linux-kernel - Re: [PATCH] io_uring: fix page leak in io_sqe_buffer

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-Id: <20250617140234.40664-1-superman.xpt@gmail.com>
Date: Tue, 17 Jun 2025 07:02:34 -0700
From: Penglei Jiang <superman.xpt@...il.com>
To: axboe@...nel.dk
Cc: io-uring@...r.kernel.org,
	linux-kernel@...r.kernel.org,
	superman.xpt@...il.com
Subject: Re: [PATCH] io_uring: fix page leak in io_sqe_buffer_register()

On Tue, 17 Jun 2025 06:53:04 -0600, Jens Axboe wrote:
> On 6/17/25 6:39 AM, Penglei Jiang wrote:
> > Add missing unpin_user_pages() in the error path
> > 
> > Fixes: d8c2237d0aa9 ("io_uring: add io_pin_pages() helper")
> > Signed-off-by: Penglei Jiang <superman.xpt@...il.com>
> > ---
> >  io_uring/rsrc.c | 4 +++-
> >  1 file changed, 3 insertions(+), 1 deletion(-)
> > 
> > diff --git a/io_uring/rsrc.c b/io_uring/rsrc.c
> > index c592ceace97d..f5ac1b530e21 100644
> > --- a/io_uring/rsrc.c
> > +++ b/io_uring/rsrc.c
> > @@ -804,8 +804,10 @@ static struct io_rsrc_node *io_sqe_buffer_register(struct io_ring_ctx *ctx,
> >  	}
> >  
> >  	imu = io_alloc_imu(ctx, nr_pages);
> > -	if (!imu)
> > +	if (!imu) {
> > +		unpin_user_pages(pages, nr_pages);
> >  		goto done;
> > +	}
> >  
> >  	imu->nr_bvecs = nr_pages;
> >  	ret = io_buffer_account_pin(ctx, pages, nr_pages, imu, last_hpage);
> 
> Wouldn't it be better to have the unpin be part of the normal error
> handling? Not sure why the pin accounting failure doesn't do that
> already.
> 
> Totally untested...
> 
> diff --git a/io_uring/rsrc.c b/io_uring/rsrc.c
> index 94a9db030e0e..a68f0cd677a3 100644
> --- a/io_uring/rsrc.c
> +++ b/io_uring/rsrc.c
> @@ -809,10 +809,8 @@ static struct io_rsrc_node *io_sqe_buffer_register(struct io_ring_ctx *ctx,
>  
>  	imu->nr_bvecs = nr_pages;
>  	ret = io_buffer_account_pin(ctx, pages, nr_pages, imu, last_hpage);
> -	if (ret) {
> -		unpin_user_pages(pages, nr_pages);
> +	if (ret)
>  		goto done;
> -	}
>  
>  	size = iov->iov_len;
>  	/* store original address for later verification */
> @@ -840,6 +838,7 @@ static struct io_rsrc_node *io_sqe_buffer_register(struct io_ring_ctx *ctx,
>  	}
>  done:
>  	if (ret) {
> +		unpin_user_pages(pages, nr_pages);
>  		if (imu)
>  			io_free_imu(ctx, imu);
>  		io_cache_free(&ctx->node_cache, node);

Thank you for taking the time to address this issue!

However, if io_pin_pages() fails, it will also jump to the done label,
but at that point, the value of nr_pages is undefined because nr_pages
is only assigned a value inside io_pin_pages() if it succeeds.

	pages = io_pin_pages((unsigned long) iov->iov_base, iov->iov_len,
				&nr_pages);
	if (IS_ERR(pages)) {
		ret = PTR_ERR(pages);
		pages = NULL;
		goto done;
	}

	...

	done:
		if (ret) {
			unpin_user_pages(NULL, undefined-value);
			...

I'm not sure what the impact of calling unpin_user_pages() in this way would be.