commit d6f11e7d91f2ac85f66194fe3ef8789b49901d64 upstream.

The MMC device tree bindings include properties used to signal various
signalling speed modes. Until now the sunxi driver was accepting them
without any further filtering, while the sunxi device trees were not
actually using them.

Since some of the H5 boards can not run at higher speed modes stably,
we are resorting to declaring the higher speed modes per-board.

Regardless, having boards declare modes and blindly following them,
even without proper support in the driver, is generally a bad thing.

Filter out all unsupported modes from the capabilities mask after
the device tree properties have been parsed.

Cc: <stable@vger.kernel.org>
Signed-off-by: NChen-Yu Tsai <wens@csie.org>
Acked-by: NMaxime Ripard <maxime.ripard@bootlin.com>
Signed-off-by: NUlf Hansson <ulf.hansson@linaro.org>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

Recent Linux versions refuse to print actual virtual kernel addresses,
to not give a hint about the location of the kernel in a randomized virtual
address space. This affects the output of the sunxi MMC controller
driver, which now produces the rather uninformative line:

[    1.482660] sunxi-mmc 1c0f000.mmc: base:0x(____ptrval____) irq:8

Since the virtual base address is not really interesting in the first
place, let's just drop this value. The same applies to Linux' notion of
the interrupt number, which is independent from the GIC SPI number.
We have the physical address as part of the DT node name, which is way
more useful for debugging purposes.
To keep a success message in the driver, we make this purpose explicit
with the word "initialized", plus print some information that is not too
obvious and that we learned while probing the device:
the maximum request size and whether it uses the new timing mode.
So the output turns into:
[    1.750626] sunxi-mmc 1c0f000.mmc: initialized, max. request size: 16384 KB, uses new timings mode
[    1.786699] sunxi-mmc 1c11000.mmc: initialized, max. request size: 2048 KB
Signed-off-by: NAndre Przywara <andre.przywara@arm.com>
Acked-by: NMaxime Ripard <maxime.ripard@bootlin.com>
Signed-off-by: NUlf Hansson <ulf.hansson@linaro.org>

Some Allwinner boards feature an on-board eMMC with fixed 3.3V voltage
(e.g. Banana Pi M2+), and in this case both the eMMC and the SoC are
capable of doing 3.3V DDR transmission.

Add capability of 3.3V DDR when DDR is available (extra clock or new
timing).
Signed-off-by: NIcenowy Zheng <icenowy@aosc.io>
Acked-by: NMaxime Ripard <maxime.ripard@bootlin.com>
Signed-off-by: NUlf Hansson <ulf.hansson@linaro.org>

When mmc host controller enters suspend state, the clocks are
disabled, but irqs are not. For some reason the irqchip emits
false interrupts, which causes system lock loop.

Debug log is:
  ...
  sunxi-mmc 1c11000.mmc: setting clk to 52000000, rounded 51200000
  sunxi-mmc 1c11000.mmc: enabling the clock
  sunxi-mmc 1c11000.mmc: cmd 13(8000014d) arg 10000 ie 0x0000bbc6 len 0
  sunxi-mmc 1c11000.mmc: irq: rq (ptrval) mi 00000004 idi 00000000
  sunxi-mmc 1c11000.mmc: cmd 6(80000146) arg 3210101 ie 0x0000bbc6 len 0
  sunxi-mmc 1c11000.mmc: irq: rq (ptrval) mi 00000004 idi 00000000
  sunxi-mmc 1c11000.mmc: cmd 13(8000014d) arg 10000 ie 0x0000bbc6 len 0
  sunxi-mmc 1c11000.mmc: irq: rq (ptrval) mi 00000004 idi 00000000
  mmc1: new DDR MMC card at address 0001
  mmcblk1: mmc1:0001 AGND3R 14.6 GiB
  mmcblk1boot0: mmc1:0001 AGND3R partition 1 4.00 MiB
  mmcblk1boot1: mmc1:0001 AGND3R partition 2 4.00 MiB
  sunxi-mmc 1c11000.mmc: cmd 18(80003352) arg 0 ie 0x0000fbc2 len 409
  sunxi-mmc 1c11000.mmc: irq: rq (ptrval) mi 00004000 idi 00000002
   mmcblk1: p1
  sunxi-mmc 1c11000.mmc: irq: rq   (null) mi 00000000 idi 00000000
  sunxi-mmc 1c11000.mmc: irq: rq   (null) mi 00000000 idi 00000000
  sunxi-mmc 1c11000.mmc: irq: rq   (null) mi 00000000 idi 00000000
  sunxi-mmc 1c11000.mmc: irq: rq   (null) mi 00000000 idi 00000000
and so on...

