fix(core): fix VMs with EFIWithSecureBoot bootloader failing to start when configured with more than 12 vCPUs

[fl64]

Description

Fix VMs with EFIWithSecureBoot bootloader failing to start when configured with more than 12 vCPUs.

• Enable persistent NVRAM for SecureBoot VMs
• Increase EDK2 limits to support up to 128 vCPUs

Why do we need it, and what problem does it solve?

VMs with SecureBoot were stuck in an infinite reboot loop or showing black screen when using more than ~12 cores.

What is the expected result?

VMs with bootloader: EFIWithSecureBoot boot successfully with any number of cores (up to 128).

Checklist

• The code is covered by unit tests.
• e2e tests passed.
• Documentation updated according to the changes.
• Changes were tested in the Kubernetes cluster manually.

Changelog entries

section: core
type: fix
summary: fix VMs with `EFIWithSecureBoot` bootloader failing to start when configured with more than 12 vCPUs.

[deckhouse-BOaTswain]

Workflow has started.
Follow the progress here: Workflow Run

The target step completed with status: failure.

[diafour]

-D DEBUG_ON_SERIAL_PORT=TRUE — is it work without -b DEBUG?

Also, these 2 new settings are commented with "TEMP" — do we need them in main?

[fl64]

I left it on purpose so that, if necessary, it could be quickly build with a debug build. But if such points are obvious, then you can remove them.

[diafour]

It is not obvious what to change to build with debug output =)

[diafour]

Do we need to enable debug build with -b DEBUG to enable output on serial port, or -D DEBUG_ON_SERIAL_PORT is enough and there is a runtime option for more verbose output for RELEASE build? Or we need rebuild ovmf to get verbose log on serial port? (I remember that ovmf rebuild may take ~40min).

[fl64]

Just put DEBUG instead of RELEASE )