EXYNOS5: Add L2 Cache Support.

This patch set adds L2 Cache Support to EXYNOS.

Signed-off-by: Arun Mankuzhi <arun.m@samsung.com>
Signed-off-by: Rajeshwari Shinde <rajeshwari.s@samsung.com>
Acked-by: Simon Glass <sjg@chromium.org>
Signed-off-by: Minkyu Kang <mk7.kang@samsung.com>

Merge branch 'u-boot-tegra/master' into 'u-boot-arm/master'

Merge branch 'u-boot-samsung/master' into 'u-boot-arm/master'

Tegra114: MMC: Enable DT MMC driver support for Tegra114 Dalmore boards

Tested on my Dalmore E1611 board, eMMC and SD-Card work fine, can load
a kernel off of an SD card OK, card detect works, and the env is now
stored in eMMC (end of the 2nd 'boot' sector, same as Tegra20/30).

Signed-off-by: Tom Warren <twarren@nvidia.com>
Reviewed-by: Stephen Warren <swarren@nvidia.com>

Tegra114: MMC: Add SD bus power-rail init routine

T114 requires SD bus power-rail bringup for the SDIO card on SDMMC3.

Signed-off-by: Tom Warren <twarren@nvidia.com>
Reviewed-by: Stephen Warren <swarren@nvidia.com>

Tegra114: Dalmore: Add SDIO3 pad config to pinctrl_config table

SDIO1 (the SD-card slot on Dalmore) needs to have its pads setup
before the MMC driver is added.

Signed-off-by: Tom Warren <twarren@nvidia.com>
Reviewed-by: Stephen Warren <swarren@nvidia.com>

Tegra114: fdt: Add SDMMC (sdhci) nodes for T114 boards (Dalmore for now)

Took these values directly from the kernel dts files.

Signed-off-by: Tom Warren <twarren@nvidia.com>
Reviewed-by: Stephen Warren <swarren@nvidia.com>

tegra114: dalmore: config: enable SPI

Turn on SPI in dalmore config file

Signed-off-by: Allen Martin <amartin@nvidia.com>
Signed-off-by: Tom Warren <twarren@nvidia.com>
Reviewed-by: Stephen Warren <swarren@nvidia.com>

tegra114: add SPI driver

Add driver for tegra114 SPI controller.  This controller is not
compatible with either the tegra20 or tegra30 controllers, so it
requires a new driver.

Signed-off-by: Allen Martin <amartin@nvidia.com>
Signed-off-by: Tom Warren <twarren@nvidia.com>
Reviewed-by: Stephen Warren <swarren@nvidia.com>

tegra114: dalmore: fdt: enable dalmore SPI controller

Dalmore has a SPI flash part attached to controller 4, so enable
controller 4 and set to 25MHz.

Signed-off-by: Allen Martin <amartin@nvidia.com>
Signed-off-by: Tom Warren <twarren@nvidia.com>
Reviewed-by: Stephen Warren <swarren@nvidia.com>

tegra114: fdt: add SPI blocks

Add nodes for t114 SPI controller hardware

Signed-off-by: Allen Martin <amartin@nvidia.com>
Signed-off-by: Tom Warren <twarren@nvidia.com>
Reviewed-by: Stephen Warren <swarren@nvidia.com>

tegra114: fdt: add apbdma block

Add node for apbdma controller hardware.

Signed-off-by: Allen Martin <amartin@nvidia.com>
Signed-off-by: Tom Warren <twarren@nvidia.com>
Reviewed-by: Stephen Warren <swarren@nvidia.com>

tegra114: fdt: add compatible string for tegra114 SPI ctrl

Add "nvidia,tegra114-spi" to represent t114 SPI controller hardware.

Signed-off-by: Allen Martin <amartin@nvidia.com>
Signed-off-by: Tom Warren <twarren@nvidia.com>
Reviewed-by: Stephen Warren <swarren@nvidia.com>

sf: winbond: add W25Q32DW

Add support for Winbond W25Q32DW 32Mbit part

Signed-off-by: Allen Martin <amartin@nvidia.com>
Signed-off-by: Tom Warren <twarren@nvidia.com>
Reviewed-by: Stephen Warren <swarren@nvidia.com>

spi: add common fdt SPI driver interface

Add a common interface to fdt based SPI drivers.  Each driver is
represented by a table entry in fdt_spi_drivers[].  If there are
multiple SPI drivers in the table, the first driver to return success
from spi_init() will be registered as the SPI driver.