This issue apears on eMMC cards, routed on MMC2 slot. The patch is
tested with A20-OLinuXino-MICRO/LIME/LIME2 boards.

Fixes: 9a8e1e8c ("mmc: sunxi: Add runtime_pm support")
Signed-off-by: NStefan Mavrodiev <stefan@olimex.com>
Acked-by: NMaxime Ripard <maxime.ripard@bootlin.com>
Signed-off-by: NUlf Hansson <ulf.hansson@linaro.org>

To be consistent with code in other mmc host drivers, convert to check the
correct PM config #ifdef in favor of using __maybe_unused.
Signed-off-by: NUlf Hansson <ulf.hansson@linaro.org>

The newly added runtime-pm functions cause a harmless warning
when CONFIG_PM is disabled:

drivers/mmc/host/sunxi-mmc.c:1452:12: error: 'sunxi_mmc_runtime_suspend' defined but not used [-Werror=unused-function]
 static int sunxi_mmc_runtime_suspend(struct device *dev)
            ^~~~~~~~~~~~~~~~~~~~~~~~~
drivers/mmc/host/sunxi-mmc.c:1435:12: error: 'sunxi_mmc_runtime_resume' defined but not used [-Werror=unused-function]
 static int sunxi_mmc_runtime_resume(struct device *dev)

This marks them as __maybe_unused to shut up the warning.

Fixes: 9a8e1e8c ("mmc: sunxi: Add runtime_pm support")
Signed-off-by: NArnd Bergmann <arnd@arndb.de>
Acked-by: NMaxime Ripard <maxime.ripard@bootlin.com>
Signed-off-by: NUlf Hansson <ulf.hansson@linaro.org>

Our set_ios hook is, when the card is power up or down, either doing a full
init or put our controller back into a reset mode.

Since we're also doing that in our runtime_pm hooks, and at possibly much
more often, we can drop it from the set_ios, and either rely on our
runtime_pm hooks or our probe to do it.
Signed-off-by: NMaxime Ripard <maxime.ripard@bootlin.com>
Signed-off-by: NUlf Hansson <ulf.hansson@linaro.org>

So far, even if our card was not in use, we didn't shut down our MMC
controller, which meant that it was still active and clocking the bus.

While this obviously means that we could save some power there, it also
creates issues when it comes to EMC control since we'll have a perfect peak
at the card clock rate.

Let's implement runtime_pm with autosuspend so that we will shut down the
controller when it's not been in use for quite some time.
Signed-off-by: NMaxime Ripard <maxime.ripard@bootlin.com>
Signed-off-by: NUlf Hansson <ulf.hansson@linaro.org>

In order to improve readibility and reusability, let's move the card setup
to a small function called by our .set_ios hook.
Signed-off-by: NMaxime Ripard <maxime.ripard@bootlin.com>
Signed-off-by: NUlf Hansson <ulf.hansson@linaro.org>

In order to improve readibility and reusability, let's move the clock setup
to a small function called by our .set_ios hook.
Signed-off-by: NMaxime Ripard <maxime.ripard@bootlin.com>
Signed-off-by: NUlf Hansson <ulf.hansson@linaro.org>

In order to improve readibility and reusability, let's move the bus width
setup to a small function called by our .set_ios hook.
Signed-off-by: NMaxime Ripard <maxime.ripard@bootlin.com>
Signed-off-by: NUlf Hansson <ulf.hansson@linaro.org>

All the other functions in the driver take a struct sunxi_mmc_host pointer.
Let's make it consistent.
Signed-off-by: NMaxime Ripard <maxime.ripard@bootlin.com>
Signed-off-by: NUlf Hansson <ulf.hansson@linaro.org>

Our headers sort algorithm has had pretty chaotic results. Let's fix that.
Signed-off-by: NMaxime Ripard <maxime.ripard@bootlin.com>
Signed-off-by: NUlf Hansson <ulf.hansson@linaro.org>

As soon as the pm_runtime_enable hook is called, our runtime_suspend and
runtime_resume hooks can be called as well. However, we only set the device
drvdata that we will use after we have registered into the MMC core. Move
that earlier so that we don't have a race that could lead to a crash.
Signed-off-by: NMaxime Ripard <maxime.ripard@bootlin.com>
Signed-off-by: NUlf Hansson <ulf.hansson@linaro.org>

As per Allwinner guidelines, the reset line should be deasserted before
turning the clocks on.

