Re: [qubes-users] Re: Installation, no AMD-vi, interrupt mapping, etc.

2018-10-04 Thread taii...@gmx.com
Most consumer mobos have broken IOMMU, I suggest instead of wasting your time trying to make it work you simply buy a KCMA-D8 or KGPE-D16 plus used cpu and install coreboot-libre. Without HVM/IOMMU your security will suck. -- You received this message because you are subscribed to the Google Gro

[qubes-users] Re: Installation, no AMD-vi, interrupt mapping, etc.

2018-10-02 Thread Sergio Matta
Em terça-feira, 2 de outubro de 2018 09:17:43 UTC-3, Sergio Matta escreveu: > Em segunda-feira, 1 de outubro de 2018 19:55:24 UTC-3, Sergio Matta escreveu: > > Em segunda-feira, 1 de outubro de 2018 16:04:34 UTC-3, naas...@gmail.com > > escreveu: > > > Installation went fine except for a libxen

[qubes-users] Re: Installation, no AMD-vi, interrupt mapping, etc.

2018-10-02 Thread Sergio Matta
Em segunda-feira, 1 de outubro de 2018 19:55:24 UTC-3, Sergio Matta escreveu: > Em segunda-feira, 1 de outubro de 2018 16:04:34 UTC-3, naas...@gmail.com > escreveu: > > Installation went fine except for a libxenlight config error of some kind. > > I still can't enable IOMMU using either of the

[qubes-users] Re: Installation, no AMD-vi, interrupt mapping, etc.

2018-10-01 Thread Sergio Matta
Em segunda-feira, 1 de outubro de 2018 16:04:34 UTC-3, naas...@gmail.com escreveu: > Installation went fine except for a libxenlight config error of some kind. I > still can't enable IOMMU using either of the approaches described in that > Ubuntu thread, even though it successfully worked with

[qubes-users] Re: Installation, no AMD-vi, interrupt mapping, etc.

2018-10-01 Thread naasking
I meant qemu-kvm not VirtualBox. I now see this table [1] that claims that Xen doesn't support IOMMU on rev.1 of our mobo have, but does on rev.2. I'll probably go with the alternative I have in mind then since, unless someone else has any ideas. If not, thanks for the suggestions everyone! Sa

[qubes-users] Re: Installation, no AMD-vi, interrupt mapping, etc.

2018-10-01 Thread naasking
Installation went fine except for a libxenlight config error of some kind. I still can't enable IOMMU using either of the approaches described in that Ubuntu thread, even though it successfully worked with raw Linux. What boot parameters did you add? I have the earlier rev.1 Sabertooth 990FX mo

[qubes-users] Re: Installation, no AMD-vi, interrupt mapping, etc.

2018-10-01 Thread Sergio Matta
Are you suggesting I just proceed with the installation regardless and try to set these parameters in the grub booting config after I get up and running? > > Sandro > Yes! And even if iommu does not works, you will be able to use Qubes 4 VM as PV. Networking will need to set up by hand and sav

[qubes-users] Re: Installation, no AMD-vi, interrupt mapping, etc.

2018-10-01 Thread naasking
Yes, I read that thread as well. The last post in that thread did not enable IOMMU on Linux, the second to last did and it's the same procedure I linked in my first post on superuser.com. However, adding these parameters while booting the Qubes installer doesn't seem to have an effect. Are you

[qubes-users] Re: Installation, no AMD-vi, interrupt mapping, etc.

2018-10-01 Thread Sergio Matta
> I updated the bios but still no luck, so I tried a manual procedure as > described at [1]: I ran a recent live linux distro from a USB key and > confirmed that interrupt remapping was disabled by default due to this BIOS > bug. I then figured out the IOMMU and SMBus addresses using the descri