Re: [PATCH 07/19] mm/filemap: Use head pages in generic_file_buffered_read

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



On Thu, Oct 29, 2020 at 08:19:09PM -0400, Kent Overstreet wrote:
> On Thu, Oct 29, 2020 at 07:33:53PM +0000, Matthew Wilcox (Oracle) wrote:
> > Add mapping_get_read_heads() which returns the head pages which
> > represent a contiguous array of bytes in the file.  It also stops
> > when encountering a page marked as Readahead or !Uptodate (but
> > does return that page) so it can be handled appropriately by
> > gfbr_get_pages().
> 
> I don't like the _heads naming - this is like find_get_pages_contig() except it
> returns compound pages instead of single pages, and the naming should reflect
> that. And, working with compound pages should be the default in the future -
> code working with individual pages should be considered a code smell in the
> future, i.e. find_get_pages() et. all. should be considered deprecated.

You're right.  It should just be called mapping_get_(something)_thps.

> Also - you're implementing new core functionality with a pecularity of the read
> path, the two probably be split out. Perhaps find_get_cpages (compound_pages)
> that can also be passed a stop condition.

It isn't clear to me _how_ to do that, so in the absence of that obvious,
I decided to implement what we need right now, and we can figure out how
to abstract it later.

So until we have a better idea, I'm going with "read" for the (something)
above.

> > +			if (writably_mapped) {
> > +				int j;
> > +
> > +				for (j = 0; j < thp_nr_pages(page); j++)
> > +					flush_dcache_page(page + j);
> > +			}
> 
> this should be a new helper

Good point.  One flush_dcache_thp() coming right up!




[Index of Archives]     [Linux ARM Kernel]     [Linux ARM]     [Linux Omap]     [Fedora ARM]     [IETF Annouce]     [Bugtraq]     [Linux OMAP]     [Linux MIPS]     [eCos]     [Asterisk Internet PBX]     [Linux API]

  Powered by Linux