I'm working on getting the Venus driver to work inside QEMU guests, but when I
tested `vkcube` in a Linux guest with `VN_DEBUG=init`, I got `stuck in ring
seqno wait with iter at {i}` with updating i at exponentially increasing
intervals. Taking a cursor glance at the code, it seems that this i
After some more checking, it seems the problem is in here:
`vn_cs_encoder_reserve_internal`, as this is the only place where a vn_cs_encoder
can grow with `vn_cs_encoder_add_buffer`. For the buffer_count to be 0, this must
mean that this is never called. The vn_cs_encoder assigned to a ring is
On the host end, I do face an error `warning: virtio_gpu_remap_udmabuf: dmabuf
mmap failed: Invalid argument`. The error is due to the fact that
`res->blob_size` is `1048576`, which is ~1.0 GB, much bigger than what a single
blob should be.
So... I found the mistake --- I forgot to actually implement the different
contexts in `virgl_cmd_context_create`.
I'm working on integrating virglrenderer with MoltenVK on MacOS. However, I
found that when vkCreateInstance is called, even though the instance is
successfully created, mesa on the guest crashes, bring down QEMU with it.
Yes, I realized that it wasn't qemu crashing, but virglrenderer just exiting if
some condition isn't satisfied.
