On 31/03/2024 14:01, Peng Fan wrote:
>> Subject: Re: [PATCH v5 4/4] clk: imx: add i.MX95 BLK CTL clk driver
>>
>> On 24/03/2024 08:52, Peng Fan (OSS) wrote:
>>> From: Peng Fan <peng.fan@xxxxxxx>
>>>
>>
>> ...
>>
>>> +
>>> +static const struct of_device_id imx95_bc_of_match[] = {
>>> + { .compatible = "nxp,imx95-camera-csr", .data = &camblk_dev_data },
>>> + { .compatible = "nxp,imx95-display-master-csr", },
>>> + { .compatible = "nxp,imx95-lvds-csr", .data = &lvds_csr_dev_data },
>>> + { .compatible = "nxp,imx95-display-csr", .data =
>> &dispmix_csr_dev_data },
>>> + { .compatible = "nxp,imx95-vpu-csr", .data = &vpublk_dev_data },
>>> + { /* Sentinel */ },
>>> +};
>>> +MODULE_DEVICE_TABLE(of, imx95_bc_of_match);
>>> +
>>> +static struct platform_driver imx95_bc_driver = {
>>> + .probe = imx95_bc_probe,
>>> + .driver = {
>>> + .name = "imx95-blk-ctl",
>>> + .of_match_table = of_match_ptr(imx95_bc_of_match),
>>
>> Drop of_match_ptr(), causes warnings. From where did you copy such code?
>> Which mainline driver has such pattern?
>
> I recall that when COMPILE_TEST is selected, OF is not selected, kernel
> robot reports warning. This may not be true now.
>
This itself was never true. Any compilation, not robots, which you can
test by yourself will report warnings if your code is not correct. Don't
use random C syntax to suppress warnings, but actually something which
will work. of_match_ptr does not make sense without maybe_unused or
ifdef. But anyway of_match_ptr is not recommended.
Best regards,
Krzysztof
