linux-kernel - Re: [PATCH] RDMA/irdma: Fix GCC 12 warning

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]

Date:   Thu, 10 Feb 2022 19:36:22 -0400
From:   Jason Gunthorpe <jgg@...pe.ca>
To:     "Saleem, Shiraz" <shiraz.saleem@...el.com>
Cc:     Victor Erminpour <victor.erminpour@...cle.com>,
        "Ismail, Mustafa" <mustafa.ismail@...el.com>,
        "linux-rdma@...r.kernel.org" <linux-rdma@...r.kernel.org>,
        "linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>,
        "trivial@...nel.org" <trivial@...nel.org>
Subject: Re: [PATCH] RDMA/irdma: Fix GCC 12 warning

On Thu, Feb 10, 2022 at 10:09:18PM +0000, Saleem, Shiraz wrote:
> > Subject: [PATCH] RDMA/irdma: Fix GCC 12 warning
> > 
> > When building with automatic stack variable initialization, GCC 12 complains about
> > variables defined outside of switch case statements.
> > Move the variable into the case that uses it, which silences the warning:
> > 
> > ./drivers/infiniband/hw/irdma/hw.c:270:47: error: statement will never be executed [-
> > Werror=switch-unreachable]
> >   270 |                         struct irdma_cm_node *cm_node;
> >       |
> > 
> > ./drivers/infiniband/hw/irdma/utils.c:1215:50: error: statement will never be executed
> > [-Werror=switch-unreachable]
> >   1215 |                         struct irdma_gen_ae_info ae_info;
> >        |
> > 
> > Signed-off-by: Victor Erminpour <victor.erminpour@...cle.com>
> >  drivers/infiniband/hw/irdma/hw.c    | 2 +-
> >  drivers/infiniband/hw/irdma/utils.c | 2 +-
> >  2 files changed, 2 insertions(+), 2 deletions(-)
> > 
> > diff --git a/drivers/infiniband/hw/irdma/hw.c b/drivers/infiniband/hw/irdma/hw.c
> > index 89234d04cc65..a41a3b128d0d 100644
> > +++ b/drivers/infiniband/hw/irdma/hw.c
> > @@ -267,8 +267,8 @@ static void irdma_process_aeq(struct irdma_pci_f *rf)
> >  		}
> > 
> >  		switch (info->ae_id) {
> > -			struct irdma_cm_node *cm_node;
> >  		case IRDMA_AE_LLP_CONNECTION_ESTABLISHED:
> > +			struct irdma_cm_node *cm_node;
> >  			cm_node = iwqp->cm_node;
> >  			if (cm_node->accept_pend) {
> >  				atomic_dec(&cm_node->listener-
> 
> This doesn't compile.
> 
> drivers/infiniband/hw/irdma/hw.c: In function \u2018irdma_process_aeq\u2019:
> drivers/infiniband/hw/irdma/hw.c:271:4: error: a label can only be part of a statement and a declaration is not a statement
>   271 |    struct irdma_cm_node *cm_node;
> 
> Seems like we are accommodating for gcc12 bug since this C code is
> legit?

It might be legit, but it is nutzo and not our coding style.

If the variable is used by many branches it should be declared at the
top of the function.

If it is used in one branch it should be as above, with the missing {}
added.

Jason