On Wed, Mar 9, 2022 at 3:06 PM YueHaibing <[email protected]> wrote:
>
> drivers/perf/marvell_cn10k_ddr_pmu.c:723:21: error: ‘cn10k_ddr_pmu_of_match’ undeclared here (not in a function); did you mean ‘cn10k_ddr_pmu_driver’?
> .of_match_table = cn10k_ddr_pmu_of_match,
> ^~~~~~~~~~~~~~~~~~~~~~
>
> Use of_match_ptr() to fix this.
>
> Fixes: 7cf83e222bce ("perf/marvell: CN10k DDR performance monitor support")
> Signed-off-by: YueHaibing <[email protected]>
It's generally better to remove the #ifdef around the match table
definition instead,
which improves readability, and has no downsides on the machines this actually
runs on, because CONFIG_OF is always enabled on arm64.
Arnd
On 2022/3/9 22:11, Arnd Bergmann wrote:
> On Wed, Mar 9, 2022 at 3:06 PM YueHaibing <[email protected]> wrote:
>>
>> drivers/perf/marvell_cn10k_ddr_pmu.c:723:21: error: ‘cn10k_ddr_pmu_of_match’ undeclared here (not in a function); did you mean ‘cn10k_ddr_pmu_driver’?
>> .of_match_table = cn10k_ddr_pmu_of_match,
>> ^~~~~~~~~~~~~~~~~~~~~~
>>
>> Use of_match_ptr() to fix this.
>>
>> Fixes: 7cf83e222bce ("perf/marvell: CN10k DDR performance monitor support")
>> Signed-off-by: YueHaibing <[email protected]>
>
> It's generally better to remove the #ifdef around the match table
> definition instead,
> which improves readability, and has no downsides on the machines this actually
> runs on, because CONFIG_OF is always enabled on arm64.
Ok， will do this in v2, thanks!
>
> Arnd
> .
>