armv7: stronger barrier for cache-maintenance operations

set-way operations need a DSB after them to ensure the
operation is complete. DMB may not be enough. Use DSB
after all operations instead of DMB.

Signed-off-by: Aneesh V <aneesh@ti.com>

omap: enable caches at system start-up

Signed-off-by: Aneesh V <aneesh@ti.com>

arm: do not force d-cache enable on all boards

c2dd0d45540397704de9b13287417d21049d34c6 added dcache_enable()
to board_init_r(). This enables d-cache for all ARM boards.
As a result some of the arm boards that are not cache-ready
are broken. Revert this change and allow platform code to
take the decision on d-cache enabling.

Also add some documentation for cache usage in ARM.

Signed-off-by: Aneesh V <aneesh@ti.com>

ORIGEN: Add MMC SPL support

Adds mmc boot support.

Signed-off-by: Chander Kashyap <chander.kashyap@linaro.org>
Signed-off-by: Minkyu Kang <mk7.kang@samsung.com>

ARMV7: Add support for Samsung ORIGEN board

Origen board is based upon S5PV310 SoC which is similiar to
S5PC210 SoC.

Signed-off-by: Chander Kashyap <chander.kashyap@linaro.org>
Signed-off-by: Minkyu Kang <mk7.kang@samsung.com>

i2c:gpio:s5p: Enable I2C GPIO on the GONI target

This patch enables the software I2C for GONI reference target.

Signed-off-by: Lukasz Majewski <l.majewski@samsung.com>
Signed-off-by: Kyungmin Park <kyungmin.park@samsung.com>
Signed-off-by: Minkyu Kang <mk7.kang@samsung.com>
Cc: Heiko Schocher <hs@denx.de>

i2c:gpio:s5p: I2C GPIO Software implementation (via soft_i2c)

This patch adds support for software I2C for GONI and Universal C210 reference targets.
It adds support for access to GPIOs by number, not as it is present,
by bank and offset.

Signed-off-by: Lukasz Majewski <l.majewski@samsung.com>
Signed-off-by: Kyungmin Park <kyungmin.park@samsung.com>
Signed-off-by: Minkyu Kang <mk7.kang@samsung.com>
Cc: Heiko Schocher <hs@denx.de>

Tegra2: Use clock and pinmux functions to simplify code

Signed-off-by: Simon Glass <sjg@chromium.org>

Tegra2: Add additional pin multiplexing features

This adds an enum for each pin and some functions for changing the pin
muxing setup.

Signed-off-by: Simon Glass <sjg@chromium.org>

Tegra2: Add more clock support

This adds functions to enable/disable clocks and reset to on-chip peripherals.

Signed-off-by: Simon Glass <sjg@chromium.org>

Tegra2: Add microsecond timer function

These functions provide access to the high resolution microsecond timer
and tidy up a global variable in the code.

Signed-off-by: Simon Glass <sjg@chromium.org>

ARM: remove broken "at91rm9200dk" board

Signed-off-by: Wolfgang Denk <wd@denx.de>

ARM: remove broken "m501sk" board

Signed-off-by: Wolfgang Denk <wd@denx.de>

ARM: remove broken "kb9202" board

Signed-off-by: Wolfgang Denk <wd@denx.de>

ARM: remove broken "csb637" board

Signed-off-by: Wolfgang Denk <wd@denx.de>

ARM: remove broken "cmc_pu2" board

Signed-off-by: Wolfgang Denk <wd@denx.de>

ARM: remove broken "at91cap9adk" board

Signed-off-by: Wolfgang Denk <wd@denx.de>
Cc: Stelian Pop <stelian.pop@leadtechdesign.com>

ARM: remove broken "voiceblue" board

Signed-off-by: Wolfgang Denk <wd@denx.de>

ARM: remove broken "smdk2400" board

Signed-off-by: Wolfgang Denk <wd@denx.de>
Cc: Gary Jennejohn <garyj@denx.de>

ARM: remove broken "sbc2410x" board

Signed-off-by: Wolfgang Denk <wd@denx.de>

ARM: remove broken "netstar" board

Signed-off-by: Wolfgang Denk <wd@denx.de>

ARM: remove broken "mx1fs2" board

Signed-off-by: Wolfgang Denk <wd@denx.de>

ARM: remove broken "lpd7a40x" boards

Remove lpd7a400 and lpd7a404 boards.

