Re: [PATCH] kexec: Move some memembers and definitions within the scope of CONFIG_KEXEC_FILE

[Andrew Morton <akpm@linux-foundation.org>]

On Tue, 22 Dec 2015 19:40:39 +0800 Xunlei Pang <xlpang@redhat.com> wrote:

> > Following functions will be used only in kexec_file. Please wrap them in
> > CONFIG_KEXEC_FILE.
> >
> > int __weak arch_kexec_kernel_image_probe(struct kimage *image, void *buf,
> > 					 unsigned long buf_len);
> > void * __weak arch_kexec_kernel_image_load(struct kimage *image);
> > int __weak arch_kimage_file_post_load_cleanup(struct kimage *image);
> > int __weak arch_kexec_kernel_verify_sig(struct kimage *image, void *buf,
> > 					unsigned long buf_len);
> > int __weak arch_kexec_apply_relocations_add(const Elf_Ehdr *ehdr,
> > 					Elf_Shdr *sechdrs, unsigned int relsec);
> > int __weak arch_kexec_apply_relocations(const Elf_Ehdr *ehdr, Elf_Shdr *sechdrs,
> > 					unsigned int relsec);
> 
> Thanks for the comment.
> 
> I noticed this as well, but seems for the function declarations we don't need do this,
> since they don't consume the actual space.
> 
> For example, in the include/linux/timekeeping.h
> /*  
>  * RTC specific
>  */ 
> extern bool timekeeping_rtc_skipsuspend(void);
> extern bool timekeeping_rtc_skipresume(void);
> 
> extern void timekeeping_inject_sleeptime64(struct timespec64 *delta);
> 
> also not embraced by the corresponding macros.

Yes.  If we add the ifdefs then a programming error will be detected at
compile time.  If we don't add the ifdefs then that error will be
detected at link time.  So the ifdefs provide a quite small advantage,
while making the code harder to read and harder to maintain.  I believe
that "no ifdefs" is the better side of this tradeoff.