| Commit message (Collapse) | Author | Age | Files | Lines |
|\
| |
| |
| |
| |
| |
| |
| |
| | |
/ptx/work/octopus/lfu/kernel/linux-2.6-git-1 into phytec-master
Conflicts:
drivers/mfd/Makefile
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
|
| |
| |
| |
| |
| |
| |
| |
| | |
mc13783_ts does three measurments for each coordinates. Comparation of these
values can indicate if there're any erroneous sampel. We now add a tolerance
limit and add comparation routines.
Signed-off-by: Luotao Fu <l.fu@pengutronix.de>
|
| |
| |
| |
| |
| |
| |
| |
| | |
The AD Converter Unit on pmic mc13783 supports also the touchscreen interface
besides "normal" conversion. To add the ADC functionality, we have to change
the touchscreen code also to avoid conflicts.
Signed-off-by: Luotao Fu <l.fu@pengutronix.de>
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
mc13783 core uses the drv_data member to store the pointer to mc13783 struct
in every platformdevice, which it registers for the subdevices. However the
devcises resets the pointer in their platform device to store the pointer to
its own priv struct as soon as it is loaded. This leads to the device driver
leaves invalid addess in its platform device once its unloaded. This way the
drivers can only loaded once since it'll need the pointer to mc13783 on
loading. We now use platform_data to store this pointer instead and let the
drivers use drv_data for their own purpose.
Signed-off-by: Luotao Fu <l.fu@pengutronix.de>
|
| |
| |
| |
| | |
Signed-off-by: Luotao Fu <l.fu@pengutronix.de>
|
| |
| |
| |
| |
| |
| | |
Signed-off-by: Jan Weitzel <J.Weitzel@phytec.de>
Acked-by: Luotao Fu <l.fu@pengutronix.de>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
|
| |
| |
| |
| |
| |
| | |
only keep mainline and the drivers we need for phytec
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
|
|\ \
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
Conflicts:
arch/arm/Kconfig
arch/arm/mach-mx2/Kconfig
arch/arm/mach-mx2/Makefile
arch/arm/mach-mx2/clock_imx27.c
arch/arm/mach-mx2/devices.c
arch/arm/mach-mx2/devices.h
arch/arm/mach-mx2/pca100.c
arch/arm/mach-mx2/pcm038.c
arch/arm/mach-mx2/pcm970-baseboard.c
arch/arm/mach-mx3/Kconfig
arch/arm/mach-mx3/Makefile
arch/arm/mach-mx3/devices.c
arch/arm/mach-mx3/devices.h
arch/arm/mach-mx3/mx31moboard.c
arch/arm/mach-mx3/pcm037.c
arch/arm/mach-mx3/pcm043.c
arch/arm/plat-mxc/Kconfig
arch/arm/plat-mxc/include/mach/debug-macro.S
drivers/mfd/mc13783-core.c
drivers/net/Kconfig
drivers/net/can/Kconfig
drivers/net/can/Makefile
drivers/net/can/dev.c
drivers/net/can/mcp251x.c
drivers/net/can/sja1000/Makefile
drivers/net/can/sja1000/ems_pci.c
drivers/net/can/sja1000/kvaser_pci.c
drivers/net/can/sja1000/sja1000.c
drivers/net/can/sja1000/sja1000.h
drivers/net/can/sja1000/sja1000_platform.c
drivers/regulator/Kconfig
drivers/usb/gadget/fsl_udc_core.c
drivers/video/imxfb.c
drivers/watchdog/Makefile
include/linux/can/dev.h
sound/arm/imx2-pcm.c
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
|
| | |
| | |
| | |
| | | |
Signed-off-by: Luotao Fu <l.fu@pengutronix.de>
|
| | |
| | |
| | |
| | | |
Signed-off-by: Luotao Fu <l.fu@pengutronix.de>
|
| | |
| | |
| | |
| | | |
Signed-off-by: Luotao Fu <l.fu@pengutronix.de>
|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
> Hi,
>
> I have mcp2515 chip (socketcan revision 970) with at91sam9261 (olimex
> board) and when I want to get state of the interface by typing:
>
> cat /sys/devices/platform/atmel_spi.0/spi0.3/net/can0/can_state
>
> I get this:
>
> BUG: scheduling while atomic: cat/387/0x00000002
this is my conjecture: what happens here is that some function in the
stack dump you posted calls mcp251x_do_get_state in an atomic context.
After some grepping I found that one culprit (but I'm not sure it's
alone) is can_dev_show doing a read_lock (I guess you have preemption
on otherwise the spinlock should vanish ... you don't have a SMP ARM,
do you?). The mcp251x works on a slow bus (SPI) which needs an
interrupt to finish register read so it sleeps until it's delivered
(we cannot busy loop since it would waste tons of CPU time). But in
Linux you can't sleep (or better, you cannot reschedule) while holing
a spinlock. To say the truth this is an unfortunate situation that is
quite common when dealing with devices over such slow buses (I2C, SPI,
etc.). So, we can:
1) rewrite the function above lockless. It's not that easy.
