[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <CAGb2v665C-EM4t1h3rQ=hNR=1Z4J2mp7DRy_zM8-OkCXzZ03iw@mail.gmail.com>
Date: Mon, 23 Nov 2015 15:41:52 +0800
From: Chen-Yu Tsai <wens@...e.org>
To: moinejf@...e.fr
Cc: Maxime Ripard <maxime.ripard@...e-electrons.com>,
Arnd Bergmann <arnd@...db.de>,
devicetree <devicetree@...r.kernel.org>,
Vishnu Patekar <vishnupatekar0510@...il.com>,
Emilio López <emilio@...pez.com.ar>,
Michael Turquette <mturquette@...libre.com>,
linux-sunxi <linux-sunxi@...glegroups.com>,
linux-kernel <linux-kernel@...r.kernel.org>,
Hans de Goede <hdegoede@...hat.com>,
Chen-Yu Tsai <wens@...e.org>, Rob Herring <robh+dt@...nel.org>,
Jens Kuske <jenskuske@...il.com>,
Philipp Zabel <p.zabel@...gutronix.de>,
Linus Walleij <linus.walleij@...aro.org>,
linux-arm-kernel <linux-arm-kernel@...ts.infradead.org>
Subject: Re: [linux-sunxi] Re: [PATCH v4 4/6] reset: sunxi: Add Allwinner H3
bus resets
On Thu, Nov 5, 2015 at 2:47 PM, Jean-Francois Moine <moinejf@...e.fr> wrote:
> On Wed, 4 Nov 2015 08:30:14 -0800
> Maxime Ripard <maxime.ripard@...e-electrons.com> wrote:
>
>> Hi Arnd,
>>
>> On Fri, Oct 30, 2015 at 09:27:03AM +0100, Arnd Bergmann wrote:
>> > On Tuesday 27 October 2015 17:50:24 Jens Kuske wrote:
>> > >
>> > > +static int sun8i_h3_bus_reset_xlate(struct reset_controller_dev *rcdev,
>> > > + const struct of_phandle_args *reset_spec)
>> > > +{
>> > > + unsigned int index = reset_spec->args[0];
>> > > +
>> > > + if (index < 96)
>> > > + return index;
>> > > + else if (index < 128)
>> > > + return index + 32;
>> > > + else if (index < 160)
>> > > + return index + 64;
>> > > + else
>> > > + return -EINVAL;
>> > > +}
>> > > +
>> > >
>> >
>> > This looks like you are doing something wrong and should either
>> > put the actual number into DT,
>>
>> This is the actual number, except that there's some useless registers
>> in between. Allwinner documents it like that:
>>
>> 0x0 Reset 0
>> 0x4 Reset 1
>> 0xc Reset 2
>>
>> So we have to adjust the offset to account with the blank register in
>> between (0x8).
>>
>> > or use a two-cell representation, with the first cell indicating the
>> > block (0, 1 or 2), and the second cell the index.
>>
>> And the missing register is not a block either.
>>
>> That would also imply either changing the bindings of that driver (and
>> all the current DTS that are using it), or introducing a whole new
>> driver just to deal with some extraordinary offset calculation.
>
> In the H3, the holes are not used, but what would occur if these holes
> would be used for some other purpose in future SoCs? Double mapping?
We'd have a different compatible string for it.
My suggestion for the resets is to just split them into 3 nodes: AHB
(since AHB1 and AHB2 devices are mixed together in the bunch), APB1,
and APB2 reset controls.
This follows what we have for existing SoCs, and gets rid of the unused
hole. We can use the existing "allwinner,sun6i-a31-clock-reset" and
"allwinner,sun6i-a31-ahb1-reset" compatibles.
Regards
ChenYu
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
Powered by blists - more mailing lists