Re: [KJ] [PATCH] fix legacy megaraid-driver to compile without CONFIG_PROC_FS

[Henne <henne@nachtwindheim.de>]

Adrian Bunk schrieb:
> On Thu, Jul 06, 2006 at 10:33:49PM +0200, Henne wrote:
>>>> From: Henrik Kretzschmar <henne@nachtwindheim.de>
>>>>
>>>> Create an empty inline function to make the legacy megaraid-driver compile
>>>> without PROC_FS.
>>>> Signed-off-by: Henrik Kretzschmar <henne@nachtwindheim.de>
>>>> ---
>>>>
>>>> --- linux-2.6.18-rc1/drivers/scsi/megaraid.h    2006-06-18 03:49:35.000000000 +0200
>>>> +++ linux/drivers/scsi/megaraid.h       2006-07-06 21:39:59.000000000 +0200
>>>> @@ -1039,6 +1039,9 @@
>>>>  static int proc_rdrv_30(char *, char **, off_t, int, int *, void *);
>>>>  static int proc_rdrv_40(char *, char **, off_t, int, int *, void *);
>>>>  static int proc_rdrv(adapter_t *, char *, int, int);
>>>> +#else
>>>> +static inline void
>>>> +mega_create_proc_entry(int index, struct proc_dir_entry *parent) {}
>>>>  #endif
>>>>
>>>>  static int mega_adapinq(adapter_t *, dma_addr_t);
>>> Already in -mm:
>>> http://www.kernel.org/pub/linux/kernel/people/akpm/patches/2.6/2.6.17/2.6.17-mm6/broken-out/drivers-scsi-megaraidc-add-a-dummy-mega_create_proc_entry-for-proc_fs=y.patch
>> Great, but isn't it better to put that define stuff into the headers?
> 
> No - the function itself is static (and has therefore itself no 
> prototype in any header).
> 
> If the function was global, I'd agree with you.
> 
>> Thanks and Greets,
>> Henne
> 
> cu
> Adrian
> 
The header I mean is /drivers/scsi/megaraid.h and is only used by megaraid.c.
And there are the prototypes, in dependency of CONFIG_PROC_FS, for all these other proc-related functions.
Thats why I decided to make the change there and not in megaraid.c .