Signed-off-by: Wolfgang Denk <wd@denx.de>

ARM: remove broken "edb93xx" boards

Remove edb9301, edb9302, edb9302a, edb9307, edb9307a, edb9312,
edb9315 and edb9315a boards.

Signed-off-by: Wolfgang Denk <wd@denx.de>

ARM: remove broken "B2" board

Signed-off-by: Wolfgang Denk <wd@denx.de>
Cc: Andrea Scian <andrea.scian@dave-tech.it>

ARM: remove broken "armadillo" board

Signed-off-by: Wolfgang Denk <wd@denx.de>
Cc: Rowel Atienza <rowel@diwalabs.com>

ARM: remove broken "assabet" board

Signed-off-by: Wolfgang Denk <wd@denx.de>
Cc: George G. Davis <gdavis@mvista.com>

ARM: versatile: drop warnings

Drop warning: "passing argument 1 of 'get_ram_size'
discards qualifiers from pointer target type"

Signed-off-by: Stefano Babic <sbabic@denx.de>
CC: Albert Aribaud <albert.u.boot@aribaud.net>

IMX: scb9328: drop warnings

Drop warning: "passing argument 1 of 'get_ram_size'
discards qualifiers from pointer target type"

Signed-off-by: Stefano Babic <sbabic@denx.de>
CC: Torsten Koschorrek <koschorrek@synertronixx.de>

MX31: imx31_litekit: make use of GPIO framework

Signed-off-by: Stefano Babic <sbabic@denx.de>

MX31: mx31ads: make use of GPIO framework

Signed-off-by: Stefano Babic <sbabic@denx.de>

MX5: mx51evk: make use of GPIO framework

Signed-off-by: Stefano Babic <sbabic@denx.de>

MX35: mx35pdk: make use of GPIO framework

Signed-off-by: Stefano Babic <sbabic@denx.de>

MX5: mx53loco: make use of GPIO framework

Signed-off-by: Stefano Babic <sbabic@denx.de>
CC: Jason Liu <r64343@freescale.com>

MX5: mx53evk: make use of GPIO framework

Signed-off-by: Stefano Babic <sbabic@denx.de>
CC: Jason Liu <r64343@freescale.com>

MX5: vision2: make use of GPIO framework

Signed-off-by: Stefano Babic <sbabic@denx.de>

MX5: mx53smd: make use of GPIO framework

Signed-off-by: Stefano Babic <sbabic@denx.de>
CC: Fabio Estevam <fabio.estevam@freescale.com>

MX5: mx53ard: make use of GPIO framework

Signed-off-by: Stefano Babic <sbabic@denx.de>
CC: Fabio Estevam <fabio.estevam@freescale.com>

MX25: zmx25: make use of GPIO framework

Signed-off-by: Stefano Babic <sbabic@denx.de>
CC: Matthias Weisser <weisserm@arcor.de>

MX5: efikamx: make use of GPIO framework

Signed-off-by: Stefano Babic <sbabic@denx.de>
CC: Marek Vasut <marek.vasut@gmail.com>

MX31: QONG: make use of GPIO framework

Signed-off-by: Stefano Babic <sbabic@denx.de>

MX35: make use of GPIO framework for MX35 processor

Signed-off-by: Stefano Babic <sbabic@denx.de>

MX5: make use of GPIO framework for MX5 processor

Signed-off-by: Stefano Babic <sbabic@denx.de>

MX31: make use of GPIO framework for MX31 processor

Signed-off-by: Stefano Babic <sbabic@denx.de>

MX25: make use of GPIO framework for MX25 processor

Signed-off-by: Stefano Babic <sbabic@denx.de>

IMX: uniform GPIO interface using GPIO framework

IMX processors has a slightly different interface
to access GPIOs and do not make use of the provided GPIO
framework. The patch substitutes mxc_ specific
functions and make use of the API in asm/gpio.h

Signed-off-by: Stefano Babic <sbabic@denx.de>

MX: MX35 / MX5: uniform clock command with powerpc

There was already a command to show the processor clocks
for PowerPC (clocks). For i.MX, the "clockinfo" command
was introduce. The patch sets the same command name used on
PowerPC.
A nasty and not needed newline is also dropped in the help for
the command.

