you are viewing a single comment's thread
view the rest of the comments
[–] 41 points 1 year ago* (29 children)

#8 reawakened my nervousness about the lack of virus protection on Linux. With every milestone we celebrate it becomes more likely that malicious people target desktop Linux with their malware, and I don't think the "Linux is inherently secure" mentality helps. I hope clamav's on access scanner is fixed and improved so it becomes commonplace before there's some big newsworthy scandal.

  • source
  • hideshow 29 child comments
  • [–] 15 points 1 year ago

    Granular permissioned access for apps from trusted supply chains is better than attempting deny lists based on signatures (AV).

    I still use it, but I put way more effort into SLSA, securing containers, flatpaks, and limiting their blow back. From there its keeping up with CVEs in ways that do not create more or break functionality.

    I will say A LOT of the Linux software ecosystem is was more secure than Window's default.

  • source
  • parent
  • [–] 15 points 1 year ago (5 children)

    People were saying the exact same thing when I first started using Linux in 1999-ish

  • source
  • parent
  • hideshow 5 child comments
  • [–] 6 points 1 year ago (2 children)

    What is survivorship bias aka gambler's fallacy?

  • source
  • parent
  • hideshow 2 child comments
  • [–] 1 point 1 year ago

    That's not it though. Linux doesn't have a problem with viruses, didn't have it before, doesn't have it now. Predicting that it's going to start right now 20 years in a row isn't a good idea refardless, and pointing it out is a right thing to do.
    Which doesn't mean it couldn't really start one day.

  • source
  • parent
  • [–] 1 point 1 year ago (1 child)

    "I have never been hacked before. Why would it start happening now?"

  • source
  • parent
  • hideshow 1 child comment
  • [–] 2 points 1 year ago

    No one said that though. What was said was that as more people use Linux we will see more malware. And they've been saying that forever.

    Here is an article talking about it from 2003 https://www.theregister.com/2003/10/06/linux_vs_windows_viruses/

    If I knew how to search Usenet these days I'm sure I could find something from the 90s

    It's not untrue, actually, is common sense that there will be more malware, but it implies Linux's only advantage is security through obscurity.

    People have been worrying Linux is going basically become Windows over the next couple of years for 2-3 decades now. Malware is a serious problem that Linux users should take precautions from, but it seems clear by now that what security advantages it has aren't based on obscurity.

  • source
  • parent
  • [–] 8 points 1 year ago (1 child)

    I don't think a Linux anti virus program would be such a big security win. Phishing is the biggest security threat to most users, and no amount of software can prevent that.

    Sure, downloading and running random shit is a concern, but people in that group are a bit of a lost cause. The best solution for that is to harden the OS, prevent running executables through the GUI, or from user folders (I think SELinux could do that), disable sudo on the user account, and only allow installing Flatpaks. The security of Flathub may not be perfect, but it's a smaller attack surface than the whole internet.

    But even if you do that, an Indian call center scam is still going to manipulate your grandma into buying Amazon gift cards, so... It's a lost cause.

  • source
  • parent
  • hideshow 1 child comment
  • [–] 4 points 1 year ago

    ... but people in that group are a bit of a lost cause.

    touche. I don't think the existence of other threats is a reason to dismiss this one. And I don't think simply prohibiting running random executables is sufficient as it isn't 'most users' who are switching to Linux. The people likely to switch to Linux are also the people likely to want to run programs that aren't yet distributed in repos. I can imagine a scenario where the malware is hidden in a program hosted on a custom flatpak repo and requires permissions for normal operation that'd make flatseal ineffective for stopping the malware.

    The ideal anti-virus in my mind would ignore programs installed from official repos and on access scan ones installed from anywhere else. It'd also keep track of critical vulnerabilities to give you a heads up about updating your system.

  • source
  • parent
  • [–] 4 points 1 year ago* (10 children)

    Why are we shouting?

    Anyway, don't waste your time with "antivirus" software. That is not how you secure a system.

  • source
  • parent
  • hideshow 10 child comments
  • [–] 6 points 1 year ago (7 children)

    Your viewer must be parsing #8 as # 8.

    You're free to not 'waste time' with anti-virus but I prefer the peace of mind.

  • source
  • parent
  • hideshow 7 child comments
  • [–] 4 points 1 year ago (6 children)

    You need to put a backslash before the hash tag. In Markdown a # is a headet

  • source
  • parent
  • hideshow 6 child comments
  • [–] 4 points 1 year ago (1 child)

    No? There needs to be a space between on piefed and that's how it works on github too.

    https://piefed.social/comment/8602660

  • source
  • parent
  • hideshow 1 child comment
  • [–] 1 point 1 year ago (3 children)

    As Lojcs said...

    The Markdown format expects a space after whatever number of #s you put at the start, for it to be a corresponding level header.
    Due to different parsers having different types of leeways, it becomes a bit difficult to make sure stuff always matches.
    e.g. I was once mistaken about the way tabs work for multi-level bullets and numbering because GitLab had more leeway. Using discount, I realised where I was being wrong.

  • source
  • parent
  • hideshow 3 child comments
  • [–] 1 point 1 year ago (2 children)

    The whole problem is because format doesn't actually expects the space there, and it's left to the interpretation of the parser

  • source
  • parent
  • hideshow 2 child comments
  • [–] 1 point 1 year ago (1 child)

    I thought the whole problem was that there was no de-facto standard and people kept on making their deviations while still calling it Markdown.

    I personally like how Doxygen implements it.

  • source
  • parent
  • hideshow 1 child comment
  • [–] 2 points 1 year ago

    Yeah. The power of MD is that it's lightweight, versatile, and not very restrictive. You don't need to remember a lot, and the parser is dirt easy to implement.
    The negatives are that it's not very restrictive, nobody remembers what's what, and the parser so easy to implement, everyone and their dog has one, and they're all slightly different

  • source
  • parent
  • [–] 2 points 1 year ago (5 children)

    AV is a joke. Best thing is ephemerality. No persistence

  • source
  • parent
  • hideshow 5 child comments
  • [–] 6 points 1 year ago (4 children)

    Immutable, ephemerable, granularly permissioned, and encrypt EVERYTHING to enforce said permissions.

    1000x better than software signature hunting

  • source
  • parent
  • hideshow 4 child comments
  • [–] 3 points 1 year ago (1 child)

    You lost me at the encryption part. How does encrypting enforce permissions?

  • source
  • parent
  • hideshow 1 child comment
  • [–] 2 points 1 year ago

    Enforces confidentiality and integrity.

    Encryption on transports protects from man in the middle and sniffing. At rest protects evil maid exploits, which for these systems is more about preventing malicious software being swapped in place of trusted software.

    The same applies to encryption of links like pcie and memory with the time of transport and rest changing.

  • source
  • parent
  • [–] -1 points 1 year ago

    Well yeah, you need to use Linux, the Linux way to make it secure.

    If you are running VSCode as root after having logged into the DE as root and installing extensions willy-nilly, it won't matter that your RHEL has SELinux installed.

  • source
  • parent