[U-Boot] [PATCH 5/7] mach-snapdragon: Introduce pinctrl driver

[Simon Glass]

Hi Ramon,

On 16 May 2018 at 02:09, Ramon Fried <ramon.fried at gmail.com> wrote:
>
>
> On Tue, May 15, 2018, 7:03 PM Simon Glass <sjg at chromium.org> wrote:
>>
>> Hi Ramon,
>>
>> On 15 May 2018 at 07:23, Ramon Fried <ramon.fried at gmail.com> wrote:
>> > On Mon, May 14, 2018 at 10:51 PM, Simon Glass <sjg at chromium.org> wrote:
>> >> Hi Ramon,
>> >>
>> >> On 14 May 2018 at 01:10, Ramon Fried <ramon.fried at gmail.com> wrote:
>> >>> On Mon, May 14, 2018 at 1:00 AM, Simon Glass <sjg at chromium.org> wrote:
>> >>>> Hi Ramon,
>> >>>>
>> >>>> On 12 May 2018 at 20:15, Ramon Fried <ramon.fried at gmail.com> wrote:
>> >>>>> This patch adds pinmux and pinctrl driver for TLMM
>> >>>>> subsystem in snapdragon chipsets.
>> >>>>> Currently, supporting only 8016, but implementation is
>> >>>>> generic and 8096 can be added easily.
>> >>>>>
>> >>>>> Driver is using the generic dt-bindings and doesn't
>> >>>>> introduce any new bindings (yet).
>> >>>>>
>> >>>>> Signed-off-by: Ramon Fried <ramon.fried at gmail.com>
>> >>>>> ---
>> >>>>>  arch/arm/mach-snapdragon/Makefile                |   2 +
>> >>>>>  arch/arm/mach-snapdragon/pinctrl-apq8016.c       | 162
>> >>>>> +++++++++++++++++++++++
>> >>>>>  arch/arm/mach-snapdragon/pinctrl-snapdragon.c    | 118
>> >>>>> +++++++++++++++++
>> >>>>>  arch/arm/mach-snapdragon/pinctrl-snapdragon.h    |  21 +++
>> >>>>>  configs/dragonboard410c_defconfig                |   5 +
>> >>>>>  include/dt-bindings/pinctrl/pinctrl-snapdragon.h |  22 +++
>> >>>>>  6 files changed, 330 insertions(+)
>> >>>>>  create mode 100644 arch/arm/mach-snapdragon/pinctrl-apq8016.c
>> >>>>>  create mode 100644 arch/arm/mach-snapdragon/pinctrl-snapdragon.c
>> >>>>>  create mode 100644 arch/arm/mach-snapdragon/pinctrl-snapdragon.h
>> >>>>>  create mode 100644 include/dt-bindings/pinctrl/pinctrl-snapdragon.h
>> >>>>>
>> >>>>> diff --git a/arch/arm/mach-snapdragon/Makefile
>> >>>>> b/arch/arm/mach-snapdragon/Makefile
>> >>>>> index 1c23dc52cf..1d35fea912 100644
>> >>>>> --- a/arch/arm/mach-snapdragon/Makefile
>> >>>>> +++ b/arch/arm/mach-snapdragon/Makefile
>> >>>>> @@ -6,4 +6,6 @@ obj-$(CONFIG_TARGET_DRAGONBOARD820C) +=
>> >>>>> clock-apq8096.o
>> >>>>>  obj-$(CONFIG_TARGET_DRAGONBOARD820C) += sysmap-apq8096.o
>> >>>>>  obj-$(CONFIG_TARGET_DRAGONBOARD410C) += clock-apq8016.o
>> >>>>>  obj-$(CONFIG_TARGET_DRAGONBOARD410C) += sysmap-apq8016.o
>> >>>>> +obj-$(CONFIG_TARGET_DRAGONBOARD410C) += pinctrl-apq8016.o
>> >>>>> +obj-$(CONFIG_TARGET_DRAGONBOARD410C) += pinctrl-snapdragon.o
>> >>>>>  obj-y += clock-snapdragon.o
>> >>>>> diff --git a/arch/arm/mach-snapdragon/pinctrl-apq8016.c
>> >>>>> b/arch/arm/mach-snapdragon/pinctrl-apq8016.c
>> >>>>> new file mode 100644
>> >>>>> index 0000000000..8e57e2338c
>> >>>>> --- /dev/null
>> >>>>> +++ b/arch/arm/mach-snapdragon/pinctrl-apq8016.c
>> >>>>> @@ -0,0 +1,162 @@
>> >>>>> +// SPDX-License-Identifier: GPL-2.0+
>> >>>>> +/*
>> >>>>> + * Qualcomm APQ8016 pinctrl
>> >>>>> + *
>> >>>>> + * (C) Copyright 2018 Ramon Fried <ramon.fried at gmail.com>
>> >>>>> + *
>> >>>>> + */
>> >>>>> +
>> >>>>> +#include "pinctrl-snapdragon.h"
>> >>>>> +#include <common.h>
>> >>>>> +
>> >>>>> +const char * const msm_pinctrl_pins[] = {
>> >>>>> +       "GPIO_0",
>> >>>>> +       "GPIO_1",
>> >>>>> +       "GPIO_2",
>> >>>>> +       "GPIO_3",
>> >>>>> +       "GPIO_4",
>> >>>>> +       "GPIO_5",
>> >>>>> +       "GPIO_6",
>> >>>>> +       "GPIO_7",
>> >>>>
>> >>>> This seems inefficient. Could you not sprintf() the name for most of
>> >>>> these values?
>> >>> The origin of this table is from the Linux kernel driver.
>> >>> I'm not sure I understand how sprintf will more efficient, do you want
>> >>> to fill up this table on runtime ?
>> >>
>> >> I think this table is only used in one function, so you could create
>> >> the string there perhaps?
>> >>
>> > Actually, it works the other way around, the generic-pinctrl needs a
>> > function to translate string to index.
>> > Basically, it reads strings from the FDT and then go over all indexes
>> > until it matches that string. this is inefficient IMHO as I think it
>> > will be easier just
>> > to be able to provide an index instead of a string in the FDT.
>>
>> OK, so long as the index is actually a known value. If someone changes
>> the DT, won't that fail?
>>
>>
>> Regards,
>> Simon
>
> I'm not sure I understand the question. If someone changes the DT to an
> invalid string? Which is not in the table. Then yes. it will fail.

I am suggesting that the table could be reduced in size since most of
the entries can be generated with sprintf().

Regards,
Simon