| Age | Commit message (Collapse) | Author |
|---|
|
Add device tree node for the OCOTP peripheral on Vybrid.
Signed-off-by: Sanchayan Maity <maitysanchayan@gmail.com>
Signed-off-by: Stefan Agner <stefan.agner@toradex.com>
|
|
Currently syscon does not provide an abstraction to access a
register from syscon reference like below
ocotp-cfg1 = <&ocotp 0x20>
syscon_regmap_read_from_offset provides a generic abstraction to
access a register from syscon reference as above. It allows to
specify the node and node name of phandle reference, reading the
offset from the node entry and providing the value from the offset
in the register map.
Signed-off-by: Sanchayan Maity <maitysanchayan@gmail.com>
|
|
This patch allows for easy integration of a custom Linux boot logo to
replace the Tux' being shown by default.
Use gimp or the like to create a raw PPM in your desired resolution.
Reduce the number of colours in the image to 224:
user@host:~$ ppmquant 224 Toradex-640x480.ppm > \
Toradex-640x480-224.ppm
ppmquant: making histogram...
ppmquant: 370 colors found
ppmquant: choosing 224 colors...
ppmquant: mapping image to new colors...
Convert it from raw PPM to ASCII format:
user@host:~$ pnmnoraw Toradex-640x480-224.ppm > \
Toradex-640x480-ascii-224.ppm
Copy it into the Linux sources:
cp Toradex-640x480-ascii-224.ppm linux-toradex/drivers/video/logo/\
logo_custom_clut224.ppm
Activate exclusively custom Linux logo in the kernel configuration:
Device Drivers -> Graphics support -> Bootup logo ->
Custom 224-color Linux logo
And re-compile the kernel.
(cherry picked from commit fa2371bff9ac03581881849d8f95678ef3992719)
Signed-off-by: Stefan Agner <stefan.agner@toradex.com>
|
|
Add a device trees for the dual Ethernet Colibri Carrier Board. The Carrier
Board has a second Ethernet PHY on board and connects it to the second Ethernet
Controller of Vybrid. Therefor the Carrier Board is only usable with Vybrid
based modules (Colibri VF50/VF61). Furthermore, the Carrier Board has two
additional UARTs beside the three default Colibri UARTs.
The schemata of the Ethernet part of a dual Ethernet Carrier
Board for Vybrid can be found online at:
http://docs.toradex.com/102058-colibri-vfxx-2nd-ethernet-rmii
Signed-off-by: Bhuvanchandra DV <bhuvanchandra.dv@toradex.com>
Signed-off-by: Stefan Agner <stefan.agner@toradex.com>
|
|
|
|
Enable TouchRevolution multitouch controller driver
Signed-off-by: Bhuvanchandra DV <bhuvanchandra.dv@toradex.com>
|
|
Add TouchRevolution multitouch controller driver which is connectable
over I2C bus. The driver is tested with 7" and 10" multitouch panels
by TouchRev.
PWM_B, PWM_C pins are used as pen down and reset interrupt, to
avoid pin conflicts the PWM functionality on these pins has to be
disabled and the relevant GPIO functionality should be enabled and
assigned to the touch driver.
Signed-off-by: Stefan Agner <stefan.agner@toradex.com>
Signed-off-by: Bhuvanchandra DV <bhuvanchandra.dv@toradex.com>
|
|
Add support for resetting the controller using the gpio reset pin
if interrupt could not be cleared. If clearing the interrupt fails,
the interrupt line stays high and no further interrupts are generated
resulting in a non functioning touchscreen until next reboot, when we
reset the controller again in probe.
Signed-off-by: Sanchayan Maity <maitysanchayan@gmail.com>
Signed-off-by: Stefan Agner <stefan.agner@toradex.com>
Signed-off-by: Bhuvanchandra DV <bhuvanchandra.dv@toradex.com>
|
|
According to multi-touch-protocol.txt the kernel knows two type of
devices:
- Type A: devices which provide anonymous contacts
- Type B: devices which are capable of tracking individual contacts
The Fusion touch screen is a Type B device: The Touch ID field allows
to differentiate between 2 different fingers. This updates the driver
to properly allocate a slot for each identified contact and use the
helper functions available from the input core system.
