Discuss: --log-level option and clarifying log levels

Alex Hung alex.hung at canonical.com
Fri Dec 11 03:36:55 UTC 2015


Hi,

I think the criteria for critical may be too restrictive:
   - if something prevent a system from working, there are good chance
that kernel won't load and fwts cannot run
   - software and firmware are rarely able to damage hardware today.

I'd like to suggest critical to include the failures of major
functions such as CPU's power management and some of UEFI RT service
(ex. uefibootpath, uefirtvariable) since CPU's PM failures can
increase temperatures and cause instability (and not
environment-friendly), and these UEFI failures can stop from booting
from bios menu and bios may be updated in the future and so on.

How about spec violation? For example, some of the new ACPI tables and
control methods aren't used when they are introduced.  From the above
definition, the failures will be "low"; however, I would suggest to
mark them as "medium" at least so the failure won't be ignored
completely.

Cheers,
Alex Hung

On Fri, Dec 11, 2015 at 10:51 AM, ivanhu <ivan.hu at canonical.com> wrote:
> Hi Colin,
>
> We also have the warnings on fwts, should we add it as well?
>
> like,
> warning: don't affect the machine and functions, the design wasn't correct
> or following the specifications and rules.
>
> Cheers,
> Ivan
>
>
> On 2015年12月10日 18:24, Colin Ian King wrote:
>>
>> Hi there,
>>
>> I'm thinking of adding a --log-level option to fwts.
>>
>> fwts --log-level=[critical|high|medium|low|info]
>>
>> It will basically only print out and account for fwts errors of the
>> given level or higher.  E.g.
>>
>> fwts --log-level=high will just account for errors that are high or
>> critical.
>>
>> The fwts team are also thinking of clarifying the meanings of these
>> levels and fixing up all the error messages to match the definition.
>>
>> So let's try and start to define this properly. I am proposing:
>>
>> critical:  errors that can stop a machine from working, or can make it
>> behave in a way detrimental to the hardware.
>>
>> high:  errors that will make the machine perform incorrectly or
>> sub-optimally
>>
>> medium:  issues that would be useful to fix as they cause kernel message
>> but the kernel can fix these up or work around the error and the machine
>> will function fine.
>>
>> low:  miscellaneous small issues but don't affect the machine, the user
>> won't notice any kernel warnings and fixing this won't make any
>> difference.
>>
>> info: not really a error, just some useful help information about a very
>> minor issue that really is worth noting to give some feedback about why
>> the kernel may print a message, e.g. "PnPBIOS Disabled by ACPI PNP".
>>
>> Feel free to comment on these definitions so we can hone these into
>> something workable.
>>
>> Colin
>>
>
>
> --
> fwts-discuss mailing list
> fwts-discuss at lists.ubuntu.com
> Modify settings or unsubscribe at:
> https://lists.ubuntu.com/mailman/listinfo/fwts-discuss



-- 
Cheers,
Alex Hung



More information about the fwts-discuss mailing list