Perform clock disable/enable in runtime suspend/resume.

Signed-off-by: Chander Kashyap <chander.kashyap@linaro.org>
Acked-by: Jaehoon Chung <jh80.chung@samsung.com>
Signed-off-by: Chris Ball <cjb@laptop.org>

Before this patch, we were using MMC_CAP2_BROKEN_VOLTAGE as a way to
avoid calling regulator_set_voltage() on a fixed regulator, but that's
just duplicating information that already exists -- we should test
whether the regulator is fixed directly, instead of via a capability.

This patch implements that test.  We can't reclaim the capability bit
just yet, since there are still boards in arch/arm/ that reference it;
those references can be removed now.

Reported-by: Tomasz Figa <t.figa@samsung.com>
Signed-off-by: Chris Ball <cjb@laptop.org>

Signed-off-by: Chris Ball <cjb@laptop.org>

Signed-off-by: Chris Ball <cjb@laptop.org>

We can use up to four bus-clocks; but on module remove, we didn't
disable the fourth bus clock.

Signed-off-by: Jaehoon Chung <jh80.chung@samsung.com>
Signed-off-by: Kyungmin Park <kyungmin.park@samsung.com>
Cc: stable <stable@vger.kernel.org>
Signed-off-by: Chris Ball <cjb@laptop.org>

On some systems, e.g., kzm9g, MMCIF interfaces can produce spurious
interrupts without any active request. To prevent the Oops, that results
in such cases, don't dereference the mmc request pointer until we make
sure, that we are indeed processing such a request.

Reported-by: Tetsuyuki Kobayashi <koba@kmckk.co.jp>
Signed-off-by: Guennadi Liakhovetski <g.liakhovetski@gmx.de>
Cc: stable <stable@vger.kernel.org>
Signed-off-by: Chris Ball <cjb@laptop.org>

Upon completion of a MMC_WRITE_MULTIPLE_BLOCK command MMCIF issues an IRQ
with the DTRANE bit set and often with one or several of CMD12 bits set.
If those interrupts are not acknowledged, an additional interrupt can be
produced and will be delivered later, possibly, when the transaction has
already been completed. To prevent this from happening, CMD12 completion
interrupt sources have to be cleared too upon reception of an DTRANE IRQ.

Signed-off-by: Guennadi Liakhovetski <g.liakhovetski@gmx.de>
Tested-by: Tetsuyuki Kobayashi <koba@kmckk.co.jp>
Signed-off-by: Chris Ball <cjb@laptop.org>

If sdhci-s3c driver is built as module, it gives following error if inserted
again after removing. This was happening as pm_runtime_use_autosuspend() is
called in sdhci_s3c_probe() function but in sdhci_s3c_remove() its
complementary pm_runtime_dont_use_autosuspend() is not called.

