Hi Christophe, thank you for your review.
There are two problems in the ucc_of_parse_tdm function.
1, NULL pointer reference
2, pdev gets modified, iomap was done with a different pdev.
We will submit a patch to repair it later.
Thanks.
132 pdev = of_find_device_by_node(np2);
......
141 utdm->si_regs = devm_ioremap_resource(&pdev->dev, res); ---> use
pdev to call devm_ioremap
......
153 pdev = of_find_device_by_node(np2); --->This line, pdev gets
modified
154 if (!pdev) {
155 ret = -EINVAL;
156 pr_err("%pOFn: failed to lookup pdev\n", np2);
157 of_node_put(np2);
158 goto err_miss_siram_property;
------> This line, if pdev is NULL, will jump to label err_miss_siram_property
159 }
......
163 utdm->siram = devm_ioremap_resource(&pdev->dev, res); --> use
changed pdev to call devm_ioremap
......
175
176 err_miss_siram_property:
177 devm_iounmap(&pdev->dev, utdm->si_regs); ------> Within the code
block, if pdev is NULL, kernel panic will be triggered; devm_iounmap() will
alse fail even when the new pdev is valid.
178 return ret;
179 }
------------------原始邮件------------------
发件人：ChristopheLEROY <christophe.le...@c-s.fr>
收件人：文洋10156314;qiang.z...@nxp.com <qiang.z...@nxp.com>
抄送人：汪翼10129963;钟卫东10001088;linux-ker...@vger.kernel.org
<linux-ker...@vger.kernel.org>leoyang...@nxp.com <leoyang...@nxp.com>Julia
Lawall <julia.law...@lip6.fr>linuxppc-dev@lists.ozlabs.org
<linuxppc-dev@lists.ozlabs.org>linux-arm-ker...@lists.infradead.org
<linux-arm-ker...@lists.infradead.org>
日 期 ：2018年11月21日 18:46
主 题 ：Re: [PATCH 2/4] soc/fsl/qe: fix potential NULL pointer dereference
inucc_of_parse_tdm
Le 21/11/2018 à 08:41, Wen Yang a écrit :
> This patch fixes a possible null pointer dereference in
> ucc_of_parse_tdm, detected by the semantic patch
> deref_null.cocci, with the following warning:
>
> ./drivers/soc/fsl/qe/qe_tdm.c:177:21-24: ERROR: pdev is NULL but dereferenced.
>
> The following code has potential null pointer references:
> pdev = of_find_device_by_node(np2);
> if (!pdev) {
> ret = -EINVAL;
> pr_err("%pOFn: failed to lookup pdev\n", np2);
> of_node_put(np2);
> goto err_miss_siram_property;
> }
> ...
> err_miss_siram_property:
> devm_iounmap(&pdev->dev, utdm->si_regs);
>
> Signed-off-by: Wen Yang <wen.yan...@zte.com.cn>
> Reviewed-by: Tan Hu <tan...@zte.com.cn>
> CC: Julia Lawall <julia.law...@lip6.fr>
> ---
> drivers/soc/fsl/qe/qe_tdm.c | 3 ++-
> 1 file changed, 2 insertions(+), 1 deletion(-)
>
> diff --git a/drivers/soc/fsl/qe/qe_tdm.c b/drivers/soc/fsl/qe/qe_tdm.c
> index f78c346..166378b 100644
> --- a/drivers/soc/fsl/qe/qe_tdm.c
> +++ b/drivers/soc/fsl/qe/qe_tdm.c
> @@ -174,7 +174,8 @@ int ucc_of_parse_tdm(struct device_node *np, struct
> ucc_tdm *utdm,
> return ret;
>
> err_miss_siram_property:
> - devm_iounmap(&pdev->dev, utdm->si_regs);
> + if (pdev)
> + devm_iounmap(&pdev->dev, utdm->si_regs);
You are just hiding the issue, not fixing it.
The problem is that pdev gets modified, so in any case that
devm_iounmap() will fail even when the new pdev is valid, because the
iomap was done with a different pdev.
Christophe
> return ret;
> }
> EXPORT_SYMBOL(ucc_of_parse_tdm);
>
