MRE request: virtualbox

[Gianfranco Costamagna]

Hello, since there might be some new SRU ongoing of virtualbox, it would be nice if you can have another look to this...
Or to let me know if something is not clear or still missing!

Have a great day

Gianfranco






Il giovedì 29 febbraio 2024 alle ore 09:44:15 CET, Gianfranco Costamagna <locutusofborg at debian.org> ha scritto: 





Hello Christopher thanks for the huge work!


>Sorry for the delay in addressing this.

No need to be sorry at all, we are all under high work load, and I appreciate you all taking the steps to finally solve this long
standing issue. Let's try to finish it! :)

>This does seem like Virtualbox fits well enough in the HWE category (and 
>probably has sufficient upstream testing) for special SRU treatment.
>
>I've taken the liberty of reworking the wiki page 
>https://wiki.ubuntu.com/VirtualboxUpdates to make it easy to review and 
>accept Virtualbox SRU bugs. I've done this to the best of my 
>understanding, but you're the domain expert, so please check my work :)

If you didn't understand something, it means I wasn't clear enough, so thanks
for adding another pair of eyes

>Particularly: it's my understanding that this does *not* require any 
>special coordination with the kernel - the Ubuntu kernel contains 
>Virtualbox guest driver modules, but these are upstream kernel modules, 
>and the Ubuntu kernel build does not build virtualbox-dkms modules at 
>kernel-build time¹

Yes and no.
Virtualbox ships both host and guest dkms modules. People might want to install
guest-dkms even if they are provided by the kernel (they shuld have an higher priority)

Reasons might be multiple, bugs in upstream dkms module, or better behaviour.
I don't recall having issues with upstream guest dkms modules, but I didn't want
to drop it because of older kernels being still around, and for safety reasons.

For *host* kernel module this is sadly a little bit trickier. Newer kernel might
break vboxdrv.ko buildability, so for each new kernel, the dkms autopkgtests
should be looked at carefully for regressions (in both host and guest).


>I've also marked a couple of places (with *******s) where the test case 
>instructions were not clear to me. The intent with test case 
>documentation is that it should be detailed enough that anyone reading 
>the bug could perform the steps and we could be confident that they have 
>exercised the expected tests.

I fully agree and sorry for not being clear enough, I do this testing in an automatic
way on my head, so I'm pretty sure I missed lots of things.

>Specifically, the questions I had were:
>
>* instructions for installing/removing the guest additions from the iso 
>pack (or pointers to documentation for that)
>
>* What the “various other tests” are - you mention changing 
>configuration and vboxmanage?
>
>* Do we need to verify that installing virtualbox guest packages 
>*outside* a VM does not interfere with the system? There seems to have 
>been at least one bug raised about that in the past.

I don't recall such issues, but in any case didn't have such reports in the last like 5-10 years...
I'll add it in case, just to be extra safe.

>Sorry again for the delay,

Sorry again from my side for not putting the page in a clear enough way, it would have made things easier for
all of us :)

Feel free to have another look, the page will be updated in some minutes.


Gianfranco

-- 
Ubuntu-release mailing list
Ubuntu-release at lists.ubuntu.com
Modify settings or unsubscribe at: https://lists.ubuntu.com/mailman/listinfo/ubuntu-release