MIME-Version: 1.0
References: <20211120201230.920082-1-shakeelb@google.com> <YZnMnyKRk8d/Ilyo@casper.infradead.org>
In-Reply-To: <YZnMnyKRk8d/Ilyo@casper.infradead.org>
From:   Shakeel Butt <shakeelb@google.com>
Date:   Sat, 20 Nov 2021 21:25:27 -0800
Message-ID: <CALvZod7TBEPAj_62WYGxFGXOkv0pu0Erv2tqj-io6LnNDt6-nw@mail.gmail.com>
Subject: Re: [PATCH] mm: split thp synchronously on MADV_DONTNEED
To:     Matthew Wilcox <willy@infradead.org>
Cc:     David Hildenbrand <david@redhat.com>,
        "Kirill A . Shutemov" <kirill.shutemov@linux.intel.com>,
        Yang Shi <shy828301@gmail.com>, Zi Yan <ziy@nvidia.com>,
        Andrew Morton <akpm@linux-foundation.org>, linux-mm@kvack.org,
        linux-kernel@vger.kernel.org
Content-Type: text/plain; charset="UTF-8"
Precedence: bulk

On Sat, Nov 20, 2021 at 8:35 PM Matthew Wilcox <willy@infradead.org> wrote:
>
> On Sat, Nov 20, 2021 at 12:12:30PM -0800, Shakeel Butt wrote:
> > This patch let such applications not worry about the low level handling
> > of THPs in the kernel and splits the THPs synchronously on
> > MADV_DONTNEED.
>
> I don't mind the synchronous split, but I have concerns about the
> implementation.  I don't think it's worth another pointer in task_struct.

Are you concerned about the size of task_struct? At least on my config
this additional pointer was just filling the holes and not increasing
the size. I can check a couple other configs as well.

> It's also the case that splitting is likely to succeed, so I think a
> better implementation would try to split and then put the page on the
> global deferred list if splitting fails.

Actually this is what this patch is doing. See the second loop in
split_local_deferred_list().