linux-kernel - Re: [RFC] rtc: Avoid a couple of -Wflex-array-member-not-at-end warnings

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] [day] [month] [year] [list]

Message-ID: <Z9EZByipsNmBoVea@google.com>
Date: Wed, 12 Mar 2025 05:17:59 +0000
From: Tzung-Bi Shih <tzungbi@...nel.org>
To: "Gustavo A. R. Silva" <gustavoars@...nel.org>
Cc: Alexandre Belloni <alexandre.belloni@...tlin.com>,
	Benson Leung <bleung@...omium.org>,
	Guenter Roeck <groeck@...omium.org>, linux-rtc@...r.kernel.org,
	chrome-platform@...ts.linux.dev, linux-kernel@...r.kernel.org,
	linux-hardening@...r.kernel.org
Subject: Re: [RFC] rtc: Avoid a couple of -Wflex-array-member-not-at-end
 warnings

On Mon, Mar 10, 2025 at 03:46:26PM +1030, Gustavo A. R. Silva wrote:
> Hi all,
> 
> As part of the efforts to globally enable -Wflex-array-member-not-at-end,
> I'm currently trying to fix the following warnings:
> 
> drivers/rtc/rtc-cros-ec.c:62:40: warning: structure containing a flexible array member is not at the end of another structure [-Wflex-array-member-not-at-end]
> drivers/rtc/rtc-cros-ec.c:40:40: warning: structure containing a flexible array member is not at the end of another structure [-Wflex-array-member-not-at-end]
> 
> The issue is that `struct cros_ec_command` is a flexible structure (which
> means that it contains a flexible-array member), and there is an object
> of this type (msg) declared within another structure but at the end.
                                                          ^ not

> It seems that the following patch would suffice, as long as the flex-array
> member in `struct cros_ec_command` is not expected to be accessed and
> overlap with `struct ec_response_rtc data` in a "controlled manner":
> 
> diff --git a/drivers/rtc/rtc-cros-ec.c b/drivers/rtc/rtc-cros-ec.c
> index 865c2e82c7a5..7e9bbab47e4c 100644
> --- a/drivers/rtc/rtc-cros-ec.c
> +++ b/drivers/rtc/rtc-cros-ec.c
> @@ -37,8 +37,8 @@ static int cros_ec_rtc_get(struct cros_ec_device *cros_ec, u32 command,
>  {
>         int ret;
>         struct {
> -               struct cros_ec_command msg;
>                 struct ec_response_rtc data;
> +               struct cros_ec_command msg;
>         } __packed msg;
>  
>         memset(&msg, 0, sizeof(msg));
> @@ -59,8 +59,8 @@ static int cros_ec_rtc_set(struct cros_ec_device *cros_ec, u32 command,
>  {
>         int ret;
>         struct {
> -               struct cros_ec_command msg;
>                 struct ec_response_rtc data;
> +               struct cros_ec_command msg;
>         } __packed msg;
>  
>         memset(&msg, 0, sizeof(msg));

This doesn't work.  The header (i.e. struct cros_ec_command) is expected to
allocate right before the payload (e.g. struct ec_response_rtc)[1][2].

[1]: https://elixir.bootlin.com/linux/v6.14-rc1/source/drivers/platform/chrome/cros_ec_proto.c#L82
[2]: https://elixir.bootlin.com/linux/v6.14-rc1/source/drivers/platform/chrome/cros_ec_i2c.c#L166

> Otherwise, we probably need to use struct_group_tagged() as follows:

I'm not a big fan of the solution.  How about using something similar to:

struct cros_ec_command *msg = kzalloc(sizeof(*msg) +
                                      sizeof(struct ec_response_rtc), ...);