Received-SPF: pass (google.com: domain of linux-kernel+bounces-127570-linux.lists.archive=gmail.com@vger.kernel.org designates 2604:1380:45d1:ec00::1 as permitted sender) client-ip=2604:1380:45d1:ec00::1;
Date: Tue, 2 Apr 2024 11:38:34 +0300
From: Leon Romanovsky <leon@kernel.org>
To: "Gustavo A. R. Silva" <gustavo@embeddedor.com>
Cc: Jason Gunthorpe <jgg@ziepe.ca>,
	"Gustavo A. R. Silva" <gustavoars@kernel.org>,
	linux-rdma@vger.kernel.org, linux-kernel@vger.kernel.org,
	linux-hardening@vger.kernel.org
Subject: Re: [PATCH][next] RDMA/cm: Avoid -Wflex-array-member-not-at-end
 warning
Message-ID: <20240402083834.GF11187@unreal>
References: <ZgHdZ15cQ7MIHsGL@neat>
 <20240325224706.GB8419@ziepe.ca>
 <5c0bb827-e5f3-4178-ad46-8ac9b99d7726@embeddedor.com>
Precedence: bulk
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <5c0bb827-e5f3-4178-ad46-8ac9b99d7726@embeddedor.com>

On Mon, Mar 25, 2024 at 08:57:08PM -0600, Gustavo A. R. Silva wrote:
> 
> 
> On 3/25/24 16:47, Jason Gunthorpe wrote:
> > On Mon, Mar 25, 2024 at 02:24:07PM -0600, Gustavo A. R. Silva wrote:
> > > -Wflex-array-member-not-at-end is coming in GCC-14, and we are getting
> > > ready to enable it globally.
> > > 
> > > Use the `struct_group_tagged()` helper to separate the flexible array
> > > from the rest of the members in flexible `struct cm_work`, and avoid
> > > embedding the flexible-array member in `struct cm_timewait_info`.
> > > 
> > > Also, use `container_of()` to retrieve a pointer to the flexible
> > > structure.
> > > 
> > > So, with these changes, fix the following warning:
> > > drivers/infiniband/core/cm.c:196:24: 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@kernel.org>
> > > ---
> > >   drivers/infiniband/core/cm.c | 21 ++++++++++++---------
> > >   1 file changed, 12 insertions(+), 9 deletions(-)
> > > 
> > > diff --git a/drivers/infiniband/core/cm.c b/drivers/infiniband/core/cm.c
> > > index bf0df6ee4f78..80c87085499c 100644
> > > --- a/drivers/infiniband/core/cm.c
> > > +++ b/drivers/infiniband/core/cm.c
> > > @@ -182,18 +182,21 @@ struct cm_av {
> > >   };
> > >   struct cm_work {
> > > -	struct delayed_work work;
> > > -	struct list_head list;
> > > -	struct cm_port *port;
> > > -	struct ib_mad_recv_wc *mad_recv_wc;	/* Received MADs */
> > > -	__be32 local_id;			/* Established / timewait */
> > > -	__be32 remote_id;
> > > -	struct ib_cm_event cm_event;
> > > +	/* New members must be added within the struct_group() macro below. */
> > > +	struct_group_tagged(cm_work_hdr, hdr,
> > > +		struct delayed_work work;
> > > +		struct list_head list;
> > > +		struct cm_port *port;
> > > +		struct ib_mad_recv_wc *mad_recv_wc;	/* Received MADs */
> > > +		__be32 local_id;			/* Established / timewait */
> > > +		__be32 remote_id;
> > > +		struct ib_cm_event cm_event;
> > > +	);
> > >   	struct sa_path_rec path[];
> > >   };
> > 
> > I didn't look, but does it make more sense to break out the path side
> > into its own type and avoid the struct_group_tagged? I seem to
> > remember only one thing used it.
> > 
> 
> I thought about that, but I'd have to change the parameter type of
> `static int cm_timewait_handler(struct cm_work *work)`, and that would
> imply also modifying the internals of function `cm_work_handler()` (and
> then I didn't look much into it). 

So let's try to invest in this direction first before we add obfuscation
with magic words to the code.

Thanks

> So, the `struct_group_tagged()` strategy is in general more cleaner and straightforward.
> 
> --
> Gustavo
> 
>