Signed-off-by: Stefano Babic <sbabic@denx.de>

MX35: MX35PDK: support additional RAM on CSD1

Modules on mx35pdk have additional 128MB
memory connected to CSD1.

Signed-off-by: Stefano Babic <sbabic@denx.de>

mx53: ddr3: Update DD3 initialization

Updated mx53 ddr3 script in order to align with the latest Freescale version from July 8, 2011:
-change ESDREF[REF_SEL]=01 (for 32KHz), from incorrect setting of 00 (64KHz)
-change DDR3 MR0 write to "setmem /32 0x63fd901c = 0x052080b0" from
"0x092080b0". This changes write recovery from 8 clocks to 6 clocks
(in line with ESDCFG1[tWR])

Signed-off-by: Lily Zhang <r58066@freescale.com>
Signed-off-by: Fabio Estevam <fabio.estevam@freescale.com>

ARM: MX51: PLL errata workaround

This is a port of the official PLL errata workaround from Freescale to
mainline u-boot.
The PLL's in the i.MX51 processor can go out of lock due to a metastable
condition in an analog flip-flop when used at high frequencies.
This workaround implements an undocumented feature in the PLL (dither
mode), which causes the effect of this failure to be much lower (in terms
of frequency deviation), avoiding system failure, or at least decreasing
the likelihood of system failure.

Signed-off-by: David Jander <david@protonic.nl>

ARM: versatilepb : drop warnings due to double definitions

CONFIG_ARCH_VERSATILE_PB  is defined twice - drop
the define from config.h.

Signed-off-by: Stefano Babic <sbabic@denx.de>

omap4: increase SRAM budget to fix build error

Signed-off-by: Aneesh V <aneesh@ti.com>
Cc: Dirk Behme <dirk.behme@googlemail.com>
Cc: Sandeep Paulraj <s-paulraj@ti.com>
Acked-by: Dirk Behme<dirk.behme@googlemail.com>

omap4: fix build warning due to signed unsigned comparison

Signed-off-by: Aneesh V <aneesh@ti.com>

mkimage: Fix 'Unknown OMAP image type - 5'

Using mkimage with e.g.

tools/mkimage -A arm -T firmware -O u-boot -d u-boot.bin foo.img

gives a warning

"Unknown OMAP image type - 5"

while it seems that the image itself is created successfully.

This does come from the patch "mkimage: Add OMAP boot image support".

The method check_image_type in image_type_params is supposed to just
return success or failure.  However, for omap it also calls fprintf:

static int omapimage_check_image_types(uint8_t type)
{
if (type == IH_TYPE_OMAPIMAGE)
return EXIT_SUCCESS;
else {
fprintf(stderr, "Unknown OMAP image type - %x", type);
return EXIT_FAILURE;
}
}

All the other image checkers and no others have this, so the fix is to
simply remove the fprintf.

Signed-off-by: Dirk Behme <dirk.behme@googlemail.com>
CC: John Rigby <john.rigby@linaro.org>
CC: Aneesh V <aneesh@ti.com>
CC: Sandeep Paulraj <s-paulraj@ti.com>

omap: fix gpio related build breaks

Signed-off-by: Aneesh V <aneesh@ti.com>
Acked-by: Dirk Behme <dirk.behme@googlemail.com>

gpio:samsung: s5p_ suffix add for GPIO functions (C210_universal)

This is a cosmetic patch, which is changing the gpio_ prefix to
s5p_gpio_.