Signed-off-by: Allen Martin <amartin@nvidia.com>
Signed-off-by: Tom Warren <twarren@nvidia.com>
Reviewed-by: Stephen Warren <swarren@nvidia.com>

tegra20: spi: move fdt probe to spi_init

Make the tegra20 SPI driver similar to the tegra30 (and soon to be
tegra114) SPI drivers in preparation of common fdt SPI driver front
end.

Signed-off-by: Allen Martin <amartin@nvidia.com>
Signed-off-by: Tom Warren <twarren@nvidia.com>
Reviewed-by: Stephen Warren <swarren@nvidia.com>

tegra: spi: pull register structs out of headers

Move register structs from headers into .c files and use common name.
This is in preparation of making common fdt front end for SPI
drivers.

Signed-off-by: Allen Martin <amartin@nvidia.com>
Signed-off-by: Tom Warren <twarren@nvidia.com>
Reviewed-by: Stephen Warren <swarren@nvidia.com>

tegra: spi: remove non fdt support

Remove non fdt support from tegra20 and tegra30 SPI drivers in
preparation of new common fdt based SPI driver front end.

Signed-off-by: Allen Martin <amartin@nvidia.com>
Signed-off-by: Tom Warren <twarren@nvidia.com>
Reviewed-by: Stephen Warren <swarren@nvidia.com>

tegra: spi: rename tegra SPI drivers

Rename tegra SPI drivers to tegra20_flash and tegra20_slink in
preparation for commonization and addition of tegra114_spi.

Signed-off-by: Allen Martin <amartin@nvidia.com>
Signed-off-by: Tom Warren <twarren@nvidia.com>
Reviewed-by: Stephen Warren <swarren@nvidia.com>

tegra: remove support for UART SPI switch

This feature was only used for tegra20 seaboard that had a pinmux
conflict on the SPI pins.  These boards were never manufactured, so
remove this support to clean up SPI driver.

Signed-off-by: Allen Martin <amartin@nvidia.com>
Signed-off-by: Tom Warren <twarren@nvidia.com>
Reviewed-by: Stephen Warren <swarren@nvidia.com>

Merge branch 'u-boot-ti/master' into 'u-boot-arm/master'

mmc: omap_hsmmc.c: only register getcd/getwp callbacks if gpio could be used

Gets rid of warnings from omap_gpio:
ERROR : check_gpio: invalid GPIO -1

(and undefined behaviour as the -1 error code is interpreted as gpio value)

Signed-off-by: Peter Korsgaard <peter.korsgaard@barco.com>

mmc: mmc_getcd/getwp: use sensible defaults

Let mmc_getcd() return true and mmc_getwp() false if mmc driver doesn't
provide handlers for them.

Signed-off-by: Peter Korsgaard <peter.korsgaard@barco.com>
[trini: Add braces around first if test in each case to fix warning]
Signed-off-by: Tom Rini <trini@ti.com>

arm: Correct CONFIG_STANDALONE_LOAD_ADDR for AM33XX/OMAP* platforms

All of these platforms have memory starting at 0x80000000, so this is
the correct CONFIG_STANDALONE_LOAD_ADDR for all of them.

Acked-by: Peter Korsgaard <jacmet@sunsite.dk>
Signed-off-by: Tom Rini <trini@ti.com>

am335x_evm: Add better timings for the new BeagleBoard DDR3 part

Tested-by: Rao Bodapati <rao@circuitco.com>
Signed-off-by: Tom Rini <trini@ti.com>

ti814x_evm: add ti814x evm board support

Add TI814X EVM board directory, config file, and MAINTAINERS
entry. Enable build.

Signed-off-by: Matt Porter <mporter@ti.com>
Reviewed-by: Tom Rini <trini@ti.com>
[trini: Adapt to recent omap_hsmmc requirements, Matt re-tested]
Signed-off-by: Tom Rini <trini@ti.com>

ns16550: enable quirks for ti814x

TI814X requires the same quirks as AM33XX to be enabled.

