win_virtio_driver_install_by_installer: support viomem test

[menli820]

Add supoport to test viomem related from
the installer.
ID: 3171, 3191, 3192, 3193, 3194

[menli820]

@xiagao @heywji please help to review it.
@xiagao for the viofs change part, please have a check, the change just make sure we can start the viomem device normally, maybe other parameters should be deleted as well.

[menli820]

@vivianQizhu A quick question is How do think we won't consider the rpm version for viomem, so the cfg will be more easier to organize for now and in the future if new feature coming ?

Currently we organize the viomem part in the cfg file considering the rpm version, the question is that as xiaogao comment "in the future when there are more drivers support, I think it's hard to extend the variants."

Both are okay for me, so just to confirm we all agree the same method before I take action to modify it again.

[heywji]

I confirmed with @menli820 about some failure test results offline discussion, and the root cause is a known product bug. @menli820 will post the new patch to avoid the effect of the product bug. And then, I will review it again.

I have my Kar cmdline here for you to refer to.

python3 ConfigTest.py --testcase=win_virtio_driver_install_by_installer..with_viomem,win_virtio_driver_installer_uninstall..with_viomem --guestname=Win11.x86_64  --machines=q35 --customsparams="cdrom_virtio = isos/windows/virtio-win-1.9.44-0.el9_5.iso" --firmware=ovmf --clone=yes  --driveformat=ahci

[menli820]

@xiagao @heywji call for your review again when you free, thanks

[heywji]

Kar encountered a product bug unrelated to this patch(the installer code related has passed) and confirmed with @menli820, so LGTM.

[xiagao]

It was defined in above.

[menli820]

Actually in my understanding it is not a conflict，the first ‘boot_with_viomem = no’ is for i386.
the second if for both x86_64 and i386.