BUG: spinlock bad magic on CPU#1, insmod/955
 lock: 0xee771368, .magic: 00000000, .owner: insmod/955, .owner_cpu: 1
[<c00147e0>] (unwind_backtrace+0x0/0xf8) from [<c0136b40>] (do_raw_spin_unlock+0xa4/0xe4)
[<c0136b40>] (do_raw_spin_unlock+0xa4/0xe4) from [<c01be508>] (_raw_spin_unlock_irqrestore+0xc/0x38)
[<c01be508>] (_raw_spin_unlock_irqrestore+0xc/0x38) from [<c01a9334>] (sdhci_runtime_suspend_host+0x54/0x80)
[<c01a9334>] (sdhci_runtime_suspend_host+0x54/0x80) from [<bf0060a8>] (sdhci_s3c_runtime_suspend+0x14/0x38 [sdhci_s3c])
[<bf0060a8>] (sdhci_s3c_runtime_suspend+0x14/0x38 [sdhci_s3c]) from [<c016cb00>] (pm_generic_runtime_suspend+0x2c/0x40)
[<c016cb00>] (pm_generic_runtime_suspend+0x2c/0x40) from [<c0170090>] (__rpm_callback+0x70/0x98)
[<c0170090>] (__rpm_callback+0x70/0x98) from [<c01703f0>] (rpm_suspend+0xf0/0x534)
[<c01703f0>] (rpm_suspend+0xf0/0x534) from [<c0171670>] (__pm_runtime_suspend+0x5c/0x74)
[<c0171670>] (__pm_runtime_suspend+0x5c/0x74) from [<c016d018>] (pm_generic_runtime_idle+0x44/0x4c)
[<c016d018>] (pm_generic_runtime_idle+0x44/0x4c) from [<c0170090>] (__rpm_callback+0x70/0x98)
[<c0170090>] (__rpm_callback+0x70/0x98) from [<c0170984>] (rpm_idle+0xdc/0x18c)
[<c0170984>] (rpm_idle+0xdc/0x18c) from [<c0171608>] (pm_runtime_set_autosuspend_delay+0x30/0x3c)
[<c0171608>] (pm_runtime_set_autosuspend_delay+0x30/0x3c) from [<bf0069c4>] (sdhci_s3c_probe+0x35c/0x52c [sdhci_s3c])
[<bf0069c4>] (sdhci_s3c_probe+0x35c/0x52c [sdhci_s3c]) from [<c016a014>] (platform_drv_probe+0x18/0x1c)

Signed-off-by: Chander Kashyap <chander.kashyap@linaro.org>
Acked-by: Jaehoon Chung <jh80.chung@samsung.com>
Signed-off-by: Chris Ball <cjb@laptop.org>

In case of multiple bus clock sources, all the clock sources were
getting enabled. As only one clock source is needed at the time hence
enable only the required bus clock.

This patch does as follows:
1.	In sdhci_s3c_probe enable only required bus clock source.

2.	Handle the disabling of old bus clock and enables the
	best clock selected in sdhci_s3c_set_clock().

Signed-off-by: Chander Kashyap <chander.kashyap@linaro.org>
Acked-by: Jaehoon Chung <jh80.chung@samsung.com>
Signed-off-by: Chris Ball <cjb@laptop.org>

This reverts commit 94c6cee9

 (Add check for IDMAC configuration).
Synopsys says that only if internal dmac is not present, optional
external dma interface is present. When internal dmac is present,
'0' value in DMA_INTERFACE of HCON is reasonable. DMA_INTERFACE
indicates external dma interface. And idmac initialization is
prohibited now.

The commit should be reverted since: the check for IDMAC is not
reliable; falling back to PIO would provide awful performance; we
wouldn't expect to see instances of this block without DMA support.

Signed-off-by: Seungwon Jeon <tgih.jun@samsung.com>
Acked-by: Will Newton <will.newton@imgtec.com>
Signed-off-by: Chris Ball <cjb@laptop.org>

The problem can be easily reproduced using a script that loops
copying a file in an SD card to another place in the same SD card
and its related to read transfers. This only happens with DMA enabled.

This is related to the fact that, when reading, an MMC irq signals
the fact that all data from the SD card has been copied to the
internal buffers. However, it doesn't signal whether the DMA transfer
that is in charge of moving data from these internal buffers to RAM
has finished or not. Thus, calling dmaengine_terminate_all() in the
MMC irq routine can cancel an ongoing DMA transfer leaving some data
in the internal buffers that produces an accumulative effect which,
in the end, blocks a read data transfer forever.

The following patch watches DMA irq for reading and MMC irqs for
writing transfers. The 'dangerous' usage of dmaengine_terminate_all()
is removed and a timeout of 10 seconds is added so that the MMC won't
block forever anymore.

