[PATCH v2 2/3] soc: rockchip: add driver handling grf setup
Olof Johansson
olof at lixom.net
Sun Jan 29 16:41:46 PST 2017
On Sun, Jan 29, 2017 at 3:36 PM, Heiko Stuebner <heiko at sntech.de> wrote:
> Hi Olof,
>
> Am Sonntag, 29. Januar 2017, 14:40:53 CET schrieb Olof Johansson:
>> On Tue, Nov 15, 2016 at 2:38 PM, Heiko Stuebner <heiko at sntech.de> wrote:
>> > The General Register Files are an area of registers containing a lot
>> > of single-bit settings for numerous components as well full components
>> > like usbphy control. Therefore all used components are accessed
>> > via the syscon provided by the grf nodes or from the sub-devices
>> > created through the simple-mfd created from the grf node.
>> >
>> > Some settings are not used by anything but will need to be set up
>> > according to expectations on the kernel side.
>> >
>> > Best example is the force_jtag setting, which defaults to on and
>> > results in the soc switching the pin-outputs between jtag and sdmmc
>> > automatically depending on the card-detect status. This conflicts
>> > heavily with how the dw_mmc driver expects to do its work and also
>> > with the clock-controller, which has most likely deactivated the
>> > jtag clock due to it being unused.
>> >
>> > So far the handling of this setting was living in the mach-rockchip
>> > code for the arm32-based rk3288 but that of course doesn't work
>> > for arm64 socs and would also look ugly for further arm32 socs.
>> >
>> > Also always disabling this setting is quite specific to linux and
>> > its subsystems, other operating systems might prefer other settings,
>> > so that the bootloader cannot really set a sane default for all.
>> >
>> > So introduce a top-level driver for the grf that handles these
>> > settings that need to be a certain way but nobody cares about.
>> >
>> > Other needed settings might surface in the future and can then
>> > be added here, but only as a last option. Ideally general GRF
>> > settings should be handled in the driver needing them.
>> >
>> > Signed-off-by: Heiko Stuebner <heiko at sntech.de>
>> > ---
>> >
>> > drivers/soc/rockchip/Kconfig | 10 ++++
>> > drivers/soc/rockchip/Makefile | 1 +
>> > drivers/soc/rockchip/grf.c | 134
>> > ++++++++++++++++++++++++++++++++++++++++++ 3 files changed, 145
>> > insertions(+)
>> > create mode 100644 drivers/soc/rockchip/grf.c
>> >
>> > diff --git a/drivers/soc/rockchip/Kconfig b/drivers/soc/rockchip/Kconfig
>> > index 7140ff8..20da55d 100644
>> > --- a/drivers/soc/rockchip/Kconfig
>> > +++ b/drivers/soc/rockchip/Kconfig
>> > @@ -3,6 +3,16 @@ if ARCH_ROCKCHIP || COMPILE_TEST
>> >
>> > #
>> > # Rockchip Soc drivers
>> > #
>> >
>> > +
>> > +config ROCKCHIP_GRF
>> > + bool
>> > + default y
>> > + help
>> > + The General Register Files are a central component providing
>> > + special additional settings registers for a lot of
>> > soc-components. + In a lot of cases there also need to be default
>> > settings initialized + to make some of them conform to
>> > expectations of the kernel. +
>> >
>> > config ROCKCHIP_PM_DOMAINS
>> >
>> > bool "Rockchip generic power domain"
>> > depends on PM
>> >
>> > diff --git a/drivers/soc/rockchip/Makefile b/drivers/soc/rockchip/Makefile
>> > index 3d73d06..c851fa0 100644
>> > --- a/drivers/soc/rockchip/Makefile
>> > +++ b/drivers/soc/rockchip/Makefile
>> > @@ -1,4 +1,5 @@
>> >
>> > #
>> > # Rockchip Soc drivers
>> > #
>> >
>> > +obj-$(CONFIG_ROCKCHIP_GRF) += grf.o
>> >
>> > obj-$(CONFIG_ROCKCHIP_PM_DOMAINS) += pm_domains.o
>> >
>> > diff --git a/drivers/soc/rockchip/grf.c b/drivers/soc/rockchip/grf.c
>> > new file mode 100644
>> > index 0000000..0c85476a
>> > --- /dev/null
>> > +++ b/drivers/soc/rockchip/grf.c
>> > @@ -0,0 +1,134 @@
>> > +/*
>> > + * Rockchip Generic Register Files setup
>> > + *
>> > + * Copyright (c) 2016 Heiko Stuebner <heiko at sntech.de>
>> > + *
>> > + * This program is free software; you can redistribute it and/or modify
>> > + * it under the terms of the GNU General Public License version 2 as
>> > + * published by the Free Software Foundation.
