[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <CAHbLzkoF+gbgopWBSP0Xe0+g-DLF_n-NQ_7gvg+0EEGpv8Q7Tw@mail.gmail.com>
Date: Wed, 25 Jan 2023 10:06:33 -0800
From: Yang Shi <shy828301@...il.com>
To: "Zach O'Keefe" <zokeefe@...gle.com>
Cc: linux-mm@...ck.org, linux-kernel@...r.kernel.org,
Andrew Morton <akpm@...ux-foundation.org>,
Hugh Dickins <hughd@...gle.com>
Subject: Re: [PATCH 1/2] mm/MADV_COLLAPSE: set EAGAIN on unexpected page refcount
On Tue, Jan 24, 2023 at 5:58 PM Zach O'Keefe <zokeefe@...gle.com> wrote:
>
> During collapse, in a few places we check to see if a given small page
> has any unaccounted references. If the refcount on the page doesn't
> match our expectations, it must be there is an unknown user concurrently
> interested in the page, and so it's not safe to move the contents
> elsewhere. However, the unaccounted pins are likely an ephemeral state.
>
> In such a situation, make MADV_COLLAPSE set EAGAIN errno, indicating that
> collapse may succeed on retry.
The page may be DMA pinned (for example, pin_user_pages()), it is not
worth retrying for such pages. But it may also not be worth optimizing
for this case at this point.
So the patch looks good to me. Reviewed-by: Yang Shi <shy828301@...il.com>
>
> Fixes: 7d8faaf15545 ("mm/madvise: introduce MADV_COLLAPSE sync hugepage collapse")
> Reported-by: Hugh Dickins <hughd@...gle.com>
> Signed-off-by: Zach O'Keefe <zokeefe@...gle.com>
>
> ---
> mm/khugepaged.c | 1 +
> 1 file changed, 1 insertion(+)
>
> diff --git a/mm/khugepaged.c b/mm/khugepaged.c
> index e23619bfecc4..fa38cae240b9 100644
> --- a/mm/khugepaged.c
> +++ b/mm/khugepaged.c
> @@ -2712,6 +2712,7 @@ static int madvise_collapse_errno(enum scan_result r)
> case SCAN_CGROUP_CHARGE_FAIL:
> return -EBUSY;
> /* Resource temporary unavailable - trying again might succeed */
> + case SCAN_PAGE_COUNT:
> case SCAN_PAGE_LOCK:
> case SCAN_PAGE_LRU:
> case SCAN_DEL_PAGE_LRU:
> --
> 2.39.1.405.gd4c25cc71f-goog
>
Powered by blists - more mailing lists