mirror of
https://github.com/aljazceru/kata-containers.git
synced 2026-01-21 07:14:25 +01:00
Augment the mock hypervisor so that we can validate that ACPI memory hotplug is carried out as expected. We'll augment the number of memory slots in the hypervisor config each time the memory of the hypervisor is changed. In this way we can ensure that large memory hotplugs are broken up into appropriately sized pieces in the unit test. Signed-off-by: Eric Ernst <eric_ernst@apple.com>
158 lines
3.8 KiB
Go
158 lines
3.8 KiB
Go
// Copyright (c) 2016 Intel Corporation
|
|
//
|
|
// SPDX-License-Identifier: Apache-2.0
|
|
//
|
|
|
|
package virtcontainers
|
|
|
|
import (
|
|
"context"
|
|
"errors"
|
|
"os"
|
|
|
|
hv "github.com/kata-containers/kata-containers/src/runtime/pkg/hypervisors"
|
|
"github.com/kata-containers/kata-containers/src/runtime/virtcontainers/types"
|
|
)
|
|
|
|
var MockHybridVSockPath = "/tmp/kata-mock-hybrid-vsock.socket"
|
|
|
|
type mockHypervisor struct {
|
|
config HypervisorConfig
|
|
mockPid int
|
|
}
|
|
|
|
func (m *mockHypervisor) Capabilities(ctx context.Context) types.Capabilities {
|
|
caps := types.Capabilities{}
|
|
caps.SetFsSharingSupport()
|
|
return caps
|
|
}
|
|
|
|
func (m *mockHypervisor) HypervisorConfig() HypervisorConfig {
|
|
return m.config
|
|
}
|
|
|
|
func (m *mockHypervisor) setConfig(config *HypervisorConfig) error {
|
|
m.config = *config
|
|
return nil
|
|
}
|
|
|
|
func (m *mockHypervisor) CreateVM(ctx context.Context, id string, network Network, hypervisorConfig *HypervisorConfig) error {
|
|
if err := m.setConfig(hypervisorConfig); err != nil {
|
|
return err
|
|
}
|
|
m.config.MemSlots = 0
|
|
return nil
|
|
}
|
|
|
|
func (m *mockHypervisor) StartVM(ctx context.Context, timeout int) error {
|
|
return nil
|
|
}
|
|
|
|
func (m *mockHypervisor) StopVM(ctx context.Context, waitOnly bool) error {
|
|
return nil
|
|
}
|
|
|
|
func (m *mockHypervisor) PauseVM(ctx context.Context) error {
|
|
return nil
|
|
}
|
|
|
|
func (m *mockHypervisor) ResumeVM(ctx context.Context) error {
|
|
return nil
|
|
}
|
|
|
|
func (m *mockHypervisor) SaveVM() error {
|
|
return nil
|
|
}
|
|
|
|
func (m *mockHypervisor) AddDevice(ctx context.Context, devInfo interface{}, devType DeviceType) error {
|
|
return nil
|
|
}
|
|
|
|
func (m *mockHypervisor) HotplugAddDevice(ctx context.Context, devInfo interface{}, devType DeviceType) (interface{}, error) {
|
|
switch devType {
|
|
case CpuDev:
|
|
return devInfo.(uint32), nil
|
|
case MemoryDev:
|
|
memdev := devInfo.(*MemoryDevice)
|
|
return memdev.SizeMB, nil
|
|
}
|
|
return nil, nil
|
|
}
|
|
|
|
func (m *mockHypervisor) HotplugRemoveDevice(ctx context.Context, devInfo interface{}, devType DeviceType) (interface{}, error) {
|
|
switch devType {
|
|
case CpuDev:
|
|
return devInfo.(uint32), nil
|
|
case MemoryDev:
|
|
return 0, nil
|
|
}
|
|
return nil, nil
|
|
}
|
|
|
|
func (m *mockHypervisor) GetVMConsole(ctx context.Context, sandboxID string) (string, string, error) {
|
|
return "", "", nil
|
|
}
|
|
|
|
func (m *mockHypervisor) ResizeMemory(ctx context.Context, memMB uint32, memorySectionSizeMB uint32, probe bool) (uint32, MemoryDevice, error) {
|
|
if m.config.MemorySize != memMB {
|
|
// For testing, we'll use MemSlots to track how many times we resized memory
|
|
m.config.MemSlots += 1
|
|
m.config.MemorySize = memMB
|
|
}
|
|
return 0, MemoryDevice{}, nil
|
|
}
|
|
func (m *mockHypervisor) ResizeVCPUs(ctx context.Context, cpus uint32) (uint32, uint32, error) {
|
|
return 0, 0, nil
|
|
}
|
|
|
|
func (m *mockHypervisor) GetTotalMemoryMB(ctx context.Context) uint32 {
|
|
return m.config.MemorySize
|
|
}
|
|
func (m *mockHypervisor) Disconnect(ctx context.Context) {
|
|
}
|
|
|
|
func (m *mockHypervisor) GetThreadIDs(ctx context.Context) (VcpuThreadIDs, error) {
|
|
vcpus := map[int]int{0: os.Getpid()}
|
|
return VcpuThreadIDs{vcpus}, nil
|
|
}
|
|
|
|
func (m *mockHypervisor) Cleanup(ctx context.Context) error {
|
|
return nil
|
|
}
|
|
|
|
func (m *mockHypervisor) GetPids() []int {
|
|
return []int{m.mockPid}
|
|
}
|
|
|
|
func (m *mockHypervisor) GetVirtioFsPid() *int {
|
|
return nil
|
|
}
|
|
|
|
func (m *mockHypervisor) fromGrpc(ctx context.Context, hypervisorConfig *HypervisorConfig, j []byte) error {
|
|
return errors.New("mockHypervisor is not supported by VM cache")
|
|
}
|
|
|
|
func (m *mockHypervisor) toGrpc(ctx context.Context) ([]byte, error) {
|
|
return nil, errors.New("mockHypervisor is not supported by VM cache")
|
|
}
|
|
|
|
func (m *mockHypervisor) Save() (s hv.HypervisorState) {
|
|
return
|
|
}
|
|
|
|
func (m *mockHypervisor) Load(s hv.HypervisorState) {}
|
|
|
|
func (m *mockHypervisor) Check() error {
|
|
return nil
|
|
}
|
|
|
|
func (m *mockHypervisor) GenerateSocket(id string) (interface{}, error) {
|
|
return types.MockHybridVSock{
|
|
UdsPath: MockHybridVSockPath,
|
|
}, nil
|
|
}
|
|
|
|
func (m *mockHypervisor) IsRateLimiterBuiltin() bool {
|
|
return false
|
|
}
|