add deepseek and gemma results

1 files changed, 28 insertions, 0 deletions

diff --git a/results/classifier/gemma3:12b/device/1794 b/results/classifier/gemma3:12b/device/1794
new file mode 100644
index 00000000..4e5d2e0f
--- /dev/null
+++ b/results/classifier/gemma3:12b/device/1794
@@ -0,0 +1,28 @@
+
+Virtio-GPU doesn't fill Response data for cursor queue
+Description of problem:
+Implementation of virtio-gpu in Qemu is likely not fill Response header in cursor commands.
+
+Inside the virtio 1.2 specification, document said:
+```
+VIRTIO_GPU_CMD_UPDATE_CURSOR
+    Update cursor. Request data is struct virtio_gpu_update_cursor. Response type is VIRTIO_GPU_RESP_OK_NODATA.
+    Full cursor update. Cursor will be loaded from the specified resource_id and will be moved to pos. The driver must 
+    transfer the cursor into the resource beforehand (using control queue commands) and make sure the commands to fill 
+    the resource are actually processed (using fencing).
+
+VIRTIO_GPU_CMD_MOVE_CURSOR
+    Move cursor. Request data is struct virtio_gpu_update_cursor. Response type is VIRTIO_GPU_RESP_OK_NODATA.
+    Move cursor to the place specified in pos. The other fields are not used and will be ignored by the device.
+```
+The cursor commands do have a response like control commands.
+
+But in [hw/display/virtio-gpu.c#L1136](https://gitlab.com/qemu-project/qemu/-/blob/master/hw/display/virtio-gpu.c#L1136), QEMU doesn't care anything about response, just fetching command and execute.
+
+It this a Implementation compromise or I missing something in the specification?
+Steps to reproduce:
+1. Write any kernel that using virtio-gpu.
+2. Run on qemu.
+3. No response on cursor command.
+Additional information:
+Specification: [virtio-v1.2-cs01.html](https://docs.oasis-open.org/virtio/virtio/v1.2/cs01/virtio-v1.2-cs01.html#x1-3650007)