Some function prototypes (in both .h and .c files) specify attributes
like __init and __exit in the prototype. gcc (at least at 3.3.3) uses
the last such attribute that is actually specified, without issuing a
warning. So we can have :-
* Prototype declarations that use one attribute and a function body
that uses another attribute.
* Functions that from the .c code appear to be normal text but the .h
file is silently setting a special attribute.
Both are potential sources of programmer confusion or bugs. I suggest
a janitor task to find all function prototypes that use __init, __exit,
__devinit, __devexit, __cpuinit or __cpuexit and remove the attribute
from the prototype. If the function body does not already specify a
suitable attribute then add the attribute to the function body.
The same task could be done for extern data declarations.
Once that is done, remove #include <linux/init.h> from all .h files.
Only .[cS] files should specify which section the data and text are
stored in, .h files should only define the C language information.
>Some function prototypes (in both .h and .c files) specify attributes
>like __init and __exit in the prototype. gcc (at least at 3.3.3) uses
>the last such attribute that is actually specified, without issuing a
>warning.
How does gcc 2.95 handle it, requiring __init in both spots?
Jan Engelhardt
--
Salı 17 Ocak 2006 18:55 tarihinde şunları yazmıştınız:
> >Some function prototypes (in both .h and .c files) specify attributes
> >like __init and __exit in the prototype. gcc (at least at 3.3.3) uses
> >the last such attribute that is actually specified, without issuing a
> >warning.
>
> How does gcc 2.95 handle it, requiring __init in both spots?
gcc 3.2 is now required for mainline kernel.
/ismail
>> >Some function prototypes (in both .h and .c files) specify attributes
>> >like __init and __exit in the prototype. gcc (at least at 3.3.3) uses
>> >the last such attribute that is actually specified, without issuing a
>> >warning.
>>
>> How does gcc 2.95 handle it, requiring __init in both spots?
>
>gcc 3.2 is now required for mainline kernel.
>
Sure, but I asked because I maintain one piece of code (ttyrpld) that
should compile for both 2.4 and 2.6. OTOH, I could use it to make people
convert to 2.6 ;-)
Jan Engelhardt
--