Signed-off-by: Javier Martin <javier.martin@vista-silicon.com>
Reviewed-by: Sascha Hauer <s.hauer@pengutronix.de>
Signed-off-by: Chris Ball <cjb@laptop.org>

In some cases mmc_suspend_host() is not able to claim the
host and proceed with the suspend process. The core returns
-EBUSY to the host controller driver. Unfortunately, the
host controller driver does not pass on this information
to the PM core and hence the system suspend process continues.

	ret = mmc_suspend_host(host->mmc);
	if (ret) {
		host->suspended = 0;
		if (host->pdata->resume) {
			ret = host->pdata->resume(dev, host->slot_id);

The return status from mmc_suspend_host() is overwritten by return
status from host->pdata->resume. So the original return status is lost.

In these cases the MMC core gets to an unexpected state
during resume and multiple issues related to MMC crop up.
1. Host controller driver starts accessing the device registers
before the clocks are enabled which leads to a prefetch abort.
2. A file copy thread which was launched before suspend gets
stuck due to the host not being reclaimed during resume.

To avoid such problems pass on the -EBUSY status to the PM core
from the host controller driver. With this change, MMC core
suspend might still fail but it does not end up making the
system unusable. Suspend gets aborted and the user can try
suspending the system again.

Signed-off-by: Vaibhav Bedia <vaibhav.bedia@ti.com>
Signed-off-by: Hebbar, Gururaja <gururaja.hebbar@ti.com>
Acked-by: Venkatraman S <svenkatr@ti.com>
Cc: stable <stable@vger.kernel.org>
Signed-off-by: Chris Ball <cjb@laptop.org>

Earlier, atmel-mci was adapted to make use of the peripheral DMA
controller (PDC), in case normal DMA wouldn't work.
( http://comments.gmane.org/gmane.linux.kernel.mmc/9403

 )

This works OK on ARM platforms (AT91), but it broke the driver
for AVR32, the AP700x.  Although the MCI has PDC support, the
connection is not done for AVR chips.

This patch makes the use of PDC depend on !CONFIG_AVR32.

Signed-off-by: Hein Tibosch <hein_tibosch@yahoo.es>
Signed-off-by: Chris Ball <cjb@laptop.org>

After the latest changes to atmel-mci, it could not be used with
DMA on the AVR32 platform. This patch will allow to use DMA again
and it will avoid access to MCI register ATMCI_DMA.

Even if the IP version is lower than v3xx and doesn't have the DMA
configuration register, DMA transfers can be used with a different
controller than the Atmel AHB DMA one. For instance, some AVR chips
use the Synopsys DesignWare AHB DMA controller.

Signed-off-by: Hein Tibosch <hein_tibosch@yahoo.es>
Acked-by: Ludovic Desroches <ludovic.desroches@atmel.com>
Signed-off-by: Chris Ball <cjb@laptop.org>

clk_prepare_enable and clk_disable_unprepare combine clk_prepare and
clk_enable, and clk_disable and clk_unprepare.  They make the code more
concise, and ensure that clk_unprepare is called when clk_enable fails.

A simplified version of the semantic patch that introduces calls to these
functions is as follows: (http://coccinelle.lip6.fr/

)

// <smpl>
@@
expression e;
@@

- clk_prepare(e);
- clk_enable(e);
+ clk_prepare_enable(e);

@@
expression e;
@@

- clk_disable(e);
- clk_unprepare(e);
+ clk_disable_unprepare(e);
// </smpl>

Signed-off-by: Julia Lawall <Julia.Lawall@lip6.fr>
Signed-off-by: Chris Ball <cjb@laptop.org>

Add device tree based discovery support for Samsung's sdhci controller.

Cc: Ben Dooks <ben-linux@fluff.org>
Cc: Kukjin Kim <kgene.kim@samsung.com>
Signed-off-by: Thomas Abraham <thomas.abraham@linaro.org>
Reviewed-by: Jaehoon Chung <jh80.chung@samsung.com>
Signed-off-by: Chris Ball <cjb@laptop.org>

Samsung Exynos SoC's extend the dw-mshc controller for additional clock
and bus control. Add support for these extensions and include provide
device tree based discovery suppory as well.