Implement it in our driver as well.
Signed-off-by: NMaxime Ripard <maxime.ripard@bootlin.com>
Signed-off-by: NUlf Hansson <ulf.hansson@linaro.org>

We've had all our resources management, and especially the clocks and reset
sequence, done directly as part of the probe.

As we want to implement runtime_pm, we'll obviously want to have that
moved outside of the probe so that we can call do it in our runtime suspend
and resume hooks without too much duplication.
Signed-off-by: NMaxime Ripard <maxime.ripard@bootlin.com>
Signed-off-by: NUlf Hansson <ulf.hansson@linaro.org>

platform_get_irq() can fail here and we must check its return value.
Signed-off-by: NArvind Yadav <arvind.yadav.cs@gmail.com>
Signed-off-by: NUlf Hansson <ulf.hansson@linaro.org>

It had an U+FFFD: not a corrupted character but a literal well-formed
replacement marker.
Signed-off-by: NAdam Borowski <kilobyte@angband.pl>
Acked-by: NMaxime Ripard <maxime.ripard@free-electrons.com>
Signed-off-by: NUlf Hansson <ulf.hansson@linaro.org>

This error message can go because a) currently nothing else than
EPROBE_DEFER is returned and b) if this is going to change a much more
detailed error message should come from mmc_regulator_get_supply()
anyhow.
Signed-off-by: NWolfram Sang <wsa+renesas@sang-engineering.com>
Signed-off-by: NUlf Hansson <ulf.hansson@linaro.org>

We might be into some troubles if the bootloader misconfigured the MMC
controller.

We currently only de-assert the reset line at probe time, which means that
if the device was already out of reset, we're going to keep whatever state
was set already.

Switch to a reset instead of the deassert to have a device in a pristine
state when we start operating.
Signed-off-by: NMaxime Ripard <maxime.ripard@free-electrons.com>
Signed-off-by: NUlf Hansson <ulf.hansson@linaro.org>

sunxi_mmc_clk_set_phase expects the actual card clock rate to be passed
to it. When the internal divider code was reworked in change ("mmc: sunxi:
Support MMC DDR52 transfer mode with new timing mode"), this requirement
was missed, and the module clock rate was passed in instead. This broke 8
bit DDR MMC on old controllers, as the module clock rate is double the
card clock rate, for which we have no valid delay settings.

Fix this by applying the internal divider to the clock rate right after
we configure it in hardware.
Signed-off-by: NChen-Yu Tsai <wens@csie.org>
Signed-off-by: NUlf Hansson <ulf.hansson@linaro.org>

The A83T MMC support code introduces the timings mode switch, however
such a switch doesn't exist on new SoCs with only new timings mode.

Only execute the switch if the SoC really have the timings mode switch,
to fix the regression shown on new timings mode only SoCs (A64, H5,
etc).
Signed-off-by: NIcenowy Zheng <icenowy@aosc.io>
Reviewed-by: NChen-Yu Tsai <wens@csie.org>
Signed-off-by: NUlf Hansson <ulf.hansson@linaro.org>

Some SoCs do not support clk delays for MMC in the clock control unit.
These include the old controllers in A10/A10s/A13/R8, and the new eMMC
controller in A64. The config structure for these controllers do not
specify clk_delays, but the check for this was replaced in change
"mmc: sunxi: Support controllers that can use both old and new timings".

This patch adds back the check for clk_delays, and also adds comments
for both checks in sunxi_mmc_clk_set_phase().
Signed-off-by: NChen-Yu Tsai <wens@csie.org>
Tested-by: NIcenowy Zheng <icenowy@aosc.io>
Signed-off-by: NUlf Hansson <ulf.hansson@linaro.org>

The third MMC controller (MMC2) on the Allwinner A83T SoC is slightly
different. It supports a wider 8-bit bus, has a dedicated controllable
reset pin for eMMC, and a "new timing mode" which is supposed to deliver
better signals and thus better performance.

Add a compatible for this one to use the new timing mode not found in the
other controllers.
Signed-off-by: NChen-Yu Tsai <wens@csie.org>
Acked-by: NRob Herring <robh@kernel.org>
Acked-by: NMaxime Ripard <maxime.ripard@free-electrons.com>
Signed-off-by: NUlf Hansson <ulf.hansson@linaro.org>

The MMC controller can support DDR52 transfers under the new timing
mode. According to the BSP kernel, the module clock has to be double
the card clock, regardless of the bus width. The default timings in
the hardware can be used.