2) you can try the patch attached (but keep in mind I haven't tested
it right now, I'm away from the hardware). But beware, the reading
could become stale. What can happen is that the bus recovers but the
entry is not updated (the other way is not possible because a bus
error triggers an interrupt and the state is update). Unfortunately we
don't have a reliable way to know if we are in an atomic context (see
http://lwn.net/Articles/274695/ ) to make the guess more rightful.
I don't see other ways for solving this problem right now. Perhaps
some more skilled kernel hacker can give a better advice. I love to
hear opinions on this since I met this problem myself a couple of
times.
Signed-off-by: Daniel Lewandowski <daniel@raciazek.eu>
|
| | |
| | |
| | |
| | | |
Signed-off-by: Luotao Fu <l.fu@pengutronix.de>
|
| | |
| | |
| | |
| | | |
Signed-off-by: Luotao Fu <l.fu@pengutronix.de>
|
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
This one is ripped of latest berlios socketcan tree (r1010) with some
minor modifications to get it run with our stack in 2.6.30-rc7-ptx-mxc1
Signed-off-by: Luotao Fu <l.fu@pengutronix.de>
|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
This is needed for MMC cards on i.MX31. Not sure why this
is needed, but otherwise we get a
mmc0: error -110 whilst initialising MMC card
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
We can now control the gpio driven cs line on the both usb phys to ensure
that they are initialized correctly. Hence there's no need to split up
the usb pin mux definition out of pca100_pins any more.
Signed-off-by: Luotao Fu <l.fu@pengutronix.de>
|
| | |
| | |
| | |
| | | |
Signed-off-by: Luotao Fu <l.fu@pengutronix.de>
|
| | |
| | |
| | |
| | |
| | |
| | | |
add clock and device definition for otg device mode usage on mx2
Signed-off-by: Luotao Fu <l.fu@pengutronix.de>
|
| | |
| | |
| | |
| | |
| | |
| | | |
fix name of usb otg interrupt and usage in device
Signed-off-by: Luotao Fu <l.fu@pengutronix.de>
|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
deactivate usb phy by pulling up the chipselect line during initialization of
usb host controller to prevent phy crash caused by undefined behaviour of mx2
usb controller during initialization
Signed-off-by: Luotao Fu <l.fu@pengutronix.de>
|
| | |
| | |
| | |
| | | |
Signed-off-by: Luotao Fu <l.fu@pengutronix.de>
|
| | |
| | |
| | |
| | |
| | |
| | | |
add ehci platform device registration and select ulpi
Signed-off-by: Luotao Fu <l.fu@pengutronix.de>
|
| | |
| | |
| | |
| | | |
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
|
| | |
| | |
| | |
| | | |
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
|
| | |
| | |
| | |
| | | |
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
|
| | |
| | |
| | |
| | | |
Signed-off-by: Luotao Fu <l.fu@pengutronix.de>
|
| | |
| | |
| | |
| | |
| | | |
Signed-off-by: Luotao Fu <l.fu@pengutronix.de>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
Add support for the MT9T031 CMOS camera sensor from Aptina to the PCM037
board. Also add two I2C iomux pin definitions, needed for pcm037
Changes since v1: removed #ifdef CONFIG_I2C_IMX* checks, added a call to
dma_free_coherent() on the error path.
Signed-off-by: Guennadi Liakhovetski <lg@denx.de>
Signed-off-by: Luotao Fu <l.fu@pengutronix.de>
|
| | |
| | |
| | |
| | |
| | | |
Signed-off-by: Valentin Longchamp <valentin.longchamp@epfl.ch>
Signed-off-by: Luotao Fu <l.fu@pengutronix.de>
|
| | |
| | |
| | |
| | | |
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
|
| | |
| | |
| | |
| | | |
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
|
| | |
| | |
| | |
| | | |
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
|
| | |
| | |
| | |
| | | |
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
|
| | |
| | |
| | |
| | | |
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
|
| | |
| | |
| | |
| | | |
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
|
| | |
| | |
| | |
| | | |
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
|
| | |
| | |
| | |
| | | |
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
|
| | |
| | |
| | |
| | | |
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
|
| | |
| | |
| | |
| | | |
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
|
| | |
| | |
| | |
| | | |
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
|
| | |
| | |
| | |
| | | |
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
|
| | |
| | |
| | |
| | | |
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
|
| | |
| | |
| | |
| | |
| | |
| | | |
By default it is not connected on the baseboard
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
|
| | |
| | |
| | |
| | | |
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
|
| | |
| | |
| | |
| | | |
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
|
| | |
| | |
| | |
| | | |
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
|
| | |
| | |
| | |
| | | |
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
HACK HACK HACK
Without this patch sending CAN packages faster then they could be
send out on the wire results in -ENOBUF. This patch increases the
default tx_queue_len to 10000 so that in this case the sender blocks.
Thus the CAN socket just behaves like congested a TCP/IP socket.
HACK HACK HACK
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
|
| | |
| | |
| | |
| | | |
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
|