linux-kernel - Re: [PATCH v2 20/46] mm/migrate: Add folio_migrate

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <YNTIr/UGVrTOZD3f@casper.infradead.org>
Date:   Thu, 24 Jun 2021 19:02:23 +0100
From:   Matthew Wilcox <willy@...radead.org>
To:     Christoph Hellwig <hch@...radead.org>
Cc:     akpm@...ux-foundation.org, linux-fsdevel@...r.kernel.org,
        linux-mm@...ck.org, linux-kernel@...r.kernel.org
Subject: Re: [PATCH v2 20/46] mm/migrate: Add folio_migrate_copy()

On Wed, Jun 23, 2021 at 10:35:00AM +0200, Christoph Hellwig wrote:
> > +void folio_migrate_copy(struct folio *newfolio, struct folio *folio)
> >  {
> > +	unsigned int i = folio_nr_pages(folio) - 1;
> >  
> > +	copy_highpage(folio_page(newfolio, i), folio_page(folio, i));
> > +	while (i-- > 0) {
> > +		cond_resched()a
> > +		/* folio_page() handles discontinuities in memmap */
> > +		copy_highpage(folio_page(newfolio, i), folio_page(folio, i));
> > +	}
> > +
> 
> What is the advantage of copying backwards here to start with?

Easier to write the loop this way?  I suppose we could do it as ...

	unsigned int i, nr = folio_nr_pages(folio);

	for (i = 0; i < nr; i++) {
		/* folio_page() handles discontinuities in memmap */
		copy_highpage(folio_page(newfolio, i), folio_page(folio, i));
		cond_resched();
	}

I'm not really bothered.  As long as we don't call folio_nr_pages() for
each iteration of the loop ... I've actually been wondering about
marking that as __pure, but I don't quite have the nerve to do it yet.