>> > + */
>> > +
>> > +#include <linux/err.h>
>> > +#include <linux/regmap.h>
>> > +#include <linux/mfd/syscon.h>
>> > +#include <linux/of_device.h>
>> > +#include <linux/platform_device.h>
>> > +#include <linux/regmap.h>
>> > +
>> > +#define HIWORD_UPDATE(val, mask, shift) \
>> > + ((val) << (shift) | (mask) << ((shift) + 16))
>> > +
>> > +struct rockchip_grf_value {
>> > + const char *desc;
>> > + u32 reg;
>> > + u32 val;
>> > +};
>> > +
>> > +struct rockchip_grf_info {
>> > + const struct rockchip_grf_value *values;
>> > + int num_values;
>> > +};
>> > +
>> > +#define RK3036_GRF_SOC_CON0 0x140
>> > +
>> > +static const struct rockchip_grf_value rk3036_defaults[] __initconst = {
>> > + /*
>> > + * Disable auto jtag/sdmmc switching that causes issues with the
>> > + * clock-framework and the mmc controllers making them unreliable.
>> > + */
>> > + { "jtag switching", RK3036_GRF_SOC_CON0, HIWORD_UPDATE(0, 1, 11)
>> > },
>> > +};
>> > +
>> > +static const struct rockchip_grf_info rk3036_grf __initconst = {
>> > + .values = rk3036_defaults,
>> > + .num_values = ARRAY_SIZE(rk3036_defaults),
>> > +};
>> > +
>> > +#define RK3288_GRF_SOC_CON0 0x244
>> > +
>> > +static const struct rockchip_grf_value rk3288_defaults[] __initconst = {
>> > + { "jtag switching", RK3288_GRF_SOC_CON0, HIWORD_UPDATE(0, 1, 12)
>> > },
>> > +};
>> > +
>> > +static const struct rockchip_grf_info rk3288_grf __initconst = {
>> > + .values = rk3288_defaults,
>> > + .num_values = ARRAY_SIZE(rk3288_defaults),
>> > +};
>> > +
>> > +#define RK3368_GRF_SOC_CON15 0x43c
>> > +
>> > +static const struct rockchip_grf_value rk3368_defaults[] __initconst = {
>> > + { "jtag switching", RK3368_GRF_SOC_CON15, HIWORD_UPDATE(0, 1, 13)
>> > }, +};
>> > +
>> > +static const struct rockchip_grf_info rk3368_grf __initconst = {
>> > + .values = rk3368_defaults,
>> > + .num_values = ARRAY_SIZE(rk3368_defaults),
>> > +};
>> > +
>> > +#define RK3399_GRF_SOC_CON7 0xe21c
>> > +
>> > +static const struct rockchip_grf_value rk3399_defaults[] __initconst = {
>> > + { "jtag switching", RK3399_GRF_SOC_CON7, HIWORD_UPDATE(0, 1, 12)
>> > },
>> > +};
>> > +
>> > +static const struct rockchip_grf_info rk3399_grf __initconst = {
>> > + .values = rk3399_defaults,
>> > + .num_values = ARRAY_SIZE(rk3399_defaults),
>> > +};
>> > +
>> > +static const struct of_device_id rockchip_grf_dt_match[] __initconst = {
>> > + {
>> > + .compatible = "rockchip,rk3036-grf",
>> > + .data = (void *)&rk3036_grf,
>> > + }, {
>> > + .compatible = "rockchip,rk3288-grf",
>> > + .data = (void *)&rk3288_grf,
>> > + }, {
>> > + .compatible = "rockchip,rk3368-grf",
>> > + .data = (void *)&rk3368_grf,
>> > + }, {
>> > + .compatible = "rockchip,rk3399-grf",
>> > + .data = (void *)&rk3399_grf,
>> > + },
>> > + { /* sentinel */ },
>> > +};
>>
>> I often come in after there's already been discussion on a topic, and
>> don't always find all of it, so let me know if this has already been
>> considered and not chosen for some reason:
>
> So far only Doug had the time to look at this, so I'm happy about discussing
> my approach more.