Signed-off-by: Matt Porter <mporter@ti.com>
Reviewed-by: Tom Rini <trini@ti.com>

am33xx: support ti814x mmc reference clock

TI814x has a 192MHz hsmmc reference clock. Select that clock rate
when building for TI814x.

Signed-off-by: Matt Porter <mporter@ti.com>

am33xx: add dmm support to emif4 library

Adds a config_dmm() routine to support TI814X DMM configuration.

Signed-off-by: Matt Porter <mporter@ti.com>
Reviewed-by: Tom Rini <trini@ti.com>

am33xx: add ti814x specific register definitions

Support the ti814x specific register definitions within
arch-am33xx.

Signed-off-by: Matt Porter <mporter@ti.com>
Reviewed-by: Tom Rini <trini@ti.com>

am33xx: refactor am33xx mux support and add ti814x support

AM33XX and TI814X have a similar mux though the pinmux register
layout and address space differ. Add a separate ti814x mux include
to support the TI814X-specific differences.

Signed-off-by: Matt Porter <mporter@ti.com>
Reviewed-by: Tom Rini <trini@ti.com>
Acked-by: Peter Korsgaard <jacmet@sunsite.dk>

am33xx: refactor am33xx clocks and add ti814x support

Split clock.c for am335x and ti814x and add ti814x specific
clock support.

Signed-off-by: Matt Porter <mporter@ti.com>

am33xx: refactor emif4/ddr to support multiple EMIF instances

The AM33xx emif4/ddr support closely matches what is need to support
TI814x except that TI814x has two EMIF instances. Refactor all the
emif4 helper calls and the config_ddr() init function to use an
additional instance number argument.

Signed-off-by: Matt Porter <mporter@ti.com>
Reviewed-by: Tom Rini <trini@ti.com>

am33xx: convert defines from am33xx-specific to generic names

Eliminate AM33xx specific names to prepare for TI814x support
within AM33xx-land.

Signed-off-by: Matt Porter <mporter@ti.com>
Reviewed-by: Tom Rini <trini@ti.com>

am33xx: Add required includes to some omap/am33xx code

- In arch/arm/cpu/armv7/omap-common/timer.c,
  drivers/mtd/nand/omap_gpmc.c and drivers/net/cpsw.c add #include files
  that the driver needs but had been relying on <config.h> to bring in.
- In arch/arm/cpu/armv7/omap-common/lowlevel_init.S add <config.h>
- In am335x_evm.h and pcm051.h don't globally include
  <asm/arch/hardware.h> and <asm/arch/cpu.h> but just <asm/arch/omap.h>
  as that is the only include which defines things the config uses.

Cc: Lars Poeschel <poeschel@lemonage.de>
Signed-off-by: Tom Rini <trini@ti.com>

igep00x0: Enable CONFIG_CMD_BOOTZ

With v3.9 and later of the Linux Kernel defaulting to multi-platform
images with omap2plus_defconfig, uImage isn't builtable anymore by
default.  Add CONFIG_CMD_BOOTZ so that we can still boot something the
kernel spits out.

Signed-off-by: Enric Balletbo i Serra <eballetbo@iseebcn.com>
Reviewed-by: Javier Martinez Canillas <javier@dowhile0.org>

ARM: AM33XX: Fix typo that causes an AM duplication in CPU name.

Just fix a typo displaying the CPU info. With CONFIG_DISPLAY_INFO we see
something like AMAM335X-GP rev 0 instead of AM335X-GP rev 0.

Signed-off-by: Enric Balletbo i Serra <eballetbo@iseebcn.com>
Reviewed-by: Javier Martinez Canillas <javier@dowhile0.org>

am335x: Enable DDR PHY dynamic power down bit for DDR3 boards

Enable DDR PHY dynamic power down bit, which enables
powering down the IO receiver when not performing read.

This also helps in reducing overall power consumption in
low power states (suspend/standby).

