-
Notifications
You must be signed in to change notification settings - Fork 1.2k
Commit
This commit does not belong to any branch on this repository, and may belong to a fork outside of the repository.
Merge branch 'main' into usama.saqib/add-ubuntu-24.04
- Loading branch information
Showing
26 changed files
with
838 additions
and
34 deletions.
There are no files selected for viewing
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -0,0 +1,31 @@ | ||
| # Report PR merged event to Datadog | ||
|
|
||
| name: Report Merged PR | ||
|
|
||
| on: | ||
| pull_request: | ||
| types: closed | ||
|
|
||
| permissions: {} | ||
|
|
||
| jobs: | ||
| if_merged: | ||
| if: github.event.pull_request.merged == true | ||
| runs-on: ubuntu-latest | ||
| steps: | ||
| - name: Checkout repository | ||
| uses: actions/checkout@0ad4b8fadaa221de15dcec353f45205ec38ea70b # v4.1.4 | ||
|
|
||
| - name: Setup Python3 | ||
| uses: actions/setup-python@f677139bbe7f9c59b41e40162b753c062f5d49a3 # v5.2.0 | ||
| with: | ||
| python-version: "3.12.6" | ||
| cache: "pip" | ||
| cache-dependency-path: '**/requirements*.txt' | ||
|
|
||
| - name: Install python dependencies | ||
| run: pip3 install -r requirements.txt | ||
|
|
||
| - name: Send merge event to Datadog | ||
| run: | | ||
| invoke -e github.pr-merge-dd-event-sender -p ${{ github.event.pull_request.number }} |
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -0,0 +1,126 @@ | ||
| // Unless explicitly stated otherwise all files in this repository are licensed | ||
| // under the Apache License Version 2.0. | ||
| // This product includes software developed at Datadog (https://www.datadoghq.com/). | ||
| // Copyright 2024-present Datadog, Inc. | ||
|
|
||
| //go:build linux_bpf | ||
|
|
||
| package cuda | ||
|
|
||
| import ( | ||
| "fmt" | ||
| "strconv" | ||
| "strings" | ||
|
|
||
| "github.com/NVIDIA/go-nvml/pkg/nvml" | ||
|
|
||
| "github.com/DataDog/datadog-agent/pkg/util/kernel" | ||
| ) | ||
|
|
||
| const cudaVisibleDevicesEnvVar = "CUDA_VISIBLE_DEVICES" | ||
|
|
||
| // GetVisibleDevicesForProcess modifies the list of GPU devices according to the | ||
| // value of the CUDA_VISIBLE_DEVICES environment variable for the specified | ||
| // process. Reference: | ||
| // https://docs.nvidia.com/cuda/cuda-c-programming-guide/index.html#env-vars. | ||
| // | ||
| // As a summary, the CUDA_VISIBLE_DEVICES environment variable should be a comma | ||
| // separated list of GPU identifiers. These can be either the index of the GPU | ||
| // (0, 1, 2) or the UUID of the GPU (GPU-<UUID>, or | ||
| // MIG-GPU-<UUID>/<instance-index>/<compute-index for multi-instance GPUs). UUID | ||
| // identifiers do not need to be the full UUID, it is enough with specifying the | ||
| // prefix that uniquely identifies the GPU. | ||
| // | ||
| // Invalid device indexes are ignored, and anything that comes after that is | ||
| // invisible, following the spec: "If one of the indices is invalid, only the | ||
| // devices whose index precedes the invalid index are visible to CUDA | ||
| // applications." If an invalid index is found, an error is returned together | ||
| // with the list of valid devices found up until that point. | ||
| func GetVisibleDevicesForProcess(systemDevices []nvml.Device, pid int, procfs string) ([]nvml.Device, error) { | ||
| cudaVisibleDevices, err := kernel.GetProcessEnvVariable(pid, procfs, cudaVisibleDevicesEnvVar) | ||