Signed-off-by: Stefan Agner <stefan.agner@toradex.com>
Signed-off-by: Bhuvanchandra DV <bhuvanchandra.dv@toradex.com>
|
|
Use level triggered interrupt which makes sure that even we miss a
rising edge (due to latencies in the kernel), the interrupt still
will be handled later, and doesn't freeze the input device.
Signed-off-by: Stefan Agner <stefan.agner@toradex.com>
|
|
Add device tree integration and add the device to the dtb.
Signed-off-by: Max Krummenacher <max.krummenacher@toradex.com>
[Add devicetree binding documentation]
Signed-off-by: Bhuvanchandra DV <bhuvanchandra.dv@toradex.com>
|
|
Use platform independent description for requested GPIOs.
Signed-off-by: Stefan Agner <stefan.agner@toradex.com>
|
|
Fix chip reset by sleeping long enought after reset. Do proper error
handling (free GPIO on failure). Use dev_* for message logging to get
similar messages for all fusion driver related errors and warnings.
Signed-off-by: Stefan Agner <stefan.agner@toradex.com>
|
|
To avoid warnings use the new I2C power management function for
suspend and resume.
Signed-off-by: Stefan Agner <stefan.agner@toradex.com>
[Use __maybe_unsed instead of ifdefs]
Signed-off-by: Bhuvanchandr DV <bhuvanchandra.dv@toradex.com>
|
|
Added platform data struct to define interrupt and reset GPIO. This
allows to initialize the touchscreen controller inside the driver
rather than in each platform and use the driver as a module.
Signed-off-by: Stefan Agner <stefan.agner@toradex.com>
|
|
When the next interrupt request apeares between the confirmation of
the previous (a write via I2C, fusion_F0710A_write_complete) and
the reenable of the GPIO interrupt, the driver hangs and no more
touch inputs are reported.
This patch moves the confirmation after the reenabling of the GPIO
interrupt.
Stefan Agner <stefan.agner@toradex.com>
|
|
Recent evdev X-Server input driver implementation complain when
only multitouch axes have been reported ("found only multitouch-axes.
That shouldn't happen."). Therefor also report the primary touch
detection with default axis.
Signed-off-by: Stefan Agner <stefan.agner@toradex.com>
|
|
This patch adds the multi-touch input driver for the TouchRevolution
Fusion 7 and 10 panels (See Fusion 7 and 10 drivers for Linux.pdf and
Linux Drivers Fusion 10.zip) as downloaded in 07.2014
Signed-off-by: Max Krummenacher <max.krummenacher@toradex.com>
Signed-off-by: Stefan Agner <stefan.agner@toradex.com>
Signed-off-by: Bhuvanchandra DV <bhuvanchandra.dv@toradex.com>
|
|
|
|
Remove DAPM widgets which are not used. It seems they have never
been used, so won't be missed by anybody. This fixes two kernel log
errors:
wm9712-codec wm9712-codec: ASoC: mux Differential Source has no paths
wm9712-codec wm9712-codec: ASoC: mux Capture Phone Mux has no paths
Signed-off-by: Stefan Agner <stefan.agner@toradex.com>
|
|
When entering and resuming from suspend-to-RAM (LPSTOP2), the
current pinmux and the automatic restore of the last GPIO state
lead to an "uncontrolled" hard reset. The wm9712 driver tries
to warm reset the codec, which succeeds due to the previous
hard reset. Since the warm reset succeeded, the wm9712 driver
will not restore the codecs registers, hence those will
accidentially stay in reset state.
To avoid this extra hard reset we need to leave the pad floating
during suspend (the external pull-down will make sure that the
line has a defined state). After suspend snd_soc_ac97_warm_reset
will request the pinctrl states "ac97-warm-reset" and "ac97-running",
we need to make sure that both states do not change the state of
the reset line since it would introduce a similar "uncontrolled"
hard reset. Therefor, remove pinctrl configurations for those
states.
