UniqAI/nvidia-container-toolkit
1
0
Fork 0
mirror of https://github.com/NVIDIA/nvidia-container-toolkit synced 2024-11-24 21:14:00 +00:00
Code Issues Actions Packages Projects Releases Wiki Activity

Only init nvml as required when generating CDI specs

Browse Source 
CDI generation modes such as management and wsl don't require
NVML. This change removes the top-level instantiation of nvmllib
and replaces it with an instanitation in the nvml CDI spec generation
code.

Signed-off-by: Evan Lezar <elezar@nvidia.com>
This commit is contained in:
Evan Lezar 2023-03-15 10:12:44 +02:00
parent 380eb8340a
commit 685802b1ce
4 changed files with 12 additions and 12 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

1
CHANGELOG.md
View File

@ -3,6 +3,7 @@
## v1.13.0-rc.3
* Prefer /run over /var/run when locating nvidia-persistenced and nvidia-fabricmanager sockets.
* Only initialize NVML for modes that require it when runing `nvidia-ctk cdi generate`
## v1.13.0-rc.2

12
cmd/nvidia-ctk/cdi/generate/generate.go
View File

@ -30,8 +30,6 @@ import (
specs "github.com/container-orchestrated-devices/container-device-interface/specs-go"
"github.com/sirupsen/logrus"
"github.com/urfave/cli/v2"
"gitlab.com/nvidia/cloud-native/go-nvlib/pkg/nvlib/device"
"gitlab.com/nvidia/cloud-native/go-nvlib/pkg/nvml"
)
const (
@ -190,21 +188,11 @@ func (m command) generateSpec(cfg *config) (spec.Interface, error) {
return nil, fmt.Errorf("failed to create device namer: %v", err)
}
nvmllib := nvml.New()
if r := nvmllib.Init(); r != nvml.SUCCESS {
return nil, r
}
defer nvmllib.Shutdown()
devicelib := device.New(device.WithNvml(nvmllib))
cdilib := nvcdi.New(
nvcdi.WithLogger(m.logger),
nvcdi.WithDriverRoot(cfg.driverRoot),
nvcdi.WithNVIDIACTKPath(cfg.nvidiaCTKPath),
nvcdi.WithDeviceNamer(deviceNamer),
nvcdi.WithDeviceLib(devicelib),
nvcdi.WithNvmlLib(nvmllib),
nvcdi.WithMode(string(cfg.mode)),
)

5
pkg/nvcdi/driver-nvml.go
View File

@ -31,6 +31,11 @@ import (
// NewDriverDiscoverer creates a discoverer for the libraries and binaries associated with a driver installation.
// The supplied NVML Library is used to query the expected driver version.
func NewDriverDiscoverer(logger *logrus.Logger, driverRoot string, nvidiaCTKPath string, nvmllib nvml.Interface) (discover.Discover, error) {
if r := nvmllib.Init(); r != nvml.SUCCESS {
return nil, fmt.Errorf("failed to initalize NVML: %v", r)
}
defer nvmllib.Shutdown()
version, r := nvmllib.SystemGetDriverVersion()
if r != nvml.SUCCESS {
return nil, fmt.Errorf("failed to determine driver version: %v", r)

6
pkg/nvcdi/lib-nvml.go
View File

@ -24,6 +24,7 @@ import (
"github.com/container-orchestrated-devices/container-device-interface/pkg/cdi"
"github.com/container-orchestrated-devices/container-device-interface/specs-go"
"gitlab.com/nvidia/cloud-native/go-nvlib/pkg/nvlib/device"
"gitlab.com/nvidia/cloud-native/go-nvlib/pkg/nvml"
)
type nvmllib nvcdilib
@ -39,6 +40,11 @@ func (l *nvmllib) GetSpec() (spec.Interface, error) {
func (l *nvmllib) GetAllDeviceSpecs() ([]specs.Device, error) {
var deviceSpecs []specs.Device
if r := l.nvmllib.Init(); r != nvml.SUCCESS {
return nil, fmt.Errorf("failed to initalize NVML: %v", r)
}
defer l.nvmllib.Shutdown()
gpuDeviceSpecs, err := l.getGPUDeviceSpecs()
if err != nil {
return nil, err