RE: [PATCH 1/3] remap: Add I3C bus support
From: Vitor Soares
Date: Tue Apr 23 2019 - 11:55:38 EST
Hi Boris,
From: Boris Brezillon <boris.brezillon@xxxxxxxxxxxxx>
Date: Tue, Apr 23, 2019 at 16:14:16
> On Tue, 23 Apr 2019 14:58:06 +0000
> Vitor Soares <Vitor.Soares@xxxxxxxxxxxx> wrote:
>
> > Hi Mark,
> >
> > From: Mark Brown <broonie@xxxxxxxxxx>
> > Date: Tue, Apr 16, 2019 at 16:39:48
> >
> > > On Mon, Apr 15, 2019 at 09:19:39PM +0200, Vitor Soares wrote:
> > >
> > > > +++ b/drivers/base/regmap/regmap-i3c.c
> > > > @@ -0,0 +1,62 @@
> > > > +// SPDX-License-Identifier: GPL-2.0
> > > > +/*
> > > > + * Copyright (c) 2018 Synopsys, Inc. and/or its affiliates.
> > >
> > > Please make the entire header C++ style so it looks more consistent.
> > > Otherwise this looks good modulo
> >
> > I will change it next drop.
> >
> > > Boris' comment; I'm fine with leaving
> > > extra modes for later so long as they can be introduced without
> > > disrupting existing users so the only question there would be if we
> > > should name the init function in some way that's specific to the I/O
> > > mode being used here.
> >
> > My concern is that booth modes (SDR/HDR) might be needed on the device.
> > e.g. use SDR to configure the device and use HDR to send/receive large
> > data.
>
> I'd say that we shouldn't use the regmap abstraction in this case or
> have a driver-specific backend implementation for it. I guess the
> common case is "regs are accessed in SDR mode", so let's keep the name
> as it is now and we'll define devm_regmap_init_i3c_hdr() if we ever
> need it. Please make it explicit in the kernel-doc that we're using SDR
> transfers here.
Where should it be documented? Can you give an example?
Thanks,
Vitor Soares