With this change, we disable the output driver of the reset signal
when entering suspend mode ("sleep" pinctrl state).
Current Colibri VF61 1.2A module have an external pull-down, hence
the reset signal will be low during suspend. With this change, the
reset signal will stay low during resume even when the GPIO state
gets restored (since the output driver is still disabled). This
will keep the codec in reset, and the wm9712 drivers warm reset will
fail. The driver will fall back to an explicit cold reset and due
to the cold reset restore the registers properly.
Future Colibri VF61 modules could use an external pull-up. This
would mean that the warm reset on resume will succeed, and the
driver would not need to restore the registers.
Note: pinctrl with the AC97 pin states is rather delicate. It seems
that when playing audio, the stack always selects the "default"
state. Before, the "running" state has been used as "default" state.
However, the "running" state is harmful during boot for initial cold
reset (maybe test mode?). Therefor we use a minimalistic "default"
pinmux. Only when the stack explicitily requests "running", we
actually mux the real functions on the affected pins.
Note 2: If there is a pull-up, we relly on the pull-up after suspend
and resume since the cold reset function does not explicitly set the
GPIO again...
Signed-off-by: Stefan Agner <stefan.agner@toradex.com>
|
|
The function snd_soc_register_card in soc-core.c selects the sleep
pinctrl state if a DAI is not marked as active. Since an AC97
controller needs the link also for bus control, we should not use
the sleep pinctrl state during normal operation. Work around this
by increment the active field on probe.
Signed-off-by: Stefan Agner <stefan.agner@toradex.com>
|
|
If the DMA has been activated but has not yet been started (e.g.
due to missing bit clock), a read or write would loop forever.
Avoid this by using a timeout counter.
The AC97 core does opportunistic reads in some cases (e.g. when
trying to determine if a cold reset is necessary, by just reading
the ID register after the warm reset).
Signed-off-by: Stefan Agner <stefan.agner@toradex.com>
|
|
Do not rely on RX/TX DMA running synchrounous. This allows to wait
until the answer frame has been received or timeout after 10 frames.
In the write case, we wait until the write request has been
transmitted to the TX FIFO. This changes should make AC97 read/write
more robust.
Signed-off-by: Stefan Agner <stefan.agner@toradex.com>
|
|
Mark the RX FIFO register as precious to avoid AC97 to get
asynachrounous (e.g. when reading regmap sysfs debug files).
Signed-off-by: Stefan Agner <stefan.agner@toradex.com>
|
|
Implement suspend and resume function which disable/enable generated
master clock explicitly and handle the audio clock too. Use regmap
cache to restore the register content on resume from lowest power
modes.
Signed-off-by: Stefan Agner <stefan.agner@toradex.com>
|
|
Lowest power modes loose the content of the registers. Use regmap
cache to restore the register entries on resume from lowest power
modes.
Signed-off-by: Stefan Agner <stefan.agner@toradex.com>
|
|
Allow playback and capture simultanously by handling two independent
instances of imx_pcm_runtime_data. Move the atomic triggers outside
of imx_pcm_runtime_data.
Signed-off-by: Stefan Agner <stefan.agner@toradex.com>
|
|
This adds initial audio capture support. A capturing sampling rate
between 8 to 48kHz is supported.
Signed-off-by: Stefan Agner <stefan.agner@toradex.com>
|
|
Add machine driver for Wolfson WM9712 AC97 codec connected to
Freescale SAI SoC audio transceiver. This combination needs
software emulated AC97 which is done by the fsl_sai_ac97
driver.
Signed-off-by: Stefan Agner <stefan.agner@toradex.com>
|
|
Add suspend/resume support for SAI AC97. Currently, both DMA
directions (RX/TX) will be disabled on suspend and completely
reinitialized on resume.
