[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <2065d0f6-660e-4647-95b4-8d1a9a7eaefe@embeddedor.com>
Date: Tue, 27 Aug 2024 13:09:46 -0600
From: "Gustavo A. R. Silva" <gustavo@...eddedor.com>
To: Kees Cook <kees@...nel.org>, "Gustavo A. R. Silva" <gustavoars@...nel.org>
Cc: Karol Herbst <kherbst@...hat.com>, Lyude Paul <lyude@...hat.com>,
Danilo Krummrich <dakr@...hat.com>, David Airlie <airlied@...il.com>,
Daniel Vetter <daniel@...ll.ch>, dri-devel@...ts.freedesktop.org,
nouveau@...ts.freedesktop.org, linux-kernel@...r.kernel.org,
linux-hardening@...r.kernel.org
Subject: Re: [PATCH][next] drm/nouveau: Avoid -Wflex-array-member-not-at-end
warning
On 22/08/24 11:27, Kees Cook wrote:
> On Wed, Aug 21, 2024 at 02:16:21PM -0600, Gustavo A. R. Silva wrote:
>> Use the `DEFINE_RAW_FLEX()` helper for an on-stack definition 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 this, fix the following warning:
>>
>> drivers/gpu/drm/nouveau/dispnv50/disp.c:779:47: 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/gpu/drm/nouveau/dispnv50/disp.c | 20 +++++++++-----------
>> 1 file changed, 9 insertions(+), 11 deletions(-)
>>
>> diff --git a/drivers/gpu/drm/nouveau/dispnv50/disp.c b/drivers/gpu/drm/nouveau/dispnv50/disp.c
>> index eed579a6c858..ddddc69640be 100644
>> --- a/drivers/gpu/drm/nouveau/dispnv50/disp.c
>> +++ b/drivers/gpu/drm/nouveau/dispnv50/disp.c
>> @@ -774,11 +774,9 @@ nv50_hdmi_enable(struct drm_encoder *encoder, struct nouveau_crtc *nv_crtc,
>> struct drm_hdmi_info *hdmi = &nv_connector->base.display_info.hdmi;
>> union hdmi_infoframe infoframe = { 0 };
>> const u8 rekey = 56; /* binary driver, and tegra, constant */
>> + DEFINE_RAW_FLEX(struct nvif_outp_infoframe_v0, args, data, 17);
>> + const u8 data_len = 17; /* same length as in DEFINE_RAW_FLEX above. */
>
> To avoid repeating the open-coded "17", this could either be a define:
>
> nv50_hdmi_enable(...)
> {
> ...
> #define data_len 17
> DEFINE_RAW_FLEX(struct nvif_outp_infoframe_v0, args, data, data_len);
> ...rest of function...
> #undef data_len
> }
>
> or an ungainly but compile-time calculated value that exposes some
> DEFINE_FLEX internals:
>
> const u8 data_len = (sizeof(args_u) - sizeof(*args)) / sizeof(*args->data);
Yeah, I actually thought of something more like just __struct_size(args) - sizeof(*args),
as the flex array member is `__u8 data[]`.
>
> (Maybe a helper is needed for that?)
>
> #define STACK_FLEX_COUNT(name, member) \
> ((sizeof(name##_u) = sizeof(*(name))) / sizeof(*(name)->member))
I don't like this `sizeof(name##_u)` part as it is detached from the DEFINE_RAW_FLEX()
internals. Probably use `__struct_size(args)` instead, as in the example above.
>
>> @@ -815,29 +813,29 @@ nv50_hdmi_enable(struct drm_encoder *encoder, struct nouveau_crtc *nv_crtc,
>> return;
>>
>> /* AVI InfoFrame. */
>> - args.infoframe.version = 0;
>> - args.infoframe.head = nv_crtc->index;
>> + args->version = 0;
>> + args->head = nv_crtc->index;
>
> The stack variable (was before and is again) already zero-initialized,
> so the "= 0" line shouldn't be needed.
>
> But neither of these comments are show-stoppers, IMO.
>
> Reviewed-by: Kees Cook <kees@...nel.org>
>
Thanks!
--
Gustavo
Powered by blists - more mailing lists