Signed-off-by: Lukasz Majewski <l.majewski@samsung.com>
Signed-off-by: Kyungmin Park <kyungmin.park@samsung.com>
Signed-off-by: Minkyu Kang <mk7.kang@samsung.com>

SMDKV310: MMC SPL: Remove unwanted dummy functions

Removed dummy functions in "mmc_spl/board/samsung/smdkv310/mmc_boot.c",
@mmc_boot.c
void do_undefined_instruction(struct pt_regs *pt_regs);
void do_software_interrupt(struct pt_regs *pt_regs);
void do_prefetch_abort(struct pt_regs *pt_regs);
void do_data_abort(struct pt_regs *pt_regs);
void do_not_used(struct pt_regs *pt_regs);
void do_fiq(struct pt_regs *pt_regs);
void do_irq(struct pt_regs *pt_regs);

not required as called conditionally in start.S
@start.S
\#ifdef CONFIG_SPL_BUILD
_undefined_instruction: .word _undefined_instruction
_software_interrupt:    .word _software_interrupt
_prefetch_abort:        .word _prefetch_abort
_data_abort:            .word _data_abort
_not_used:              .word _not_used
_irq:                   .word _irq
_fiq:                   .word _fiq
_pad:                   .word 0x12345678 /* now 16*4=64 */
\#else
_undefined_instruction: .word undefined_instruction
_software_interrupt:    .word software_interrupt
_prefetch_abort:        .word prefetch_abort
_data_abort:            .word data_abort
_not_used:              .word not_used
_irq:                   .word irq
_fiq:                   .word fiq
_pad:                   .word 0x12345678 /* now 16*4=64 */
\#endif
e.g.
undefined_instruction:
get_bad_stack
bad_save_user_regs
bl      do_undefined_instruction

Signed-off-by: Chander Kashyap <chander.kashyap@linaro.org>

SMDKV310: Fix undefined reference error

Fix buld error:
undefined reference to '__image_copy_end' and `save_boot_params'.

start.o: In function `_image_copy_end_ofs':
mmc_spl/board/samsung/smdkv310/start.S:44: undefined reference to `__image_copy_end'
start.o: In function `reset':
mmc_spl/board/samsung/smdkv310/start.S:137: undefined reference to `save_boot_params'

Signed-off-by: Chander Kashyap <chander.kashyap@linaro.org>
Signed-off-by: Minkyu Kang <mk7.kang@samsung.com>

SMDKV310: Fix build error for smdkv310 board

Fix build error for smdkv310 board:

board/samsung/smdkv310/smdkv310.c:126: undefined reference to `gpio_set_pull'

Signed-off-by: Chander Kashyap <chander.kashyap@linaro.org>
Signed-off-by: Minkyu Kang <mk7.kang@samsung.com>

gpio:samsung s5p_ suffix add for GPIO functions

This change is driven by need of general gpio_* functions,
which as their parameter are accepting the GPIO pin number, NOT
block and pin.

This makes the code alike to omap, and allows for using more
generic frameworks (e.g. software I2C).

Signed-off-by: Lukasz Majewski <l.majewski@samsung.com>
Signed-off-by: Kyungmin Park <kyungmin.park@samsung.com>
Signed-off-by: Minkyu Kang <mk7.kang@samsung.com>
Acked-by: Mike Frysinger <vapier@gentoo.org>

mmc: S5P: Support DMA restarts at buffer boundaries

Currently if a DMA buffer straddles a buffer alignment boundary
(512KiB) then the DMA engine will pause and generate a DMA
interrupt.  Since the DMA interrupt is not enabled it will hang
the MMC driver.

This patch adds support for restarting the DMA transfer.  The
SYSTEM_ADDRESS register contains the next address that would have
been read/written when a boundary is hit.  So we can read that
and write it back.  The write triggers the resumption of the
transfer.

Signed-off-by: Anton Staaf <robotboy@chromium.org>
Signed-off-by: Minkyu Kang <mk7.kang@samsung.com>
Tested-by : Jaehoon Chung <jh80.chung@samsung.com>
Cc: Albert ARIBAUD <albert.u.boot@aribaud.net>

SMDKV310: Fix host compilation of mkv310_image

Fix compilation of mkv310_image host tool

tools/mkv310_image.c: In function 'main':
tools/mkv310_image.c:67: error: 'S_IRUSR' undeclared (first use in this function)
tools/mkv310_image.c:67: error: (Each undeclared identifier is reported only once
tools/mkv310_image.c:67: error: for each function it appears in.)
tools/mkv310_image.c:67: error: 'S_IWUSR' undeclared (first use in this function)
tools/mkv310_image.c:67: error: 'S_IRGRP' undeclared (first use in this function)
tools/mkv310_image.c:67: error: 'S_IWGRP' undeclared (first use in this function)
tools/mkv310_image.c:67: error: 'S_IROTH' undeclared (first use in this function)
tools/mkv310_image.c:67: error: 'S_IWOTH' undeclared (first use in this function)

resulting from a 'make smdkv310_config'.

