Re: [PATCH v2] soc: ti: k3-ringacc: include platform_device header

[Nishanth Menon]

On 16:08-20260225, Siddharth Vadapalli wrote:
> On 25/02/26 15:58, Nishanth Menon wrote:
> > On 15:33-20260225, Siddharth Vadapalli wrote:
> > > On 25/02/26 13:18, Vinod Koul wrote:
> > > > On 20-02-26, 12:32, Siddharth Vadapalli wrote:
> > > > > Users of "k3-ringacc" APIs may not necessarily be platform devices. Such
> > > > > users therefore will not include "platform_device.h", resulting in
> > > > > compilation warnings for the APIs declared in "k3-ringacc.h" which assume
> > > > > inclusion of "platform_device.h" by its users.
> > > > > 
> > > > > Fix the compilation warnings by including the "platform_device" header
> > > > > file, which should ideally have been included since commit under Fixes.
> > > > 
> > > > Which tree is this intended to get picked by?
> > > > 
> > > 
> > > I assume it is the SoC tree since 'get_maintainer.pl' doesn't output a
> > > mailing list apart from 'linux-kernel@xxxxxxxxxxxxxxx'.
> > > 
> > 
> > drivers/soc/ti/k3-ringacc.c already includes the header.. so who needs
> > it?
> > 
> 
> At the point in time that the v1 patch was posted, this was needed for the
> patch in an RFC series:
> https://lore.kernel.org/r/20240518124234.2671651-19-s-vadapalli@xxxxxx/
> Although the RFC series hasn't been merged, since it is still possible for
> non platform-device users to call the DMA APIs which in turn depend on the
> k3-ringacc APIs being exported, I posted the v2 patch.
> 
> If you believe that this patch should be posted along with the dependent
> series, I will do so and the patch can be ignored.

The right thing, IMHO is to include the header in the files that need it.
We don't need a single kernel.h with everything piled on it.

-- 
Regards,
Nishanth Menon
Key (0xDDB5849D1736249D) / Fingerprint: F8A2 8693 54EB 8232 17A3  1A34 DDB5 849D 1736 249D
https://ti.com/opensource