[PATCH] spi: microchip-core: fix passing zero to PTR_ERR warning

[PATCH] spi: microchip-core: fix passing zero to PTR_ERR warning

[Conor Dooley]

It is possible that the error case for devm_clk_get() returns NULL,
in which case zero will be passed to PTR_ERR() as shown by the Smatch
static checker warning:
drivers/spi/spi-microchip-core.c:557 mchp_corespi_probe()
warn: passing zero to 'PTR_ERR'

Fix the warning by explicitly returning an error if devm_clk_get()
returns NULL.

Fixes: 9ac8d17694b6 ("spi: add support for microchip fpga spi controllers")
Reported-by: Dan Carpenter <dan.carpenter@oracle.com>
Link: https://lore.kernel.org/linux-spi/20220615091633.GI2168@kadam/
Signed-off-by: Conor Dooley <conor.dooley@microchip.com>
---
 drivers/spi/spi-microchip-core.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/spi/spi-microchip-core.c b/drivers/spi/spi-microchip-core.c
index 5b2aee30fa04..d44663cca071 100644
--- a/drivers/spi/spi-microchip-core.c
+++ b/drivers/spi/spi-microchip-core.c
@@ -554,7 +554,7 @@ static int mchp_corespi_probe(struct platform_device *pdev)
 
 	spi->clk = devm_clk_get(&pdev->dev, NULL);
 	if (!spi->clk || IS_ERR(spi->clk)) {
-		ret = PTR_ERR(spi->clk);
+		ret = !spi->clk ? -ENXIO : PTR_ERR(spi->clk);
 		dev_err(&pdev->dev, "could not get clk: %d\n", ret);
 		goto error_release_master;
 	}
-- 
2.36.1

Re: [PATCH] spi: microchip-core: fix passing zero to PTR_ERR warning

[Mark Brown]

[-- Attachment #1: Type: text/plain, Size: 196 bytes --]

On Wed, Jun 15, 2022 at 12:30:22PM +0100, Conor Dooley wrote:

> -		ret = PTR_ERR(spi->clk);
> +		ret = !spi->clk ? -ENXIO : PTR_ERR(spi->clk);

I think you're looking for PTR_ERR_OR_ZERO() here?

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 488 bytes --]

Re: [PATCH] spi: microchip-core: fix passing zero to PTR_ERR warning

[Conor.Dooley]

On 15/06/2022 12:40, Mark Brown wrote:
> EXTERNAL EMAIL: Do not click links or open attachments unless you know the content is safe
> 
> On Wed, Jun 15, 2022 at 12:30:22PM +0100, Conor Dooley wrote:
> 
>> -             ret = PTR_ERR(spi->clk);
>> +             ret = !spi->clk ? -ENXIO : PTR_ERR(spi->clk);
> 
> I think you're looking for PTR_ERR_OR_ZERO() here?

Maybe I don't understand, so let me explain what I think you're
suggesting & maybe you can correct me:
> -             ret = PTR_ERR(spi->clk);
> +             ret = PTR_ERR_OR_ZERO(spi->clk);

But if spi->clk is NULL, this will return 0 from the probe
rather than returning an error?
If that's not what you meant, lmk
Thanks,
Conor.

> 
> 
> _______________________________________________
> linux-riscv mailing list
> linux-riscv@lists.infradead.org
> http://lists.infradead.org/mailman/listinfo/linux-riscv

Re: [PATCH] spi: microchip-core: fix passing zero to PTR_ERR warning

[Mark Brown]

[-- Attachment #1: Type: text/plain, Size: 885 bytes --]

On Wed, Jun 15, 2022 at 11:57:37AM +0000, Conor.Dooley@microchip.com wrote:
> On 15/06/2022 12:40, Mark Brown wrote:
> > EXTERNAL EMAIL: Do not click links or open attachments unless you know the content is safe
> > On Wed, Jun 15, 2022 at 12:30:22PM +0100, Conor Dooley wrote:

> >> -             ret = PTR_ERR(spi->clk);
> >> +             ret = !spi->clk ? -ENXIO : PTR_ERR(spi->clk);

> > I think you're looking for PTR_ERR_OR_ZERO() here?