Signed-off-by: Vaibhav Hiremath <hvaibhav@ti.com>
Signed-off-by: Satyanarayana, Sandhya <sandhya.satyanarayana@ti.com>
Cc: Tom Rini <trini@ti.com>
Reviewed-by: Tom Rini <trini@ti.com>

ARM: tegra: enable workaround for ARM erratum 716044

Tegra20 requires the workaround for this erratum. Enable it.

Signed-off-by: Stephen Warren <swarren@nvidia.com>

ARM: implement erratum 716044 workaround

Add common code to enable the workaround for ARM erratum 716044. This
will be enabled for Tegra.

Signed-off-by: Stephen Warren <swarren@nvidia.com>

am33xx:ddr:Fix config_sdram to work for all DDR

The original write to sdram_config is correct for DDR3 but incorrect
for DDR2 so SPL was hanging. For DDR2, the write to sdram_config
should be after the writes to ref_ctrl. This was working for DDR3
because there was a write of 0x2800 to ref_ctrl before a write
to sdram_config.

Tested on: GP EVM 1.1A (DDR2), GP EVM 1.5A (DDR3),
           Beaglebone A6 (DDR2), Beagleone Blacd A4A (DDR3)

Signed-off-by: Steve Kipisz <s-kipisz2@ti.com>

am335x_evm: Add more variables and switch to DT booting.

Make bootcmd run findfdt so that we know what dtb file to load.  Add a
loadfdt command to load this file in.  Make mmcboot pass in ${fdtaddr}
and make the mmc section of bootcmd run loadfdt.

Signed-off-by: Koen Kooi <koen@dominion.thruhere.net>
Signed-off-by: Tom Rini <trini@ti.com>
Acked-by: Peter Korsgaard <jacmet@sunsite.dk>

am335x_evm: Enable CMD_EXT4 and CMD_FS_GENERIC, add bootpart to env

The kernel is loaded from some form of ext[234] or FAT, depending on the
distribution used.  We add a bootpart variable to the environment so
that we can load from the correct mmc partition as well.  We leave
CONFIG_CMD_EXT2 for existing scripts that use ext2load.

Signed-off-by: Koen Kooi <koen@dominion.thruhere.net>
Signed-off-by: Tom Rini <trini@ti.com>
Acked-by: Peter Korsgaard <jacmet@sunsite.dk>

am335x_evm: add support for BeagleBone Black DT name

Cc: Matt Porter <mporter@ti.com>
Cc: Nishanth Menon <nm@ti.com>
Signed-off-by: Koen Kooi <koen@dominion.thruhere.net>
Signed-off-by: Tom Rini <trini@ti.com>
Acked-by: Matt Porter <mporter@ti.com>
Acked-by: Peter Korsgaard <jacmet@sunsite.dk>
Acked-by: Nishanth Menon <nm@ti.com>

Initialise correct GPMC WAITx irq for AM33xx

Currently WAIT0 irq is reset and then WAIT1 irq is enabled.
Fix it such that WAIT0 irq is enabled instead.

Signed-off-by: Mark Jackson <mpfj@newflow.co.uk>
Acked-by: Peter Korsgaard <jacmet@sunsite.dk>

Allow AM335x MPU core clock speed to be specified in the board config file

Allow AM335x MPU core clock speed to be specified in the board config file.
To use, add the following to the board's config file:-

#define CONFIG_SYS_MPUCLK <desired clock freq in MHz>

Signed-off-by: Mark Jackson <mpfj@newflow.co.uk>
Acked-by: Peter Korsgaard <jacmet@sunsite.dk>

cm-t35: add support for loading splash image from NAND

Add support for loading splash image from NAND

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

ARM: rpi_b: enable SD controller, add related env/cmds

Enable the SD controller driver for the Raspberry Pi. Enable a number
of useful MMC, partition, and filesystem-related commands. Set up the
environment to provide standard locations for loading a kernel, DTB,
etc. Provide a boot command that loads and executes boot.scr.uimg from
the SD card; this is written considering future extensibilty to USB
storage.

Signed-off-by: Stephen Warren <swarren@wwwdotorg.org>

mmc: add bcm2835 driver

This adds a simple driver for the BCM2835's SD controller.

Workarounds are implemented for:
* Register writes can't be too close to each-other in time, or they will
  be lost.
* Register accesses must all be 32-bit, so implement custom accessors.