Signed-off-by: Stefan Agner <stefan.agner@toradex.com>
|
|
Initial support for AC97 using SAI peripheral. The SAI peripheral
does not support AC97 completely: Only the frame synchronization
and different slot length is taken care of, but all AC97 slots are
concatenated in one stream in a single FIFO. Hence the AC97 control
slots need to be handled completely in software. To do this in a
efficient manor the driver uses a ring buffer and prepares a whole
buffer at a time.
Due to the nature how variable sample rates are handled in AC97,
this software AC97 only supports AC97 native sample rate which is
48000Hz. However, the device announces this circumstances and the
upper stack should handle that correctly. Measurements showed a
overhead of about 3% CPU when playing a file with 44.1kHz sample
rate.
Signed-off-by: Stefan Agner <stefan.agner@toradex.com>
|
|
This adds a clock driver to use the bitclock (TX_BCLK) as a master
clock.
[4.4: added clk_prepare_enable call for mclk1]
Signed-off-by: Stefan Agner <stefan.agner@toradex.com>
|
|
Activate SAI0/SAI2 according to our needs and create a sound node
which enables the Wolfson WM9712 compatible AC97 codec.
[4.4: squashed capture audio routing fix]
Signed-off-by: Stefan Agner <stefan.agner@toradex.com>
|
|
This adds the remaining SAI instances SAI0, SAI1 and SAI3. All
instances are very similar, except that the DMA channel of SAI3
is available on MUX1 (compared to MUX0 for SAI0-SAI2). Also,
SAI3 has a slightly different memory map due to a deeper FIFO,
however in practice the current driver works for SAI3 fine.
Signed-off-by: Stefan Agner <stefan@agner.ch>
Signed-off-by: Shawn Guo <shawnguo@kernel.org>
(cherry picked from commit 4b336f1be2adc6dd2a1ad7d6fa7d4e16b2cd74e1)
|
|
|
|
The DMA implementation has been and is still affected by bugs and
race conditions. To bridge the time until those have been addressed,
add fsl-lpuart.nodma module parameter which allows to disable the
UART DMA using a kernel argument. Let the DMA to be disabled by
default.
Signed-off-by: Stefan Agner <stefan.agner@toradex.com>
|
|
Commit 8e4934c6d6c6 ("tty: serial: fsl_lpuart: clear receive flag
on FIFO flush") implemented clearing of the receive flag by reading
the status register only. It turned out that even though we flush
the FIFO afterwards, a explicit read of the data register is still
required.
This leads to a FIFO underrun. To avoid this, follow the advice in
the overrun "Operation section": Unconditionally clear RXUF after
using RXFLUSH.
Signed-off-by: Stefan Agner <stefan.agner@toradex.com>
|
|
Currently the tx_empty callback only considers the Transmit Complete
Flag (TC). The reference manual is not quite clear if the TC flag
covers the TX FIFO too. Debug prints on real hardware have shown that
from time to time the TC flag is asserted (indicating Transmitter
idle) while there are still data in the TX FIFO. Hence, in this case
the serial core will call the shutdown callback even though there are
data remaining in the TX FIFO buffers.
Avoid early shutdowns by considering the TX FIFO empty flag too. Also
avoid theoretical race conditions between DMA and the driver by
checking whether the TX DMA is in progress too.
Signed-off-by: Stefan Agner <stefan.agner@toradex.com>
|
|
Add pinmux for UART_A RTS, CTS pin's.
Signed-off-by: Bhuvanchandra DV <bhuvanchandra.dv@toradex.com>
Acked-by: Stefan Agner <stefan@agner.ch>
Signed-off-by: Shawn Guo <shawnguo@kernel.org>
(cherry picked from commit 6914a64eef2c6a03768738525ed0ebbe637234a0)
|
|
The Freescale Vybrid SoC has GPIO capabilities as part of the
IOMUXC. To enable GPIO's, the gpio_request_enable callback has
been implemented, however the corsponding gpio_disable_free
callback is missing. So far, disabling (unexporting) a GPIO left
the pin in its last state.
