Ryan passed this along early this morning (really early for Ryan, seeing that we’re not even in the double digits of the morning hours!), and while this issue is only likely to affect a very small subset of users, anything that can completely brick a laptop in a matter of seconds is worrisome enough that we wanted to pass it along. File this one under the "things that should never happen" category.

You can read more of what is known at H-Online, but the short summary is this: Samsung’s UEFI implementation appears to be faulty. It was most likely tested with Windows only and found to work, but thorough testing with other operating systems doesn’t appear to have been a priority—or perhaps a consideration at all. At present, the bug appears to affect Samsung 530U3C, 300E5C, NP700Z5C, NP700Z7C, and NP900X4C series laptops; if you have one of those laptops, we recommend you exercise extreme caution if you have a need to boot into a Linux environment.

The bigger picture here is that this is what happens in a race to the bottom: corners get cut, which means less testing and validation, which can in turn lead to some catastrophic failures in specific circumstances.  What's really scary is that these Samsung laptops aren't even budget offerings, so the budget race ended up impacting higher priced offerings! Granted, in the increasingly complex world of computer hardware it can be difficult to test all of the likely scenarios. UEFI represents a fundamental change in how many low-level aspects of the computer function, however, so it needs to be thoroughly tested; not properly testing any OS besides Windows would be a gross oversight.

Long-term, we expect Samsung to release BIOS and firmware updates for the affected laptops, though how long that might take is unknown. Short-term, the workaround is for Linux to boot these Samsung models using the Compatibility Support Module (CSM), which basically bypasses the UEFI bootloader, but dual-booting via CSM appears to be a bit complex. Ubuntu’s development team has worked with Samsung and identified the kernel’s Samsung-laptop driver as the prime suspect, and there are other workarounds proposed already to address the issue. However, these fixes have not yet been merged into the main Linux development tree, so again we recommend Samsung laptop owners who use Linux exercise caution.

Update: It appears the problem stems from NVRAM corruption. Removing power, opening the laptop up, and disconnecting the CMOS battery appears like it will clear the problem, but that's a pretty serious set of steps to take for most laptops.

Source: H-Online

Comments Locked

42 Comments

View All Comments

  • Alexvrb - Friday, February 1, 2013 - link

    Unless they tested those USB chips using these Samsung laptops, then his point still stands. :D If you want a Linux-friendly laptop, get a Linux-friendly laptop. Check with the manufacturer and make sure they've tested Linux on it, or at least make sure it is supported and/or covered by warranty.

    I mean, did Samsung promise Linux would work on these? I could say that a Galaxy S III is a "general purpose computer" but that doesn't mean Samsung will care if I brick it by installing Linux (yes, it is possible :P).

    Should they fix it? Yes, because it is a bugged implementation and needs to be fixed. Should people be shocked and outraged? In my opinion no. But the precautionary article is nevertheless appreciated.
  • powerarmour - Thursday, January 31, 2013 - link

    If they don't test their machines with Linux, then they only have themselves to blame if it goes wrong. How about they get their hardware implementation correct in the first place, is that too much to ask?
  • sherlockwing - Thursday, January 31, 2013 - link

    It have a Windows sticker on it, suffice to say they don't care about linux at all.
  • tuxRoller - Thursday, January 31, 2013 - link

    Per http://lkml.indiana.edu/hypermail/linux/kernel/130... the problem is related to the samsung-laptop compatibility driver. That should be disabled if booting into a uefi environment b/c it works by assuming a bios is at work.
    The quick fix is just to have the module check, upon loading, to see if it is being booted from a uefi machine.
  • JarredWalton - Thursday, January 31, 2013 - link

    The fact that the driver can basically corrupt the NVRAM is scary, nonetheless. This is one of those new "features" I guess we get with UEFI?
  • tuxRoller - Friday, February 1, 2013 - link

    We do get a lot of new features with uefi, but the spec is a mess (like ACPI!). Lots of needless complications (like ACPI!).
    Matthew Garrett has been the main guy dealing with this on Linux and has written extensively about the ordeals, including the many issues with the older efi spec used by apple.
    Here's his blog if you're interested.
    http://mjg59.dreamwidth.org/
  • yankeeDDL - Thursday, January 31, 2013 - link

    I don't really see this as a budget vs mid range issue.
    Considering that UEFI is designed not to have non-allower SW installed, I'd say that it is basic testing to check that:
    a) it lets install what it should be possible to install
    b) blocks what it shouldn't let you install
    So I see this as a big no-no for the quality dep of Samsung: somebody forgot to add some checks as part of their OQC ...
  • lord_beavis - Thursday, January 31, 2013 - link

    From the beginning, this has been Microsoft's goal: to limit the potential impact of Linux. If it is difficult to install/run it, the less people will want to deal with it.

    I can see it now:

    Me: Well to install Linux we need to disable this "security" feature

    Luser: Do what now? No.
  • neothe0ne - Thursday, January 31, 2013 - link

    Take off your tin-foil hat and understand that Windows can't be installed in UEFI on these laptops either.
  • sherlockwing - Thursday, January 31, 2013 - link

    Windows have already been installed, or do you not see that Windows sticker?

Log in

Don't have an account? Sign up now