This code was extracted from:
git://github.com/gonzoua/u-boot-pi.git master
which was created by Oleksandr Tymoshenko.

Portions of the code there were obviously based on the Linux kernel at:
git://github.com/raspberrypi/linux.git rpi-3.6.y
commit f5b930b "Main bcm2708 linux port" signed-off-by Dom Cobley.

swarren changed the following for upstream:
* Removed hack udelay()s in bcm2835_sdhci_raw_writel(); setting
  SDHCI_QUIRK_WAIT_SEND_CMD appears to solve the issues.
* Remove register logging from read*/write* functions.
* Sort out confusion with min/max_freq values passed to add_sdhci().
* Use more descriptive variable names and calculations in IO accessors.
* Simplified and commented twoticks_delay calculation.
* checkpatch fixes.

Cc: Andy Fleming <afleming@gmail.com>
Signed-off-by: Oleksandr Tymoshenko <gonzo@bluezbox.com>
Signed-off-by: Stephen Warren <swarren@wwwdotorg.org>
Acked-by: Andy Fleming <afleming@gmail.com>

video: add a driver for the bcm2835

The firmware running on the bcm2835 SoC's VideoCore CPU manages the
display controller. Add a simple "LCD" driver that communicates with the
firmware using the property mailbox protocol. This configures the
display and frame-buffer to match whatever physical resolution the
firmware chosen when booting, which is typically the native resolution
of the attached display device, presumably unless otherwise specified
in config.txt on the boot media.

Enable this driver in the Raspberry Pi board configuration.

Signed-off-by: Stephen Warren <swarren@wwwdotorg.org>
Acked-by: Anatolij Gustschin <agust@denx.de>

ARM: rpi_b: disable rpi_b dcache explicitly

There appears to be no implementation of flush_dcache_range() for
ARM1176, so explicitly disable dcache support to avoid references to
that function from the LCD core in the next patch. This was presumably
not noticed before simply because no drivers for the rpi_b were
attempting DMA.

Signed-off-by: Stephen Warren <swarren@wwwdotorg.org>

lcd: calculate line_length after lcd_ctrl_init()

When an LCD driver is actually driving a regular external display, e.g.
an HDMI monitor, the display resolution might not be known until the
display controller has initialized, i.e. during lcd_ctrl_init(). However,
lcd.c calculates lcd_line_length before calling this function, thus
relying on a hard-coded resolution in struct panel_info.

Instead, defer this calculation until after lcd_ctrl_init() has had the
chance to dynamically determine the resolution. This needs to happen
before lcd_clear(), since the value is used there.

grep indicates that no code outside lcd.c uses this lcd_line_length; in
particular, no lcd_ctrl_init() implementations read it.

Signed-off-by: Stephen Warren <swarren@wwwdotorg.org>
Acked-by: Anatolij Gustschin <agust@denx.de>

ARM: rpi_b: use bcm2835 mbox driver to get memory size

The firmware running on the bcm2835 SoC's VideoCore CPU determines how
much of the system RAM is available for use by the ARM CPU. Previously,
U-Boot assumed that only 128MB was available, since this was the
smallest value configured by any public firmware. However, we can now
query the actual value at run-time from the firmware using the mbox
property protocol.

Signed-off-by: Stephen Warren <swarren@wwwdotorg.org>

ARM: bcm2835: add mailbox driver

The BCM2835 SoC contains (at least) two CPUs; the VideoCore (a/k/a "GPU")
and the ARM CPU. The ARM CPU is often thought of as the main CPU.
However, the VideoCore actually controls the initial SoC boot, and hides
much of the hardware behind a protocol. This protocol is transported
using the SoC's mailbox hardware module.

Here, we add a very simplistic driver for the mailbox module, and define
a few structures for the property messages.

Signed-off-by: Stephen Warren <swarren@wwwdotorg.org>

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

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

Albert's rework of the linker scripts conflicted with Simon's making
everyone use __bss_end.  We also had a minor conflict over
README.scrapyard being added to in mainline and enhanced in
u-boot-arm/master with proper formatting.