Signed-off-by: Thomas Abraham <thomas.abraham@linaro.org>
Acked-by: Will Newton <will.newton@imgtec.com>
Signed-off-by: Chris Ball <cjb@laptop.org>

The core dw-mshc controller driver can let platform specific
implementations of the dw-mshc controller to control the hardware
as required by such implementations. This is acheived by invoking
implementation specific (optional) callbacks. Define the list of
callbacks supported the add invocation points for the same.

Signed-off-by: Thomas Abraham <thomas.abraham@linaro.org>
Acked-by: Will Newton <will.newton@imgtec.com>
Signed-off-by: Chris Ball <cjb@laptop.org>

Platform implementations of dw-mshc controller may choose to extend the
features of the standard dw-mshc controller such as adding additional
clocking options or modifying the bus interface. Support for such
implementation specific extensions can be incorporated into dw_mmc-pltfm,
but including multiple such platform specific implementations would
convolute the existing dw_mmc-pltfm code.

Instead, it would be better to create implementation specific platform
drivers to support implementation specific features. Such platforms
drivers can reuse the existing infrastructure in dw_mmc-pltfm for
resource identification and controller registration and provide support
for implementation specific features. So, allow the infrastructure in
dw_mmc-pltfm to be reused by other implementation specific platform
drivers.

Signed-off-by: Thomas Abraham <thomas.abraham@linaro.org>
Acked-by: Will Newton <will.newton@imgtec.com>
Signed-off-by: Chris Ball <cjb@laptop.org>

Add device tree based discovery support.

Signed-off-by: Thomas Abraham <thomas.abraham@linaro.org>
Acked-by: Will Newton <will.newton@imgtec.com>
Signed-off-by: Chris Ball <cjb@laptop.org>

If the write protect pad of the controller is not connected to the write
protect pin of the slot, the driver should be notified of this condition
so that incorrect check for write protection by reading the WRTORT
register can avoided. The get_ro platform callback can be used for in
such cases, but with device tree support enabled, such platform callbacks
cannot be supported.

Add a new quirk for notifying the driver about the missing write protect
line so the driver can assume that the card write protection is disabled.

Signed-off-by: Thomas Abraham <thomas.abraham@linaro.org>
Acked-by: Will Newton <will.newton@imgtec.com>
Signed-off-by: Chris Ball <cjb@laptop.org>

Some platforms allow for clock gating and control of bus interface unit
clock and card interface unit clock. Add support for clock lookup of
optional biu and ciu clocks for clock gating and clock speed
determination.

Signed-off-by: Abhilash Kesavan <a.kesavan@samsung.com>
Signed-off-by: Thomas Abraham <thomas.abraham@linaro.org>
Acked-by: Will Newton <will.newton@imgtec.com>
Signed-off-by: Chris Ball <cjb@laptop.org>

Instead of aborting the probe when a slot initialization fails, allow
initialization of as many slots as possible. If there is at least one
instance of a slot that is successfully initialized, allow the driver
probe to succeed.

Signed-off-by: Thomas Abraham <thomas.abraham@linaro.org>
Acked-by: Will Newton <will.newton@imgtec.com>
Signed-off-by: Chris Ball <cjb@laptop.org>

Use devm_* managed functions for simpler error handling.

Signed-off-by: Thomas Abraham <thomas.abraham@linaro.org>
Acked-by: Will Newton <will.newton@imgtec.com>
Signed-off-by: Chris Ball <cjb@laptop.org>

The 'struct dw_mci' maintains a copy of the pdev->dev instance instead of
maintaining a reference to that 'struct device' instance. Any resource
allocated using the device resource management kernel API with the instance
of 'struct device' in 'struct dw_mci' is then incorrect. Fix this by
converting the copy of 'struct device' in 'struct dw_mci' to a reference.

