MCIFIFOCNT register behaviour on Qcom chips is very different than the other
pl180 integrations. MCIFIFOCNT register contains the number of
words that are still waiting to be transferred through the FIFO. It keeps
decrementing once the host CPU reads the MCIFIFO. With the existing logic and
the MCIFIFOCNT behaviour, mmci_pio_read will loop forever, as the FIFOCNT
register will always return transfer size before reading the FIFO.

Also the data sheet states that "This register is only useful for debug
purposes and should not be used for normal operation since it does not reflect
data which may or may not be in the pipeline".

This patch implements a qcom specific get_rx_fifocnt function which is
implemented based on status register flags. Based on qcom_fifo flag in
variant data structure, the corresponding get_rx_fifocnt function is selected.
Signed-off-by: NSrinivas Kandagatla <srinivas.kandagatla@linaro.org>
Reviewed-by: NLinus Walleij <linus.walleij@linaro.org>
Signed-off-by: NUlf Hansson <ulf.hansson@linaro.org>

On Controllers like Qcom SD card controller where cclk is mclk and mclk should
be directly controlled by the driver.

This patch adds support to control mclk directly in the driver, and also
adds explicit_mclk_control flag in variant structure giving more flexibility
to the driver.
Signed-off-by: NSrinivas Kandagatla <srinivas.kandagatla@linaro.org>
Reviewed-by: NLinus Walleij <linus.walleij@linaro.org>
[Ulf Hansson] Fixed checkpatch warning
Signed-off-by: NUlf Hansson <ulf.hansson@linaro.org>

Some of the controller have maximum supported frequency, This patch adds
support in variant data structure to specify such restrictions. This
gives more flexibility in calculating the f_max before passing it to
mmc-core.
Signed-off-by: NSrinivas Kandagatla <srinivas.kandagatla@linaro.org>
Reviewed-by: NLinus Walleij <linus.walleij@linaro.org>
Signed-off-by: NUlf Hansson <ulf.hansson@linaro.org>

On some SOCs like Qcom there are explicit bits in the command register
to specify if its a data transfer command or not. So this patch adds
support to such bits in variant data, giving more flexibility to the
driver.
Signed-off-by: NSrinivas Kandagatla <srinivas.kandagatla@linaro.org>
Reviewed-by: NLinus Walleij <linus.walleij@linaro.org>
Signed-off-by: NUlf Hansson <ulf.hansson@linaro.org>

This patch adds edge support for data and command out to variant structure
giving more flexibility to the driver to support more SOCs which have
different clock register layout.

Without this patch other new SOCs like Qcom will have to add more code to
special case them
Signed-off-by: NSrinivas Kandagatla <srinivas.kandagatla@linaro.org>
Reviewed-by: NLinus Walleij <linus.walleij@linaro.org>
[Ulf Hansson] Resolved conflict
Signed-off-by: NUlf Hansson <ulf.hansson@linaro.org>

This patch adds 8bit bus enable to variant structure giving more flexibility
to the driver to support more SOCs which have different clock register layout.

Without this patch other new SOCs like Qcom will have to add more code
to special case them.
Signed-off-by: NSrinivas Kandagatla <srinivas.kandagatla@linaro.org>
Reviewed-by: NLinus Walleij <linus.walleij@linaro.org>
[Ulf Hansson] Resolved conflict
Signed-off-by: NUlf Hansson <ulf.hansson@linaro.org>

This patch adds ddrmode mask to variant structure giving more flexibility
to the driver to support more SOCs which have different datactrl register
layout.

Without this patch datactrl register is updated with incorrect ddrmode mask,
resulting in failures on Qualcomm SD Card Controller.
Signed-off-by: NSrinivas Kandagatla <srinivas.kandagatla@linaro.org>
Reviewed-by: NLinus Walleij <linus.walleij@linaro.org>
[Ulf Hansson] Resolved conflict
Signed-off-by: NUlf Hansson <ulf.hansson@linaro.org>

Instance of this IP on Qualcomm's SOCs has bit different layout for datactrl
register. Bit position datactrl[16:4] hold the true block size instead of power
of 2.
Signed-off-by: NSrinivas Kandagatla <srinivas.kandagatla@linaro.org>
Reviewed-by: NLinus Walleij <linus.walleij@linaro.org>
Signed-off-by: NUlf Hansson <ulf.hansson@linaro.org>

On Qcom SD Card controller POWER, CLKCTRL, DATACTRL and COMMAND registers
should be updated in MCLK domain, and writes to these registers must be
separated by three MCLK cycles. This resitriction is not applicable for
other registers. Any subsequent writes to these register will be ignored
until 3 MCLK have passed.

