On Mon Dec 16, 2024 at 2:56 PM CET, Bruce Ashfield via lists.openembedded.org
wrote:
> On Mon, Dec 16, 2024 at 8:50 AM Antonin Godard <antonin.god...@bootlin.com>
> wrote:
>
>> Hi Bruce,
>>
>> On Mon Dec 16, 2024 at 2:29 PM CET, Bruce Ashfield via
>> lists.openembedded.org wrote:
>> > On Mon, Dec 16, 2024 at 8:21 AM Bruce Ashfield <bruce.ashfi...@gmail.com
>> >
>> > wrote:
>> >
>> >>
>> >>
>> >> On Mon, Dec 16, 2024 at 6:35 AM Antonin Godard <
>> antonin.god...@bootlin.com>
>> >> wrote:
>> >>
>> >>> Following commit a39a1f7cf78ad1ca07438bce634a47e970f25047
>> >>> ("kernel-yocto: allow early exit to configuration audit"), we can
>> >>> disable auditing of the kernel configuration by setting KMETA_AUDIT
>> >>> empty.
>> >>>
>> >>> Since the default value of this variables is "yes", use
>> >>> oe.types.boolean() so that it will also be disabled when setting it to
>> >>> "no", and avoid potential confusion.
>> >>>
>> >>
>> >> That isn't the intent of the variable. It is being used as a flag here,
>> >> but it can also be used to encode other information.
>> >>
>> >> Quite literally, if it is set to anything, the audit should happen. Only
>> >> when it is unset should it not.
>> >>
>> >
>> > I hit send too soon. It's Monday.
>> >
>> > What I wanted to add was this question: Is that still the behaviour with
>> > the boolean call ? If so, no concerns from me.
>>
>> It seems that this is the only place where the KMETA_AUDIT variable is
>> used so I
>> figured its role was to enable/disable the audit and only that. Maybe the
>> "audit_flag" name is a bit misleading then?
>>
>
> Maybe, but I'd rather not change the name if nothing is broken.
>
> I normally just put a comment above the variable to indicate
> that sort of behaviour (i.e. # if this variable is set to anything non-zero
> an audit will be performed), but I for some reason didn't do that
> here!
>
> There are some audit modes for in-tree configs that I'm working
> on that will use that variable to change modes (different from the
> allnoconfig/alldefconfig flags that are already used).
Ok. boolean() will error if it doesn't find 'yes', 'no', '1', '0' or any such
value, so we need to NACK this patch then. Thanks for the added context.
I will work on documenting this in yocto-docs when I have the time (unless
you've got some free time to document these variables :) ). There's an open bug
on this here https://bugzilla.yoctoproject.org/show_bug.cgi?id=13835, which is
the initial reason why I was looking into these.
Antonin
--
Antonin Godard, Bootlin
Embedded Linux and Kernel engineering
https://bootlin.com
-=-=-=-=-=-=-=-=-=-=-=-
Links: You receive all messages sent to this group.
View/Reply Online (#208766):
https://lists.openembedded.org/g/openembedded-core/message/208766
Mute This Topic: https://lists.openembedded.org/mt/110142731/21656
Group Owner: openembedded-core+ow...@lists.openembedded.org
Unsubscribe: https://lists.openembedded.org/g/openembedded-core/unsub
[arch...@mail-archive.com]
-=-=-=-=-=-=-=-=-=-=-=-
