Re: [PATCH] hibernate: Allow uswsusp to write to swap

From: "Darrick J. Wong" <darrick.wong@oracle.com>
To: Domenico Andreoli <domenico.andreoli@linux.com>
Cc: "Rafael J. Wysocki" <rafael@kernel.org>,
	Linux PM <linux-pm@vger.kernel.org>,
	Linux Memory Management List <linux-mm@kvack.org>,
	linux-fsdevel@vger.kernel.org, mkleinsoft@gmail.com,
	Christoph Hellwig <hch@lst.de>,
	Andrew Morton <akpm@linux-foundation.org>,
	"Rafael J. Wysocki" <rjw@rjwysocki.net>,
	Len Brown <len.brown@intel.com>, Pavel Machek <pavel@ucw.cz>
Subject: Re: [PATCH] hibernate: Allow uswsusp to write to swap
Date: Wed, 4 Mar 2020 07:45:26 -0800	[thread overview]
Message-ID: <20200304154526.GH8036@magnolia> (raw)
In-Reply-To: <20200304082327.GA14236@dumbo>

On Wed, Mar 04, 2020 at 09:23:27AM +0100, Domenico Andreoli wrote:
> From: Domenico Andreoli <domenico.andreoli@linux.com>
> 
> It turns out that there is one use case for programs being able to
> write to swap devices, and that is the userspace hibernation code.

You might want to start a separate thread for this... :)

> Quick fix: disable the S_SWAPFILE check if hibernation is configured.
> 
> Fixes: 1638045c3677 ("mm: set S_SWAPFILE on blockdev swap devices")

Also, this should be

Fixes: dc617f29dbe5 ("vfs: don't allow writes to swap files")

since we still want to set S_SWAPFILE on active swap devices.

> Reported-by: Domenico Andreoli <domenico.andreoli@linux.com>
> Reported-by: Marian Klein <mkleinsoft@gmail.com>
> Signed-off-by: Domenico Andreoli <domenico.andreoli@linux.com>
> 
> ---
>  fs/block_dev.c |    3 ++-
>  1 file changed, 2 insertions(+), 1 deletion(-)
> 
> Index: b/fs/block_dev.c
> ===================================================================
> --- a/fs/block_dev.c
> +++ b/fs/block_dev.c
> @@ -2001,7 +2001,8 @@ ssize_t blkdev_write_iter(struct kiocb *
>  	if (bdev_read_only(I_BDEV(bd_inode)))
>  		return -EPERM;
>  
> -	if (IS_SWAPFILE(bd_inode))
> +	/* uswsusp needs to write to the swap */
> +	if (IS_SWAPFILE(bd_inode) && !IS_ENABLED(CONFIG_HIBERNATION))

...&& hibernation_available() ?  That way we can configure this
dynamically.

Thanks for taking this on, I'll go self-NAK the revert patch I sent
yesterday.

--D

>  		return -ETXTBSY;
>  
>  	if (!iov_iter_count(from))