Appreciate the additional context! Have thankfully not needed to use the safetynet module with microg either.
grow a plant, hug your dog, lift heavy, eat healthy, be a nerd, play a game and help each other out
Appreciate the additional context! Have thankfully not needed to use the safetynet module with microg either.
I appreciate that you’re trying to inform me but if you make such a claim, you should be able to prove it.
A friend was able to provide some context, regardless:
The one binary I’m aware of microG downloading (assuming it still does) is the SafetyNet “DroidGuard” thing, which it only does if you explicitly enable SafetyNet, which is not on by default. There is no other way to provide it.
microG only has privileged access if you install it as a privileged app, which is up to you / your distribution, as microG works fine as a user app (provided signature spoofing is available to it). Also, being privileged itself really doesn’t mean giving privileges to “Google”.
Apps needing Google services may indeed contain all sorts of binaries, generally including Google ones, which doesn’t mean they contain Google services themselves. Anyway, they are proprietary apps and as such will certainly contain proprietary things, and it’s all to you to install them or not. It’s not like microG includes them.
Its also just a reimplementation of a small handful of useful Google services, such as push notifications, or the maps (not the spyware stuff like advertising) and each can be toggled on/off.
Also all apps on android are sandboxed
I appreciate the info. For my own learning, could you provide a link to some context around the types of official binaries leveraged by microG? The only firm info I have of its behaviour is that it will pseudonomise as much user information as possible.
I’m familiar with sandboxed google play on grapheneOS and have used it in the past.
Can you elaborate on being misled there?
As for google devices - yes, there’s irony in the notion that the most de-googleable phones are theirs, sure. They’re often sold at a loss around the holiday season, though.
I also use calyx but I’ll agree that graphene is technologically superior of the two. I’m more comfortable with the idea of using MicroG as opposed to sandboxes google play but that’s not to slant the implementation in any way.
Best of luck with this, let us know how it goes
It’s kind of crazy to me how well it works! It’s hard for me to wrap my head around it sometimes.
My end goal is to not have to eventually not need to use windows at all but I’m still very impressed with how this behaves.
Very welcome! Yes, exactly as you described. The nice thing is that you have greater control over Windows in this virtualized environment, particularly with regards to limiting device and network access.
I gather that display dummy plugs are pretty common in the looking glass community.
KVM/QEMU via virt-manager. I would imagine that your use case would work if you pass the USB device or the entire usb host controller through to the VM, but I’m not sure. Please check the video linked in my other comment for more information on the single GPU setup
Hey there, just using a single GPU in this system. If you have multiple adapters, you can try something like LookingGlass instead. In my case, I would need a single GPU that supports SRIOV, which is typically relegated to data centre products (I believe someone actually managed this with an Intel iGPU + and experimental sriov driver!).
I’m just passing my GPU through to a virtual machine; it takes precedence over the graphical session, leverages all connected displays and relevant peripherals, and gracefully resumes back into GDM / GNOME once the VM is powered off (can do this conventionally within W10).
I mostly followed this video:
https://www.youtube.com/watch?v=eTWf5D092VY
key thing for AMD gfx is to set ROMBAR = 0 in virt config, this will allow you to actually get functioning display output once the VM is started up.
As for your buying choices, consumer AMD GPUs have issues with GPU reset (unlike Intel or Nvidia). I think your experience with nvidia graphics here will be better than mine here with amd.
Byt yeah, since you have multiple gfx adapters at your disposal, it should be possible to get started with LookingGlass (a VM in a movable, resizable window that is fully hw accelerated with shared memory). The Level1Techs forum for LG is very helpful, though I believe the creator of the video above also has a relevant guide for this.
None whatsoever. You’re good to use heliboard with the gesture input library. There is also the FUTO keyboard, which somehow has a gesture input library built in. I would presume theirs is FOSS but I’ve not done my homework on this.
E: as someone else mentioned, you can deny network permission for gboard with ROMs like LineageOS, GrapheneOS, CalyxOS etc. This is what I did for many years until heliboard came along.
I got VFIO/IOMMU + single GPU pass-through working on Fedora 40 with my RX 6800 xt into a win10 VM.
More of a see if I could sort of thing, I don’t imagine I will actually need it much, but it may help if any of my friends are curious about switching over.
Speaking from my experience with fedora and windows 10 and 11 within the same system.
As others have stated here, If you can, please keep each operating system on it’s own physical disk. Disconnect others if you perform a new Windows install on any, as it’ll attempt to store its bootloader on disk 0 regardless of the OS destination drive.
LUKS2 is part of the fedora workstation setup, I imagine it will be presented to you upon install with Mint. If you’re on separate physical disks, you shouldn’t have much to worry about, but as far as I’m aware, you’re okay to use disk encryption on drives partitioned with two systems.
There’s a Dropbox .deb and .rpm for linux as far as I can tell, but I cannot attest to its quality or how well it integrates with a given file manager. Cloud accounts are generally well supported amongst the key desktop environments, for which I’d consider Cinnamon to be a part of.
Modern, mainstream distributions are pretty GUI friendly. I fully expect you to be able to get by on Mint without needing to touch the command line much if at all. That said, I grab CLI oriented tools from the terminal and graphical apps from the app store. Enabling flathub will give you access to a broad selection of graphical software so by all means, go for it.
I’m not wise so I’ll hold back here. I will say that Fedora has allowed me to approach linux as an absolute casual for nearly 6 years now.
Very key points! Some distros will also accommodate window’s default timekeeping if a win install is detected, and also need to be changed retroactively to prevent wonky behaviour with DST
That’s fair enough, though one of the characteristics I had in mind was also battery life (that said, it would be at a given level of performance so either way).
Also definitely not thrilled about things like ME, Pluton and so on.
We’re a ways away from reaching equivalent performance characteristics of the currently available options they have with RISC V, but I would also love to see that as well.
Does hyprland work for your setup?
Yes, exactly this, though I know Jeremy Soller from System76 has been working on Coreboot for AMD UMA platforms a couple years back.
There’s hope, still 😅
Thank you for this info! I expected this to be more feasible on the Intel models for whatever reason. Glad to see a glimmer of hope for the AMD platforms.
And the Linux / Unix-specific ecosystem & technology arguments therein.