Currently we process the first i2c_dw_xfer_msg() in i2c_dw_xfer(),
but in this case there is a possibility to be interrupted by certain
interrupts.  As described before in this patchset, we need to keep
providing new transmit data within a given time period, otherwise Tx
FIFO underrun takes place and STOP condition will be generated on the
bus, even if we have more bytes to be written.

In order to exclude all such possibilities, change TX_EMPTY interrupt
usage as below:

* DW_IC_INTR_DEFAULT_MASK: Define a default interrupt mask set, and
  put TX_EMPTY there.

* i2c_dw_xfer_init: Enable DW_IC_INTR_DEFAULT_MASK prior to initiating
  a new I2C transaction.  The first TX_EMPTY will be triggered shortly.
  With the help of it, we can make the first call to i2c_dw_xfer_msg()
  in the interrupt handler.

* i2c_dw_xfer_msg: Fixup intr_mask operation accordingly.  Make sure
  that TX_EMPTY operations need to be reversed.

* request_irq: Set IRQF_DISABLED so that we could load transmit data
  into Tx FIFO without being distracted by other interrupts.

* Remove i2c_dw_xfer_msg() in i2c_dw_xfer().
Signed-off-by: NShinya Kuribayashi <shinya.kuribayashi@necel.com>
Signed-off-by: NBen Dooks <ben-linux@fluff.org>

I2c_dw_xfer_msg() also has the same target address inconsistency check,
and furthermore it checks across all i2c_msg messages, while
i2c_dw_read() walks through i2c_msg messages only with_ I2C_M_RD flag.
That is, target address check in i2c_dw_read() is redundant and useless.
Signed-off-by: NShinya Kuribayashi <shinya.kuribayashi@necel.com>
Signed-off-by: NBen Dooks <ben-linux@fluff.org>

Set proper I2C_FUNC_SMBUS_* bits so that the driver could be used with
some utilities requiring SMBus functionalities, such as i2c-tools.

Note that DW I2C core doesn't support I2C_FUNC_SMBUS_QUICK, as it's not
capable of zero-length data transactions.
Signed-off-by: NShinya Kuribayashi <shinya.kuribayashi@necel.com>
Signed-off-by: NBen Dooks <ben-linux@fluff.org>

As the driver and hardware always process the given data in parallel,
then it would be better to initialize tx_limit, rx_limit and rx_valid
variables just prior to being used.

This will help us to send / receive as much data as possible.
Signed-off-by: NShinya Kuribayashi <shinya.kuribayashi@necel.com>
Acked-by: NBaruch Siach <baruch@tkos.co.il>
Signed-off-by: NBen Dooks <ben-linux@fluff.org>

While we have a local variable "buf_len" for dev->tx_buf_len, we don't
have such local variable for dev->tx_buf pointer.  While "buf_len" is
restored at first then updated when we start processing a new i2c_msg
(determined by STATUS_WRITE_IN_PROGRESS flag), ->tx_buf is different.

Such inconsistency makes the code slightly hard to follow.
Signed-off-by: NShinya Kuribayashi <shinya.kuribayashi@necel.com>
Acked-by: NBaruch Siach <baruch@tkos.co.il>
Signed-off-by: NBen Dooks <ben-linux@fluff.org>

We have some steps at the top of i2c_dw_xfer_msg() to set up a slave
address and enable DW I2C core.  And it's executed only when we don't
have STATUS_WRITE_IN_PROGRESS.

But we need to make sure that STATUS_WRITE_IN_PROGRESS only indicates
that we have a pending i2c_msg to process.  In other words, even if
STATUS_WRITE_IN_PROGRESS is not set, that doesn't mean we're at initial
state in the I2C transaction.

Since i2c_dw_xfer_msg() will be invoked again and again during a
transaction, those init steps have a possibility to be re-processed
needlessly.  For example, this issue easily takes place when processing
a combined transaction with a certain condition (the number of tx bytes
in the first i2c_msg, equals to the Tx FIFO depth).

