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: <202504071203.1024EFE5@keescook>
Date: Mon, 7 Apr 2025 12:09:50 -0700
From: Kees Cook <kees@...nel.org>
To: "Gustavo A. R. Silva" <gustavoars@...nel.org>
Cc: Krzysztof Kozlowski <krzk@...nel.org>, linux-kernel@...r.kernel.org,
	linux-hardening@...r.kernel.org
Subject: Re: [PATCH][next] w1: Avoid -Wflex-array-member-not-at-end warnings

On Thu, Mar 27, 2025 at 10:59:04AM -0600, Gustavo A. R. Silva wrote:
> -Wflex-array-member-not-at-end was introduced in GCC-14, and we are
> getting ready to enable it, globally.
> 
> Use the `DEFINE_RAW_FLEX()` helper for on-stack definitions of
> a flexible structure where the size of the flexible-array member
> is known at compile-time, and refactor the rest of the code,
> accordingly.
> 
> So, with these changes, fix the following warnings:
> 
> drivers/w1/w1_netlink.c:198:31: warning: structure containing a flexible array member is not at the end of another structure [-Wflex-array-member-not-at-end]
> drivers/w1/w1_netlink.c:219:31: warning: structure containing a flexible array member is not at the end of another structure [-Wflex-array-member-not-at-end]
> 
> Signed-off-by: Gustavo A. R. Silva <gustavoars@...nel.org>
> ---
>  drivers/w1/w1_netlink.c | 41 +++++++++++++++++++----------------------
>  1 file changed, 19 insertions(+), 22 deletions(-)
> 
> diff --git a/drivers/w1/w1_netlink.c b/drivers/w1/w1_netlink.c
> index 691978cddab7..845d66ab7e89 100644
> --- a/drivers/w1/w1_netlink.c
> +++ b/drivers/w1/w1_netlink.c
> @@ -194,16 +194,16 @@ static void w1_netlink_queue_status(struct w1_cb_block *block,
>  static void w1_netlink_send_error(struct cn_msg *cn, struct w1_netlink_msg *msg,
>  	int portid, int error)
>  {
> -	struct {
> -		struct cn_msg cn;
> -		struct w1_netlink_msg msg;
> -	} packet;
> -	memcpy(&packet.cn, cn, sizeof(packet.cn));
> -	memcpy(&packet.msg, msg, sizeof(packet.msg));
> -	packet.cn.len = sizeof(packet.msg);
> -	packet.msg.len = 0;
> -	packet.msg.status = (u8)-error;
> -	cn_netlink_send(&packet.cn, portid, 0, GFP_KERNEL);
> +	DEFINE_RAW_FLEX(struct cn_msg, packet, data,
> +			sizeof(struct w1_netlink_msg));
> +	struct w1_netlink_msg *pkt_msg = (struct w1_netlink_msg *)packet->data;

I'm starting to think we need a helper for "DEFINE_RAW_FLEX with a
trailing structure" for these. :)

Anyway, conversion looks good... structs are packed, so alignment issues
are unchanged.

> +
> +	memcpy(packet, cn, sizeof(*packet));
> +	memcpy(pkt_msg, msg, sizeof(*pkt_msg));

These could just be:

	*packet = *cn;
	*pkg_msg = *msg;

But that was always true. The memcpy() style is retained. But it would
catch type mismatches (like is accidentally introduced below).

> +	packet->len = sizeof(*pkt_msg);
> +	pkt_msg->len = 0;
> +	pkt_msg->status = (u8)-error;
> +	cn_netlink_send(packet, portid, 0, GFP_KERNEL);
>  }
>  
>  /**
> @@ -215,22 +215,19 @@ static void w1_netlink_send_error(struct cn_msg *cn, struct w1_netlink_msg *msg,
>   */
>  void w1_netlink_send(struct w1_master *dev, struct w1_netlink_msg *msg)
>  {
> -	struct {
> -		struct cn_msg cn;
> -		struct w1_netlink_msg msg;
> -	} packet;
> -	memset(&packet, 0, sizeof(packet));
> +	DEFINE_RAW_FLEX(struct cn_msg, packet, data,
> +			sizeof(struct w1_netlink_msg));
>  
> -	packet.cn.id.idx = CN_W1_IDX;
> -	packet.cn.id.val = CN_W1_VAL;
> +	packet->id.idx = CN_W1_IDX;
> +	packet->id.val = CN_W1_VAL;
>  
> -	packet.cn.seq = dev->seq++;
> -	packet.cn.len = sizeof(*msg);
> +	packet->seq = dev->seq++;
> +	packet->len = sizeof(*msg);
>  
> -	memcpy(&packet.msg, msg, sizeof(*msg));
> -	packet.msg.len = 0;
> +	memcpy(packet, msg, sizeof(*msg));

This memcpy() is wrong. It should be targeting packet->data.

> +	((struct w1_netlink_msg *)packet->data)->len = 0;

And since you need it again here, I'd recommend defining a struct
w1_netlink_msg pointer similar to the first hunk.

-Kees

>  
> -	cn_netlink_send(&packet.cn, 0, 0, GFP_KERNEL);
> +	cn_netlink_send(packet, 0, 0, GFP_KERNEL);
>  }
>  
>  static void w1_send_slave(struct w1_master *dev, u64 rn)
> -- 
> 2.43.0
> 

-- 
Kees Cook

Powered by blists - more mailing lists

Powered by Openwall GNU/*/Linux Powered by OpenVZ