If we call clk_get() from arch_reset we get ugly messages before
reboot.
Signed-off-by: Tony Lindgren <tony@atomide.com>

Backport patch 68aa2c0d.
Without this patch sysrq does not work with 8250 serial console.

"clocks" should be static. It is currenly needed only for
CONFIG_OMAP_RESET_CLOCKS. A follow-up patch will fix
CONFIG_OMAP_RESET_CLOCKS while keeping clocks static.
Signed-off-by: Tony Lindgren <tony@atomide.com>

Some time critical work may be put onto the keventd workqueue so let's
not block it with msleep, use a dedicated workqueue instead.
Signed-off-by: Imre Deak <imre.deak@solidboot.com>
Signed-off-by: Juha Yrjola <juha.yrjola@solidboot.com>

ls041y3 needs the PCLK,HS,VS signals for a while after sending the sleep
in command, so sleep 50 msec.
Signed-off-by: Imre Deak <imre.deak@solidboot.com>
Signed-off-by: Juha Yrjola <juha.yrjola@solidboot.com>

Otherwise it will spin forever.
Signed-off-by: Imre Deak <imre.deak@solidboot.com>
Signed-off-by: Juha Yrjola <juha.yrjola@solidboot.com>

Add support for assigning memory regions dynamically to peripherals
attached to GPMC interface. Platform specific code should now call
gpmc_cs_request to get a free GPMC memory region instead of using
a fixed address.

Make the H4 and Apollon platform initialization use the new API.
Signed-off-by: Imre Deak <imre.deak@solidboot.com>
Signed-off-by: Juha Yrjola <juha.yrjola@solidboot.com>

Reprogramming takes places before putting the CPU into idle mode if
the dynamic tick option is enabled. The timer is then set to expire
at the next pending timer event. Because some time has already passed
since the last reported jiffy we have to wait less than the time
specified in jiffies.

Also make sure we don't set a load value of 0 whose outcome is
unspecified according to the TRM.
Signed-off-by: Imre Deak <imre.deak@solidboot.com>
Signed-off-by: Juha Yrjola <juha.yrjola@solidboot.com>

FB kernel virtual address was left unset causing the FB console crash on
OMAP1.
Signed-off-by: Imre Deak <imre.deak@solidboot.com>

The PWM block on the TI OMAP CPU can drive some external GPIOs.  LEDs
can be connected to these GPIOs.
Signed-off-by: Juha Yrjola <juha.yrjola@solidboot.com>

On some LCDs leaving the Vref on provides much better readings.
Signed-off-by: Jarkko Oikarinen <jarkko.oikarinen@nokia.com>
Signed-off-by: Imre Deak <imre.deak@solidboot.com>
Signed-off-by: Juha Yrjola <juha.yrjola@solidboot.com>

Some LCDs like the LS041Y3 require a customized filtering
logic for reliable readings, so make the filtering function
replacable through platform specific hooks.
Signed-off-by: Imre Deak <imre.deak@solidboot.com>
Signed-off-by: Juha Yrjola <juha.yrjola@solidboot.com>

Use hrtimer instead of the normal timer, since it provides better
sampling resolution. This will:

- avoid a problem where we have a 1 jiffy poll period and dynamic tick on
- utilize high resolution HW clocks when they are added to the hrtimer
  framework
Signed-off-by: Imre Deak <imre.deak@solidboot.com>
Signed-off-by: Juha Yrjola <juha.yrjola@solidboot.com>

We can't depend on the pressure value to determine when the pen was
lifted, so use the IRQ line state instead.
Signed-off-by: Imre Deak <imre.deak@solidboot.com>
Signed-off-by: Juha Yrjola <juha.yrjola@solidboot.com>

DISPC DMA can, in certain conditions, underflow which will trigger an error
condition and halt the graphics pipeline. The only way to recover from
this would be to reset the related HW modules, which would abort the
current transfer and cause unwanted visual side-effects. The problem
arises mainly when enabling multiple planes with higher resolutions.

