01 Aug, 2011
40 commits
-
get_ctrl_reg() returns -EINVAL so the error handling won't work here
if reg is a u8.Signed-off-by: Dan Carpenter
Signed-off-by: Samuel Ortiz -
There are a couple of situations where we leak init_data in
drivers/mfd/tps65910.c:tps65910_i2c_probe() - this patch should take
care of them.Signed-off-by: Jesper Juhl
Signed-off-by: Samuel Ortiz -
Add regulator driver for AnalogicTech AAT2870.
Signed-off-by: Jin Park
Acked-by: Mark Brown
Signed-off-by: Samuel Ortiz -
Add backlight driver for AnalogicTech AAT2870.
Signed-off-by: Jin Park
Signed-off-by: Samuel Ortiz -
Add mfd core driver for AnalogicTech AAT2870.
The AAT2870 is communication through I2C and contains backlight and
regulator components.Signed-off-by: Jin Park
Signed-off-by: Samuel Ortiz -
We either hit one of the case's or the default in the switch statement
in get_i2c(), so the 'return ERR_PTR(-EINVAL);' at the end of the
function is just dead code - remove it.Signed-off-by: Jesper Juhl
Acked-by: MyungJoo Ham
Signed-off-by: Samuel Ortiz -
Having another TPS chip at the end of the Kconfig when all it's
relatives are grouped together in their own section seems totally
counter-intuitive. Move it, also in the Makefile.Signed-off-by: Linus Walleij
Signed-off-by: Samuel Ortiz -
Generating kerneldoc for STMPE result in warnings, so fix this by
adding missing documentation.Signed-off-by: Om Prakash
Reviewed-by: Rabin Vincent
Reviewed-by: Jonas Aberg
Reviewed-by: Srinidhi Kasagar
Signed-off-by: Linus Walleij
Signed-off-by: Samuel Ortiz -
If bytes == (TPS6591X_MAX_REGISTER + 1), we have a buffer overflow when
doing memcpy(&msg[1], src, bytes).Signed-off-by: Axel Lin
Signed-off-by: Samuel Ortiz -
With the new generic clk API that should appear at some point we should be
able to support the clocking sensibly in Linux.Signed-off-by: Mark Brown
Signed-off-by: Samuel Ortiz -
Signed-off-by: Axel Lin
Signed-off-by: Samuel Ortiz -
The user has to select the I2C and SPI drivers individually and they select
the core driver for the device so there's no point in presenting the user
with an option for the core driver.Signed-off-by: Mark Brown
Acked-by: Randy Dunlap
Signed-off-by: Samuel Ortiz -
The GPIO IRQs aren't the first IRQs defined, we need to subtract the base
for the GPIOs as well to use them for array indexes.Signed-off-by: Dimitris Papastamos
Signed-off-by: Mark Brown
Signed-off-by: Samuel Ortiz -
Very similar to TPS65920
List of differences: http://www.ti.com/litv/pdf/swcu066bSigned-off-by: Oleg Drokin
Signed-off-by: Samuel Ortiz -
Signed-off-by: Mark Brown
Signed-off-by: Samuel Ortiz -
Purely for defensiveness.
Signed-off-by: Mark Brown
Signed-off-by: Samuel Ortiz -
The tps65910_irq_exit() cleanup function was generating a warning from
sparse due to the lack of a prototype. This wasn't causing GCC warnings
as the driver wasn't cleaning up its IRQs on exit at all so there was no
use of an unprototyped function.Signed-off-by: Mark Brown
Signed-off-by: Samuel Ortiz -
Tunnel Creek has an additional watchdog core.
Signed-off-by: Alexander Stein
Signed-off-by: Samuel Ortiz -
This adds a previously undefined test register and removed a
number of double-defined accessory detect registers (they are
already defined higher up in the file.Signed-off-by: Linus Walleij
Signed-off-by: Samuel Ortiz -
This synchronize the subdevice entries for the AB8500 MFD driver
with the latest development of subdrivers for things like battery
charging and temperature monitoring.Signed-off-by: Linus Walleij
Signed-off-by: Samuel Ortiz -
Internal MFD device structs are marked as __devinitdata since the kernel
will allocate memory for the same when calling mfd_add_devices.Signed-off-by: Robert Rosengren
Reviewed-by: Mattias Wallin
Reviewed-by: Johan Palsson
Signed-off-by: Linus Walleij
Signed-off-by: Samuel Ortiz -
The tps65912 consist of 4 DCDCs and 10 LDOs. The output voltages can be
configured by the SPI or I2C interface, they are meant to supply power
to the main processor and other components.Signed-off-by: Margarita Olaya Cabrera
Acked-by: Mark Brown
Acked-by: Liam Girdwood
Signed-off-by: Samuel Ortiz -
TPS65912 has five GPIOs that can be configured for different
purposes.Signed-off-by: Margarita Olaya Cabrera
Acked-by: Grant Likely
Acked-by: Liam Girdwood
Signed-off-by: Samuel Ortiz -
This module controls the interrupt handling for the tps65912.