One usec delay between two CMD register writes is not sufficient in the
card identification phase where the CCLK is very low. This patch replaces
a static 1 usec delay to use mmci_reg_delay function which can provide
correct delay depending on the cclk frequency.

Without this patch the card is not detected.
Signed-off-by: NSrinivas Kandagatla <srinivas.kandagatla@linaro.org>
Reviewed-by: NLinus Walleij <linus.walleij@linaro.org>
Signed-off-by: NUlf Hansson <ulf.hansson@linaro.org>

This patch replaces a constant used in calculating timeout with a proper
macro. This is make code more readable.
Signed-off-by: NSrinivas Kandagatla <srinivas.kandagatla@linaro.org>
Reviewed-by: NLinus Walleij <linus.walleij@linaro.org>
Signed-off-by: NUlf Hansson <ulf.hansson@linaro.org>

Remove the option to provide DMA configuration as platform data,
enforce it through DT.
Signed-off-by: NUlf Hansson <ulf.hansson@linaro.org>
Cc: Russell King <linux@arm.linux.org.uk>
Cc: Roland Stigge <stigge@antcom.de>
Acked-by: NArnd Bergmann <arnd@arndb.de>

Remove the option to provide a maximum frequency as platform data,
enforce it through DT.
Signed-off-by: NUlf Hansson <ulf.hansson@linaro.org>

Remove the option to provide the flags for mmc capabilities as platform
data, enforce it through DT.
Signed-off-by: NUlf Hansson <ulf.hansson@linaro.org>

Remove the option to provide signal direction configuration and
feeback clock as platform data, enforce it through DT.
Signed-off-by: NUlf Hansson <ulf.hansson@linaro.org>

This is pure software configuration, which mmci has been supporting for
a while. Let's enable it as default so we can take benefit from it.
Signed-off-by: NUlf Hansson <ulf.hansson@linaro.org>

Let mmci DT parser only handle the specific bindings related to mmci
and extend the DT support by converting to the common mmc DT parser.

While both DT and platform data exist, DT takes precedence. If there
are supplied DT data, the card detect and write protect GPIOS are
enforced to be provided through it.
Signed-off-by: NUlf Hansson <ulf.hansson@linaro.org>

The ST Micro variant supports the option of using a feedback clock signal in
favor of the clockout pin when latching incoming signals on the data bus.

Since this is matter of how pins are being routed we need to provide a new DT
binding to be able to configure this through DT.
Signed-off-by: NUlf Hansson <ulf.hansson@linaro.org>
Reviewed-by: NLinus Walleij <linus.walleij@linaro.org>

Some variants have support for indicating the bus signal directions,
which currently are configured through platform data.

Add corresponding DT bindings to enable us to move away from using the
platform data.
Signed-off-by: NUlf Hansson <ulf.hansson@linaro.org>
Reviewed-by: NLinus Walleij <linus.walleij@linaro.org>

Converting to devm functions to simplify error handling in ->probe() and
to cleanup ->remove().
Signed-off-by: NUlf Hansson <ulf.hansson@linaro.org>

To avoid duplication of code while handling card detect and write
protect GPIO pins/irqs, let's convert to use the mmc gpio API.
Signed-off-by: NUlf Hansson <ulf.hansson@linaro.org>

For CONFIG_PM_SLEEP, the device were always left in full power state
after system suspend.

We solely relied on a power domain to put it into low power state,
which is an unreasonable requirement to put on SOCs to implement.
Especially for those SOCs not supporting power domains at all.

Use pm_runtime_force_suspend|resume() as the system suspend callbacks,
to resolve the issue.

Cc: Russell King <linux@arm.linux.org.uk>
Signed-off-by: NUlf Hansson <ulf.hansson@linaro.org>

Convert to the SET_PM_RUNTIME_PM macro while defining the runtime PM
callbacks. This means the callbacks becomes available for both
CONFIG_PM_SLEEP and CONFIG_PM_RUNTIME, which is needed to handle the
combinations of these scenarios.

