Re: [PATCH 3/4] mfd: tps65910: Add device-tree support

[Mark Brown]

On Thu, Apr 19, 2012 at 09:35:49AM -0600, Stephen Warren wrote:

> That's not right - the idea is that pdata should override device tree so
> that if there's a platform where the DT is known to contain incorrect
> data, then some early platform code can add pdata to the device to fix
> the problem, and that will be used in preference to the DT data.

> At least, that's the last I heard Grant say on the subject, and that's
> how I've been writing all the Tegra-related drivers, and I've seen
> others do the same for other platforms.

Ugh, this is just leading to horrible code here (and why on earth aren't
we just implementing fixups for DT data?).  Though that said I don't
understand why the code here isn't just checking that there's platform
data rather than setting a flag that says there's platform data, the
driver already needs to use the presence of platform data as a check
anyway.

Attachment: signature.asc
Description: Digital signature