> Maybe I don't understand, so let me explain what I think you're
> suggesting & maybe you can correct me:

> > -             ret = PTR_ERR(spi->clk);
> > +             ret = PTR_ERR_OR_ZERO(spi->clk);

> But if spi->clk is NULL, this will return 0 from the probe
> rather than returning an error?
> If that's not what you meant, lmk

Oh, hang on - what error conditions can clk_get() return 0 in?  The
documentation doesn't mention any...

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 488 bytes --]

Re: [PATCH] spi: microchip-core: fix passing zero to PTR_ERR warning

[Conor.Dooley]

On 15/06/2022 13:37, Mark Brown wrote:
> EXTERNAL EMAIL: Do not click links or open attachments unless you know the content is safe
> 
> On Wed, Jun 15, 2022 at 11:57:37AM +0000, Conor.Dooley@microchip.com wrote:
>> On 15/06/2022 12:40, Mark Brown wrote:
>>> EXTERNAL EMAIL: Do not click links or open attachments unless you know the content is safe
>>> On Wed, Jun 15, 2022 at 12:30:22PM +0100, Conor Dooley wrote:
> 
>>>> -             ret = PTR_ERR(spi->clk);
>>>> +             ret = !spi->clk ? -ENXIO : PTR_ERR(spi->clk);
> 
>>> I think you're looking for PTR_ERR_OR_ZERO() here?
> 
>> Maybe I don't understand, so let me explain what I think you're
>> suggesting & maybe you can correct me:
> 
>>> -             ret = PTR_ERR(spi->clk);
>>> +             ret = PTR_ERR_OR_ZERO(spi->clk);
> 
>> But if spi->clk is NULL, this will return 0 from the probe
>> rather than returning an error?
>> If that's not what you meant, lmk
> 
> Oh, hang on - what error conditions can clk_get() return 0 in?  The
> documentation doesn't mention any...

If !CONFIG_HAVE_CLK, (without which it won't boot on the coreplex)
but I don't think I can be sure that CONFIG_HAVE_CLK will /always/
be enabled for other uses of the FPGA.

Re: [PATCH] spi: microchip-core: fix passing zero to PTR_ERR warning

[Mark Brown]

[-- Attachment #1: Type: text/plain, Size: 689 bytes --]

On Wed, Jun 15, 2022 at 12:42:27PM +0000, Conor.Dooley@microchip.com wrote:
> On 15/06/2022 13:37, Mark Brown wrote:

> >> But if spi->clk is NULL, this will return 0 from the probe
> >> rather than returning an error?
> >> If that's not what you meant, lmk

> > Oh, hang on - what error conditions can clk_get() return 0 in?  The
> > documentation doesn't mention any...

> If !CONFIG_HAVE_CLK, (without which it won't boot on the coreplex)
> but I don't think I can be sure that CONFIG_HAVE_CLK will /always/
> be enabled for other uses of the FPGA.

That's not an error, that's returning NULL as a dummy clock.  The
expectation is that the driver will proceed as though it has a clock.

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 488 bytes --]

Re: [PATCH] spi: microchip-core: fix passing zero to PTR_ERR warning

[Conor.Dooley]

On 15/06/2022 13:49, Mark Brown wrote:
> On Wed, Jun 15, 2022 at 12:42:27PM +0000, Conor.Dooley@microchip.com wrote:
>> On 15/06/2022 13:37, Mark Brown wrote:
> 
>>>> But if spi->clk is NULL, this will return 0 from the probe
>>>> rather than returning an error?
>>>> If that's not what you meant, lmk
> 
>>> Oh, hang on - what error conditions can clk_get() return 0 in?  The
>>> documentation doesn't mention any...
> 
>> If !CONFIG_HAVE_CLK, (without which it won't boot on the coreplex)
>> but I don't think I can be sure that CONFIG_HAVE_CLK will /always/
>> be enabled for other uses of the FPGA.
> 
> That's not an error, that's returning NULL as a dummy clock.  The
> expectation is that the driver will proceed as though it has a clock.

Ahh, cool. I guess I'll just drop the null check entirely so & guard
against a div zero when I actually go to use the clock.
Thanks!
Conor