Conflicts:
arch/arm/cpu/ixp/u-boot.lds
arch/arm/cpu/u-boot.lds
arch/arm/lib/Makefile
board/actux1/u-boot.lds
board/actux2/u-boot.lds
board/actux3/u-boot.lds
board/dvlhost/u-boot.lds
board/freescale/mx31ads/u-boot.lds
doc/README.scrapyard
include/configs/tegra-common.h

Build tested for all of ARM and run-time tested on am335x_evm.

Signed-off-by: Tom Rini <trini@ti.com>

usb: Add multiple controllers support for EHCI PCI

Use the ability to have several active EHCI controller on a system
in the PCI EHCI controller implementation.

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

x86: Enable USB features for coreboot

Enable PCI EHCI, storage, keyboard and Ethernet for USB.

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

usb: usbeth: smsc95xx: remove EEPROM loaded check

[port of Linux kernel commit bcd218be5aeb by Steve Glendinning]

The eeprom read & write commands currently check the E2P_CMD_LOADED_ bit is
set before allowing any operations.  This prevents any reading or writing
unless a correctly programmed EEPROM is installed.

Signed-off-by: Michael Spang <spang@chromium.org>
Signed-off-by: Simon Glass <sjg@chromium.org>
Acked-by: Marek Vasut <marex@denx.de>

usb: ehci: Fix aliasing issue in EHCI interrupt code

The interrupt endpoint handling code stores the buffer pointer in the QH
padding field. We need to make it the size of a pointer to avoid strict
aliasing issue with the compiler.

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

usb: ehci: Support interrupt transfers via periodic list

Interrupt transfers aren't meant to be used from the async list
(the EHCI spec indicates trouble with low/full-speed intr on async).

Build a periodic list instead, and provide an API to make use of it.
Then, use that API from the existing interrupt transfer API.

This provides support for USB keyboards using EHCI.

Use timeouts to ensure we cannot get stuck in the keyboard scanning
if something wrong happens (USB device unplugged or fatal I/O error)

Signed-off-by: Vincent Palatin <vpalatin@chromium.org>
Signed-off-by: Julius Werner <jwerner@chromium.org>
Signed-off-by: Simon Glass <sjg@chromium.org>

usb: ehci: exynos: Enable non-dt path

Enabling the non-dt path for the driver so that
we don't get any build errors for non-dt configuration.

Signed-off-by: Vivek Gautam <gautam.vivek@samsung.com>
Acked-by: Simon Glass <sjg@chromium.org>

usb: ehci: exynos: Fix multiple FDT decode

With current FDT support driver tries to parse device node
twice in ehci_hcd_init() and ehci_hcd_stop(), which shouldn't
happen ideally.
Making provision to store data in a global structure and thereby
passing its pointer when needed.

Signed-off-by: Vivek Gautam <gautam.vivek@samsung.com>

arm:trats: Use new ums command

This patch enables new "ums" command on Trats board

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

usb:gadget: USB Mass Storage Gadget support

This patch adds the USB Mass Storage Gadget to u-boot
New command called "ums" is implemented to provide access
to on-device embedded persistent memory.

USB Mass Storage is supposed to work on top of the USB
Gadget framework

Signed-off-by: Lukasz Majewski <l.majewski@samsung.com>
Signed-off-by: Piotr Wilczek <p.wilczek@samsung.com>
Signed-off-by: Kyungmin Park <kyungmin.park@samsung.com>
CC: Marek Vasut <marek.vasut@gmail.com>

usb:composite: USB Mass Storage - f_mass_storage.c from Linux kernel

The f_mass_storage.c source file from v2.6.36 Linux kernel.

commit 8876f5e7d3b2a320777dd4f6f5301d474c97a06c
Author: Michal Nazarewicz <m.nazarewicz@samsung.com>
Date:   Mon Jun 21 13:57:09 2010 +0200

USB: gadget: f_mass_storage: added eject callback

Signed-off-by: Lukasz Majewski <l.majewski@samsung.com>
Signed-off-by: Piotr Wilczek <p.wilczek@samsung.com>
Signed-off-by: Kyungmin Park <kyungmin.park@samsung.com>
CC: Marek Vasut <marek.vasut@gmail.com>

usb:composite: USB Mass Storage - storage_common.c from Linux kernel

The storage_common.c source file from v2.6.36 Linux kernel.