Cc: Russell King <linux@arm.linux.org.uk>
Signed-off-by: NUlf Hansson <ulf.hansson@linaro.org>

In runtime suspended state, we are not expecting IRQs and thus we can
safely mask them, not only for pwrreg_nopower variants but for all.

Obviously we then also need to make sure we restore the IRQ mask while
becoming runtime resumed.

Cc: Russell King <linux@arm.linux.org.uk>
Signed-off-by: NUlf Hansson <ulf.hansson@linaro.org>

Added MMC_DDR52 as eMMC's DDR mode distinguished from SD-UHS.

CC: Russell King <linux@arm.linux.org.uk>
Signed-off-by: NSeungwon Jeon <tgih.jun@samsung.com>
Signed-off-by: NUlf Hansson <ulf.hansson@linaro.org>
Signed-off-by: NChris Ball <chris@printf.net>

The ux500 variants have HW busy detection support, which is indicated
by the busy_detect flag. For these variants let's enable the
MMC_CAP_WAIT_WHILE_BUSY flag and add the support for it.

The mmc core will provide the RSP_BUSY command flag for those requests
we should care about busy detection. Regarding the max_busy_timeout,
the HW don't support busy detection timeouts so at this initial step
let's make it simple and set it to zero to indicate we are able to
support any timeout.

Cc: Russell King <linux@arm.linux.org.uk>
Cc: Johan Rudholm <jrudholm@gmail.com>
Signed-off-by: NUlf Hansson <ulf.hansson@linaro.org>
Signed-off-by: NChris Ball <chris@printf.net>

In case of a read operation both MCI_CMDRESPEND and MCI_DATAEND can be
set in the status register when entering the interrupt handler. This is
due to that the card start sending data before the host has
acknowledged the command response.

To resolve the issue for this scenario, we must start by handling the
CMD irq instead of the DATA irq. The reason is beacuse the completion
of the DATA irq will not respect the current command and then causing
it to be garbled.

Cc: Russell King <linux@arm.linux.org.uk>
Cc: Johan Rudholm <jrudholm@gmail.com>
Signed-off-by: NUlf Hansson <ulf.hansson@linaro.org>
Signed-off-by: NChris Ball <chris@printf.net>

Driver core clears the driver data to NULL after device_release
or on probe failure, so just remove it from here.

Driver core change:
"device-core: Ensure drvdata = NULL when no driver is bound"
(sha1: 0998d063)
Signed-off-by: NMichal Simek <michal.simek@xilinx.com>
Signed-off-by: NRussell King <rmk+kernel@arm.linux.org.uk>

Suspend and resume of cards are being handled from the protocol layer
and consequently the mmc_suspend|resume_host APIs are deprecated.

This means we can simplify the suspend|resume callbacks by removing the
use of the deprecated APIs.

Cc: Russell King <linux@arm.linux.org.uk>
Signed-off-by: NUlf Hansson <ulf.hansson@linaro.org>
Signed-off-by: NChris Ball <cjb@laptop.org>

If a corresponding power domain exists for the device and it manages
to cut the domain regulator while the device is runtime suspended,
the IP loses it's registers context. We restore the context in the
.runtime_resume callback from the existing register caches to adapt
to this situation.

We also want to make sure the registers are in a known state while
restoring context in the case when the power domain did not drop the
power, since there are restrictions for the order of writing to these
registers. To handle this, we clear the registers in the
.runtime_suspend callback.
Signed-off-by: NUlf Hansson <ulf.hansson@linaro.org>
Acked-by: NRickard Andersson <rickard.andersson@stericsson.com>
Reviewed-by: NDaniel Lezcano <daniel.lezcano@linaro.org>
Signed-off-by: NRussell King <rmk+kernel@arm.linux.org.uk>

After a write to the MMCICLOCK register data cannot be written to this
register for three feedback clock cycles. Writes to the MMCIPOWER
register must be separated by three MCLK cycles. Previously no issues
has been observered, but using higher ARM clock frequencies on STE-
platforms has triggered this problem.

The MMCICLOCK register is written to in .set_ios and for some data
transmissions for SDIO. We do not need a delay at the data transmission
path, because sending and receiving data will require more than three
clock cycles. Then we use a simple logic to only delay in .set_ios and
thus we don't affect throughput performance.
Signed-off-by: NJohan Rudholm <jrudholm@gmail.com>
Signed-off-by: NUlf Hansson <ulf.hansson@linaro.org>
Acked-by: NRickard Andersson <rickard.andersson@stericsson.com>
Reviewed-by: NDaniel Lezcano <daniel.lezcano@linaro.org>
Signed-off-by: NRussell King <rmk+kernel@arm.linux.org.uk>

By optionally putting the pins into sleep state in the .runtime_suspend
callback we can accomplish two things. One is to minimize current leakage
from pins and thus save power, second we can prevent the IP from driving
pins output in an uncontrolled manner, which may happen if the power domain
drops the domain regulator.