Signed-off-by: Dirk Behme <dirk.behme@googlemail.com>
Signed-off-by: Minkyu Kang <mk7.kang@samsung.com>

arm: fix bd pointer dereference prior initialization

gd->bd pointer has been used prior been initialized.
Move the relevant code after the initialization.

Signed-off-by: Igor Grinberg <grinberg@compulab.co.il>

arm, lib/board.c: use gd->ram_size instead of bd->bi_memsize

Signed-off-by: Heiko Schocher <hs@denx.de>
cc: Albert Aribaud <albert.u.boot@aribaud.net>

mx5: Remove CONFIG_L2_OFF and CONFIG_SYS_L2CACHE_OFF

CONFIG_L2_OFF is obsolete after the following commit:

e47f2db5371047eb9bcd115fee084e6a8a92a239
armv7: rename cache related CONFIG flags
Replace the cache related CONFIG flags with more meaningful
names. Following are the changes:
CONFIG_L2_OFF      -> CONFIG_SYS_L2CACHE_OFF

Since imx5 does not provide L2 cache operations(Enable/Disable)
Simply remove CONFIG_L2_OFF and CONFIG_SYS_L2CACHE_OFF

Signed-off-by: Jason Liu <jason.hui@linaro.org>
Cc:Stefano Babic <sbabic@denx.de>
Acked-by: Stefano Babic <sbabic@denx.de>

MX31: removed warnings due to clock.h

Signed-off-by: Stefano Babic <sbabic@denx.de>

integrator: convert to new build system

This deletes the integrator split_by_variant.sh script and
defines a number of unique board types for the core modules
that are meaningful to support for the Integrator AP/CP, i.e.
the ones that did not just say "unsupported core module" in
split_by_variant.sh. If more core modules need to be supported
they are easy to add.

We delete all the old cruft in Makefile and MAKEALL that was
working around the old way of building boards. We create a
unique config file per board to satisfy the build system, but
they are just oneliners that include the existing
integratorap.h and integratorcp.h configs.

Signed-off-by: Linus Walleij <linus.walleij@linaro.org>

integratorcp: make the board compile

This defines the requires CONFIG_SYS_* variables to make the
Integrator CP board compile.

Signed-off-by: Linus Walleij <linus.walleij@linaro.org>

integratorap: remove hardcoded 32MB memory cmdline

The default configuration for the Integrator AP forces memory to be
32 MB on the command line to the kernel, while we have perfect
information and detection of the actual memory size in the ATAGs.
Delete the confusion.

Signed-off-by: Linus Walleij <linus.walleij@linaro.org>

integratorap: support some rudimentary commands

This adds support for a subset of the default commands for the
Integrator, however since the card does not have Ethernet (unless
you plug in a PCI card) we can not use the default command set.

Signed-off-by: Linus Walleij <linus.walleij@linaro.org>

integratorap: support the hush shell

Give us some kind of sane shell environment so the bootloader can
be used.

Signed-off-by: Linus Walleij <linus.walleij@linaro.org>

integratorap: support relocation

The integrator board was apparently never converted over to support
relocation until now. After this the integrator u-boot both compiles
and boots on the Integrator AP.

This also fixes the SDRAM memory size detection.

Signed-off-by: Linus Walleij <linus.walleij@linaro.org>

integratorap: make the compile work again

The integratorap/cp config for u-boot was outdated and would not
even compile, so fix the obvious missing bits for it to start
building. After this "make ap920t_config/make all" starts working
again.

Signed-off-by: Linus Walleij <linus.walleij@linaro.org>

arm: lib: memcpy: Do not copy to same address

In some cases (e.g. bootm with a elf payload which is already at the right
position) there is a in place copy of data to the same address. Catching this
saves some ms while booting.

Signed-off-by: Matthias Weisser <weisserm@arcor.de>

arm: omap: innovator: use common code for machine type

Innovator and H2 boards used machine_is_* macros for setting the machine
type. These macros are expanded in compile time and thus leaves
unreachable code (though gcc might optimize it).
Switch them to use common code for machine type setting.

