Re: [PATCH v1] regmap: add generic indirect regmap support

[Mark Brown <broonie@kernel.org>]

[-- Attachment #1: Type: text/plain, Size: 1502 bytes --]

On Thu, Jun 09, 2022 at 05:30:43PM -0700, matthew.gerlach@linux.intel.com wrote:
> On Thu, 9 Jun 2022, Mark Brown wrote:
> > On Wed, Jun 08, 2022 at 11:54:26PM +0000, Zhang, Tianfei wrote:

> > > > ...that this appears to be entirely specific to some particular device, it's got
> > > > things like hard coded register addresses and timeouts which mean it can't be
> > > > reused.

> I think part of the confusion is that this patch should have been included
> in a patch set that actually uses this regmap.  This patch really should be
> included with the following:

> https://lore.kernel.org/all/20220607032833.3482-1-tianfei.zhang@intel.com

> The hard coded register definitions are offsets to the passed in void
> __iomem base address.  This set of registers provides the semantics of
> indirect register read/write to whatever the register set is connected to on
> the back end.  Conceptually this could be considered a specific type
> indirect register access controller, but currently we have very different
> backend implementations in RTL.  Part of our intent is to have consistent
> register interfaces for our FPGA IP so multiple drivers can reuse this
> regmap.

...which is all specific to your particular implmentation of this and
not a generic thing that will work for anyone else who implements the
concept of indirected register access and happens to use a different
control mechanism.  If you have to use the same register interfaces to
reuse this code then the code isn't generic.

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 488 bytes --]