For example I don’t know if a cheap mini PC like the GMKtec G5 would be more likely to have firmware/driver related security flaws over a more standard PC like an older Dell Optiplex from Staples.
Maybe there is something else entirely I am overlooking. I’m not sure. Most of the complaints I’ve heard are in regards to mini PCs containing bloat and potentially malicious things installed.
Everyone in this thread focusing on malware put inside the BIOS by the manufacturer.
But the more realistic issue is just a hopelessly outdated BIOS with known vulnerabilities that won’t ever get an update.Do most motherboards receive long term support updates? I feel like I’ve only ever gotten a few updates for the ones I’ve used but maybe that was because I was buying older boards.
Server hardware does.
I think dell Rx30 are only just getting to EOL, and it was released in 2015.Although, buying an Rx30 before 5 years ago would be in the 10s of thousands.
Refurbished Rx40 and Rx50 are somewhat affordable.
I mean… what is your threat profile? Are you a LastPass engineer with an unpatched Plex install and access to the company file server? (edit: LastPass not LinkedIn)
Are you going to do something that would attract the attention of law enforcement or nation-state threat actors?
Are you going to be using this mini PC to do your taxes?
Is it going to be in a DMZ with open access to the Internet?
Are you going to use it as an authentication server for other critical assets?
If you aren’t assessing your risk level with some realistic idea of what threats actually apply to you and weighing that against the possible consequences of a breach, then you’re pointlessly worrying about low-probability scenarios. Operational Risk Management right? Judge your risk by probability of occurrence and severity of impact and then make decisions based on that.
These questions are the correct answer
I’m considering getting one for surfing the web on a secondary monitor. Nothing important is occurring but I asked this question for two reasons.
- I’m a nerd who finds a lot of this neat
- It can be hard to create a threat profile if you don’t know a threat exists.
Plenty of people put some level of work into making their PC’s OS more secure using YouTube tutorials and tips but they may not think to update their BIOS for example
It depends on your exposure profile.
Installing malware and bloatware into an OS is relatively easy. Doing the same to a bios is doable, but a LOT harder.
If you’re after a mini PC for home use or even a small business, wiping the os is likely fine. The concern would mostly kick in with larger organisations or government level targets.
It’s a question of how many man hours of effort hacking you is worth. Even if they are compromised, they are unlikely to risk outing the breach for anything less than a high value target.
It’s hard to tell.
You install your own windows? Do you really know if the pre-installed windows gets deleted?
You install your own linux? There is still the BIOS that could carry some unwanted “evil” software.
Then there is the possibility of unwanted “evil” hardware or firmware parts inside some of the chips…
You install your own windows? Do you really know if the pre-installed windows gets deleted?
This one seems to be easy to manage. Formatting the disc seems easy to do.
Deleting all existing partitions via the installer is already enough. Once a new partition is created, all files are essentially wiped.
Sure, repartitioning works too.
Once a new partition is created, all files are essentially wiped.
Partitioning does not wipe things.
It rewrites the partition tables, obviously, but no data elsewhere.
That’s why I said “essentially”. The data is no longer accessible by the system, which is all that’s needed to prevent malicious files from being executed.
is all that’s needed to prevent malicious files
Hmmm. So why do tools for “safe erase” even exist? ;-)
I doubt it’d have anything malicious in the firmware. I’d be more concerned about getting ripped off with fake parts or having it break on you.
There have been several cases of the default OS installation being infected with malware. I would recommend always wiping the drive and installing your own OS. Firmware is less likely to contain malware targeted at individuals (or anyone else that is irrelevant to APT’s), but it might be outdated and have unpatched vulnerabilities.
Probably about the same, you still have stuff like the intel management engine backdoor.
My only concern to these machines is that if theyre running an intel chip, its most likely to have intel ME which is a QOL stuff but a potential backdoor. Luckily, these cheap mini pcs has most of the time, an unlocked bios, so it is easier to pull the bios bin, patch it with intel ME removal and then reflash the patched bios.
I find the US branded ones (HP, DELL, etc.) more pain in the ass to patch the bios because most of them has locked bios. Doable but a pain in the ass for sure.
Worth noting that this stands in addition to the concerns already shared by OP.
Also, that differing generations of Intel chips yield varying degrees of “disabled ME” with the me_cleaner tools and/or HAP bit flipping. With the newer stuff being more tyrannical.
For AMD, as far as I know, there exist no such efforts like me_cleaner and users are left to simply to trust the benevolence of AMD’s vendor spec for “disabling” PSP in EFI settings. Assuming that does anything at all other than being a sugar pill.
While it is outside of the scope of most people’s abilities, the bios is on a flash chip that can be removed, read, and disassembled. I’m no expert here by any stretch. That said, my usual check with software is to simply check for
httpin strings. Even with a binary like a bios ROM, I can pass it through the$ stringscommand to look for any addresses. No matter what kind of malicious nonsense the software is doing, it has very low value unless it can dial out.My lack of a complete understanding in this area is why I use a whitelist firewall for most of my devices. It is also the ultimate ad and tracker blocker as I only visit the places I chose to access. I don’t conform to the lowest developer’s ethics and will simply stop using any site or service that fails to be direct and transparent.
The thing is, even most whitelist firewalls are inadequate. They only filter incoming packets. That is really an inadequate model in most cases now, especially with local large language models where it is impossible to verify their capabilities. My reason for all thus bla bla bla is to say, a whitelist on a trusted 3rd party device is a PITA but an effective low barrier way to prevent any bad actor from communicating with the questionable device. It still leaves you open to a potential situation where the device could be sending a packet stream to the outside world over something like UDP.
Otherwise, the main thing I would be concerned with, if it is a UEFI device, are the UEFI secure boot keys. Whomever holds these keys has a lower ring access than the operating system kernel. Anything happening in kernel or user space is effectively under their control.
Anyways, the main way to monitor and check the device is a trusted 3rd party router that blocks any unauthorized connections. This can be challenging to setup with something like OpenWRT. There is a forked OpenWRT device running a version that makes a lot of this easier called PC WRT. That can make a whitelist fw a little easier than sorting out NF Tables and scripting a whitelist firewall.
I’d at least get one with a motherboard from a reputable company and install my own Win/Linux build. Too much chance for bad actors otherwise.
