCESS_CAN_SPECULATE | ACCESS_NON_WRITEABLE)
07:15 karolherbst: mareko: I think so unless I missed anything
08:31 karolherbst: zmike: so I kinda of have a very nasty bug and I'm not really sure how to address it properly. Imagine a client doing a host_ptr allocation, it maybe enqueues a shader using it, but at some point it flushes the context waits and considers being done with the buffer. So it frees the host memory backing it. _but_ with zink I'm seeing the same
08:31 karolherbst: allocation is still referenced. The pipe_resources refcounts drops to 0 and that one gets nuked, but zink_resource::obj still has a ref count of 1, so that's a bit annoying. But I'm more interested in figuring out where it's still used
10:08 karolherbst: okay.. I think I'm getting an idea on what's going on here and it's... annoying...
10:38 karolherbst: uhhhh.. what an annoying issue, I think I understand what's going on and it's ugly
14:09 zmike: probably it's still alive until the buffer is no longer in use
19:23 glehmann: mareko: I don't think it would have ACCESS_CAN_SPECULATE