[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <YqiJaOiGnUzzB1+W@casper.infradead.org>
Date: Tue, 14 Jun 2022 14:13:12 +0100
From: Matthew Wilcox <willy@...radead.org>
To: Miaohe Lin <linmiaohe@...wei.com>
Cc: Joao Martins <joao.m.martins@...cle.com>,
Andrew Morton <akpm@...ux-foundation.org>, linux-mm@...ck.org,
linux-kernel@...r.kernel.org,
Dan Williams <dan.j.williams@...el.com>
Subject: Re: [PATCH] mm/page_alloc: make calling prep_compound_head more
reliable
On Wed, Jun 08, 2022 at 08:17:35PM +0800, Miaohe Lin wrote:
> +++ b/mm/page_alloc.c
> @@ -6771,13 +6771,18 @@ static void __ref memmap_init_compound(struct page *head,
> set_page_count(page, 0);
>
> /*
> - * The first tail page stores compound_mapcount_ptr() and
> - * compound_order() and the second tail page stores
> - * compound_pincount_ptr(). Call prep_compound_head() after
> - * the first and second tail pages have been initialized to
> - * not have the data overwritten.
> + * The first tail page stores compound_mapcount_ptr(),
> + * compound_order() and compound_pincount_ptr(). Call
> + * prep_compound_head() after the first tail page have
> + * been initialized to not have the data overwritten.
> + *
> + * Note the idea to make this right after we initialize
> + * the offending tail pages is trying to take advantage
> + * of the likelihood of those tail struct pages being
> + * cached given that we will read them right after in
> + * prep_compound_head().
It's not that we'll read them again, it's that the cacheline will still
be in cache, and therefore dirty.
Honestly, I don't think we need this extra explanation in a comment.
Just change the first paragraph to reflect reality and leave it at that.
> */
> - if (pfn == head_pfn + 2)
> + if (unlikely(pfn == head_pfn + 1))
We definitely don't need the unlikely here.
> prep_compound_head(head, order);
> }
> }
>
> Or am I miss something?
>
> Thanks!
>
> > .
> >
>
Powered by blists - more mailing lists