Implement a proper gpio_disable_free function which clears the
three enable bits which influence the state (IBE, OBE and PUE).
Signed-off-by: Stefan Agner <stefan.agner@toradex.com>
|
|
Specify suspend resume functions for the PIT block. Without these
functions which take care of enabling the PIT block and restoring
the clock value, suspend resume does not work when using PIT as
the clocksource instead of the ARM Global Timer.
Signed-off-by: Sanchayan Maity <maitysanchayan@gmail.com>
Signed-off-by: Stefan Agner <stefan.agner@toradex.com>
|
|
Add alias for FEC ethernet on Vybrid to allow bootloaders (like U-Boot)
patch-in the MAC address using this alias.
Signed-off-by: Stefan Agner <stefan@agner.ch>
Signed-off-by: Shawn Guo <shawnguo@kernel.org>
(cherry picked from commit ba734935937cfe29463b4e03906238944876b6de)
|
|
All Freescale Vybrid SoC include a Cortex-A5 core which supports
ARM's standard PMU (performance monitoring unit). Include the
monitoring unit into the Cortex-A5 base device tree vf500.dtsi.
Signed-off-by: Stefan Agner <stefan.agner@toradex.com>
|
|
The Colibri standard does not define a pin for SD-Card write-
protection. Use the disable-wp property to indicate that there
is no physical WP line present.
Signed-off-by: Stefan Agner <stefan.agner@toradex.com>
|
|
The calls to iounmap require a pointer to __iomem. The element
vbase is already a pointer to __iomem, there is no need to
dereference the pointer again.
Signed-off-by: Stefan Agner <stefan.agner@toradex.com>
|
|
The current probe code ignores a missing L2 cache, but still returns
ENODEV. Avoid returning ENODEV if a L2 cache is missing by setting
ret to 0.
Signed-off-by: Stefan Agner <stefan.agner@toradex.com>
|
|
The second argument of the mutex_lock_nested() helper is only
evaluated if CONFIG_DEBUG_LOCK_ALLOC is set. Otherwise we
get this build warning for the new regulator_lock_supply
function:
drivers/regulator/core.c: In function 'regulator_lock_supply':
drivers/regulator/core.c:142:6: warning: unused variable 'i' [-Wunused-variable]
To avoid the warning, this restructures the code to make it
both simpler and to move the 'i++' outside of the mutex_lock_nested
call, where it is now always used and the variable is not
flagged as unused.
We had some discussion about changing mutex_lock_nested to an
inline function, which would make the code do the right thing here,
but in the end decided against it, in order to guarantee that
mutex_lock_nested() does not introduced overhead without
CONFIG_DEBUG_LOCK_ALLOC.
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Fixes: 9f01cd4a915 ("regulator: core: introduce function to lock regulators and its supplies")
Link: http://permalink.gmane.org/gmane.linux.kernel/2068900
Signed-off-by: Mark Brown <broonie@kernel.org>
(cherry picked from commit fa731ac7ea04a7d3a5c6d2f568132478c02a83b3)
Signed-off-by: Stefan Agner <stefan.agner@toradex.com>
|
|
In some architectures the L2 cache controller is integrated in the
processor's block itself and it doesn't use any external cache
controller. This means that an entry in the board's dtb related
to the l2c is not necessary.
Distinguish between error codes and do not print anything in case
l2x0_of_init() doesn't find any L2C DTB entry and returns -ENODEV.
This patch mutes the following error message:
L2C: failed to init: -19
on boards like odroid-xu4, cortex A7/A15, which don't have
external cache controller.
Signed-off-by: Andi Shyti <andi.shyti@samsung.com>
Reported-by: Krzysztof Kozlowski <k.kozlowski@samsung.com>
Reviewed-by: Krzysztof Kozlowski <k.kozlowski@samsung.com>
Signed-off-by: Stefan Agner <stefan.agner@toradex.com>
|
|
Add Linux 4.4 default kernel configuration for Colibri VF50/VF61
modules.
Signed-off-by: Stefan Agner <stefan.agner@toradex.com>
|
