|
| 1 | +//go:build !darwin && !windows |
| 2 | + |
| 3 | +/* |
| 4 | + * Copyright (c) 2025, NVIDIA CORPORATION. All rights reserved. |
| 5 | + * |
| 6 | + * Licensed under the Apache License, Version 2.0 (the "License"); |
| 7 | + * you may not use this file except in compliance with the License. |
| 8 | + * You may obtain a copy of the License at |
| 9 | + * |
| 10 | + * http://www.apache.org/licenses/LICENSE-2.0 |
| 11 | + * |
| 12 | + * Unless required by applicable law or agreed to in writing, software |
| 13 | + * distributed under the License is distributed on an "AS IS" BASIS, |
| 14 | + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. |
| 15 | + * See the License for the specific language governing permissions and |
| 16 | + * limitations under the License. |
| 17 | + */ |
| 18 | + |
| 19 | +package main |
| 20 | + |
| 21 | +import ( |
| 22 | + "fmt" |
| 23 | + |
| 24 | + "github.com/sirupsen/logrus" |
| 25 | + "github.com/urfave/cli/v2" |
| 26 | + |
| 27 | + "github.com/NVIDIA/k8s-driver-manager/internal/nvpci" |
| 28 | +) |
| 29 | + |
| 30 | +type unbindCommand struct { |
| 31 | + logger *logrus.Logger |
| 32 | + nvpciLib nvpci.Interface |
| 33 | +} |
| 34 | + |
| 35 | +type unbindOptions struct { |
| 36 | + all bool |
| 37 | + deviceID string |
| 38 | +} |
| 39 | + |
| 40 | +// newUnbindCommand constructs an unbind command with the specified logger |
| 41 | +func newUnbindCommand(logger *logrus.Logger) *cli.Command { |
| 42 | + c := unbindCommand{ |
| 43 | + logger: logger, |
| 44 | + nvpciLib: nvpci.New(), |
| 45 | + } |
| 46 | + return c.build() |
| 47 | +} |
| 48 | + |
| 49 | +// build the unbind command |
| 50 | +func (m unbindCommand) build() *cli.Command { |
| 51 | + cfg := unbindOptions{} |
| 52 | + |
| 53 | + // Create the 'unbind' command |
| 54 | + c := cli.Command{ |
| 55 | + Name: "unbind", |
| 56 | + Usage: "Unbind device(s) from their current driver", |
| 57 | + Before: func(c *cli.Context) error { |
| 58 | + return m.validateFlags(&cfg) |
| 59 | + }, |
| 60 | + Action: func(c *cli.Context) error { |
| 61 | + return m.run(&cfg) |
| 62 | + }, |
| 63 | + Flags: []cli.Flag{ |
| 64 | + &cli.BoolFlag{ |
| 65 | + Name: "all", |
| 66 | + Aliases: []string{"a"}, |
| 67 | + Destination: &cfg.all, |
| 68 | + Usage: "Bind all NVIDIA devices to vfio-pci", |
| 69 | + }, |
| 70 | + &cli.StringFlag{ |
| 71 | + Name: "device-id", |
| 72 | + Aliases: []string{"d"}, |
| 73 | + Destination: &cfg.deviceID, |
| 74 | + Usage: "Specific device ID to bind (e.g., 0000:01:00.0)", |
| 75 | + }, |
| 76 | + }, |
| 77 | + } |
| 78 | + |
| 79 | + return &c |
| 80 | +} |
| 81 | + |
| 82 | +func (m unbindCommand) validateFlags(cfg *unbindOptions) error { |
| 83 | + if !cfg.all && cfg.deviceID == "" { |
| 84 | + return fmt.Errorf("either --all or --device-id must be specified") |
| 85 | + } |
| 86 | + |
| 87 | + if cfg.all && cfg.deviceID != "" { |
| 88 | + return fmt.Errorf("cannot specify both --all and --device-id") |
| 89 | + } |
| 90 | + |
| 91 | + return nil |
| 92 | +} |
| 93 | + |
| 94 | +func (m unbindCommand) run(cfg *unbindOptions) error { |
| 95 | + if cfg.deviceID != "" { |
| 96 | + return m.unbindDevice(cfg.deviceID) |
| 97 | + } |
| 98 | + |
| 99 | + return m.unbindAll() |
| 100 | +} |
| 101 | + |
| 102 | +func (m unbindCommand) unbindAll() error { |
| 103 | + devices, err := m.nvpciLib.GetGPUs() |
| 104 | + if err != nil { |
| 105 | + return fmt.Errorf("failed to get NVIDIA GPUs: %w", err) |
| 106 | + } |
| 107 | + |
| 108 | + for _, dev := range devices { |
| 109 | + m.logger.Infof("Unbinding device %s", dev.Address) |
| 110 | + // (cdesiniotis) ideally this should be replaced by a call to nvdev.UnbindFromDriver() |
| 111 | + if err := m.nvpciLib.UnbindFromDriver(dev); err != nil { |
| 112 | + m.logger.Warnf("Failed to unbind device %s: %v", dev.Address, err) |
| 113 | + } |
| 114 | + } |
| 115 | + return nil |
| 116 | +} |
| 117 | + |
| 118 | +func (m unbindCommand) unbindDevice(device string) error { |
| 119 | + nvdev, err := m.nvpciLib.GetGPUByPciBusID(device) |
| 120 | + if err != nil { |
| 121 | + return fmt.Errorf("failed to get NVIDIA GPU device: %w", err) |
| 122 | + } |
| 123 | + if nvdev == nil || !nvdev.IsGPU() { |
| 124 | + m.logger.Infof("Device %s is not a GPU", device) |
| 125 | + return nil |
| 126 | + } |
| 127 | + |
| 128 | + m.logger.Infof("Unbinding device %s", device) |
| 129 | + |
| 130 | + // (cdesiniotis) ideally this should be replaced by a call to nvdev.UnbindFromDriver() |
| 131 | + if err := m.nvpciLib.UnbindFromDriver(nvdev); err != nil { |
| 132 | + return fmt.Errorf("failed to unbind device %s from driver: %w", device, err) |
| 133 | + } |
| 134 | + |
| 135 | + return nil |
| 136 | +} |
0 commit comments