Signed-off-by: Thomas Abraham <thomas.abraham@linaro.org>
Acked-by: Will Newton <will.newton@imgtec.com>
Signed-off-by: Chris Ball <cjb@laptop.org>

Enable eMMC background operations (BKOPS) feature.

If URGENT_BKOPS is set after a response, note that BKOPS are required.
Immediately run BKOPS if required.  Read/write operations should be
requested during BKOPS(LEVEL-1), then issue HPI to interrupt the
ongoing BKOPS and service the foreground operation.
(This patch only controls the LEVEL2/3.)

When repeating the writing 1GB data, at a certain time, performance is
decreased.  At that time, card triggers the Level-3 or Level-2.  After
running bkops, performance is recovered.

Future considerations:
 * Check BKOPS_LEVEL=1 and start BKOPS in a preventive manner.
 * Interrupt ongoing BKOPS before powering off the card.
 * How do we get BKOPS_STATUS value (periodically send ext_csd command)?
 * If using periodic bkops, also consider runtime_pm control.

Signed-off-by: Jaehoon Chung <jh80.chung@samsung.com>
Signed-off-by: Kyungmin Park <kyungmin.park@samsung.com>
Signed-off-by: Konstantin Dorfman <kdorfman@codeaurora.org>
Reviewed-by: Maya Erez <merez@codeaurora.org>
Signed-off-by: Chris Ball <cjb@laptop.org>

Previously to this patch, an SDHCI platform that uses a GPIO for
card detection instead of the internal SDHCI_CARD_PRESENT bit on
the presence register would fail to detect a new card.

Some drivers worked around this in various ways: esdhc-imx defines
an IO accessor to fake the presence bit being true, s3c turns on
polling (which stops the SDHCI driver from checking the bit) after
a card's inserted.  But none of this should be necessary; the real
fix is to check whether we're using a GPIO and avoid relying on
the presence bit if so, as this patch implements.

Signed-off-by: Guennadi Liakhovetski <g.liakhovetski@gmx.de>
Signed-off-by: Chris Ball <cjb@laptop.org>

Removes use of the DaVinci EDMA private DMA API and replaces
it with use of the DMA engine API.

Signed-off-by: Matt Porter <mporter@ti.com>
Tested-by: Koen Kooi <koen@dominion.thruhere.net>
Signed-off-by: Chris Ball <cjb@laptop.org>

Power management callbacks defined by SIMPLE_DEV_PM_OPS are only used if
the PM_SLEEP Kconfig symbol has been defined. If not, the compiler will
complain about them being unused. However, since the callback for this
driver doesn't do anything it can just as well be dropped.

Signed-off-by: Thierry Reding <thierry.reding@avionic-design.de>
Acked-by: Rafael J. Wysocki <rjw@sisk.pl>
Signed-off-by: Chris Ball <cjb@laptop.org>

It is expected that Extended CSD register (the size of this register
is larger than CID/CSD) will be referenced more frequently as more
fields have been added to Extended CSD and it seems that it is not
a good option to double the memory used.

This patch is intended to avoid the use of bounce buffer for reading
Extended CSD register in mmc_send_cxd_data(). It will provide a better
performance gain by removing memcpy() overhead for a half KiB and
a redundant bounce buffer allocated repeatedly at the cost of providing
DMA-capable buffer from upper caller (but on-stack buffer is allowed
with no performance gain).

Signed-off-by: Kyungsik Lee <kyungsik.lee@lge.com>
Reviewed-by: Venkatraman S <svenkatr@ti.com>
Signed-off-by: Chris Ball <cjb@laptop.org>

Tested on OLPC XO-4/MMP3, where the card detection pin for one of
the controllers is a sideband GPIO.  The third cell in the cd-gpios
property controls whether the GPIO is active high/active low.

(Also, pass host_caps2 through from platdata to the mmc host.)