This also reworks the code setting the internal divider, getting rid
of a extra conditional.
Signed-off-by: NChen-Yu Tsai <wens@csie.org>
Acked-by: NMaxime Ripard <maxime.ripard@free-electrons.com>
Signed-off-by: NUlf Hansson <ulf.hansson@linaro.org>

On the SoCs that introduced the new timing mode for MMC controllers,
both the old (where the clock delays are set in the CCU) and new
(where the clock delays are set in the MMC controller) timing modes
are available, and we have to support them both. However there are
two bits that control which mode is active. One is in the CCU, the
other is in the MMC controller. The settings on both sides must be
the same, or nothing will work.

The sunxi-ng clock driver provides an API to query and set the
active timing mode. At probe time, we try to set the active mode
to the "new timing mode". If it succeeds, we can then use the MMC
controller in the new mode. If not, we fall back to the old mode.
Signed-off-by: NChen-Yu Tsai <wens@csie.org>
Acked-by: NMaxime Ripard <maxime.ripard@free-electrons.com>
Signed-off-by: NUlf Hansson <ulf.hansson@linaro.org>

The mmc_host_ops structure is only stored in the ops field of an
mmc_host structure, which is declared as const.  Thus the mmc_host_ops
structure itself can be const.

Done with the help of Coccinelle.

// <smpl>
@r disable optional_qualifier@
identifier i;
position p;
@@
static struct mmc_host_ops i@p = { ... };

@ok1@
struct mmc_host *mmc;
identifier r.i;
position p;
@@
mmc->ops = &i@p

@bad@
position p != {r.p,ok1.p};
identifier r.i;
struct mmc_host_ops e;
@@
e@i@p

@depends on !bad disable optional_qualifier@
identifier r.i;
@@
static
+const
 struct mmc_host_ops i = { ... };
// </smpl>
Signed-off-by: NJulia Lawall <Julia.Lawall@lip6.fr>
Signed-off-by: NUlf Hansson <ulf.hansson@linaro.org>

Commit a53e35db ("reset: Ensure drivers are explicit when requesting
reset lines") started to transition the reset control request API calls
to explicitly state whether the driver needs exclusive or shared reset
control behavior. Convert all drivers requesting exclusive resets to the
explicit API call so the temporary transition helpers can be removed.

No functional changes.

Cc: Ulf Hansson <ulf.hansson@linaro.org>
Cc: Maxime Ripard <maxime.ripard@free-electrons.com>
Cc: Chen-Yu Tsai <wens@csie.org>
Cc: linux-mmc@vger.kernel.org
Signed-off-by: NPhilipp Zabel <p.zabel@pengutronix.de>
Signed-off-by: NUlf Hansson <ulf.hansson@linaro.org>

The register for the "new timing mode" also has bit fields for setting
output and sample timing phases. According to comments in Allwinner's
BSP kernel, the default values are good enough.

Keep the default values already in the hardware when setting new timing
mode, instead of overwriting the whole register.

Fixes: 9a37e53e ("mmc: sunxi: Enable the new timings for the A64 MMC
controllers")
Signed-off-by: NChen-Yu Tsai <wens@csie.org>
Acked-by: NMaxime Ripard <maxime.ripard@free-electrons.com>
Cc: <stable@vger.kernel.org>
Signed-off-by: NUlf Hansson <ulf.hansson@linaro.org>

Use new core function mmc_get_dma_dir().
Signed-off-by: NHeiner Kallweit <hkallweit1@gmail.com>
Signed-off-by: NUlf Hansson <ulf.hansson@linaro.org>

The controller's errors are usually normal (for example, for MMC or SDIO
cards, some errors are expected to happen; and for boards without a
dedicated card detect pin the error info will even flood console and
hide other normal messages) and hard to understand.

Change their print level to debug, thus it won't be shown to generic
users.
Signed-off-by: NIcenowy Zheng <icenowy@aosc.xyz>
Acked-by: NMaxime Ripard <maxime.ripard@free-electrons.com>
Signed-off-by: NUlf Hansson <ulf.hansson@linaro.org>

Add a bit more debug messages that can be helpful when debugging the clock
setup.

Also fill the actual_clock field in struct mmc_host to report properly the
current frequency in debugfs.
Signed-off-by: NMaxime Ripard <maxime.ripard@free-electrons.com>
Tested-by: NFlorian Vaussard <florian.vaussard@heig-vd.ch>
Acked-by: NChen-Yu Tsai <wens@csie.org>
Signed-off-by: NUlf Hansson <ulf.hansson@linaro.org>

The MMC2 controller on the A64 is kind of a special beast.

