Re: [alsa-devel] [PATCH] ASoC: rt5645: fix build warning

[Jarkko Nikula <jarkko.nikula@linux.intel.com>]

On 10/08/2015 11:55 AM, Sudip Mukherjee wrote:
> On Wed, Oct 07, 2015 at 03:05:27PM +0100, Mark Brown wrote:
>> On Wed, Oct 07, 2015 at 05:52:12PM +0530, Sudip Mukherjee wrote:
>>> We were getting build warning about "Section mismatch".
>>> dmi_platform_intel_broadwell is being referenced from the probe function
>>> rt5645_i2c_probe(), but dmi_platform_intel_broadwell was marked with
>>> __initdata.
>>>
>>> Signed-off-by: Sudip Mukherjee <sudip@vectorindia.org>
>>
>>> -static struct dmi_system_id dmi_platform_intel_broadwell[] __initdata = {
>>> +static struct dmi_system_id dmi_platform_intel_broadwell[] = {
>>
>> This doesn't seem like the obvious fix - why are we not annotating the
>> probe function suitably (or alternatively if we can't why does
>> __initdata still exist)?
>
> probe function should not be __init. probe can be called anytime after
> the module has been loaded.
> __initdata still exists as that part of the code was added by
> e9159e7577cf ("ASoC: rt5645: Add dmi for Broadwell") which is a very
> recent modification and I think that has been added by mistake.
> One more argument in my favor:
> The use in probe function is
> if (dmi_check_system(dmi_platform_intel_braswell) ||
> 		dmi_check_system(dmi_platform_intel_broadwell))
>
> dmi_platform_intel_braswell is not marked as __initdata but
> dmi_platform_intel_broadwell is marked but they both have same use at
> the same place.
>
If one really wants to save a few bytes then one could annotate this it 
with __initdata_or_module (very low use in kernel) but I don't think 
it's worth of effort and probably should be done for other sections here 
too by another patch.

To me this patch looks a right thing to do at the moment.

Reviewed-by: Jarkko Nikula <jarkko.nikula@linux.intel.com>