Signed-off-by: Igor Grinberg <grinberg@compulab.co.il>

arm: nvidia and smdk6400: use common code for machine type

NVIDIA boards and Samsung SMDK6400 already use a local variant of
CONFIG_MACH_TYPE option.
Switch to use the new common code.

Signed-off-by: Igor Grinberg <grinberg@compulab.co.il>

MAKEALL ARMv7: Use boards.cfg

Use the boards from boards.cfg for building ./MAKEALL ARMV7.

Signed-off-by: Dirk Behme <dirk.behme@googlemail.com>

atmel: Update support of board AT91SAM9M10G45-EK to new style

Based on earlier work by Alex Waterman <awaterman@dawning.com>.

Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

AT91: fix at91sam9g45.h to include USB Host defines

Signed-off-by: Reinhard Meyer <u-boot@emk-elektronik.de>

make pm9g45 buildable for v2011.06 release

Signed-off-by: Asen Chavdarov Dimov <dimov@ronetix.at>
Signed-off-by: Reinhard Meyer <u-boot@emk-elektronik.de>
changed at91_serial_hw_init to at91_seriald_hw_init

make pm9263 buildable for v2011.06 release

Signed-off-by: Asen Chavdarov Dimov <dimov@ronetix.at>

atmel: update at91sam9m10g45 SoC support to new style

Based on earlier work by Alex Waterman <awaterman@dawning.com>.

Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

pm9261: compiles with the AT91 reworked scheme

Signed-off-by: Asen Chavdarov Dimov <dimov@ronetix.at>

tny_a9260/tny_a9g20: update board to the new AT91 organization

Cc: Albin Tonnerre <tonnerrealbin@gmail.com>
CC: Gregory Hermant <gregory.hermant@calao-systems.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

sbc35_a9g20: update board to the new AT91 organization

Cc: Albin Tonnerre <tonnerrealbin@gmail.com>
Cc: Gregory Hermant <gregory.hermant@calao-systems.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Reinhard Meyer <u-boot@emk-elektronik.de>
Removed SBC35 from MAKEALL

at91: reworked support for meesc board

The meesc board support was broken. Within this opportunity, I completely
reworked the board files.

Signed-off-by: Daniel Gorsulowski <Daniel.Gorsulowski@esd.eu>
Signed-off-by: Matthias Fuchs <Matthias.Fuchs@esd.eu>

MPC8xx: fix build problem for ETX094 board

Commit 7616e78 "Add Ethernet hardware MAC address framework to usbnet"
increased the code size and broke building for the ETX094 board.
Adjust the linker script to make it build again.

While we are at it, remove unused u-boot.lds.debug

Signed-off-by: Wolfgang Denk <wd@denx.de>

phy.c: make less verbose - turn printf() into debug()

The PHY driver was too verbose and corrupted the boot message display
like this:

...
Net:   TSEC0 connected to Marvell 88E1111S
TSEC1 connected to Marvell 88E1111S
TSEC0, TSEC1
...

Turn printf() into debug() so we het the expected output again:

...
Net:   TSEC0, TSEC1
...

Signed-off-by: Wolfgang Denk <wd@denx.de>
Cc: Kumar Gala <galak@kernel.crashing.org>

Merge branch 'master' of git://git.denx.de/u-boot-sh

* 'master' of git://git.denx.de/u-boot-sh:
  sh: add calling mmc_initialize in board.c
  sh: sh7757lcr: Add KEEP order to start.o section
  usb: r8a66597: Fix argument mistake of inl
  sh: Clean up rsk7264 board settings
  sh: sh2a: Add sh2a optimize to config.mk
  sh: Fix rsk7203 alignment problem
  sh: Add support for SH2A freestanding build
  sh: Add Renesas rsk7264 board

Merge branch 'master' of git://git.denx.de/u-boot-ubi

* 'master' of git://git.denx.de/u-boot-ubi:
  ubifs: Fix bad free() sequence in ubifs_finddir()

sh: add calling mmc_initialize in board.c

Some SH have MMC controller. So, if we need it, we have to call
the mmc_initialize().