>
>
>> I get a little worried when you see these per-SoC tables build up in
>> the kernel. It means there'll need to be additions here for a bunch of
>> different SoCs.
>
> We also add clock drivers (which are essentially only tables listing the
> clock-hirarchy) on a per-soc basis. In clk-land the move was the other way
> around, because it was deemed that defining clocks in the devicetree does not
> scale :-) .
Yeah, I think the case of clocks is one where we've decided that it
makes sense to add a new clock driver per major SoC, too much changes
between each revision and doing it in DT (which some platforms have
attempted) isn't overall cleaner.
>> Would it be possible to describe this directly in the DT instead? If
>> nothing else, something like
>>
>> function-regs = < array of regs >
>> function-values = < array of values >
>> function-names = < array of names >
>>
>> Not ideal either, especially if abused into a random poke table, but
>> it won't require per-SoC changes to the kernel. If we keep the binding
>> under close eyes we can hopefully avoid abuse.
>>
>> Given that Rockchip is still working on new SoCs, this list will just
>> grow and doing it in-kernel will stop scaling at some point.
>
> As explained in the commit message, this is meant as a last resort for
> settings that no driver can be responsible for in a sensible way and that are
> Linux-specific and thus cannot be set in firmware for everybody. Most GRF-
> specific settings are already done by the specific drivers needing them.
>
> I.e. you see the mmc/jtag thing, which is getting disabled because the Linux
> mmc subsystem gets a hickup if the pinmux is changed by the soc to early after
> card eject. Other OSes might like that behaviour.
>
> As the dts is meant to be for everyone I don't see how that could be sanely
> defined for things.
But every OS that wants to use JTAG instead of MMC, or enable JTAG
around MMC, will need to know about this setting, right? So we're not
necessarily leaking linux implementation details into the DT here.
Whether each OS applies the setting is up to them.
> Another option would be to burden the rockchip mmc-driver with that specific
> setting, but then the mmc-driver (and possible other drivers if needed) would
> need per-soc additions, resulting in the same scaling problem.
> Having one per-soc list at least keeps in one place :-) .
That's exactly what I'm looking to avoid -- i.e. I don't want the GRF
driver to need patching for every SoC either, if we can avoid it.
Once the list gets too long, you'll want to add #ifdef SOCNAME and
only build in the tables for some of the SoCs. And then we've all lost
:-)
Or, if you're firmly deciding to keep updating kernel code for all
SoCs, could also just add one platform quirk file in
drivers/soc/rockchip with the postcore_initcall that matches toplevel
compatible per SoC, finds the device node, maps the address range,
sets the value.
That'll give you a place for other platform quirks, if they ever come
up. In one place, even if the other quirks might be in other register
ranges.
> If you really dislike the current approach, we can of course postpone it for
> now though.
It's a bit bikesheddy, but it just seems like adding a driver that
will always need updates is a bit of a heavy solution.
-Olof
More information about the Linux-rockchip
mailing list