Acked-by: Haojian Zhuang <haojian.zhuang@gmail.com>
[kliu5@marvell.com: Compile fix]
Signed-off-by: Chris Ball <cjb@laptop.org>

mmc_gpio_request_ro() doesn't store the requested gpio in ctx->ro_gpio.
As a result, subsequent calls to mmc_gpio_get_ro() will always fail
with -ENOSYS because the gpio number isn't available to that function.

Cc: stable <stable@vger.kernel.org>
Acked-by: Guennadi Liakhovetski <g.liakhovetski@gmx.de>
Signed-off-by: Chris Ball <cjb@laptop.org>

module_pci_driver makes the code simpler by eliminating
module_init and module_exit calls.

Signed-off-by: Sachin Kamat <sachin.kamat@linaro.org>
Acked-by: Harald Welte <laforge@gnumonks.org>
Signed-off-by: Chris Ball <cjb@laptop.org>

module_pci_driver makes the code simpler by eliminating module_init
and module_exit calls.

Signed-off-by: Sachin Kamat <sachin.kamat@linaro.org>
Signed-off-by: Chris Ball <cjb@laptop.org>

module_pci_driver makes the code simpler by eliminating
module_init and module_exit calls.

Signed-off-by: Sachin Kamat <sachin.kamat@linaro.org>
Signed-off-by: Chris Ball <cjb@laptop.org>

module_spi_driver makes the code simpler by eliminating
module_init and module_exit calls.

Signed-off-by: Sachin Kamat <sachin.kamat@linaro.org>
Signed-off-by: Chris Ball <cjb@laptop.org>

This patch fixes the checkpatch error and warnings listed below:

ERROR: return is not a function, parentheses are not required
WARNING: Prefer pr_err(... to printk(KERN_ERR, ...
WARNING: quoted string split across lines
WARNING: line over 80 characters
WARNING: braces {} are not necessary for single statement blocks

Signed-off-by: Jingoo Han <jg1.han@samsung.com>
Acked-by: Jaehoon Chung <jh80.chung@samsung.com>
Signed-off-by: Chris Ball <cjb@laptop.org>

If MMC_CAP_NONREMOVABLE is set, only issue a detect job on init.

Signed-off-by: Johan Rudholm <johan.rudholm@stericsson.com>
Acked-by: Ulf Hansson <ulf.hansson@linaro.org>
Acked-by: Linus Walleij <linus.walleij@linaro.org>
Signed-off-by: Chris Ball <cjb@laptop.org>

Fixed the below message.

WARNING: drivers/mmc/host/built-in.o(.text+0x5ee8): Section mismatch in reference from the function dw_mci_probe() to the function .init.text:dw_mci_init_slot()
The function dw_mci_probe() references
the function __init dw_mci_init_slot().
This is often because dw_mci_probe lacks a __init
annotation or the annotation of dw_mci_init_slot is wrong

Signed-off-by: Jaehoon Chung <jh80.chung@samsung.com>
Signed-off-by: Kyungmin Park <kyungmin.park@samsung.com>
Acked-by: Will Newton <will.newton@imgtec.com>
Signed-off-by: Chris Ball <cjb@laptop.org>

When compiling a kernel that supports only Tegra30 and not Tegra20,
the SDHCI driver will fail to compile since the of_match_device() failure
fallback code unconditinally references soc_data_tegra20, which is only
compiled in when Tegra20 support is enabled.

In Kernel 3.7, both Tegra20 and Tegra30 will only support booting using
device tree, Hence, there is never a need to fall back to using
soc_data_tegra20 if of_mach_device() fails. Instead, make this case an
error. This removes the reference to soc_data_tegra20, and hence solves
the compile failure.

Signed-off-by: Joseph Lo <josephl@nvidia.com>
Tested-by: Stephen Warren <swarren@wwwdotorg.org>
Signed-off-by: Chris Ball <cjb@laptop.org>