From mboxrd@z Thu Jan  1 00:00:00 1970
From: Nathan Chancellor <natechancellor@gmail.com>
Subject: Re: [PATCH] scsi: gdth: Only call dma_free_coherent when buf is not
 NULL in ioc_general
Date: Fri, 8 Mar 2019 14:31:43 -0700
Message-ID: <20190308213143.GA27509@archlinux-ryzen>
References: <20190307231839.3330-1-natechancellor@gmail.com>
 <CAKwvOd=PZUW1U0KiBgn8z8ReHeRX=s8xEtS7Gebf9-AXxguNNQ@mail.gmail.com>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Return-path: <linux-kernel-owner@vger.kernel.org>
Content-Disposition: inline
In-Reply-To: <CAKwvOd=PZUW1U0KiBgn8z8ReHeRX=s8xEtS7Gebf9-AXxguNNQ@mail.gmail.com>
Sender: linux-kernel-owner@vger.kernel.org
To: Nick Desaulniers <ndesaulniers@google.com>
Cc: "James E.J. Bottomley" <jejb@linux.ibm.com>, "Martin K. Petersen" <martin.petersen@oracle.com>, Achim Leubner <achim_leubner@adaptec.com>, linux-scsi@vger.kernel.org, LKML <linux-kernel@vger.kernel.org>, clang-built-linux@googlegroups.com, Christoph Hellwig <hch@lst.de>
List-Id: linux-scsi@vger.kernel.org

On Fri, Mar 08, 2019 at 01:14:25PM -0800, Nick Desaulniers wrote:
> On Thu, Mar 7, 2019 at 3:19 PM Nathan Chancellor
> <natechancellor@gmail.com> wrote:
> >
> > When building with -Wsometimes-uninitialized, Clang warns:
> >
> > drivers/scsi/gdth.c:3662:6: warning: variable 'paddr' is used
> > uninitialized whenever 'if' condition is false
> > [-Wsometimes-uninitialized]
> >
> > Don't attempt to call dma_free_coherent when buf is NULL (meaning that
> > we never called dma_alloc_coherent and initialized paddr), which avoids
> > this warning.
> >
> > Link: https://github.com/ClangBuiltLinux/linux/issues/402
> > Signed-off-by: Nathan Chancellor <natechancellor@gmail.com>
> > ---
> >  drivers/scsi/gdth.c | 5 +++--
> >  1 file changed, 3 insertions(+), 2 deletions(-)
> >
> > diff --git a/drivers/scsi/gdth.c b/drivers/scsi/gdth.c
> > index e7f1dd4f3b66..0ca9b4393770 100644
> > --- a/drivers/scsi/gdth.c
> > +++ b/drivers/scsi/gdth.c
> > @@ -3697,8 +3697,9 @@ static int ioc_general(void __user *arg, char *cmnd)
> >
> >         rval = 0;
> >  out_free_buf:
> > -       dma_free_coherent(&ha->pdev->dev, gen.data_len + gen.sense_len, buf,
> > -                       paddr);
> > +       if (buf)
> > +               dma_free_coherent(&ha->pdev->dev, gen.data_len + gen.sense_len,
> > +                                 buf, paddr);
> >         return rval;
> >  }
> >
> > --
> > 2.21.0
> >
> 
> Alternatively, paddr is a dma_addr_t defined in include/linux/types.h:
> 
> #ifdef CONFIG_ARCH_DMA_ADDR_T_64BIT
> typedef u64 dma_addr_t;
> #else
> typedef u32 dma_addr_t;
> #endif
> 
> Just initializing it to zero might be simpler than complicating the
> control flow of this function further. Thoughts?
> 
> diff --git a/drivers/scsi/gdth.c b/drivers/scsi/gdth.c
> index e7f1dd4f3b66..5a3f849ebf64 100644
> --- a/drivers/scsi/gdth.c
> +++ b/drivers/scsi/gdth.c
> @@ -3643,7 +3643,7 @@ static int ioc_general(void __user *arg, char *cmnd)
>         gdth_ioctl_general gen;
>         gdth_ha_str *ha;
>         char *buf = NULL;
> -       dma_addr_t paddr;
> +       dma_addr_t paddr = 0U;
>         int rval;
> 
>         if (copy_from_user(&gen, arg, sizeof(gdth_ioctl_general)))
> -- 
> Thanks,
> ~Nick Desaulniers

I suppose it depends on if it's okay to call dma_free_coherent without
dma_alloc_coherent. I did a scan of the tree before sending this out
and pretty much all sites that have error handling check that the third
parameter is not NULL before calling it.

I should have added Christoph to this thread when I initially sent it
out since commit 9f475ebff8e4 ("scsi: gdth: refactor ioc_general")
introduced this. Done now.

Cheers,
Nathan