Re: [PATCH, RFC] map multiple blocks at a time in mpage_readpage(s)

[Suparna Bhattacharya <suparna@in.ibm.com>]

On Wed, Oct 05, 2005 at 09:10:37AM +0200, Christoph Hellwig wrote:
> On Wed, Oct 05, 2005 at 10:04:48AM +0400, Nikita Danilov wrote:
> >  >  					page->index, GFP_KERNEL)) {
> >  >  			bio = do_mpage_readpage(bio, page,
> >  >  					nr_pages - page_idx,
> >  > -					&last_block_in_bio, get_block);
> >  > +					&last_block_in_bio, &map_bh,
> >  > +					&first_logical_block, &map_valid,
> >  > +					get_blocks, get_block);
> > 
> > Oh, function taking 9 arguments, please... this is not windows.
> 
> I'm happy about suggestion to reduce the count.  One of the real
> problems is that the buffer_head structure is extremly badly suited for
> the get_blocks callback.  Having information like in the xfs_iomap_t
> structure (which is XFS bmap information container) would simplify lots
> of these things.

In the writepages case, I ultimately ended up defining an mpageio
structure somewhat like the DIO code.

> 
> 
> > By the way, I cannot see where get_block callback is used by mpage.c
> > code, why is it needed?
> 
> It's only used by block_read_full_page which is called in the 'confused'
> case.

Having a readpage_fn callback may be a little bit better than this, but
yes this is when one really starts wanting to get rid of bufferheads ...

> -
> To unsubscribe from this list: send the line "unsubscribe linux-fsdevel" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html

Regards
Suparna

-- 
Suparna Bhattacharya (suparna@in.ibm.com)
Linux Technology Center
IBM Software Lab, India