The arbitration priority of the DISPC DMA cannot be raised higher, thus -
so far - the only workaround is to keep the DMA FIFO full as much as
possible to minimize the risk of it being totally drained. The patch does
this for external LCD controllers connected to RFBI, where the new FIFO
setting doesn't degrade power consumption significantly.
Signed-off-by: Imre Deak <imre.deak@solidboot.com>
Signed-off-by: Juha Yrjola <juha.yrjola@solidboot.com>

Currently, IDLETIMER timer routine calls kobject_uevent, which might sleep.
Since we are in interrupt context, we now schedule a work when the timer
expires. This calls sysfs_notify() for notifying userspace.
Signed-off-by: Samuel Ortiz <samuel.ortiz@solidboot.com>
Signed-off-by: Juha Yrjola <juha.yrjola@solidboot.com>

In ipt_IDLETIMER.c, we are using an old prototype for xt_target hooks. We need
to pass one additional argument, a pointer to xt_target.
We also need to specify the targetsize value, otherwise registration fails.
Signed-off-by: Samuel Ortiz <samuel.ortiz@solidboot.com>
Signed-off-by: Juha Yrjola <juha.yrjola@solidboot.com>

Signed-off-by: Kevin Hilman <khilman@deeprooted.net>

Signed-off-by: Samuel Ortiz <samuel.ortiz@solidboot.com>
Signed-off-by: Juha Yrjola <juha.yrjola@solidboot.com>

Signed-off-by: Juha Yrjola <juha.yrjola@solidboot.com>

Signed-off-by: Juha Yrjola <juha.yrjola@solidboot.com>

When filtering is disabled the driver will ignore all samples and
never detect the pen up event.
Signed-off-by: Imre Deak <imre.deak@nokia.com>

Various fixes for the OHCI OMAP glue:

 - Properly use the usbcore HCD lifecycle entry points

 - Remove calls to non-exported usbcore routines.

 - Stop inappropriate export of some bus glue calls

Plus remove some extra whitespace.

Fix warnings

WARNING: drivers/pcmcia/omap_cf.o - Section mismatch:
reference to .init.text: from .data between '' (at offset
0x80) and 'omap_cf_ops'

WARNING: sound/arm/omap/snd-omap-alsa-aic23.o - Section
mismatch: reference to .init.text: from .data between ''
(at offset 0x68) and 'aic23_hw_constraints_rates'

Signed-off-by: Dirk Behme <dirk.behme_at_gmail.com>
Signed-off-by: Tony Lindgren <tony@atomide.com>

Signed-off-by: Juha Yrjola <juha.yrjola@solidboot.com>

Fix GPMC_CS_CONFIG1 writing and typos
Signed-off-by: Tony Lindgren <tony@atomide.com>

Signed-off-by: Samuel Ortiz <samuel.ortiz@solidboot.com>
Signed-off-by: Juha Yrjola <juha.yrjola@solidboot.com>

Signed-off-by: Samuel Ortiz <samuel.ortiz@solidboot.com>
Signed-off-by: Juha Yrjola <juha.yrjola@solidboot.com>

Fixes compilation errors on OMAP1.
Signed-off-by: Timo Teras <timo.teras@solidboot.com>
Signed-off-by: Juha Yrjola <juha.yrjola@solidboot.com>

- replace custom debugging stuff with the default kernel one
- pass a pointer to the LCD object instance to member functions
- rename lcd_lph8923 to the more generic lcd_mipid
- use SPI_MODE_1 instead of SPI_MODE_0 in lcd_mipid
- add ESD spike recovery support to lcd_mipid
Signed-off-by: Imre Deak <ext-imre.deak@nokia.com>
Signed-off-by: Juha Yrjola <juha.yrjola@solidboot.com>

- Support for
        - overlays through separate /dev/fbX nodes
        - scaling and color space conversion
        - per-plane memory configuration either in SRAM or SDRAM

- Replace custom debug stuff with the default kernel one
Signed-off-by: Imre Deak <ext-imre.deak@nokia.com>
Signed-off-by: Juha Yrjola <juha.yrjola@solidboot.com>

- OMAP ATAG changes:
        FBMEM - support for per plane memory configurations
        LCD   -  add data_lines, reset GPIO line params

- FB memory configuration:
        Support for per-plane memory configurations with FB location
        either in SRAM or SDRAM.

- DMA consistent memory size:
        Make the ARM DMA consistent memory size overridable. The default
        is 2MB, but for larger frame buffers we may need more. Kconfig
        option is added in the next patch.
Signed-off-by: Imre Deak <ext-imre.deak@nokia.com>
Signed-off-by: Juha Yrjola <juha.yrjola@solidboot.com>

Multiplexing for 24xx GPMC wait pin monitoring
Signed-off-by: Tony Lindgren <tony@atomide.com>

Signed-off-by: Kevin Hilman <khilman@deeprooted.net>

Signed-off-by: Kevin Hilman <khilman@deeprooted.net>

MT_MEMORY is needed in order to execute code in SRAM. Also
need to use section mapping.

Based on a patch by Richard Woodruff.
Signed-off-by: Tony Lindgren <tony@atomide.com>