When returning from idle, entering .runtime_resume callback, the pins
are restored to default state.
Signed-off-by: NUlf Hansson <ulf.hansson@linaro.org>
Acked-by: NRickard Andersson <rickard.andersson@stericsson.com>
Reviewed-by: NLinus Walleij <linus.walleij@linaro.org>
Signed-off-by: NRussell King <rmk+kernel@arm.linux.org.uk>

There is no need for every driver to fetch a pinctrl handle and to
select the default state. Instead this is handled by the device driver
core, thus we can remove this piece of code from mmci.
Signed-off-by: NUlf Hansson <ulf.hansson@linaro.org>
Reviewed-by: NLinus Walleij <linus.walleij@linaro.org>
Signed-off-by: NRussell King <rmk+kernel@arm.linux.org.uk>

To verify a signal voltage switch at initialization of UHS cards the
.card_busy callback is used. For some of the ST-variants, card busy
detection on the DAT0 pin is supported.

We extend the variant struct with a busy_detect flag to indicate
support for it. A corresponding busy detect function, which polls the
busy status bit, is then set to the .card_busy callback.
Signed-off-by: NUlf Hansson <ulf.hansson@linaro.org>
Signed-off-by: NRussell King <rmk+kernel@arm.linux.org.uk>

Add a cache variable in the host struct that reflects the current data in
the MMCIDATACTRL register. This patch will not introduce any functional
change but instead provide an easy option to keep specific bits in the
register between each data transfer.
Signed-off-by: NUlf Hansson <ulf.hansson@linaro.org>
Signed-off-by: NRussell King <rmk+kernel@arm.linux.org.uk>

Add .start_signal_voltage_switch callback to be able to support UHS cards.
The voltage switch requires the optional vqmmc regulator to exist since
the actual voltage switch will be performed directly on it.
Signed-off-by: NUlf Hansson <ulf.hansson@linaro.org>
Signed-off-by: NRussell King <rmk+kernel@arm.linux.org.uk>

We can not rely on regulator_is_enabled to decide whether to
enable|disable the regulator. It would mean that the reference
counter for it is not balanced properly.

Instead keep track of our internal state by using a new flag in
the host struct, so we can take correct decisions.
Signed-off-by: NUlf Hansson <ulf.hansson@linaro.org>
Signed-off-by: NRussell King <rmk+kernel@arm.linux.org.uk>

Currently, if DMA information isn't passed from platform data, then DMA
will not be used. This patch allows DMA information obtained though Device
Tree to be used as well.

Cc: Chris Ball <cjb@laptop.org>
Cc: linux-mmc@vger.kernel.org
Signed-off-by: NLee Jones <lee.jones@linaro.org>
Signed-off-by: NRussell King <rmk+kernel@arm.linux.org.uk>

Support added for transmission of CMD23 during multi block read or
write. In order to activate this feature, MMC_CAP_CMD23 flag needs
to be enabled in the capabilities field. Note that CMD23 support is
mandatory to support features like reliable write, data tag, context
ID, packed command.

This patch is based upon a patch from Saugata Das.
Signed-off-by: NUlf Hansson <ulf.hansson@linaro.org>
Acked-by: NLinus Walleij <linus.walleij@linaro.org>
Signed-off-by: NRussell King <rmk+kernel@arm.linux.org.uk>

Update cclk to the acutal used value and copy it to the actual_clock
variable in the mmc host for debug purpose.
Signed-off-by: NFredrik Soderstedt <fredrik.soderstedt@stericsson.com>
Signed-off-by: NUlf Hansson <ulf.hansson@linaro.org>
Signed-off-by: NRussell King <rmk+kernel@arm.linux.org.uk>

Converting to devm_clk_get simplifies error handling in probe
and we can remove other corresponding calls to clk_put.
Signed-off-by: NUlf Hansson <ulf.hansson@linaro.org>
Signed-off-by: NRussell King <rmk+kernel@arm.linux.org.uk>