commit d26a6aa08b9f12b44fb1ee65625e7480d3d5bb81
Author: Michal Nazarewicz <m.nazarewicz@samsung.com>
Date:   Mon Nov 9 14:15:23 2009 +0100

USB: g_mass_storage: code cleaned up and comments updated

Signed-off-by: Lukasz Majewski <l.majewski@samsung.com>
Signed-off-by: Andrzej Pietrasiewicz <andrzej.p@samsung.com>
Signed-off-by: Kyungmin Park <kyungmin.park@samsung.com>
CC: Marek Vasut <marek.vasut@gmail.com>

usb:composite:fix Provide function data when addressing device with only one interface

This commit fixes problems with some non-standard requests send with
device address instead of interface address (bmRequestType.Receipent field).

This happens with dfu-util (debian version: 0.5), which address non standard
requests (like w_value=0x21 and bRequest=GET_DESCRIPTOR) to device.
Without this fix, the above request is STALLED, and hence causes dfu-util
to assume some standard configuration (packet size = 1024B instead of 4096B)
In turn it displays following errors:
Error obtaining DFU functional descriptor
Warning: Assuming DFU version 1.0
Warning: Transfer size can not be detected
...
Warning: Trying default transfer size 1024

This fix allows passing non-standard request to function setup code, where
it shall be handled.

Tested at:  Trats (exynos4210)
Tested with:DFU and UMS gadgets

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

usbtty: adapt buffers for large packet support

Increase buffer sizes at driver and tty level to accommodate kermit
large packet support.

Signed-off-by: Shiraz Hashim <shiraz.hashim@st.com>

usb/host/ehci: Add support for EHCI on spear

Add EHCI support for spear boards

Signed-off-by: Armando Visconti <armando.visconti@st.com>
Signed-off-by: Vipin Kumar <vipin.kumar@st.com>

usb: Fix bug when both DFU & ETHER are defined

When both CONFIG_USB_GADGET & CONFIG_USB_ETHER are defined
the makefile links objects twice.

This patch uses a Makefile specific idiom of
'if defined(CONFIG_USB_GADGET) || defined(CONFIG_USB_ETHER)'
to handle the case.

Signed-off-by: Pantelis Antoniou <panto@antoniou-consulting.com>

dfu: Add a partition type target

Dealing with raw block numbers with the dfu is very annoying.
Introduce a partition method.

Signed-off-by: Pantelis Antoniou <panto@antoniou-consulting.com>

dfu: Properly zero out timeout value

Zero out timeout value; letting it filled with undefined values
ends up with the dfu host hanging.

Signed-off-by: Pantelis Antoniou <panto@antoniou-consulting.com>

dfu: Fix crash when wrong number of arguments given

Fix obvious crash when not enough arguments are given to the dfu
command.

Signed-off-by: Pantelis Antoniou <panto@antoniou-consulting.com>

dfu: Only perform DFU board_usb_init() for TRATS

USB initialization shouldn't happen for all the boards.

Signed-off-by: Pantelis Antoniou <panto@antoniou-consulting.com>

usb: ehci: Add 64-bit controller support

On EHCI controller with 64-bit address space support, we must initialize
properly the high word for the PCI bus master accesses.

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

usb: ehci: generic PCI support

Instead of hardcoding the PCI IDs on the USB controller, use the PCI
class to detect them.

Ensure the busmaster bit is properly set in the PCI configuration.

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

Use 'unsigned int' for global_data's baudrate

We decided to used unsigned int here, rather than unsigned long. But
for the generic global_data it is still unsigned long. So change it
over.

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

sparc: Fix build warnings in serial.c

These macros are already defined in io.h so should not be declared in
serial.c.

serial.c:38:0: warning: "READ_BYTE" redefined
/home/sjg/c/src/third_party/u-boot/files/include/asm/io.h:36:0: note: this is the location of the previous definition
serial.c:39:0: warning: "READ_HWORD" redefined
/home/sjg/c/src/third_party/u-boot/files/include/asm/io.h:37:0: note: this is the location of the previous definition
serial.c:40:0: warning: "READ_WORD" redefined
/home/sjg/c/src/third_party/u-boot/files/include/asm/io.h:38:0: note: this is the location of the previous definition
serial.c:41:0: warning: "READ_DWORD" redefined
/home/sjg/c/src/third_party/u-boot/files/include/asm/io.h:39:0: note: this is the location of the previous definition

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

sparc: Fix out-of-tree building

It isn't clear why the sparc cpu Makefile has its own compile line, but
it does not work correctly with an out-of-tree build. Removing it fixes
this problem. Perhaps it does not introduce others.

