Re: [PATCH -next] clk: hisilicon: hi3660：Fix potential NULL dereference in hi3660_stub_clk_probe()

[Leo Yan <leo.yan@linaro.org>]

On Fri, Jan 05, 2018 at 12:59:38PM +0900, Masahiro Yamada wrote:
> 2018-01-04 15:36 GMT+09:00 Wei Yongjun <weiyongjun1@huawei.com>:
> > platform_get_resource() may return NULL, add proper check to
> > avoid potential NULL dereferencing.
> >
> > This is detected by Coccinelle semantic patch.
> >
> > @@
> > expression pdev, res, n, t, e, e1, e2;
> > @@
> >
> > res = platform_get_resource(pdev, t, n);
> > + if (!res)
> > +   return -EINVAL;
> > ... when != res == NULL
> > e = devm_ioremap(e1, res->start, e2);
> >
> > Fixes: 4f16f7ff3bc0 ("clk: hisilicon: Add support for Hi3660 stub clocks")
> > Signed-off-by: Wei Yongjun <weiyongjun1@huawei.com>
> > ---
> >  drivers/clk/hisilicon/clk-hi3660-stub.c | 2 ++
> >  1 file changed, 2 insertions(+)
> >
> > diff --git a/drivers/clk/hisilicon/clk-hi3660-stub.c b/drivers/clk/hisilicon/clk-hi3660-stub.c
> > index 9b6c72b..e8b2c43 100644
> > --- a/drivers/clk/hisilicon/clk-hi3660-stub.c
> > +++ b/drivers/clk/hisilicon/clk-hi3660-stub.c
> > @@ -149,6 +149,8 @@ static int hi3660_stub_clk_probe(struct platform_device *pdev)
> >                 return PTR_ERR(stub_clk_chan.mbox);
> >
> >         res = platform_get_resource(pdev, IORESOURCE_MEM, 0);
> > +       if (!res)
> > +               return -EINVAL;
> >         freq_reg = devm_ioremap(dev, res->start, resource_size(res));
> >         if (!freq_reg)
> >                 return -ENOMEM;
> 
> 
> Probably, it is better to use devm_ioremap_resource().
> This checks NULL input.

Thanks for suggestion, Masahiro. This is good suggestion.

Yongjun, could you spin new patch by using devm_ioremap_resource()
to replace devm_ioremap()?

Thanks,
Leo Yan