fix nvidia-k8s-device-plugin linker flags

[bcressey]

Issue number:
N/A

Description of changes:
The flags to compile the nvidia-k8s-device-plugin have to be adjusted for the new SDK.

Specifically, the default GOLDFLAGS now includes -extldflags set to the stock LDFLAGS value. Builds that override CGO_LDFLAGS also need to override GOLDFLAGS so that the linker flags match.

Testing done:
Before:

bash-5.1# systemctl status nvidia-k8s-device-plugin
● nvidia-k8s-device-plugin.service - Start NVIDIA kubernetes device plugin
     Loaded: loaded (/aarch64-bottlerocket-linux-gnu/sys-root/usr/lib/systemd/system/nvidia-k8s-device-plugin.service; enabled; preset: enabled)
     Active: activating (auto-restart) (Result: exit-code) since Sat 2024-04-27 15:40:42 UTC; 1s ago
    Process: 6080 ExecStart=/usr/bin/nvidia-device-plugin --device-list-strategy volume-mounts --device-id-strategy index --pass-device-specs=true (code=exited, status=127)
   Main PID: 6080 (code=exited, status=127)
        CPU: 2ms

bash-5.1# /usr/bin/nvidia-device-plugin --device-list-strategy volume-mounts --device-id-strategy index --pass-device-specs=true
/usr/bin/nvidia-device-plugin: symbol lookup error: /usr/bin/nvidia-device-plugin: undefined symbol: nvmlGpuInstanceGetComputeInstanceProfileInfoV
bash-5.1# systemctl status nvidia-k8s-device-plugin 

After:

bash-5.1# systemctl status nvidia-k8s-device-plugin
● nvidia-k8s-device-plugin.service - Start NVIDIA kubernetes device plugin
     Loaded: loaded (/aarch64-bottlerocket-linux-gnu/sys-root/usr/lib/systemd/system/nvidia-k8s-device-plugin.service; enabled; preset: enabled)
     Active: active (running) since Sat 2024-04-27 15:50:30 UTC; 3min 22s ago
   Main PID: 1390 (nvidia-device-p)
      Tasks: 9 (limit: 9232)
     Memory: 56.6M
        CPU: 5.956s
     CGroup: /system.slice/nvidia-k8s-device-plugin.service
             └─1390 /usr/bin/nvidia-device-plugin --device-list-strategy volume-mounts --device-id-strategy index --pass-device-specs=true

Terms of contribution:

By submitting this pull request, I agree that this contribution is dual-licensed under the terms of both the Apache License, version 2.0, and the MIT license.