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

serial: ns16550: Remove unwanted cast

This cast does not seem correct, since we should be writing to a pointer,
not a ulong.

This fixes the following warning on nds32:

-ns16550.c:49: warning: passing argument 2 of 'writeb' makes pointer from integer without a cast

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

avr32: Fix cast warning in board.c

The frame buffer pointer in global_data is not a pointer, so we should
remove these casts.

Signed-off-by: Simon Glass <sjg@chromium.org>
Acked-by: Andreas Bießmann <andreas.devel@googlemail.com>

blackfin: Remove noreturn attribute from cpu_init_f()

Now that board_init_f() is not marked as noreturn, we need to do the same
to blackfin's cpu_init_f() function.

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

blackfin: Fix up board_type global data

This should be in arch_global_data but was dropped in the change-over.

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

exynos5: Enable generic board for Exynos5 device tree boards

Enable CONFIG_SYS_GENERIC_BOARD for all Exynos5 boards that use
CONFIG_OF_CONTROL.

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

x86: Enable generic board

This will be used exclusively on x86, so enable it for the whole
architecture.

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

tegra: Enable generic board for Tegra

Bravely enable CONFIG_SYS_GENERIC_BOARD for all Tegra boards.

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

x86: Enable generic board support

This enables generic board support so that x86 boards can define
CONFIG_SYS_GENERIC_BOARD.

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

x86: Use sections header to obtain link symbols

These are defined in asm-generic/sections.h, so remove them from
architecture-specific files.

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

x86: Adjust board_r.c for x86

For x86 the global_data is managed entirely by the start.S code so we do
not need to touch it. However, we do have some more initcalls to add.

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

x86: Adjust board_f.c for x86

For x86, things have adjusted somewhat since this series was originally
written. It has its own way of running through initcalls which is actually
nicer than others archs.

Unfortunately this does introduce exceptions. We will soon require use of
generic board on x86, but until then we need to fit in with what is there,
and treat x86 as a special case.

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

ppc: Enable generic board support

This enables generic board support so that ppc boards can define
CONFIG_SYS_GENERIC_BOARD.

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

Adjust board_r.c for ppc

This adds ppc features to the generic post-relocation board init.

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

Adjust board_f.c for ppc

This adds ppc features to the generic pre-relocation board init.

This is a separate commit so that these features are clearly shown.

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

Add CONFIG_SYS_SYM_OFFSETS to support offset symbols

Link symbols as created by the link script can either be absolute or
relative to the text start. This option switches between the two options
so that we can support both.

As we convert architectures over to generic board, we can see if this
option is actually needed, or whether it is possible to unify this feature
also.

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

arm: Enable generic board support

This enables generic board support so that ARM boards can define
CONFIG_SYS_GENERIC_BOARD.

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

arm: Remove use of board_early_init_r/last_stage_init()

These boards define CONFIG_LAST_STAGE_INIT and CONFIG_BOARD_EARLY_INIT_R
but these options are not available on ARM. Move them into the powerpc
common file instead.

This change affects: km_kirkwood_pci, mgcoge3un, kmnusa, kmcoge5un,
km_kirkwood and portl2.

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

Add spl load feature

This adds secondary program loader support to the generic board.

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

Introduce generic post-relocation board_r.c

This file handles common post-relocation init for boards which use
the generic framework.

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

Introduce generic pre-relocation board_f.c

This file handles common pre-relocation init for boards which use
the generic framework.

It starts up the console, DRAM, performs relocation and then jumps
to post-relocation init.

Signed-off-by: Simon Glass <sjg@chromium.org>
Tested-by: Wolfgang Denk <wd@denx.de>
Acked-by: Wolfgang Denk <wd@denx.de>