The interrupt sources can be the following:- GPIO
- PWRON signal
- PWRHOLD signal
- Temperature detectionSigned-off-by: Margarita Olaya Cabrera
Acked-by: Samuel Ortiz
Acked-by: Liam Girdwood
Signed-off-by: Samuel Ortiz -
The tps65912 chip is a power management IC. It contains the following
components:- Regulators
- GPIO controllerThe core driver is registered as a platform driver, it provides communication
through I2C and SPI interfaces.Signed-off-by: Margarita Olaya Cabrera
Acked-by: Samuel Ortiz
Acked-by: Liam Girdwood
Signed-off-by: Samuel Ortiz -
This patch replaces the code for getting an unsigned long from a
userspace buffer by a simple call to kstroul_from_user.
This makes it easier to read and less error prone.Signed-off-by: Peter Huewe
Signed-off-by: Samuel Ortiz -
This patch replaces the code for getting an unsigned long from a
userspace buffer by a simple call to kstroul_from_user.
This makes it easier to read and less error prone.Signed-off-by: Peter Huewe
Signed-off-by: Samuel Ortiz -
Use the generic irq chip framework for implementing the irq chip for
the jz4740-adc driver.Signed-off-by: Lars-Peter Clausen
Cc: Thomas Gleixner
Signed-off-by: Samuel Ortiz -
Even if we would've BUG()ed we should still tidy up after ourselves if that
isn't enabled in the kernel config.Signed-off-by: Mark Brown
Signed-off-by: Samuel Ortiz -
When the byte swap was factored out into the per-register I/O functions
the register restore for the IRQ mask cache (which we use and store in
CPU native format for the interrupt handler) was not updated to do a byte
swap when it uses the bulk I/O. Fix this by writing the cache out one
register at a time.Signed-off-by: Mark Brown
Signed-off-by: Samuel Ortiz -
The WM831x AUXADC hardware can schedule multiple conversions at once,
allowing higher performance when more than one source is in use as we
can have the hardware start new conversions without having to wait for
a register write.Take advantage of this in the interrupt driven case, maintaining a list of
callers that are waiting for AUXADC conversions and completing them all
simultaneously. The external interface of the AUXADC is not changed so
there will be limited use of the feature immediately.Signed-off-by: Mark Brown
Signed-off-by: Samuel Ortiz -
Switch on the device type before revision since anything we do here will
be device as well as revision specific.Signed-off-by: Mark Brown
Signed-off-by: Samuel Ortiz -
This allows boards to leave the irq_base field unitialized and
prevents them having to reserve irqs in the platform.
pdata can be optional for irq support now. Without pdata the
driver allocates some free irq range. With pdata and irq_base > 0
the driver allocates exactly the specified irq.
Without pdata the irq defaults to IRQF_TRIGGER_LOW.Signed-off-by: Sascha Hauer
Signed-off-by: Mark Brown
Signed-off-by: Samuel Ortiz -
In preparation for some additional work on the wm831x AUXADC code move the
support into a separate file. This is a simple code motion patch, there
should be no functional changes.Signed-off-by: Mark Brown
Signed-off-by: Samuel Ortiz -
Ensure that there's no possibility of loosing an AUXADC interrupt by reading
the conversion result in the IRQ handler when using interrupts. Otherwise
it's possible that under very heavy load a new conversion could be initiated
before the acknowledgement for a previous interrupt has happened.Signed-off-by: Mark Brown
Signed-off-by: Samuel Ortiz -
Use irq_allocate_desc() to get the IRQ range, which turns into a noop on
non-sparse systems. Since all existing users are non-sparse there should
be no compatibility issues.Signed-off-by: Mark Brown
Signed-off-by: Samuel Ortiz -
The RTC uses the 32.768kHz crystal so if it's not enabled (and it can only
be enabled via OTP or InstantConfig, not runtime software) the RTC can't
function.Signed-off-by: Mark Brown
Signed-off-by: Samuel Ortiz -
Allow platform data to flag the touchscreen as disabled so that if the
touch driver is built in we don't end up causing lots of work by spuriously
detecting touchscreen activity on systems where it isn't in use.Signed-off-by: Mark Brown
Signed-off-by: Samuel Ortiz -
Systems using this functionality will be uncommon.
Signed-off-by: Mark Brown
Signed-off-by: Samuel Ortiz -
The WM831x IRQ set_type() operation is doing a direct register write when
called but since set_type() is called with the bus lock held this isn't
legal and could cause deadlocks in the IRQ core.Fix this by posting the updates into an array and syncing in the
bus_sync_unlock() callback.Signed-off-by: Mark Brown
Signed-off-by: Samuel Ortiz