While the general controller design is the same than the other MMC
controllers in the SoC, it also has a bunch of features and changes that
prevent it to be driven in the same way.

It has for example a different bus width limit, a different maximum
frequency, and, for some reason, the maximum buffer size of a DMA
descriptor.

Add a new compatible specifically for this controller.
Acked-by: NRob Herring <robh@kernel.org>
Signed-off-by: NMaxime Ripard <maxime.ripard@free-electrons.com>
Tested-by: NFlorian Vaussard <florian.vaussard@heig-vd.ch>
Acked-by: NChen-Yu Tsai <wens@csie.org>
Signed-off-by: NUlf Hansson <ulf.hansson@linaro.org>

The A64 MMC controllers need DATA0 to be masked while updating the clock,
otherwise any subsequent command will result in a timeout.

It's not really clear at this point what DATA0 is exactly, but this
behaviour is present in Allwinner's tree, and has been suggested by
Allwinner engineers as fixes for the timeout.
Signed-off-by: NMaxime Ripard <maxime.ripard@free-electrons.com>
Tested-by: NFlorian Vaussard <florian.vaussard@heig-vd.ch>
Acked-by: NChen-Yu Tsai <wens@csie.org>
Signed-off-by: NUlf Hansson <ulf.hansson@linaro.org>

The A64 MMC controllers need to set a "new timings" bit when a new rate is
set.

The actual meaning of that bit is not clear yet, but not setting it leads
to some corner-case issues, like the CMD53 failing, which is used to
implement SDIO packet aggregation.
Signed-off-by: NMaxime Ripard <maxime.ripard@free-electrons.com>
Tested-by: NFlorian Vaussard <florian.vaussard@heig-vd.ch>
Acked-by: NChen-Yu Tsai <wens@csie.org>
Signed-off-by: NUlf Hansson <ulf.hansson@linaro.org>

Experience have shown that the using the  autocalibration could severely
degrade the performances of the MMC bus.

Allwinner is using in its BSP a delay set to 0 for all the modes but HS400.
Remove the calibration code for now, and add comments to document our
findings.
Reviewed-by: NAndre Przywara <andre.przywara@arm.com>
Signed-off-by: NMaxime Ripard <maxime.ripard@free-electrons.com>
Tested-by: NFlorian Vaussard <florian.vaussard@heig-vd.ch>
Acked-by: NChen-Yu Tsai <wens@csie.org>
Signed-off-by: NUlf Hansson <ulf.hansson@linaro.org>

The MMC core assumes that the code will gate the clock when the bus
frequency is set to 0, which we've been ignoring so far.

Handle that.
Signed-off-by: NMaxime Ripard <maxime.ripard@free-electrons.com>
Tested-by: NFlorian Vaussard <florian.vaussard@heig-vd.ch>
Acked-by: NChen-Yu Tsai <wens@csie.org>
Signed-off-by: NUlf Hansson <ulf.hansson@linaro.org>

The SD specification documents that the clock frequency should only be
changed once gated (Section 3.2.3 - SD Clock Frequency Change Sequence).

The current code first modifies the parent clock, gates it and then
modifies the internal divider. This means that since the parent clock rate
might be changed, the bus clock might be changed as well before it is
gated, which breaks the specification.

Move the gating before the parent rate modification.
Signed-off-by: NMaxime Ripard <maxime.ripard@free-electrons.com>
Tested-by: NFlorian Vaussard <florian.vaussard@heig-vd.ch>
Acked-by: NChen-Yu Tsai <wens@csie.org>
Signed-off-by: NUlf Hansson <ulf.hansson@linaro.org>

VMMC is an optional regulator, which means that mmc_regulator_get_supply
will only return an error in case of a deferred probe, but not when the
regulator is not set in the DT.

However, the sunxi driver assumes that VMMC is always there, and doesn't
check the value of the regulator pointer before using it, which obviously
leads to a (close to) null pointer dereference.

Add proper checks to prevent that.
Signed-off-by: NMaxime Ripard <maxime.ripard@free-electrons.com>
Acked-by: NChen-Yu Tsai <wens@csie.org>
Signed-off-by: NUlf Hansson <ulf.hansson@linaro.org>

A64 SoC features a MMC controller which need only the mod clock, and can
calibrate delay by itself. This patch adds support for the new MMC
controller IP core.
Signed-off-by: NIcenowy Zheng <icenowy@aosc.xyz>
Signed-off-by: NUlf Hansson <ulf.hansson@linaro.org>