Re: [PATCH] [fuse] alloc_page nofs avoid deadlock

From: Michal Hocko <mhocko@suse.com>
To: chenguanyou <chenguanyou9338@gmail.com>
Cc: linux-kernel@vger.kernel.org, akpm@linux-foundation.org,
	keescook@chromium.org, lukas.bulwahn@gmail.com, vbabka@suse.cz,
	gpiccoli@canonical.com, chenguanyou <chenguanyou@xiaomi.com>
Subject: Re: [PATCH] [fuse] alloc_page nofs avoid deadlock
Date: Thu, 3 Jun 2021 14:56:12 +0200	[thread overview]
Message-ID: <YLjRbEekFbjZXqiI@dhcp22.suse.cz> (raw)
In-Reply-To: <20210603124105.29052-1-chenguanyou@xiaomi.com>

On Thu 03-06-21 20:41:05, chenguanyou wrote:
> ABA deadlock
> 
> PID: 17172 TASK: ffffffc0c162c000 CPU: 6 COMMAND: "Thread-21"
> 
> PID: 9652 TASK: ffffffc0c9ce0000 CPU: 4 COMMAND: "kworker/u16:8"

Please be much more specific about the underlying issue and what the
actual lock dependency during the reclaim is. The above is essentially
void of any relevant information.

> Signed-off-by: chenguanyou <chenguanyou@xiaomi.com>
> ---
>  fs/fuse/dev.c | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/fs/fuse/dev.c b/fs/fuse/dev.c
> index c0fee830a34e..d36125ff0405 100644
> --- a/fs/fuse/dev.c
> +++ b/fs/fuse/dev.c
> @@ -721,7 +721,7 @@ static int fuse_copy_fill(struct fuse_copy_state *cs)
>  			if (cs->nr_segs >= cs->pipe->max_usage)
>  				return -EIO;
>  
> -			page = alloc_page(GFP_HIGHUSER);
> +			page = alloc_page(GFP_NOFS | __GFP_HIGHMEM);
>  			if (!page)
>  				return -ENOMEM;
>  
> -- 
> 2.17.1

-- 
Michal Hocko
SUSE Labs