IS_ENABLED() vs CONFIG()

Dear all,

Now, when we run "coreboot$ ./util/lint/kconfig_lint " , we can see :

#!!!!! Warning: CONFIG() used on unknown value (option) at src/include/kconfig.h:21.
#!!!!! Warning: CONFIG() used on unknown value (%02x) at src/northbridge/amd/amdfam10/util.c:138.

B. Regards,

Elyes

[Julius Werner]

> Now, when we run "coreboot$ ./util/lint/kconfig_lint " , we can see :
> #!!!!! Warning: CONFIG() used on unknown value (option) at src/include/kconfig.h:21.
> #!!!!! Warning: CONFIG() used on unknown value (%02x) at src/northbridge/amd/amdfam10/util.c:138.

Yes... that's why it's a warning and not an error. ;) It's hard to
completely avoid false positives with the limited parsing capabilities
of a Perl script.

I still think this is valuable enough as a warning because for new CLs
it should point out a real problem most of the time and will help
people avoid errors. We have plenty of other spurious warnings when
you run the linter across the whole tree right now (e.g. a page full
of "Unused symbol" that someone should probably clean up at some
point), so I hope adding two more isn't going to be a huge problem?

Hi Julius,

Yes, you are right.

I've send a new patch #31848 & 31810, would you have a look on it when you have time please?

BTW: I did  some cleaning : https://review.coreboot.org/#/q/status:merged+project:coreboot+branch:master... :p

Thank you

---- Message d'origine ----
De : "Julius Werner" 
À : [e-mail]
Objet : Re: [coreboot] IS_ENABLED() vs CONFIG()
Date : 11/03/2019 19:54:23 CET
Copie à : "Coreboot"
> Now, when we run "coreboot$ ./util/lint/kconfig_lint " , we can see :
> #!!!!! Warning: CONFIG() used on unknown value (option) at src/include/kconfig.h:21.
> #!!!!! Warning: CONFIG() used on unknown value (%02x) at src/northbridge/amd/amdfam10/util.c:138.

Yes... that's why it's a warning and not an error. ;) It's hard to
completely avoid false positives with the limited parsing capabilities
of a Perl script.

I still think this is valuable enough as a warning because for new CLs
it should point out a real problem most of the time and will help
people avoid errors. We have plenty of other spurious warnings when
you run the linter across the whole tree right now (e.g. a page full
of "Unused symbol" that someone should probably clean up at some
point), so I hope adding two more isn't going to be a huge problem?