Consequently we should not use STATUS_WRITE_IN_PROGRESS to determine
where we're at in an I2C transaction.  It would be better to separate
those initialization steps from i2c_dw_xfer_msg().
Signed-off-by: NShinya Kuribayashi <shinya.kuribayashi@necel.com>
Acked-by: NBaruch Siach <baruch@tkos.co.il>
Signed-off-by: NBen Dooks <ben-linux@fluff.org>

Enable RX_FULL interrupt mask by default, and hook it in the interrupt
handler.  If requested amount of rx data (defined by IC_RX_TL) is not
available, we don't have to process i2c_dw_read().
Signed-off-by: NShinya Kuribayashi <shinya.kuribayashi@necel.com>
Acked-by: NBaruch Siach <baruch@tkos.co.il>
Signed-off-by: NBen Dooks <ben-linux@fluff.org>

As a hardware feature, DW I2C core generates a STOP condition whenever
the Tx FIFO becomes empty (strictly speaking, whenever the last byte in
the Tx FIFO is sent out), even if we have more bytes to be written.

In other words, we must never make "Tx FIFO underrun" happen during
a transaction, except for the last byte.  For the safety's sake, we'd
make TX_EMPTY interrupt get triggered every time one byte is processed.

The Rx FIFO threshold needs to be set as well.
Signed-off-by: NShinya Kuribayashi <shinya.kuribayashi@necel.com>
Acked-by: NBaruch Siach <baruch@tkos.co.il>
Signed-off-by: NBen Dooks <ben-linux@fluff.org>

Symptom:
--------
When we're going to send/receive the longer size of data than the Tx
FIFO length, the I2C transaction will be divided into several separated
transactions, limited by the Tx FIFO length.

Details:
--------
As a hardware feature, DW I2C core generates a STOP condition whenever
the Tx FIFO becomes empty (strictly speaking, whenever the last byte in
the Tx FIFO is sent out), even if we have more bytes to be written.
Then, once a new transmit data is written to the Tx FIFO, DW I2C core
will initiate a new transaction, which leads to another START condition.

This explains how the transaction in question goes, and implies that
current tasklet-based dw_i2c_pump_msg() strategy couldn't meet the
timing constraint required for avoiding Tx FIFO underrun.

To avoid this scenario, we must keep providing new transmit data within
a given time period.  In case of Fast-mode + 32-byte Tx FIFO, for
instance, it takes about 22.5[us] to process single byte, and 720[us] in
total.

This patch removes the existing tasklet-based "pump" system, and move
its jobs into the interrupt handler.
Signed-off-by: NShinya Kuribayashi <shinya.kuribayashi@necel.com>
Acked-by: NBaruch Siach <baruch@tkos.co.il>
Signed-off-by: NBen Dooks <ben-linux@fluff.org>

In case a work-in-progress i2c_msg has more bytes to be written, we
need to set STATUS_WRITE_IN_PROGRESS and exit from the msg_write_idx-
searching loop.  Otherwise, we will overtake the current msg_write_idx
without waiting for its transmission to be processed.
Signed-off-by: NShinya Kuribayashi <shinya.kuribayashi@necel.com>
Signed-off-by: NBen Dooks <ben-linux@fluff.org>

* Calculate with accurate conditional expressions from DW manuals.
* Round ic_clk by adding 0.5 as it's important at high ic_clk rate.
* Take into account "tHD;STA" issue for _HCNT calculation.
* Take into account "tf" for _LCNT calculation.
* Add "cond" and "offset" fot further correction requirements.

For _HCNT calculation, there's one issue needs to be carefully
considered; DesignWare I2C core doesn't seem to have solid strategy
to meet the tHD;STA timing spec.  If you configure _HCNT based on the
tHIGH timing spec, it easily results in violation of the tHD;STA spec.

After many trials, we came to the conclusion that the tHD;STA period
is proportional to (_HCNT + 3).  For the safety's sake, this should be
selected by default.

As for _LCNT calculation, DW I2C core has one characteristic behavior;
he starts counting the SCL CNTs for the LOW period of the SCL clock
(tLOW) as soon as it pulls the SCL line.  At that time, he doesn't take
into account the fall time of SCL signal (tf), IOW, he starts counting
CNTs without confirming the SCL input voltage has dropped to below VIL.

This characteristics becomes a problem on some platforms where tf is
considerably long, and results in violation of the tLOW timing spec.

To make the driver configurable as much as possible for various cases,
we'd have separated arguments "tf" and "offset", and for safety default
values should be 0.3 us and 0, respectively.
Signed-off-by: NShinya Kuribayashi <shinya.kuribayashi@necel.com>
Acked-by: NBaruch Siach <baruch@tkos.co.il>
Signed-off-by: NBen Dooks <ben-linux@fluff.org>

We couldn't know the original intent for this variable, but at this
point it's useless.
Signed-off-by: NShinya Kuribayashi <shinya.kuribayashi@necel.com>
Acked-by: NBaruch Siach <baruch@tkos.co.il>
Signed-off-by: NBen Dooks <ben-linux@fluff.org>

We don't have to use "struct i2c_adapter" pointer here.
Let's use a local "struct dw_i2c_dev" pointer, instead.
Signed-off-by: NShinya Kuribayashi <shinya.kuribayashi@necel.com>
Acked-by: NBaruch Siach <baruch@tkos.co.il>
Signed-off-by: NBen Dooks <ben-linux@fluff.org>

We don't have to use "struct i2c_adapter" pointer here.
Let's use a local "struct dw_i2c_dev" pointer, instead.
Signed-off-by: NShinya Kuribayashi <shinya.kuribayashi@necel.com>
Acked-by: NBaruch Siach <baruch@tkos.co.il>
Signed-off-by: NBen Dooks <ben-linux@fluff.org>

Signed-off-by: NShinya Kuribayashi <shinya.kuribayashi@necel.com>
Acked-by: NBaruch Siach <baruch@tkos.co.il>
Signed-off-by: NBen Dooks <ben-linux@fluff.org>

We're strongly discouraged from using the IC_CLR_INTR register because
it clears all software-clearable interrupts asserted at the moment.

  stat = readl(IC_INTR_STAT);
    :
    :  <=== Interrupts asserted during this period will be lost
    :
  readl(IC_CLR_INTR);

Instead, use the separately-prepared IC_CLR_* registers.

At the same time, this patch adds all remaining interrupt definitions
available in the DesignWare I2C hardware.
Signed-off-by: NShinya Kuribayashi <shinya.kuribayashi@necel.com>
Signed-off-by: NBen Dooks <ben-linux@fluff.org>

This driver looks originally meant for armel machines where readw()/
writew() works perfectly fine with this hardware.  But that doens't
work for big-endian systems.

This patch converts all 8/16-bit-aware usages to 32-bit variants.
Signed-off-by: NShinya Kuribayashi <shinya.kuribayashi@necel.com>
Acked-by: NBaruch Siach <baruch@tkos.co.il>
Signed-off-by: NBen Dooks <ben-linux@fluff.org>

This reverts commit f20941f3.

Sergei Shtylyov notes "You call min() on uncomparables [in
mwdma_clip_to_pio()], i.e. mwdma_to_pio[] contains XFER_PIO_* and
adev->pio_mode - XFER_PIO_0 yields you a mode number.  Thus the second
argument will always "win" as a minimal one"

Bartlomiej Zolnierkiewicz adds "There are more issues with the patch related
to mwdma_clip_to_pio().  The function can return values between 0 and
4 which obviously won't work well for the new code below for values
>2 (i.e. resulting in out-of-bounds array access for the common-case
of dev->pio_mode == XFER_PIO_4)."

Bartlomiej Zolnierkiewicz also notes the patch is incomplete, failing to
update MWDMA mode masks.
Signed-off-by: NJeff Garzik <jgarzik@redhat.com>

Jonathan Cameron reports that building PCMCIA as modules doesn't work:

As module get a load of undefined symbols:
ERROR: "soc_pcmcia_request_irqs" [drivers/pcmcia/pxa2xx_stargate2.ko] undefined!
ERROR: "soc_pcmcia_free_irqs" [drivers/pcmcia/pxa2xx_stargate2.ko] undefined!
ERROR: "soc_pcmcia_enable_irqs" [drivers/pcmcia/pxa2xx_stargate2.ko] undefined!
ERROR: "soc_pcmcia_disable_irqs" [drivers/pcmcia/pxa2xx_stargate2.ko] undefined!
ERROR: "soc_pcmcia_add_one" [drivers/pcmcia/pxa2xx_base.ko] undefined!
ERROR: "soc_common_pcmcia_get_timing" [drivers/pcmcia/pxa2xx_base.ko] undefined!
ERROR: "soc_pcmcia_remove_one" [drivers/pcmcia/pxa2xx_base.ko] undefined!
make[1]: *** [__modpost] Error 1
make: *** [modules] Error 2

This is because soc_common tries to be built-in, but it should be a module.
Allow soc_common to be a module.
Reported-by: NJonathan Cameron <jic23@cam.ac.uk>
Signed-off-by: NRussell King <rmk+kernel@arm.linux.org.uk>

Both iPAQs h3600 and h3100 share the same control
GPIOs for PCMCIA, so driver can be reused.
Signed-off-by: NDmitry Artamonow <mad_soft@inbox.ru>
Signed-off-by: NRussell King <rmk+kernel@arm.linux.org.uk>

Combine both headers into one, rename to h3xxx.h and change all
users accordingly.
Signed-off-by: NDmitry Artamonow <mad_soft@inbox.ru>
Signed-off-by: NRussell King <rmk+kernel@arm.linux.org.uk>

Use of gpio_request/gpio_free in some callbacks may look ugly, but
corresponding drivers (sa1100_serial and sa1100_fb) don't provide (yet)
init/exit hooks and registering these gpios in *_mach_init is also
not possible, because htc-gpio driver starts a bit later...
Signed-off-by: NDmitry Artamonow <mad_soft@inbox.ru>
Signed-off-by: NRussell King <rmk+kernel@arm.linux.org.uk>

Convert all operations with GPLR/GPCR/GPSR to gpiolibs calls.
Also change all IRQ_GPIO* to gpio_to_irq(*GPIO*)
Signed-off-by: NDmitry Artamonow <mad_soft@inbox.ru>
Signed-off-by: NRussell King <rmk+kernel@arm.linux.org.uk>

Add a module parameter to override the functionality bitfield. This
lets the user disable some commands. This can be used to force a chip
driver to take different code paths.
Signed-off-by: NJean Delvare <khali@linux-fr.org>

This is required to test some drivers, for example at24.
Signed-off-by: NJean Delvare <khali@linux-fr.org>

Functions i2c_do_add_adapter() and __attach_adapter() do essentially
the same thing, differing only in how the parameters are passed. Same
for i2c_do_add_adapter() and __detach_adapter(). Introduce wrappers to
normalize the parameters, so that we do not have to duplicate the
code.
Signed-off-by: NJean Delvare <khali@linux-fr.org>
Cc: David Brownell <dbrownell@users.sourceforge.net>

The Intel 82801 is sometimes used on systems with a BMC connected. The
BMC can access the SMBus, resulting in lost arbitration for the 82801.
We should let i2c-core retry transactions for us in this case.
Signed-off-by: NJean Delvare <khali@linux-fr.org>

The BKL is held over a kmalloc so cannot protect anything beyond that.
The two calls before the kmalloc have their own locking.
Improve device open function by removing the now unnecessary ret variable
Signed-off-by: NVincent Sanders <vince@simtec.co.uk>
Signed-off-by: NThomas Gleixner <tglx@linutronix.de>
Signed-off-by: NJean Delvare <khali@linux-fr.org>

As kind is now hard-coded to -1, there is room for code clean-ups.
Signed-off-by: NJean Delvare <khali@linux-fr.org>
Acked-by: N"Darrick J. Wong" <djwong@us.ibm.com>

The kind parameter of i2c_detect_address() always has value -1, so we
can get rid of it.

Next step is to update all i2c detect callback functions to get rid of
this now useless parameter.
Signed-off-by: NJean Delvare <khali@linux-fr.org>

The legacy probe and force module parameters are obsolete now, the
same can be achieved using the new_device sysfs interface, which is
both more flexible and cheaper (it is implemented by i2c-core rather
than replicated in every driver module.)

The legacy ignore module parameters can be dropped as well. Ignoring
can be done by instantiating a "dummy" device at the problematic
address.

This is the first step of a huge cleanup to i2c-core's i2c_detect
function, i2c.h's I2C_CLIENT_INSMOD* macros, and all drivers that made
use of them.
Signed-off-by: NJean Delvare <khali@linux-fr.org>

These _setup functions are called from _probe so they can be marked
__devinit.
Signed-off-by: NJean Delvare <khali@linux-fr.org>

I2C bus drivers don't have to support I2C_M_REV_DIR_ADDR. It is a
deviation from the I2C specification, which only makes sense to
implement when really needed.
Signed-off-by: NJean Delvare <khali@linux-fr.org>
Cc: Ben Dooks <ben-linux@fluff.org>

Low priority thread holding the i2c bus mutex could block higher
priority threads to access the bus resulting in unacceptable
latencies. Change the mutex type to rt_mutex preventing priority
inversion.
Tested-by: NPeter Ujfalusi <peter.ujfalusi@nokia.com>
Signed-off-by: NMika Kuoppala <mika.kuoppala@nokia.com>
Signed-off-by: NJean Delvare <khali@linux-fr.org>

Superseded by tdfxfb. I2C/DDC support used to live in a separate
driver but this caused driver conflicts.
Signed-off-by: NJean Delvare <khali@linux-fr.org>
Cc: Krzysztof Helt <krzysztof.h1@wp.pl>

We no longer need to write the adapter name to a temporary buffer.
We can write it directly to the i2c_adapter's name field. This is
more efficient.
Signed-off-by: NJean Delvare <khali@linux-fr.org>
Tested-by: NMichel Daenzer <michel@daenzer.net>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>

Include the i2c_adapter in struct pmac_i2c_bus. This avoids memory
fragmentation and allows for several code cleanups.
Signed-off-by: NJean Delvare <khali@linux-fr.org>
Tested-by: NMichel Daenzer <michel@daenzer.net>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>

Log errors when they happen, otherwise we have no idea what went
wrong.
Signed-off-by: NJean Delvare <khali@linux-fr.org>
Tested-by: NMichel Daenzer <michel@daenzer.net>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Cc: Paul Mackerras <paulus@samba.org>

I wanted to add some error logging to the i2c-powermac driver, but
found that it was very difficult due to the way the
i2c_powermac_smbus_xfer function is organized. Refactor the code in
this function so that each low-level function is only called once.
Signed-off-by: NJean Delvare <khali@linux-fr.org>
Tested-by: NMichel Daenzer <michel@daenzer.net>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Cc: Paul Mackerras <paulus@samba.org>

The i2c-powermac driver doesn't support arbitrary multi-message I2C
transactions, only SMBus ones. Make it clear by returning an error if
a multi-message I2C transaction is attempted. This is better than only
processing the first message, because most callers won't recover from
the short transaction. Anyone wishing to issue multi-message
transactions should use the SMBus API instead of the raw I2C API.
Signed-off-by: NJean Delvare <khali@linux-fr.org>
Tested-by: NMichel Daenzer <michel@daenzer.net>
Acked-by: NBenjamin Herrenschmidt <benh@kernel.crashing.org>
Cc: Paul Mackerras <paulus@samba.org>