pci-exynos.c coverity issue #1437287

pci-exynos.c coverity issue #1437287

[Bjorn Helgaas]

Hi Jingoo,

Please take a look at this issue reported by Coverity:

222 static void exynos_pcie_assert_reset(struct exynos_pcie *ep)
223 {
224        struct dw_pcie *pci = ep->pci;
225        struct device *dev = pci->dev;
226
   	1. Condition ep->reset_gpio >= 0, taking true branch.
227        if (ep->reset_gpio >= 0)

CID 1437287 (#1 of 1): Unchecked return value (CHECKED_RETURN) 2.
check_return: Calling devm_gpio_request_one without checking return
value (as is done elsewhere 112 out of 113 times).

228                devm_gpio_request_one(dev, ep->reset_gpio,
229                                GPIOF_OUT_INIT_HIGH, "RESET");


Maybe you can guarantee that it's safe to ignore the return value by
doing a lot of code analysis, but I think it would be better to check
like all the other callers do.

Re: pci-exynos.c coverity issue #1437287

[Krzysztof Kozlowski]

On Tue, Oct 27, 2020 at 11:34:46AM -0500, Bjorn Helgaas wrote:
> Hi Jingoo,
> 
> Please take a look at this issue reported by Coverity:
> 
> 222 static void exynos_pcie_assert_reset(struct exynos_pcie *ep)
> 223 {
> 224        struct dw_pcie *pci = ep->pci;
> 225        struct device *dev = pci->dev;
> 226
>    	1. Condition ep->reset_gpio >= 0, taking true branch.
> 227        if (ep->reset_gpio >= 0)
> 
> CID 1437287 (#1 of 1): Unchecked return value (CHECKED_RETURN) 2.
> check_return: Calling devm_gpio_request_one without checking return
> value (as is done elsewhere 112 out of 113 times).
> 
> 228                devm_gpio_request_one(dev, ep->reset_gpio,
> 229                                GPIOF_OUT_INIT_HIGH, "RESET");
> 
> 
> Maybe you can guarantee that it's safe to ignore the return value by
> doing a lot of code analysis, but I think it would be better to check
> like all the other callers do.

+Cc Marek,

Hi Bjorn,

I think this is being removed by Marek's patchset:
https://lore.kernel.org/linux-samsung-soc/20201023075744.26200-1-m.szyprowski@samsung.com/

Best regards,
Krzysztof

Re: pci-exynos.c coverity issue #1437287

[Bjorn Helgaas]

On Tue, Oct 27, 2020 at 05:40:42PM +0100, Krzysztof Kozlowski wrote:
> On Tue, Oct 27, 2020 at 11:34:46AM -0500, Bjorn Helgaas wrote:
> > Hi Jingoo,
> > 
> > Please take a look at this issue reported by Coverity:
> > 
> > 222 static void exynos_pcie_assert_reset(struct exynos_pcie *ep)
> > 223 {
> > 224        struct dw_pcie *pci = ep->pci;
> > 225        struct device *dev = pci->dev;
> > 226
> >    	1. Condition ep->reset_gpio >= 0, taking true branch.
> > 227        if (ep->reset_gpio >= 0)
> > 
> > CID 1437287 (#1 of 1): Unchecked return value (CHECKED_RETURN) 2.
> > check_return: Calling devm_gpio_request_one without checking return
> > value (as is done elsewhere 112 out of 113 times).
> > 
> > 228                devm_gpio_request_one(dev, ep->reset_gpio,
> > 229                                GPIOF_OUT_INIT_HIGH, "RESET");
> > 
> > 
> > Maybe you can guarantee that it's safe to ignore the return value by
> > doing a lot of code analysis, but I think it would be better to check
> > like all the other callers do.
> 
> +Cc Marek,
> 
> Hi Bjorn,
> 
> I think this is being removed by Marek's patchset:
> https://lore.kernel.org/linux-samsung-soc/20201023075744.26200-1-m.szyprowski@samsung.com/

Perfect, thanks!