| if err != nil { | ||
| return nil, fmt.Errorf("cannot get env var %s for process %d: %w", cudaVisibleDevicesEnvVar, pid, err) | ||
| } | ||
|
|
||
| return getVisibleDevices(systemDevices, cudaVisibleDevices) | ||
| } | ||
|
|
||
| // getVisibleDevices processes the list of GPU devices according to the value of | ||
| // the CUDA_VISIBLE_DEVICES environment variable | ||
| func getVisibleDevices(systemDevices []nvml.Device, cudaVisibleDevices string) ([]nvml.Device, error) { | ||
| if cudaVisibleDevices == "" { | ||
| return systemDevices, nil | ||
| } | ||
|
|
||
| var filteredDevices []nvml.Device | ||
| visibleDevicesList := strings.Split(cudaVisibleDevices, ",") | ||
|
|
||
| for _, visibleDevice := range visibleDevicesList { | ||
| var matchingDevice nvml.Device | ||
| var err error | ||
| switch { | ||
| case strings.HasPrefix(visibleDevice, "GPU-"): | ||
| matchingDevice, err = getDeviceWithMatchingUUIDPrefix(systemDevices, visibleDevice) | ||
| if err != nil { | ||
| return filteredDevices, err | ||
| } | ||
| case strings.HasPrefix(visibleDevice, "MIG-GPU"): | ||
| // MIG (Multi Instance GPUs) devices require extra parsing and data | ||
| // about the MIG instance assignment, which is not supported yet. | ||
| return filteredDevices, fmt.Errorf("MIG devices are not supported") | ||
| default: | ||
| matchingDevice, err = getDeviceWithIndex(systemDevices, visibleDevice) | ||
| if err != nil { | ||
| return filteredDevices, err | ||
| } | ||
| } | ||
|
|
||
| filteredDevices = append(filteredDevices, matchingDevice) | ||
| } | ||
|
|
||
| return filteredDevices, nil | ||
| } | ||
|
|
||
| // getDeviceWithMatchingUUIDPrefix returns the first device with a UUID that | ||
| // matches the given prefix. If there are multiple devices with the same prefix | ||
| // or the device is not found, an error is returned. | ||
| func getDeviceWithMatchingUUIDPrefix(systemDevices []nvml.Device, uuidPrefix string) (nvml.Device, error) { | ||
| var matchingDevice nvml.Device | ||
| var matchingDeviceUUID string | ||
|
|
||
| for _, device := range systemDevices { | ||
| uuid, ret := device.GetUUID() | ||
| if ret != nvml.SUCCESS { | ||
| return nil, fmt.Errorf("cannot get UUID for device: %s", nvml.ErrorString(ret)) | ||
| } | ||
|
|
||
| if strings.HasPrefix(uuid, uuidPrefix) { | ||
| if matchingDevice != nil { | ||
| return nil, fmt.Errorf("non-unique UUID prefix %s, found UUIDs %s and %s", uuidPrefix, matchingDeviceUUID, uuid) | ||
| } | ||
| matchingDevice = device | ||
| matchingDeviceUUID = uuid | ||
| } | ||
| } | ||
|
|
||
| if matchingDevice == nil { | ||
| return nil, fmt.Errorf("device with UUID prefix %s not found", uuidPrefix) | ||
| } | ||
|
|
||
| return matchingDevice, nil | ||
| } | ||
|
|
||
| // getDeviceWithIndex returns the device with the given index. If the index is | ||
| // out of range or the index is not a number, an error is returned. | ||
| func getDeviceWithIndex(systemDevices []nvml.Device, visibleDevice string) (nvml.Device, error) { | ||
| idx, err := strconv.Atoi(visibleDevice) | ||
| if err != nil { | ||
| return nil, fmt.Errorf("invalid device index %s: %w", visibleDevice, err) | ||
| } | ||
|
|
||
| if idx < 0 || idx >= len(systemDevices) { | ||
| return nil, fmt.Errorf("device index %d is out of range [0, %d]", idx, len(systemDevices)-1) | ||
| } | ||
|
|
||
| return systemDevices[idx], nil | ||
| } |
Oops, something went wrong.