swagger: fix schema for PATCH /machine-config

The description for PATCH /machine-config was reusing the schema for the
PUT request, which specified default values and required fields that do
not apply to the PATCH request. This was wrong, because the defaults for
the PATCH are `null` for all fields (e.g. "leave as is), and no field is
required.

Signed-off-by: Patrick Roy <[email protected]>

Author: roypat

src/firecracker/swagger/firecracker.yaml

@@ -365,7 +365,7 @@ paths:
           in: body
           description: A subset of Machine Configuration Parameters
           schema:
-            $ref: "#/definitions/MachineConfiguration"
+            $ref: "#/definitions/MachineConfigurationPatch"
       responses:
         204:
           description: Machine Configuration created/updated
@@ -1064,6 +1064,47 @@ definitions:
           - 2M
         description: Which huge pages configuration (if any) should be used to back guest memory.
 
+  MachineConfigurationPatch:
+    type: object
+    description:
+      Describes an update to the number of vCPUs, memory size, SMT capabilities, huge page configuration or
+      the CPU template. Fields set to null will remain unchanged.
+    properties:
+      cpu_template:
+        $ref: "#/definitions/CpuTemplate"
+      # gdb_socket_path:
+      #   type: string
+      #   description: Path to the GDB socket. Requires the gdb feature to be enabled.
+      smt:
+        type: boolean
+        description: Flag for enabling/disabling simultaneous multithreading. Can be enabled only on x86.
+        default: null
+      mem_size_mib:
+        type: integer
+        description: Memory size of VM
+        default: null
+      track_dirty_pages:
+        type: boolean
+        description:
+          Enable dirty page tracking. If this is enabled, then incremental guest memory
+          snapshots can be created. These belong to diff snapshots, which contain, besides
+          the microVM state, only the memory dirtied since a previous snapshot. Full snapshots
+          each contain a full copy of the guest memory.
+        default: null
+      vcpu_count:
+        type: integer
+        minimum: 1
+        maximum: 32
+        default: null
+        description: Number of vCPUs (either 1 or an even number)
+      huge_pages:
+        type: string
+        enum:
+          - None
+          - 2M
+        default: null
+        description: Which huge pages configuration (if any) should be used to back guest memory.
+
   MemoryBackend:
     type: object
     required: