project/bcm63xx/u-boot.git
8 years agoMerge branch 'master' of git://git.denx.de/u-boot-i2c
Tom Rini [Tue, 26 Jul 2016 12:29:30 +0000 (08:29 -0400)]
Merge branch 'master' of git://git.denx.de/u-boot-i2c

8 years agoi2c: mvtwsi: Add documentation
mario.six@gdsys.cc [Thu, 21 Jul 2016 09:57:13 +0000 (11:57 +0200)]
i2c: mvtwsi: Add documentation

Add full documentation to all driver functions.

Signed-off-by: Mario Six <mario.six@gdsys.cc>
Reviewed-by: Stefan Roese <sr@denx.de>
8 years agoi2c: mvtwsi: Make delay times frequency-dependent
mario.six@gdsys.cc [Thu, 21 Jul 2016 09:57:12 +0000 (11:57 +0200)]
i2c: mvtwsi: Make delay times frequency-dependent

Some devices using the MVTWSI driver have the option to run at speeds
faster than Standard Mode (100kHZ). On the Armada 38x controllers, this
is actually necessary, since due to erratum FE-8471889, a timing
violation concerning repeated starts prevents the controller from
working correctly in Standard Mode. One of the workarounds recommended
in the erratum is to set the bus to Fast Mode (400kHZ) operation and
ensure all connected devices are set to Fast Mode.

In the current version of the driver, however, the delay times are
hard-coded to 10ms, corresponding to Standard Mode operation. To take
full advantage of the faster modes, we would need to either keep the
currently configured I2C speed in a globally accessible variable, or
pass it to the necessary functions as a parameter. For DM, the first
option is not a problem, and we can simply keep the speed in the private
data of the driver. For the legacy interface, however, we would need to
introduce a static variable, which would cause problems with boots from
NOR flashes; see commit d6b7757 "i2c: mvtwsi: Eliminate
twsi_control_flags."

As to not clutter the interface with yet another parameter, we therefore
keep the default 10ms delays for the legacy functions.

In DM mode, we make the delay time dependant on the frequency to allow
taking full advantage of faster modes of operation (tested with up to
1MHZ frequency on Armada MV88F6820).

Signed-off-by: Mario Six <mario.six@gdsys.cc>
Reviewed-by: Stefan Roese <sr@denx.de>
8 years agoi2c: mvtwsi: Handle zero-length offsets properly
mario.six@gdsys.cc [Thu, 21 Jul 2016 09:57:11 +0000 (11:57 +0200)]
i2c: mvtwsi: Handle zero-length offsets properly

Zero-length offsets are not properly handled by the driver. When a read
operation with a zero-length offset is started, a START condition is
asserted, and since no offset bytes are transferred, a repeated START is
issued immediately after, which confuses the controller.

To fix this, we send the first START only if any address bytes need to
be sent, and keep track of the expected start status accordingly.

Signed-off-by: Mario Six <mario.six@gdsys.cc>
Reviewed-by: Stefan Roese <sr@denx.de>
8 years agoi2c: mvtwsi: Add compatibility to DM
mario.six@gdsys.cc [Thu, 21 Jul 2016 09:57:10 +0000 (11:57 +0200)]
i2c: mvtwsi: Add compatibility to DM

This patch adds the necessary functions and Kconfig entry to make the
MVTWSI I2C driver compatible with the driver model.

A possible device tree entry might look like this:

i2c@11100 {
compatible = "marvell,mv64xxx-i2c";
reg = <0x11000 0x20>;
clock-frequency = <100000>;
u-boot,i2c-slave-addr = <0x0>;
};

Signed-off-by: Mario Six <mario.six@gdsys.cc>
Reviewed-by: Stefan Roese <sr@denx.de>
8 years agoi2c: mvtwsi: Make address length variable
mario.six@gdsys.cc [Thu, 21 Jul 2016 09:57:09 +0000 (11:57 +0200)]
i2c: mvtwsi: Make address length variable

The length of the address parameter of the __twsi_i2c_read and
__twsi_i2c_write functions is fixed to four bytes.

As a final step in the preparation of the DM conversion, we make the
length of this parameter variable by turning it into an array of bytes,
and convert the 32 bit value that's passed to the legacy functions into
a four-byte-array on the fly.

Signed-off-by: Mario Six <mario.six@gdsys.cc>
Reviewed-by: Stefan Roese <sr@denx.de>
8 years agoi2c: mvtwsi: Factor out adap parameter
mario.six@gdsys.cc [Thu, 21 Jul 2016 09:57:08 +0000 (11:57 +0200)]
i2c: mvtwsi: Factor out adap parameter

To be able to use the compatibility layer from the DM functions, we
factor the adap parameter out of all functions, and pass the actual
register base instead.

Signed-off-by: Mario Six <mario.six@gdsys.cc>
Reviewed-by: Stefan Roese <sr@denx.de>
8 years agoi2c: mvtwsi: Add compatibility functions
mario.six@gdsys.cc [Thu, 21 Jul 2016 09:57:07 +0000 (11:57 +0200)]
i2c: mvtwsi: Add compatibility functions

To prepare for the DM conversion, we add a layer of compatibility
functions to be used by both the legacy and the DM functions.

Signed-off-by: Mario Six <mario.six@gdsys.cc>
Reviewed-by: Stefan Roese <sr@denx.de>
8 years agoi2c: mvtwsi: Use 'uint' instead of 'unsigned int'
mario.six@gdsys.cc [Thu, 21 Jul 2016 09:57:06 +0000 (11:57 +0200)]
i2c: mvtwsi: Use 'uint' instead of 'unsigned int'

Since some additional parameters will be added in the course of this
patch series (especially with the addition of DM support), we replace
the longer "unsigned int" declarations with "uint" declarations to keep
the parameter lists more readable.

Signed-off-by: Mario Six <mario.six@gdsys.cc>
Reviewed-by: Stefan Roese <sr@denx.de>
8 years agoi2c: mvtwsi: Get rid of status parameter
mario.six@gdsys.cc [Thu, 21 Jul 2016 09:57:05 +0000 (11:57 +0200)]
i2c: mvtwsi: Get rid of status parameter

The twsi_stop function contains a parameter "status," which is used to
pass in the current exit status of the function calling twsi_stop, and
either return this status unchanged if it indicates an error, or return
twsi_stop's exit status if it does not indicate an error.

While not massively complicated, this adds another purpose to the
twsi_stop function, which should have the sole purpose of asserting a
STOP condition on the bus (and not manage the exit status of its
caller).

Therefore, we move the exit status management into the caller functions
by introducing a "stop_status" variable and returning either the status
before the twsi_stop call (kept in the "status" variable), or the status
from the twsi_stop call, depending on which indicates an error.

Signed-off-by: Mario Six <mario.six@gdsys.cc>
Reviewed-by: Stefan Roese <sr@denx.de>
8 years agoi2c: mvtwsi: Eliminate flags parameter
mario.six@gdsys.cc [Thu, 21 Jul 2016 09:57:04 +0000 (11:57 +0200)]
i2c: mvtwsi: Eliminate flags parameter

Due to breaking boots from NOR flashes, commit d6b7757 ("i2c: mvtwsi:
Eliminate twsi_control_flags") removed the static global
twsi_control_flags variable, which kept a set of default flags that were
always or'd to the control register when writing. It was replaced with a
flags parameter, which was passed around between the functions that
needed it.

Since the twsi_control_flags variable was used just for the purposes of
a) setting the MVTWSI_CONTROL_TWSIEN on every control register write,
   and
b) setting the MVTWSI_CONTROL_ACK from twsi_i2c_read if needed,
anyway, the added overhead of another variable being passed around is no
longer justified, and we are better off implementing this flag setting
logic locally in the functions that actually write to the control
register.

Therefore, this patch sets MVTWSI_CONTROL_TWSIEN on every control
register write, replaces the twsi_i2c_read's flags parameter with a
ack_flag parameter, which tells the function whether to acknowledge the
read or not, and removes every other instance of the flags variable.
This has the added benefit that now every notion of "global default
flags" is gone, and it's much easier to see which control flags are
actually set at which point in time.

Signed-off-by: Mario Six <mario.six@gdsys.cc>
Reviewed-by: Stefan Roese <sr@denx.de>
8 years agoi2c: mvtwsi: Improve and fix comments
mario.six@gdsys.cc [Thu, 21 Jul 2016 09:57:03 +0000 (11:57 +0200)]
i2c: mvtwsi: Improve and fix comments

This patch fixes only comments/documentation: Streamline capitalization
and improve grammar/punctuation.

Signed-off-by: Mario Six <mario.six@gdsys.cc>
Reviewed-by: Stefan Roese <sr@denx.de>
8 years agoi2c: mvtwsi: Streamline code and add documentation
mario.six@gdsys.cc [Thu, 21 Jul 2016 09:57:02 +0000 (11:57 +0200)]
i2c: mvtwsi: Streamline code and add documentation

Convert groups of logically connected preprocessor defines into proper
enums, one macro into an inline function, and add documentation
to/extend existing documentation of these items.

Signed-off-by: Mario Six <mario.six@gdsys.cc>
Reviewed-by: Stefan Roese <sr@denx.de>
8 years agoi2c: mvtwsi: Fix style violations
mario.six@gdsys.cc [Thu, 21 Jul 2016 09:57:01 +0000 (11:57 +0200)]
i2c: mvtwsi: Fix style violations

This patch fixes seven style violations: Six superfluous spaces after
casts, and one logical continuation violation.

Signed-off-by: Mario Six <mario.six@gdsys.cc>
Reviewed-by: Stefan Roese <sr@denx.de>
8 years agodefconfig: am57xx_hs_evm: enable i2c driver model
Mugunthan V N [Mon, 18 Jul 2016 09:41:09 +0000 (15:11 +0530)]
defconfig: am57xx_hs_evm: enable i2c driver model

Enable i2c driver model for am57xx_hs_evm as omap i2c
supports driver model.

Signed-off-by: Mugunthan V N <mugunthanvnm@ti.com>
8 years agodefconfig: am57xx_evm: enable i2c driver model
Mugunthan V N [Mon, 18 Jul 2016 09:41:08 +0000 (15:11 +0530)]
defconfig: am57xx_evm: enable i2c driver model

Enable i2c driver model for am57xx_evm as omap i2c
supports driver model.

Signed-off-by: Mugunthan V N <mugunthanvnm@ti.com>
Reviewed-by: Tom Rini <trini@konsulko.com>
8 years agodefconfig: dra7xx_hs_evm: enable i2c driver model
Mugunthan V N [Mon, 18 Jul 2016 09:41:07 +0000 (15:11 +0530)]
defconfig: dra7xx_hs_evm: enable i2c driver model

Enable i2c driver model for dra7xx_hs_evm as omap i2c
supports driver model.

Signed-off-by: Mugunthan V N <mugunthanvnm@ti.com>
Reviewed-by: Tom Rini <trini@konsulko.com>
8 years agodefconfig: dra7xx_evm: enable i2c driver model
Mugunthan V N [Mon, 18 Jul 2016 09:41:06 +0000 (15:11 +0530)]
defconfig: dra7xx_evm: enable i2c driver model

Enable i2c driver model for dra7xx_evm as omap i2c
supports driver model.

Signed-off-by: Mugunthan V N <mugunthanvnm@ti.com>
Reviewed-by: Tom Rini <trini@konsulko.com>
8 years agodefconfig: am43xx_hs_evm: enable i2c driver model
Mugunthan V N [Mon, 18 Jul 2016 09:41:05 +0000 (15:11 +0530)]
defconfig: am43xx_hs_evm: enable i2c driver model

Enable i2c driver model for am43xx_hs_evm as omap i2c
supports driver model.

Signed-off-by: Mugunthan V N <mugunthanvnm@ti.com>
Reviewed-by: Tom Rini <trini@konsulko.com>
8 years agodefconfig: am43xx_evm: enable i2c driver model
Mugunthan V N [Mon, 18 Jul 2016 09:41:04 +0000 (15:11 +0530)]
defconfig: am43xx_evm: enable i2c driver model

Enable i2c driver model for am43xx_evm as omap i2c
supports driver model.

Signed-off-by: Mugunthan V N <mugunthanvnm@ti.com>
Reviewed-by: Tom Rini <trini@konsulko.com>
8 years agodefconfig: am335x_evm: enable i2c driver model
Mugunthan V N [Mon, 18 Jul 2016 09:41:03 +0000 (15:11 +0530)]
defconfig: am335x_evm: enable i2c driver model

Enable i2c driver model for am335x_evm as omap i2c
supports driver model.

Signed-off-by: Mugunthan V N <mugunthanvnm@ti.com>
Reviewed-by: Tom Rini <trini@konsulko.com>
8 years agodefconfig: am335x_boneblack_vboot: enable i2c driver model
Mugunthan V N [Mon, 18 Jul 2016 09:41:02 +0000 (15:11 +0530)]
defconfig: am335x_boneblack_vboot: enable i2c driver model

Enable i2c driver model for am335x_boneblack_vboot as omap i2c
supports driver model. Also enable CONFIG_DM_I2C_COMPAT for
legacy drivers of i2c devices.

Signed-off-by: Mugunthan V N <mugunthanvnm@ti.com>
Reviewed-by: Tom Rini <trini@konsulko.com>
8 years agodrivers: i2c: omap24xx_i2c: adopt omap_i2c driver to driver model
Mugunthan V N [Mon, 18 Jul 2016 09:41:01 +0000 (15:11 +0530)]
drivers: i2c: omap24xx_i2c: adopt omap_i2c driver to driver model

Convert omap i2c driver to adopt i2c driver model

Signed-off-by: Mugunthan V N <mugunthanvnm@ti.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
8 years agodrivers: i2c: omap24xx_i2c: prepare driver for DM conversion
Mugunthan V N [Mon, 18 Jul 2016 09:41:00 +0000 (15:11 +0530)]
drivers: i2c: omap24xx_i2c: prepare driver for DM conversion

Prepare the driver for DM conversion.

Signed-off-by: Mugunthan V N <mugunthanvnm@ti.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
8 years agoti_armv7_common: i2c: do not define DM_I2C for spl
Mugunthan V N [Mon, 18 Jul 2016 09:40:59 +0000 (15:10 +0530)]
ti_armv7_common: i2c: do not define DM_I2C for spl

Since omap's spl doesn't support DM currently, do not define
DM_I2C for spl build.

Signed-off-by: Mugunthan V N <mugunthanvnm@ti.com>
Reviewed-by: Tom Rini <trini@konsulko.com>
8 years agodrivers: i2c: uclass: parse dt parameters only when CONFIG_OF_CONTROL is enable
Mugunthan V N [Mon, 18 Jul 2016 09:40:58 +0000 (15:10 +0530)]
drivers: i2c: uclass: parse dt parameters only when CONFIG_OF_CONTROL is enable

parse dt parameter of i2c devices only when CONFIG_OF_CONTROL
is enabled.

Signed-off-by: Mugunthan V N <mugunthanvnm@ti.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
8 years agoomap5/dra7: i2c: correct register offset for sync register
Mugunthan V N [Mon, 18 Jul 2016 09:40:57 +0000 (15:10 +0530)]
omap5/dra7: i2c: correct register offset for sync register

The register offset of i2c_sysc offset is not correct as per
omap5[1]/dra7[2] TRM, correct the offsets as per the
documentation.

[1] - http://www.ti.com/lit/pdf/swpu249
[2] - http://www.ti.com/lit/pdf/spruhz6

Signed-off-by: Mugunthan V N <mugunthanvnm@ti.com>
Reviewed-by: Tom Rini <trini@konsulko.com>
8 years agoomap4: i2c: correct register offset for sync register
Mugunthan V N [Mon, 18 Jul 2016 09:40:56 +0000 (15:10 +0530)]
omap4: i2c: correct register offset for sync register

The register offset of i2c_sysc offset is not correct as per
omap4 TRM [1], correct the offsets as per the documentation.

[1] - http://www.ti.com/lit/ug/swpu235ab/swpu235ab.pdf

Signed-off-by: Mugunthan V N <mugunthanvnm@ti.com>
Reviewed-by: Tom Rini <trini@konsulko.com>
8 years agotools: env: Fix format warnings in debug
Marcin Niestroj [Fri, 6 May 2016 12:58:29 +0000 (14:58 +0200)]
tools: env: Fix format warnings in debug

Format warnings (-Wformat) were shown in printf() calls after defining
DEBUG macro.

Update format string and explicitly cast variables to suppress all
warnings.

Signed-off-by: Marcin Niestroj <m.niestroj@grinn-global.com>
8 years agorockchip: sdram: Fix register layout for Linux
John Keeping [Fri, 15 Jul 2016 16:33:23 +0000 (17:33 +0100)]
rockchip: sdram: Fix register layout for Linux

The ChromeOS kernel reads the RAM settings from PMU_SYS_REG2 and expects
the bootloader to store the necessary information there.  We're using
the same register to pass the same information between the SPL and
U-Boot but in a slightly different format.

Change this to use the format expected by the Linux DMC driver so that
the system doesn't hang in Linux by misconfiguring the RAM.

This is almost the same as commit b5788dc ("rockchip: rk3288: correct
sdram setting") which was reverted in commit b525556 ("Revert "rockchip:
rk3288: correct sdram setting"") but parenthese have been added to apply
the mask correctly when reading the "bw" setting and a couple of minor
style issues have been fixed to keep check_patch.pl happy.

Signed-off-by: John Keeping <john@metanate.com>
Reviewed-by: Tom Rini <trini@konsulko.com>
Acked-by: Simon Glass <sjg@chromium.org>
8 years agommc: rockchip: add SDHCI driver support for rockchip soc
Kever Yang [Mon, 18 Jul 2016 09:00:58 +0000 (17:00 +0800)]
mmc: rockchip: add SDHCI driver support for rockchip soc

Rockchip rk3399 using arasan sdhci-5.1 controller.
This patch add the controller support to enable mmc device
with full driver-model support, tested on rk3399 evb board.

According to my test result, this driver should be OK,
the command "part list mmc 0" can result in a right output,
but all the mmc command failed like this:
=> mmc info
No MMC device available
Command failed, result=1

The result of get_mmc_num in cmd/mmc.c is always 0?

Signed-off-by: Kever Yang <kever.yang@rock-chips.com>
Acked-by: Simon Glass <sjg@chromium.org>
8 years agoARM64: evb-rk3399: add a README for this board setup
Kever Yang [Tue, 19 Jul 2016 13:17:01 +0000 (21:17 +0800)]
ARM64: evb-rk3399: add a README for this board setup

Add a README to guide people flash the ATF and U-Boot
with Rockchip tools to bring up to board.

Signed-off-by: Kever Yang <kever.yang@rock-chips.com>
Acked-by: Simon Glass <sjg@chromium.org>
8 years agoconfig: add config file for evb-rk3399
Kever Yang [Tue, 19 Jul 2016 13:17:00 +0000 (21:17 +0800)]
config: add config file for evb-rk3399

This patch add basic config option for evb-rk3399 board.

Signed-off-by: Kever Yang <kever.yang@rock-chips.com>
Acked-by: Simon Glass <sjg@chromium.org>
8 years agoARM64: rockchip: add support for rk3399 SoC based evb
Kever Yang [Tue, 19 Jul 2016 13:16:59 +0000 (21:16 +0800)]
ARM64: rockchip: add support for rk3399 SoC based evb

RK3399 is a SoC from Rockchip with dual-core Cortex-A72
and quad-core Cortex-A53 CPU. It supports two USB3.0
type-C ports and two USB2.0 EHCI ports. Other interfaces
are very much like RK3288, the DRAM are 32bit width address
and support address from 0 to 4GB-128MB range.

Signed-off-by: Kever Yang <kever.yang@rock-chips.com>
Acked-by: Simon Glass <sjg@chromium.org>
8 years agodts: add support for Rockchip rk3399 soc
Kever Yang [Tue, 19 Jul 2016 13:16:58 +0000 (21:16 +0800)]
dts: add support for Rockchip rk3399 soc

These files are from kernel upstream:
"649a371 Add linux-next specific files for 20160616"
with some modification need by U-Boot:
- chosen with stdout-path to uart2.
- add clock-frequency for uart2

Signed-off-by: Kever Yang <kever.yang@rock-chips.com>
Acked-by: Simon Glass <sjg@chromium.org>
8 years agorockchip: update fastboot usage
Xu Ziyuan [Mon, 18 Jul 2016 01:56:46 +0000 (09:56 +0800)]
rockchip: update fastboot usage

Introduce how to use fastboot feature on rk3288.

Signed-off-by: Ziyuan Xu <xzy.xu@rock-chips.com>
Acked-by: Simon Glass <sjg@chromium.org>
8 years agomkimage: rockchip: add suport for rk33 serial
Kever Yang [Mon, 18 Jul 2016 01:35:26 +0000 (09:35 +0800)]
mkimage: rockchip: add suport for rk33 serial

Add support for rockchip rk33 series Soc like rk3368 and rk3399

Signed-off-by: Kever Yang <kever.yang@rock-chips.com>
Acked-by: Simon Glass <sjg@chromium.org>
8 years agorockchip: Use rockchip_get_clk() to obtain the SoC clock
Simon Glass [Sun, 17 Jul 2016 21:23:17 +0000 (15:23 -0600)]
rockchip: Use rockchip_get_clk() to obtain the SoC clock

The current code picks the first available clock. In U-Boot proper this is
the oscillator device, not the SoC clock device. As a result the HDMI display
does not work.

Fix this by calling rockchip_get_clk() instead.

Fixes: 135aa950 (clk: convert API to match reset/mailbox style)
Signed-off-by: Simon Glass <sjg@chromium.org>
Acked-by: Anatolij Gustschin <agust@denx.de>
8 years agorockchip: Add a way to obtain the main clock device
Simon Glass [Sun, 17 Jul 2016 21:23:16 +0000 (15:23 -0600)]
rockchip: Add a way to obtain the main clock device

On Rockchip SoCs we typically have a main clock device that uses the Soc
clock driver. There is also a fixed clock for the oscillator. Add a function
to obtain the core clock.

Signed-off-by: Simon Glass <sjg@chromium.org>
8 years agodm: core: Add a way to find a device by its driver
Simon Glass [Sun, 17 Jul 2016 21:23:15 +0000 (15:23 -0600)]
dm: core: Add a way to find a device by its driver

Some SoCs have a single clock device. Provide a way to find it given its
driver name. This is handled by the linker so will fail if the name is not
found, avoiding strange errors when names change and do not match. It is
also faster than a string comparison.

Signed-off-by: Simon Glass <sjg@chromium.org>
8 years agorockchip: rk3288: fix FREF_MIN_HZ constant
Heiko Stübner [Fri, 15 Jul 2016 22:17:17 +0000 (00:17 +0200)]
rockchip: rk3288: fix FREF_MIN_HZ constant

According to the TRM the minimum FREF frequency is 269kHz not MHz.
Adapt the constant accordingly.

Signed-off-by: Heiko Stuebner <heiko@sntech.de>
Acked-by: Simon Glass <sjg@chromium.org>
8 years agocosmetic: rockchip: rk3288: rename rkclk_configure_cpu
Heiko Stübner [Fri, 15 Jul 2016 22:17:16 +0000 (00:17 +0200)]
cosmetic: rockchip: rk3288: rename rkclk_configure_cpu

The function is very specific to the rk3288 in its arguments
referencing the rk3288 cru and grf and every other rockchip soc
has differing cru and grf registers. So make that function naming
explicit.

Signed-off-by: Heiko Stuebner <heiko@sntech.de>
Acked-by: Simon Glass <sjg@chromium.org>
8 years agocosmetic: rockchip: sort socs according to numbers
Heiko Stübner [Fri, 15 Jul 2016 22:17:15 +0000 (00:17 +0200)]
cosmetic: rockchip: sort socs according to numbers

Having some sort of ordering proofed helpful in a lot of other places
already. So for a larger number of rockchip socs it might be helpful
as well instead of an ever increasing unsorted list.

Signed-off-by: Heiko Stuebner <heiko@sntech.de>
Reviewed-by: Andreas Färber <afaerber@suse.de>
Acked-by: Simon Glass <sjg@chromium.org>
8 years agocosmetic: rockchip: rk3036: pinctrl: fix config symbol naming
Heiko Stübner [Fri, 15 Jul 2016 22:17:14 +0000 (00:17 +0200)]
cosmetic: rockchip: rk3036: pinctrl: fix config symbol naming

Rockchip socs are always named rkxxxx in all places, as also shown
by the naming of the rk3036 pinctrl file itself.
Therefore also name the config symbol according to this scheme.

Signed-off-by: Heiko Stuebner <heiko@sntech.de>
Acked-by: Simon Glass <sjg@chromium.org>
8 years agocosmetic: rockchip: rk3288: pinctrl: fix config symbol naming
Heiko Stübner [Fri, 15 Jul 2016 22:17:13 +0000 (00:17 +0200)]
cosmetic: rockchip: rk3288: pinctrl: fix config symbol naming

The rk3288 pinctrl is very specific to this soc, so should
not hog the generic rockchip naming.

Signed-off-by: Heiko Stuebner <heiko@sntech.de>
Acked-by: Simon Glass <sjg@chromium.org>
8 years agorockchip: rk3288: add fastboot support
Xu Ziyuan [Thu, 14 Jul 2016 16:26:59 +0000 (00:26 +0800)]
rockchip: rk3288: add fastboot support

Enable fastboot feature on rk3288.

This path doesn't support the fastboot flash function command entirely.
We will hit "cannot find partition" assertion without specified
partition environment. Define gpt partition layout in specified board
such as firefly-rk3288, then enjoy it!

Signed-off-by: Ziyuan Xu <xzy.xu@rock-chips.com>
Acked-by: Simon Glass <sjg@chromium.org>
8 years agousb: dwc2 : invalidate dcache before starting DMA
Xu Ziyuan [Thu, 14 Jul 2016 06:52:35 +0000 (14:52 +0800)]
usb: dwc2 : invalidate dcache before starting DMA

Invalidate dcache before starting the DMA to ensure coherency. In case
there are any dirty lines from the DMA buffer in the cache, subsequent
cache-line replacements may corrupt the buffer in memory while the DMA
is still going on. Cache-line replacement can happen if the CPU tries to
bring some other memory locations into the cache while the DMA is going
on.

Signed-off-by: Ziyuan Xu <xzy.xu@rock-chips.com>
Acked-by: Simon Glass <sjg@chromium.org>
8 years agousb: dwc2-otg: adjust fifo size via platform data
Xu Ziyuan [Thu, 14 Jul 2016 06:52:33 +0000 (14:52 +0800)]
usb: dwc2-otg: adjust fifo size via platform data

The total FIFO size of some SoCs may be different from the existen, this
patch supports fifo size setting from platform data.

Signed-off-by: Ziyuan Xu <xzy.xu@rock-chips.com>
Acked-by: Simon Glass <sjg@chromium.org>
8 years agousb: rockchip-phy: implement USB2.0 phy control
Xu Ziyuan [Thu, 14 Jul 2016 06:52:32 +0000 (14:52 +0800)]
usb: rockchip-phy: implement USB2.0 phy control

So far, Rockchip SoCs have two kinds of USB2.0 phy, such as Synopsys and
Innosilicon. This patch applys dwc2 usb driver framework to implement
phy_init() and phy_off() methods for Synopsys phy on Rockchip platform.

Signed-off-by: Ziyuan Xu <xzy.xu@rock-chips.com>
Acked-by: Simon Glass <sjg@chromium.org>
8 years agorockchip: Exclude rk_timer for ARM64
Andreas Färber [Thu, 14 Jul 2016 04:22:09 +0000 (06:22 +0200)]
rockchip: Exclude rk_timer for ARM64

It conflicts with the generic_timer.

Cc: Kever Yang <kever.yang@rock-chips.com>
Signed-off-by: Andreas Färber <afaerber@suse.de>
Acked-by: Simon Glass <sjg@chromium.org>
8 years agomkimage: rockchip: add suport for rk33 serial
Kever Yang [Thu, 14 Jul 2016 03:51:05 +0000 (11:51 +0800)]
mkimage: rockchip: add suport for rk33 serial

Add support for rockchip rk33 series Soc like rk3368 and rk3399

Signed-off-by: Kever Yang <kever.yang@rock-chips.com>
Acked-by: Simon Glass <sjg@chromium.org>
8 years agorockchip: Clean up CPU selection
Andreas Färber [Thu, 14 Jul 2016 03:09:26 +0000 (05:09 +0200)]
rockchip: Clean up CPU selection

In preparation for RK3368 and RK3399, which need to select ARM64, don't
select CPU_V7 at the ARCH_ROCKCHIP level but at the SoC level instead.

Cc: Kever Yang <kever.yang@rock-chips.com>
Signed-off-by: Andreas Färber <afaerber@suse.de>
Acked-by: Simon Glass <sjg@chromium.org>
8 years agoboard: move all the rockchip board in one folder
Kever Yang [Fri, 8 Jul 2016 03:30:58 +0000 (11:30 +0800)]
board: move all the rockchip board in one folder

The 'evb_rk3036' and 'kylin' is not a vendor name, let's replace them
to 'rockchip' which is a real _vendor_ name, and meet the architecure
'board/<vendor>/<board-name>/'.

More boards from rockchip like evb_rk3288, evb_rk3399 will comes later.

Signed-off-by: Kever Yang <kever.yang@rock-chips.com>
Reviewed-by: Eddie Cai <eddie.cai.kernel@gmail.com>
8 years agorockchip: add basic support for evb-rk3288 board
Xu Ziyuan [Tue, 5 Jul 2016 10:06:30 +0000 (18:06 +0800)]
rockchip: add basic support for evb-rk3288 board

evb-3288 board RK3288-based development board with 2 USB ports, HDMI,
VGA, micro-SD card, audio, WiFi and Gigabit Ethernet. It also includes
on-board 8G eMMC and 2GB of SDRAM. Expansion connector provide access to
display pins, I2C, SPI, UART and GPIOs. This add some basic files
required to allow the board to output serial messaged and can run
command(mmc info etc).

evb-rk3288 also supports booting from eMMC or SD card, the default is eMMC.

Signed-off-by: Ziyuan Xu <xzy.xu@rock-chips.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
8 years agorockchip: add option to change method of loading u-boot
Xu Ziyuan [Tue, 12 Jul 2016 11:09:49 +0000 (19:09 +0800)]
rockchip: add option to change method of loading u-boot

If we would like to boot from SD card, we have to implement mmc driver
in SPL stage, and get a slightly large SPL binary. Rockchip SoC's
bootrom code has the ability to load spl and u-boot, then boot.

If CONFIG_ROCKCHIP_SPL_BACK_TO_BROM is enabled, the spl will return to
bootrom in board_init_f(), then bootrom loads u-boot binary.

Loading sequence after rework:
bootrom ==> spl ==> bootrom ==> u-boot

Signed-off-by: Ziyuan Xu <xzy.xu@rock-chips.com>
Acked-by: Simon Glass <sjg@chromium.org>
Fixed up spelling of U-Boot, boorom, opinion->option, Rochchip:
Signed-off-by: Simon Glass <sjg@chromium.org>
8 years agoPrepare v2016.09-rc1
Tom Rini [Tue, 26 Jul 2016 02:25:52 +0000 (22:25 -0400)]
Prepare v2016.09-rc1

Signed-off-by: Tom Rini <trini@konsulko.com>
8 years agosandbox: Migrate CONFIG_I2C_EEPROM
Tom Rini [Mon, 25 Jul 2016 22:18:15 +0000 (18:18 -0400)]
sandbox: Migrate CONFIG_I2C_EEPROM

Most users of CONFIG_I2C_EEPROM were migrated to defconfig a while ago,
but sandbox was skipped.  Leave it off for sandbox_spl where it does not
build, but does not need to be either.

Cc: Simon Glass <sjg@chromium.org>
Signed-off-by: Tom Rini <trini@konsulko.com>
8 years agoMerge git://git.denx.de/u-boot-nand-flash
Tom Rini [Mon, 25 Jul 2016 18:49:54 +0000 (14:49 -0400)]
Merge git://git.denx.de/u-boot-nand-flash

8 years agoarm64: thunderx_88xx_defconfig: remove unneeded CONFIG_SYS_EXTRA_OPTIONS
Masahiro Yamada [Mon, 25 Jul 2016 13:06:08 +0000 (22:06 +0900)]
arm64: thunderx_88xx_defconfig: remove unneeded CONFIG_SYS_EXTRA_OPTIONS

ARM64 is correctly select'ed in arch/arm/Kconfig, so this line in
the defconfig is unneeded.

Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
Reviewed-by: Tom Rini <trini@konsulko.com>
8 years agodtoc: Correct the type widening code in fdt_fallback
Simon Glass [Fri, 22 Jul 2016 15:22:49 +0000 (09:22 -0600)]
dtoc: Correct the type widening code in fdt_fallback

This code does not match the fdt version in fdt.py. When dtoc is unable to
use the Python libfdt library, it uses the fallback version, which does not
widen arrays correctly.

Fix this to avoid a warning 'excess elements in array initialize' in
dt-platdata.c which happens on some platforms.

Reported-by: Tom Rini <trini@konsulko.com>
Signed-off-by: Simon Glass <sjg@chromium.org>
Tested-by: Tom Rini <trini@konsulko.com>
8 years agohashtable: Fix compiler warning on 32-bit sandbox
Simon Glass [Fri, 22 Jul 2016 15:22:48 +0000 (09:22 -0600)]
hashtable: Fix compiler warning on 32-bit sandbox

This fixes a mismatch between the %zu format and the type used on sandbox.

Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Tom Rini <trini@konsulko.com>
8 years agopart_efi: Fix compiler warning on 32-bit sandbox
Simon Glass [Fri, 22 Jul 2016 15:22:47 +0000 (09:22 -0600)]
part_efi: Fix compiler warning on 32-bit sandbox

This fixes a mismatch between the %zu format and the type used on sandbox.

Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Tom Rini <trini@konsulko.com>
8 years agolzmadec: Use the same type as the lzma call
Simon Glass [Fri, 22 Jul 2016 15:22:46 +0000 (09:22 -0600)]
lzmadec: Use the same type as the lzma call

With sandbox on 32-bit the size_t type can be a little inconsistent. Use
the same type as the caller expects to avoid a compiler warning.

Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Tom Rini <trini@konsulko.com>
8 years agosandbox: Add instructions about building on 32-bit machines
Simon Glass [Fri, 22 Jul 2016 15:22:45 +0000 (09:22 -0600)]
sandbox: Add instructions about building on 32-bit machines

Sandbox is built with 64-bit ints by default. This doesn't work properly on
32-bit machines.

Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Tom Rini <trini@konsulko.com>
8 years agotools, rsa: Further minor cleanups on top of c236ebd and 2b9ec7
mario.six@gdsys.cc [Fri, 22 Jul 2016 06:58:40 +0000 (08:58 +0200)]
tools, rsa: Further minor cleanups on top of c236ebd and 2b9ec7

[NOTE: I took v1 of these patches in, and then v2 came out, this commit
is squashing the minor deltas from v1 -> v2 of updates to c236ebd and
2b9ec76 into this commit - trini]

- Added an additional NULL check, as suggested by Simon Glass to
  fit_image_process_sig
- Re-formatted the comment blocks

Signed-off-by: Mario Six <mario.six@gdsys.cc>
Reviewed-by: Simon Glass <sjg@chromium.org>
[For merging the chnages from v2 back onto v1]
Signed-off-by: Tom Rini <trini@konsulko.com>
8 years agoARM: am33xx: Always inhibit init/refresh during DDR phy init
Russ Dill [Thu, 21 Jul 2016 11:28:32 +0000 (04:28 -0700)]
ARM: am33xx: Always inhibit init/refresh during DDR phy init

A couple of commits have modified the am33xx/am437x ddr2/ddr3
initialization path to fix certain issues, but have had the side effect
of causing L3 noc errors during initialization. The two commits are:

69b918 "am33xx,ddr3: fix ddr3 sdram configuration"
fc46ba "arm: am437x: Enable hardware leveling for EMIF"

The EMIF_REG_INITREF_DIS_MASK bit still needs to be set for all
platforms. This delays initialization and refresh until a later stage.
The 500us timer can be programmed for platforms that require it
and for platforms that don't require it. It is currently hardcoded
for 400MHz systems. For systems with a higher memory frequency
this needs to be a larger value, and for systems with a lower
memory frequency this can be a lower value. This can be
considered a separate issue and corrected in a later commit.

Signed-off-by: Russ Dill <Russ.Dill@ti.com>
Reviewed-by: Tom Rini <trini@konsulko.com>
8 years agoARM: am33xx: Fix DDR init delay placement
Russ Dill [Thu, 21 Jul 2016 11:28:31 +0000 (04:28 -0700)]
ARM: am33xx: Fix DDR init delay placement

The delay needs to be before the write to ref_ctrl register
which initiates refreshes. An improper initialization sequence
generates an L3 noc error.

Signed-off-by: Russ Dill <Russ.Dill@ti.com>
Reviewed-by: Tom Rini <trini@konsulko.com>
8 years agoefi_loader: Make exposed image loader path absolute
Alexander Graf [Wed, 20 Jul 2016 23:44:46 +0000 (01:44 +0200)]
efi_loader: Make exposed image loader path absolute

When loading an efi image, we pass it the location it was loaded from.

On file system backends, there are no relative paths, so we should always
pass in absolute ones. For network paths, we may be relative.

This fixes distro booting with grub2 for me when it fetches the grub2 config
file from the loader partition.

Reported-by: york sun <york.sun@nxp.com>
Signed-off-by: Alexander Graf <agraf@suse.de>
8 years agocommon: fit: Allow U-Boot images to be booted
mario.six@gdsys.cc [Wed, 20 Jul 2016 06:32:50 +0000 (08:32 +0200)]
common: fit: Allow U-Boot images to be booted

In certain circumstances it comes in handy to be able to boot into a second
U-Boot. But as of now it is not possible to boot a U-Boot binary that is inside
a FIT image, which is problematic for projects that e.g. need to guarantee a
unbroken chain of trust from SOC all the way into the OS, since the FIT signing
mechanism cannot be used.

This patch adds the capability to load such FIT images.

An example .its snippet (utilizing signature verification) might look
like the following:

images {
firmware@1 {
description = "2nd stage U-Boot image";
data = /incbin/("u-boot-dtb.img.gz");
type = "firmware";
arch = "arm";
os = "u-boot";
compression = "gzip";
load = <0x8FFFC0>;
entry = <0x900000>;
signature@1 {
algo = "sha256,rsa4096";
key-name-hint = "key";
};
};
};

Signed-off-by: Mario Six <mario.six@gdsys.cc>
Reviewed-by: Tom Rini <trini@konsulko.com>
8 years agokeystone: k2h/e/l: Fix DMA coherency for QM PDSP
Karicheri, Muralidharan [Tue, 19 Jul 2016 18:39:14 +0000 (14:39 -0400)]
keystone: k2h/e/l: Fix DMA coherency for QM PDSP

commit 1f807a9f32aa ("ARM: keystone2: Refactor MSMC macros to avoid
left under a macro KS2_MSMC_SEGMENT_QM_PDSP which is no longer valid.
This, in effect disabled DMA coherency for QM PDSP.

Given that msmc_k2hkle_common_setup is valid for all K2H/K/L/E SoCs,
the #ifdef should been removed in the first place. Do the same.

Fixes: 1f807a9f32aa ("ARM: keystone2: Refactor MSMC macros to avoid #ifdeffery")
Signed-off-by: Murali Karicheri <m-karicheri2@ti.com>
Acked-by: Nishanth Menon <nm@ti.com>
Reviewed-by: Tom Rini <trini@konsulko.com>
8 years agocmd: misc: Add support for fractions in sleep
mario.six@gdsys.cc [Tue, 19 Jul 2016 14:20:13 +0000 (16:20 +0200)]
cmd: misc: Add support for fractions in sleep

A feasible way to communicate certain errors for devices that have no
other way of signalling besides LEDs is to flash these LEDs. For errors
in U-Boot, a script that utilizes the led and sleep commands would be a
practicable way, but currently the sleep command can only delay for an
integral amount of seconds, which is too slow to create an easily
noticeable pattern for flashing LEDs.

Therefore, this patch adds support for fractions (down to .001 seconds)
to the sleep command.

The parsing is kept minimal, simplistic and as robust as possible: After
converting the passed string using simple_strtoul and multiplying it
with 1000, we search for the first dot, convert the three characters
after that into a number (if they are not numbers, we ignore the
fractional part and just use the delay we got from simple_strtoul), and
add this number to the delay.

Signed-off-by: Mario Six <mario.six@gdsys.cc>
8 years agomtd: fix compiler warnings
Steve Rae [Wed, 29 Jun 2016 20:50:59 +0000 (13:50 -0700)]
mtd: fix compiler warnings

- add missing declaration
- update debug output format specifiers

Signed-off-by: Steve Rae <steve.rae@raedomain.com>
8 years agomtd: nand: fix bug writing 1 byte less than page size
Hector Palacios [Mon, 18 Jul 2016 07:37:41 +0000 (09:37 +0200)]
mtd: nand: fix bug writing 1 byte less than page size

nand_do_write_ops() determines if it is writing a partial page with the
formula:
part_pagewr = (column || writelen < (mtd->writesize - 1))

When 'writelen' is exactly 1 byte less than the NAND page size the formula
equates to zero, so the code doesn't process it as a partial write, although
it should.
As a consequence the function remains in the while(1) loop with 'writelen'
becoming 0xffffffff and iterating until the watchdog timeout triggers.

To reproduce the issue on a NAND with 2K page (0x800):
=> nand erase.part <partition>
=> nand write $loadaddr <partition> 7ff

Signed-off-by: Hector Palacios <hector.palacios@digi.com>
8 years agosunxi: Enable NAND controller on the CHIP
Boris Brezillon [Wed, 15 Jun 2016 19:09:28 +0000 (21:09 +0200)]
sunxi: Enable NAND controller on the CHIP

Enable the NAND controller in the sun5i-r8-chip.dts.

Signed-off-by: Boris Brezillon <boris.brezillon@free-electrons.com>
Acked-by: Hans de Goede <hdegoede@redhat.com>
8 years agosunxi: nand: Increase CONFIG_SYS_NAND_MAX_ECCPOS value
Boris Brezillon [Wed, 15 Jun 2016 19:09:27 +0000 (21:09 +0200)]
sunxi: nand: Increase CONFIG_SYS_NAND_MAX_ECCPOS value

On some sunxi boards we have NANDs exposing 1664 OOB bytes per page.
Define the CONFIG_SYS_NAND_MAX_ECCPOS value accordingly.

Signed-off-by: Boris Brezillon <boris.brezillon@free-electrons.com>
8 years agomtd: nand: Increase the max OOB size
Boris Brezillon [Wed, 15 Jun 2016 19:09:26 +0000 (21:09 +0200)]
mtd: nand: Increase the max OOB size

Some NANDs are now exposing 1664 OOB bytes per page. Adjust the
NAND_MAX_OOBSIZE value accordingly.

Signed-off-by: Boris Brezillon <boris.brezillon@free-electrons.com>
8 years agomtd: nand: Add a full-id entry for the H27QCG8T2E5R‐BCF NAND
Boris Brezillon [Wed, 15 Jun 2016 19:09:25 +0000 (21:09 +0200)]
mtd: nand: Add a full-id entry for the H27QCG8T2E5R‐BCF NAND

Add a full-id entry for the H27QCG8T2E5R‐BCF NAND.

Signed-off-by: Boris Brezillon <boris.brezillon@free-electrons.com>
Acked-by: Hans de Goede <hdegoede@redhat.com>
8 years agosun5i: Add NAND controller to the sun5i DTSI
Maxime Ripard [Wed, 15 Jun 2016 19:09:24 +0000 (21:09 +0200)]
sun5i: Add NAND controller to the sun5i DTSI

Add the NAND controller definition to sun5i.dtsi.

Signed-off-by: Maxime Ripard <maxime.ripard@free-electrons.com>
Signed-off-by: Boris Brezillon <boris.brezillon@free-electrons.com>
8 years agomtd: nand: Add the sunxi NAND controller driver
Boris Brezillon [Wed, 15 Jun 2016 19:09:23 +0000 (21:09 +0200)]
mtd: nand: Add the sunxi NAND controller driver

We already have an SPL driver for the sunxi NAND controller, now add
the normal/standard one.

The source has been copied from Linux 4.6 with a few changes to make
it work in u-boot.

Signed-off-by: Boris Brezillon <boris.brezillon@free-electrons.com>
Acked-by: Hans de Goede <hdegoede@redhat.com>
8 years agomtd: nand: add common DT init code
Brian Norris [Wed, 15 Jun 2016 19:09:22 +0000 (21:09 +0200)]
mtd: nand: add common DT init code

These are already-documented common bindings for NAND chips. Let's
handle them in nand_base.

If NAND controller drivers need to act on this data before bringing up
the NAND chip (e.g., fill out ECC callback functions, change HW modes,
etc.), then they can do so between calling nand_scan_ident() and
nand_scan_tail().

The original commit has been slightly reworked to use the fdtdec_xxx()
helpers (instead of the of_xxxx() ones).

Signed-off-by: Brian Norris <computersforpeace@gmail.com>
Signed-off-by: Boris Brezillon <boris.brezillon@free-electrons.com>
Acked-by: Hans de Goede <hdegoede@redhat.com>
8 years agosunxi: Add missing macros to configure the NAND controller clk
Boris Brezillon [Wed, 15 Jun 2016 19:09:21 +0000 (21:09 +0200)]
sunxi: Add missing macros to configure the NAND controller clk

We need some macros to manipulate the NAND controller clock.

Signed-off-by: Boris Brezillon <boris.brezillon@free-electrons.com>
Acked-by: Hans de Goede <hdegoede@redhat.com>
8 years agocmd, nand: add an option to disable the verification when writing in raw mode
Boris Brezillon [Wed, 15 Jun 2016 08:42:18 +0000 (10:42 +0200)]
cmd, nand: add an option to disable the verification when writing in raw mode

Modern NANDs do not guarantee that data written in raw mode will not
contain bitflips just after writing them. This is fine since the number
of bitflips should be rather low and thus fixable by the ECC engine,
but since we are reading data in raw mode to verify if they match the
input data we cannot prevent failures if some bits are flipped.

The option of using standard mode to verify the data is not acceptable
either, since one of the usage of raw mode is to allow flashing images
that do not respect the standard NAND page layout or the default ECC
config (this is the case on Allwinner platforms, where the ROM code
tests several hardcoded configs, which are not necessarily matching the
NAND characteristics).

Add an extension to the nand write.raw command allowing one to disable
the verification step.

Signed-off-by: Boris Brezillon <boris.brezillon@free-electrons.com>
Reviewed-by: Tom Rini <trini@konsulko.com>
8 years agoARM: uniphier: add clock/reset settings for xHCI of ProXstream2
Masahiro Yamada [Fri, 22 Jul 2016 11:20:11 +0000 (20:20 +0900)]
ARM: uniphier: add clock/reset settings for xHCI of ProXstream2

Deassert resets and enable clock signals of xHCI blocks if the
corresponding CONFIG is enabled.

Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
8 years agoARM: uniphier: add PH1-LD21 board data
Masahiro Yamada [Fri, 22 Jul 2016 04:38:33 +0000 (13:38 +0900)]
ARM: uniphier: add PH1-LD21 board data

This has the same silicon die as PH1-LD20, but includes DRAM chips
in its package.

Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
8 years agoARM: uniphier: introduce flags to uniphier_board_data structure
Masahiro Yamada [Fri, 22 Jul 2016 04:38:32 +0000 (13:38 +0900)]
ARM: uniphier: introduce flags to uniphier_board_data structure

I need to add more board attributes, so the "flags" member will be
handier than separate boolean ones.

Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
8 years agoARM: uniphier: rename outer-cache register macros
Masahiro Yamada [Fri, 22 Jul 2016 04:38:31 +0000 (13:38 +0900)]
ARM: uniphier: rename outer-cache register macros

Sync register macros with Linux code.  This will be helpful to
develop the counterpart of Linux.

Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
8 years agoARM: uniphier: clear notification flag before L2 operation
Masahiro Yamada [Fri, 22 Jul 2016 04:38:30 +0000 (13:38 +0900)]
ARM: uniphier: clear notification flag before L2 operation

Clear the flag immediately before cache operation to not depend on
the previous state.

Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
8 years agoARM: uniphier: use (devm_)ioremap() instead of map_sysmem()
Masahiro Yamada [Tue, 19 Jul 2016 12:56:13 +0000 (21:56 +0900)]
ARM: uniphier: use (devm_)ioremap() instead of map_sysmem()

This does not have much impact on behavior, but makes code look more
more like Linux.  The use of devm_ioremap() often helps to delete
.remove callbacks entirely.

Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
8 years agoARM: uniphier: fix doubled tftpboot commands
Masahiro Yamada [Tue, 19 Jul 2016 02:36:53 +0000 (11:36 +0900)]
ARM: uniphier: fix doubled tftpboot commands

This downloads the same file twice for nothing.

Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
8 years agoARM: uniphier: select CONFIG_ARMV8_SPIN_TABLE
Masahiro Yamada [Sat, 16 Jul 2016 16:38:21 +0000 (01:38 +0900)]
ARM: uniphier: select CONFIG_ARMV8_SPIN_TABLE

This is needed when booting Linux without ARM Trusted Firmware.

Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
8 years agoARM: dts: uniphier: renumber serial aliases for Gentil/Vodka boards
Masahiro Yamada [Wed, 6 Jul 2016 10:24:09 +0000 (19:24 +0900)]
ARM: dts: uniphier: renumber serial aliases for Gentil/Vodka boards

On these two boards, the serial0 is used for inter-chip connection,
so cannot be used for login console.  The serial2 is used instead
for them, but it is tedious to use because upper level deployment
projects must switch login console per board.

[ Linux commit: 2a4a2aadbaad9dffdb564a2895348f3d8e825416 ]

Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
Signed-off-by: Olof Johansson <olof@lixom.net>
8 years agotest/py: vboot can be run only at Sandbox
Michal Simek [Mon, 18 Jul 2016 06:49:08 +0000 (08:49 +0200)]
test/py: vboot can be run only at Sandbox

Getting this error:
Zynq> sb load hostfs - 100
/home/monstr/data/disk/u-boot/build-zynq_zc706/test.fit
Unknown command 'sb' - try 'help'

because sb command is present only for Sandbox
obj-$(CONFIG_SANDBOX) += host.o

that's why mark this test to be run only at Sandbox

Signed-off-by: Michal Simek <michal.simek@xilinx.com>
Acked-by: Simon Glass <sjg@chromium.org>
8 years agoarm: omap5: fix build dependency for secure devices
Andreas Dannenberg [Wed, 20 Jul 2016 18:55:59 +0000 (13:55 -0500)]
arm: omap5: fix build dependency for secure devices

Commit 17c2987 introduces an undesired dependency on CONFIG_SPL_LOAD_FIT
when building U-Boot for AM57xx and DRA7xx high-security (HS) devices that
causes the build to break when that option is not active. Fix this issue
by only building the u-boot_HS.img target when building U-Boot into an
actual FIT image.

Signed-off-by: Andreas Dannenberg <dannenberg@ti.com>
Reviewed-by: Tom Rini <trini@konsulko.com>
8 years agoarm: am4x: fix build dependency for secure devices
Andreas Dannenberg [Wed, 20 Jul 2016 18:55:58 +0000 (13:55 -0500)]
arm: am4x: fix build dependency for secure devices

Commit e29878f introduces an undesired dependency on CONFIG_SPL_LOAD_FIT
when building U-Boot for AM43xx high-security (HS) devices that causes the
build to break when that option is not active. Fix this issue by only
building the u-boot_HS.img target when building U-Boot into an actual
FIT image.

Signed-off-by: Andreas Dannenberg <dannenberg@ti.com>
Reviewed-by: Tom Rini <trini@konsulko.com>
8 years agofixing typo error in README file. CPU15 -> CP15
yeongjun Kim [Wed, 20 Jul 2016 13:56:12 +0000 (22:56 +0900)]
fixing typo error in README file. CPU15 -> CP15

It looks typo error.
Not CPU15, CP15(CoProcessor15)

Signed-off-by: yeongjun Kim <iam.yeongjunkim@gmail.com>
8 years agospl: fit: Fix the number of bytes read in raw mode
Lokesh Vutla [Tue, 19 Jul 2016 09:26:14 +0000 (14:56 +0530)]
spl: fit: Fix the number of bytes read in raw mode

In raw mode a full sector is to be read even if image covers part of
a sector. Number of sectors are calculated as ROUND_UP(size)/sec_size by FIT
framework. This calculation assumes that image is at the 0th offset of a sector,
which is not true always in FIT case. So, include the image offset while
calculating number of sectors.

Signed-off-by: Lokesh Vutla <lokeshvutla@ti.com>
Reviewed-by: Tom Rini <trini@konsulko.com>
8 years agorsa: Fix return value and masked error
mario.six@gdsys.cc [Tue, 19 Jul 2016 09:07:07 +0000 (11:07 +0200)]
rsa: Fix return value and masked error

When signing images, we repeatedly call fit_add_file_data() with
successively increasing size values to include the keys in the DTB.

Unfortunately, if large keys are used (such as 4096 bit RSA keys), this
process fails sometimes, and mkimage needs to be called repeatedly to
integrate the keys into the DTB.

This is because fit_add_file_data actually returns the wrong error
code, and the loop terminates prematurely, instead of trying again with
a larger size value.

This patch corrects the return value by fixing the return value of
fdt_add_bignum, fixes a case where an error is masked by a unconditional
setting of a return value variable, and also removes a error message,
which is misleading, since we actually allow the function to fail. A
(hopefully helpful) comment is also added to explain the lack of error
message.

This is probably related to 1152a05 ("tools: Correct error handling in
fit_image_process_hash()") and the corresponding error reported here:

https://www.mail-archive.com/u-boot@lists.denx.de/msg217417.html

Signed-off-by: Mario Six <mario.six@gdsys.cc>
8 years agotools: Fix return code of fit_image_process_sig()
mario.six@gdsys.cc [Tue, 19 Jul 2016 09:07:06 +0000 (11:07 +0200)]
tools: Fix return code of fit_image_process_sig()

When signing images, we repeatedly call fit_add_file_data() with
successively increasing size values to include the keys in the DTB.

Unfortunately, if large keys are used (such as 4096 bit RSA keys), this
process fails sometimes, and mkimage needs to be called repeatedly to
integrate the keys into the DTB.

This is because fit_add_file_data actually returns the wrong error
code, and the loop terminates prematurely, instead of trying again with
a larger size value.

This patch corrects the return value and also removes a error message,
which is misleading, since we actually allow the function to fail. A
(hopefully helpful) comment is also added to explain the lack of error
message.

This is probably related to 1152a05 ("tools: Correct error handling in
fit_image_process_hash()") and the corresponding error reported here:

https://www.mail-archive.com/u-boot@lists.denx.de/msg217417.html

Signed-off-by: Mario Six <mario.six@gdsys.cc>
8 years agotest/py: use absolute dts path in vboot test
Stephen Warren [Mon, 18 Jul 2016 16:07:25 +0000 (10:07 -0600)]
test/py: use absolute dts path in vboot test

Without this, the test fails if the test is run with a cwd other than the
root of the U-Boot source tree.

Fixes: 8729d582595d ("test: Convert the vboot test to test/py")
Signed-off-by: Stephen Warren <swarren@nvidia.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
8 years agoarmv8: spl: Call board_init_r from crt0_64 in SPL
Jeremy Hunt [Mon, 18 Jul 2016 16:01:02 +0000 (12:01 -0400)]
armv8: spl: Call board_init_r from crt0_64 in SPL

As part of the startup process for boards using the SPL, the
meaning of board_init_f changed such that it should return normally
rather than calling board_init_r directly. (see
db910353a126d84fe8dff7a694ea792f50fcfb6a )
This was fixed in 32-bit arm, but broke when SPL was added to
64 bit arm. This fixes crt0_64 so that it calls board_init_r
during the SPL and removes the direct call from board_init_f
from the arm SPL example.

Signed-off-by: Jeremy Hunt <Jeremy.Hunt@DEShawResearch.com>
Acked-by: Simon Glass <sjg@chromium.org>