Signed-off-by: Yoshihiro Shimoda <yoshihiro.shimoda.uh@renesas.com>
Signed-off-by: Nobuhiro Iwamatsu <iwamatsu@nigauri.org>

sh: sh7757lcr: Add KEEP order to start.o section

The sh7757lcr has a local u-boot.lds because the sh7757lcr is only
supported the SPI booting.
This patch refers from the commit "sh: Add KEEP order to start.o section"
(commit ID: b52da2aed8c2c388661f369052a97d5b5c9ed00a).

Signed-off-by: Yoshihiro Shimoda <yoshihiro.shimoda.uh@renesas.com>
Signed-off-by: Nobuhiro Iwamatsu <iwamatsu@nigauri.org>

usb: r8a66597: Fix argument mistake of inl

Fail in build, because argument of inl used in r8a66597_read_fifo is wrong.

r8a66597.h:441:35: error: macro "inl" passed 2 arguments, but takes just 1
In file included from r8a66597-hcd.c:25:
r8a66597.h: In function ‘r8a66597_read_fifo’:
r8a66597.h:441: error: ‘inl’ undeclared (first use in this function)
r8a66597.h:441: error: (Each undeclared identifier is reported only once
r8a66597.h:441: error: for each function it appears in.)

Signed-off-by: Nobuhiro Iwamatsu <iwamatsu@nigauri.org>

sh: Clean up rsk7264 board settings

Adjusted default settings so that we can boot zImages and uImages.
Removed unused settings, use default commands and where possible
calculate all other settings.

Signed-off-by: Phil Edworthy <phil.edworthy@renesas.com>
Cc: Nobuhiro Iwamatsu <iwamatsu@nigauri.org>
Signed-off-by: Nobuhiro Iwamatsu <iwamatsu@nigauri.org>

sh: sh2a: Add sh2a optimize to config.mk

Only the optimization of sh2 had been supported up.
This adds the optimization of sh2a.

Signed-off-by: Nobuhiro Iwamatsu <nobuhiro.iwamatsu.yj@renesas.com>
Signed-off-by: Nobuhiro Iwamatsu <iwamatsu@nigauri.org>

sh: Fix rsk7203 alignment problem

Signed-off-by: Phil Edworthy <phil.edworthy@renesas.com>
Signed-off-by: Nobuhiro Iwamatsu <iwamatsu@nigauri.org>

sh: Add support for SH2A freestanding build

SH2A toolchains often only provide an fdpic version of libgcc. This
can't be used with bare-metal software like U-Boot, so this patch
provides the necessary functions extracted from libgcc.

Signed-off-by: Phil Edworthy <phil.edworthy@renesas.com>
Cc: Nobuhiro Iwamatsu <iwamatsu@nigauri.org>
Signed-off-by: Nobuhiro Iwamatsu <iwamatsu@nigauri.org>

sh: Add Renesas rsk7264 board

The rsk7264 (also know as rsk2+sh7264) is an SH2A based board
with 64MB NAND flash and 64MB SDRAM. It is very similar to the
rsk7203 board.

Signed-off-by: Phil Edworthy <phil.edworthy@renesas.com>
Cc: Nobuhiro Iwamatsu <iwamatsu@nigauri.org>
Signed-off-by: Nobuhiro Iwamatsu <iwamatsu@nigauri.org>

ubifs: Fix bad free() sequence in ubifs_finddir()

Free private_data member element before freeing file structure.
This was causing malloc to crash. Also remove unnecessary variable
assigments as file structure gets free'd as well.

Signed-off-by: Rod Boyce <uboot@teamboyce.co.uk>
Signed-off-by: Wolfgang Denk <wd@denx.de>
Cc: Stefan Roese <sr@denx.de>
Signed-off-by: Stefan Roese <sr@denx.de>

usb: increase non-bulk timeout for slow chipsets.

If you take a look at 96820a35, you'll see the original timeout was
CONFIG_SYS_HZ.  Which is 1000.  After the mentioned change, non-bulk timeout
was changed to 100.  This causes timeout failures on the dreamplug platform
when trying to initialize the usb microsd reader.

Signed-off-by: Jason Cooper <u-boot@lakedaemon.net>
Acked-by: Mike Frysinger <vapier@gentoo.org>