Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 23.128.96.18 as permitted sender) client-ip=23.128.96.18;
Date:   Tue, 1 Jun 2021 17:49:22 +0100
From:   Matthew Wilcox <willy@infradead.org>
To:     Xu Yu <xuyu@linux.alibaba.com>
Cc:     linux-mm@kvack.org, linux-kernel@vger.kernel.org,
        akpm@linux-foundation.org, gavin.dg@linux.alibaba.com,
        Hugh Dickins <hughd@google.com>,
        Nicholas Piggin <npiggin@gmail.com>,
        Vlastimil Babka <vbabka@suse.cz>
Subject: Re: [PATCH] mm, thp: relax migration wait when failed to get tail
 page
Message-ID: <YLZlEqHlX18/eKhn@casper.infradead.org>
References: <bc8567d7a2c08ab6fdbb8e94008157265d5d28a3.1622564942.git.xuyu@linux.alibaba.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <bc8567d7a2c08ab6fdbb8e94008157265d5d28a3.1622564942.git.xuyu@linux.alibaba.com>
Precedence: bulk

On Wed, Jun 02, 2021 at 12:31:41AM +0800, Xu Yu wrote:
> We notice that hung task happens in a conner but practical scenario when
> CONFIG_PREEMPT_NONE is enabled, as follows.
> 
> Process 0                       Process 1                     Process 2..Inf
> split_huge_page_to_list
>     unmap_page
>         split_huge_pmd_address
>                                 __migration_entry_wait(head)
>                                                               __migration_entry_wait(tail)
>     remap_page (roll back)
>         remove_migration_ptes
>             rmap_walk_anon
>                 cond_resched
> 
> Where __migration_entry_wait(tail) is occurred in kernel space, e.g.,
> copy_to_user, which will immediately fault again without rescheduling,
> and thus occupy the cpu fully.
> 
> When there are too many processes performing __migration_entry_wait on
> tail page, remap_page will never be done after cond_resched.
> 
> This relaxes __migration_entry_wait on tail page, thus gives remap_page
> a chance to complete.

I don't like this at all.  If only we could call wait_on_page_locked()
without having a reference to the page.

I /think/ the only obstacle to doing that is that Slab may reuse
the PG_locked bit without following the pagecache wakeup protocol.
That wouldn't be a hard thing to change; I think it's only slub that
uses that bit, and it really could use any bit in the page flags.

> Signed-off-by: Gang Deng <gavin.dg@linux.alibaba.com>
> Signed-off-by: Xu Yu <xuyu@linux.alibaba.com>
> ---
>  mm/migrate.c | 7 +++++--
>  1 file changed, 5 insertions(+), 2 deletions(-)
> 
> diff --git a/mm/migrate.c b/mm/migrate.c
> index b234c3f3acb7..df2dc39fe566 100644
> --- a/mm/migrate.c
> +++ b/mm/migrate.c
> @@ -301,8 +301,11 @@ void __migration_entry_wait(struct mm_struct *mm, pte_t *ptep,
>  	 * is zero; but we must not call put_and_wait_on_page_locked() without
>  	 * a ref. Use get_page_unless_zero(), and just fault again if it fails.
>  	 */
> -	if (!get_page_unless_zero(page))
> -		goto out;
> +	if (!get_page_unless_zero(page)) {
> +		pte_unmap_unlock(ptep, ptl);
> +		cond_resched();
> +		return;
> +	}
>  	pte_unmap_unlock(ptep, ptl);
>  	put_and_wait_on_page_locked(page, TASK_UNINTERRUPTIBLE);
>  	return;
> -- 
> 2.20.1.2432.ga663e714
> 
>