Re: [PATCH] mtd: fix mtd_blkdevs problem with caches on some architectures (2.6.31)

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

On Sun, 22 Nov 2009 23:55:39 +0300
Ilya Loginov <isloginov@gmail.com> wrote:

> > > I tried this way from the beginning, and to avoid redefinitions I enclose
> > > everything in #if #endif in *.c files. I think it is not very good.
> > > So there are two possibilities.
> > > 
> 
> Ah. I think the first choice is always the best choise. =)
> 
> > In .h:
> > 
> > #if ARCH_IMPLEMENTS_FLUSH_DCACHE_PAGE
> > extern void rq_flush_dcache_pages(struct request *rq);
> > #else
> > static inline void bio_flush_dcache_pages(struct bio *bi)
> > {
> > }
> > #endif
> > 
> > In .c:
> > 	
> > #if ARCH_IMPLEMENTS_FLUSH_DCACHE_PAGE
> > void rq_flush_dcache_pages(struct request *rq)
> > {
> > 	struct req_iterator iter;
> > 	struct bio_vec* bvec;
> > 
> > 	rq_for_each_segment(bvec, rq, iter)
> > 		flush_dcache_page(bvec->bv_page);
> > }
> > EXPORT_SYMBOL(rq_flush_dcache_pages);
> > #endif
> 
> This is 2) =)
> 
> Thanks a lot for your help and comments!
> 
> Signed-off-by: Ilya Loginov <isloginov@gmail.com>

Looks good to me, but...

Patch doesn't vaguely apply to linux-next because of arch/sh changes. 
I was going to fix that but afaict your assertion that sh doesn't
implement flush_dcache_page() appears to be untrue in linux-next and a
bit of thought is needed there.

Also, please provide a nice chagnelog whcih describes the change, and
our reasons for making it.