[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: Debian derivatives census: Inquisitor: status?



Hi there,

>>> BTW, Yocto-based LUVos (Linux UEFI Validation) distro, especially it's
>>> LUV-live release is probably the best hardware diagnostic distro these
>>> days. It bundles BIOS BITS, FTWS, CHIPSEC, and has additional tests.
>>> Created -- and actively maintained -- by Intel, targets BIOS/UEFI
>>> systems. Linaro is in the middle of porting it -- and its bundled tools
>>> -- to AArch64. Debian needs CHIPSEC package at least.
>>> https://01.org/linux-uefi-validation
>>
>> I might be biased, but from what I see in LUVos description, it
>> pursues a completely different goal that Inquisitor. LUV concenrates
>> on firmware validation, combatting various security issues. It's for
>> situation when you suspect that your kernel configuration or something
>> like that might be off and thus you just lost some sort of
>> functionality you desire in a certain device - not for situations when
>> you have 1000s of a given device model with 1% of them having faulty
>> hardware.
>
> Yes, I think Intel's initial goal with LUV was to build something that
> helped OEMs build systems with proper UEFI. It is also useful for
> sysadmins and security researchers.
>
> However, I also consider security issues part of basic hardware
> functionality. If I'm buying a new server, I want to see CHIPSEC logs
> first, if the hardware is known to be vulnerable, I'm going to return
> it, just like if it had a bad mobo.

CHIPSEC checks are not a matter of stability or something - it's just
a matter of identification. It's binary - vulnerable / not vulnerable,
that's not some analogue signal / cooling system performance /
temperature graphs / etc.

>> Talking about "competitors", I'd say that primary competitor of
>> Inquisitor nowadays is PTS[1]. It indeed follows the same basic
>> principles (i.e. detecting configuration, running tests, reporting
>> results), it indeed has a relatively big user base, but PTS's author
>> (Michael Larabel) concenrates mostly on benchmarking stuff, while
>> Inquisitor is mostly suited for big-scale burn-in tests.
>
> Agreed, Phoronix is nice.

Well, can't really agree on this (however, I repeat, I'm biased ;)) -
from my POV, 95% of Phoronix stuff is mostly for the show and bring no
real short or long-term benefit. Most testing pre-conditions and
methodics are screwed as hell and might show or not show anything that
resembles reality.

On the other hand, it's hard to argue that Inquisitor includes ~34
tests, while PTS sports ~240.

> Inquisitor and Phoronix have many hardware testing abilities that LUV
> doesn't have, so they're very useful, LUV isn't touching those features,
> as far as I can tell.

Yeah, that's kind of my point.

-- 
WBR, Mikhail Yakshin


Reply to: