Writing messages larger than the FIFO size results in a hang, rendering
the machine unusable. This is because the RXD status flag is set on the
first interrupt which results in bcm2835_drain_rxfifo() stealing bytes
from the buffer. The controller continues to trigger interrupts waiting
for the missing bytes, but bcm2835_fill_txfifo() has none to give.
In this situation wait_for_completion_timeout() apparently is unable to
stop the madness.

The BCM2835 ARM Peripherals datasheet has this to say about the flags:
  TXD: is set when the FIFO has space for at least one byte of data.
  RXD: is set when the FIFO contains at least one byte of data.
  TXW: is set during a write transfer and the FIFO is less than full.
  RXR: is set during a read transfer and the FIFO is or more full.

Implementing the logic from the downstream i2c-bcm2708 driver solved
the hang problem.
Signed-off-by: NNoralf Trønnes <noralf@tronnes.org>
Reviewed-by: NEric Anholt <eric@anholt.net>
Reviewed-by: NMartin Sperl <kernel@martin.sperl.org>
Signed-off-by: NWolfram Sang <wsa@the-dreams.de>

Fixes dmesg spam when we just need to wait a moment for the clock
driver to probe.
Signed-off-by: NEric Anholt <eric@anholt.net>
Signed-off-by: NWolfram Sang <wsa@the-dreams.de>

As reported in the links given below. the BCM2835 has a hardware bug in
its i2c module which prevents a correct clock stretching. This patch
adds the I2C_AQ_NO_CLK_STRETCH quirk flag to i2c-bcm2835.
Signed-off-by: NNicola Corna <nicola@corna.info>
[wsa: put the links into the code as comments]
Signed-off-by: NWolfram Sang <wsa@the-dreams.de>

Checks if the cdiv value is in between min (0x2) and max (0xFFFE)
supported values by the bcm2835. If not, it returns -ENODEV.

See page 33/34 of BCM2835-ARM-Peripherals.pdf for the DIV register.
https://www.raspberrypi.org/wp-content/uploads/2012/02/BCM2835-ARM-Peripherals.pdfSigned-off-by: NSilvan Wicki <linux_wi@tinag.ch>
[wsa: resolved a merge conflict]
Signed-off-by: NWolfram Sang <wsa@the-dreams.de>

The datasheet mentions on page 31 that the bits 10-31 must be read as
don't care and written as 0.
https://www.raspberrypi.org/wp-content/uploads/2012/02/BCM2835-ARM-Peripherals.pdf

We cannot guarantee that we read bits 10-31 as always 0 (because the
datasheet says read as don't care). We clear the bits with a bitmask to
prevent writing back unknown data at the reserved bits.
Signed-off-by: NSilvan Wicki <linux_wi@tinag.ch>
Acked-by: NStephen Warren <swarren@wwwdotorg.org>
Signed-off-by: NWolfram Sang <wsa@the-dreams.de>

return type of wait_for_completion_timeout is unsigned long not int. as
time_left is used for wait_for_completion_timeout exclusively here its
type is simply changed to unsigned long.
Signed-off-by: NNicholas Mc Guire <hofrat@osadl.org>
Acked-by: NLee Jones <lee.jones@linaro.org>
Signed-off-by: NWolfram Sang <wsa@the-dreams.de>

A platform_driver does not need to set an owner, it will be populated by the
driver core.
Signed-off-by: NWolfram Sang <wsa@the-dreams.de>

This driver has been flagged to drop class based instantiation. The removal
improves boot-up time and is unneeded for embedded controllers. Users have been
warned to switch for some time now, so we can actually do the removal. Keep the
DEPRECATED flag, so the core can inform users that the behaviour finally
changed now. After another transition period, this flag can go, too.
Signed-off-by: NWolfram Sang <wsa@the-dreams.de>

The site-specific OOM messages are unnecessary, because they
duplicate the MM subsystem generic OOM message. For example,
k.alloc and v.alloc failures use dump_stack().
Signed-off-by: NJingoo Han <jg1.han@samsung.com>
Acked-by: NGuenter Roeck <linux@roeck-us.net>
Reviewed-by: NJean Delvare <jdelvare@suse.de>
Acked-by: NUwe Kleine-König <u.kleine-koenig@pengutronix.de>
Acked-by: NFelipe Balbi <balbi@ti.com>
Acked-by: NLaurent Pinchart <laurent.pinchart@ideasonboard.com>
Acked-by: NLinus Walleij <linus.walleij@linaro.org>
Signed-off-by: NWolfram Sang <wsa@the-dreams.de>

Warn users that class based instantiation is going away soon in favour
of more robust probing and faster bootup times.
Signed-off-by: NWolfram Sang <wsa@the-dreams.de>

Use devm_ioremap_resource() in order to make the code simpler,
and remove redundant return value check of platform_get_resource()
because the value is checked by devm_ioremap_resource().
Signed-off-by: NJingoo Han <jg1.han@samsung.com>
Acked-by: NStephen Warren <swarren@wwwdotorg.org>
Signed-off-by: NWolfram Sang <wsa@the-dreams.de>

In order to find I2C devices in the device tree, the platform nodes
have to be known by the I2C core. This requires setting the
dev.of_node parameter of the adapter.
Signed-off-by: NFlorian Meier <florian.meier@koalo.de>
Tested-by: NStephen Warren <swarren@wwwdotorg.org>
Signed-off-by: NWolfram Sang <wsa@the-dreams.de>

Use this new function to make code more comprehensible, since we are
reinitialzing the completion, not initializing.

[akpm@linux-foundation.org: linux-next resyncs]
Signed-off-by: NWolfram Sang <wsa@the-dreams.de>
Acked-by: Linus Walleij <linus.walleij@linaro.org> (personally at LCE13)
Cc: Ingo Molnar <mingo@kernel.org>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

This implements a very basic I2C host driver for the BCM2835 SoC. Missing
features so far are:

* 10-bit addressing.
* DMA.
Reviewed-by: NGrant Likely <grant.likely@secretlab.ca>
Signed-off-by: NStephen Warren <swarren@wwwdotorg.org>
Signed-off-by: NWolfram Sang <wolfram@the-dreams.de>