On Wed, Aug 19, 2009 at 10:09 AM, Garrett D'Amore <[email protected]>wrote:
> Reang Su wrote:
>
>> I've a file of function definitions say xxx.h. This file is included and
>> used in other n files. Earlier this definition was static inline yyyy(){}.
>> Now after removing static this complains as "static unused"
>>
>
> So you have a *definition* (as opposed to a declaration) in the header?
>
> And using inline?
>
> You might need to eliminate all the *.ln files to generate a fresh set,
> otherwise you might get this lint error on pass 2 as a result of the static
> function being included into a .ln file that was previously generated from a
> different .c.
>
Looks like eliminating old *.ln files doesn't work. After separating those
definitions to a separate file and keeping only declarations on xxx.h file
I'm able to isolate the problem.
Thank you again.
---Reang.
>
> In general, I find the practice of defining functions in header files (as
> opposed to declaring them) distasteful. And I never put static declarations
> in a header file. After all, the point of static is to limit their scope to
> just one file, and (IMO) the main purpose of a header file is to provide
> declarations that can be shared amongst multiple source files.
>
> Putting such static declarations in the header file only serves (IMO) to
> add confusion as to what functions are available where.
>
> (I'll also note that at this point I usually try to keep simple drivers to
> a single source file. It allows the compiler some additional flexibility
> when performing optimizations, and makes it easier to find functions and
> declarations if you only have to look in a single file.)
>
> -- Garrett
>
>> ---Reang.
>>
>> On Wed, Aug 19, 2009 at 9:48 AM, Garrett D'Amore <[email protected]<mailto:
>> [email protected]>> wrote:
>>
>> Reang Su wrote:
>>
>> Hi Garrett,
>> I tried without using /*LINTED */ initially, it didn't work.
>> After removing static from my function definition I was still
>> getting "warning: static unused", So I tried with /*LINTED */.
>> Unfortunately this also didn't work.
>>
>>
>> Lint pass two over a whole program might notice that the function
>> is not used. But if its a device driver, it should get by if you
>> remove the static, without a LINTED flag. The behavior you're
>> describing is not one I've ever seen.
>>
>> But again, if the function is static, and *not* used, then why
>> have it in the first place?
>>
>> - Garrett
>>
>> Thank you,
>> --Reang.
>>
>>
>> On Tue, Aug 18, 2009 at 7:23 PM, Garrett D'Amore
>> <[email protected] <mailto:[email protected]>
>> <mailto:[email protected] <mailto:[email protected]>>> wrote:
>>
>> Reang Su wrote:
>>
>> Hi ,
>> I'm stuck while removing E_STATIC_UNUSED LINTTAG from my
>> driver Makefile.
>> Once I removed this LINTTAG in my Makefile I get following
>> warning.
>> xxx.h", line 85: warning: static unused:
>> __hal_ifmsg_is_manager_up (E_STATIC_UNUSED)
>> I removed static from my function definition and added
>> /*LINTED */ before it. Now I get
>> E_SUPPRESSION_DIRECTIVE_UNUSED on that line.
>> xxx.h", line 85: warning: lint suppression directive
>> not used
>> (E_SUPPRESSION_DIRECTIVE_UNUSED)
>>
>>
>> You shouldn't have also added /*LINTED*/. Since you
>> removed the
>> "static", you shouldn't have got the
>> E_STATIC_UNUSED_LINTTAG anymore.
>>
>> Of course, if the function really was static, and not used
>> anywhere, maybe you should have just removed it, and saved a
>> little space in the code and in the binary.
>>
>> - Garrett
>>
>> Please suggest.
>> Thank you,
>> -Reang,
>>
>> ------------------------------------------------------------------------
>>
>> _______________________________________________
>> driver-discuss mailing list
>> [email protected]
>> <mailto:[email protected]>
>> <mailto:[email protected]
>> <mailto:[email protected]>>
>>
>> http://mail.opensolaris.org/mailman/listinfo/driver-discuss
>>
>>
>>
>>
>> ------------------------------------------------------------------------
>>
>> _______________________________________________
>> driver-discuss mailing list
>> [email protected]
>> http://mail.opensolaris.org/mailman/listinfo/driver-discuss
>>
>>
>
>
_______________________________________________
driver-discuss mailing list
[email protected]
http://mail.opensolaris.org/mailman/listinfo/driver-discuss
