platform/kernel/u-boot.git
9 years agodm: Init device tree as well as driver model in SPL
Simon Glass [Sat, 28 Feb 2015 05:06:42 +0000 (22:06 -0700)]
dm: Init device tree as well as driver model in SPL

If enabled, make sure that the device tree is available in SPL before
setting up driver model.

Signed-off-by: Simon Glass <sjg@chromium.org>
9 years agodm: core: Select device tree control correctly for SPL
Simon Glass [Sat, 28 Feb 2015 05:06:41 +0000 (22:06 -0700)]
dm: core: Select device tree control correctly for SPL

Some boards will not use device tree for SPL even with driver model. Add
the logic to support this.

Signed-off-by: Simon Glass <sjg@chromium.org>
9 years agofdt: Allow FDT functions to be built for SPL
Simon Glass [Sat, 28 Feb 2015 05:06:40 +0000 (22:06 -0700)]
fdt: Allow FDT functions to be built for SPL

Remove the implicit assumption that SPL does not support device tree.

Signed-off-by: Simon Glass <sjg@chromium.org>
9 years agofdt: Add an option to disable device tree in SPL
Simon Glass [Sat, 28 Feb 2015 05:06:38 +0000 (22:06 -0700)]
fdt: Add an option to disable device tree in SPL

Some boards cannot support device tree due to lack of memory. Add an option
to allow these boards to continue to work (and even use driver model).
This is a 'negative' option since most boards are expected to support device
tree in SPL.

Signed-off-by: Simon Glass <sjg@chromium.org>
9 years agoCorrect malloc_limit value for pre-relocation malloc()
Simon Glass [Sat, 28 Feb 2015 05:06:37 +0000 (22:06 -0700)]
Correct malloc_limit value for pre-relocation malloc()

The limit is measured from the start of the malloc() area and is not an
absolute address. Correct this.

Signed-off-by: Simon Glass <sjg@chromium.org>
9 years agoMove initf_malloc() to a common place
Simon Glass [Sat, 28 Feb 2015 05:06:36 +0000 (22:06 -0700)]
Move initf_malloc() to a common place

To allow this function to be used from SPL, move it to the malloc()
code.

Signed-off-by: Simon Glass <sjg@chromium.org>
9 years agofdt: Rename setup_fdt() and make it prepare also
Simon Glass [Sat, 28 Feb 2015 05:06:35 +0000 (22:06 -0700)]
fdt: Rename setup_fdt() and make it prepare also

There is little reason to split these two functions. Bring them together
which simplifies the init sequence.

Signed-off-by: Simon Glass <sjg@chromium.org>
9 years agofdt: sandbox: Move setup code from board_f to fdtdec
Simon Glass [Sat, 28 Feb 2015 05:06:34 +0000 (22:06 -0700)]
fdt: sandbox: Move setup code from board_f to fdtdec

We want to be able to set up the device tree in SPL, so move this code
to a common place.

Signed-off-by: Simon Glass <sjg@chromium.org>
9 years agodm: core: Drop device removal error path when not supported
Simon Glass [Sat, 28 Feb 2015 05:06:33 +0000 (22:06 -0700)]
dm: core: Drop device removal error path when not supported

When CONFIG_DM_DEVICE_REMOVE is not enabled, such as in SPL, we cannot
remove or unbind devices and do not expect to get errors when binding
and probing devices. So drop the error path to reduce code size.

Signed-off-by: Simon Glass <sjg@chromium.org>
9 years agodm: Add a panic_str() function to reduce code size
Simon Glass [Sat, 28 Feb 2015 05:06:32 +0000 (22:06 -0700)]
dm: Add a panic_str() function to reduce code size

The printf() in panic() adds about 1.5KB of code size to SPL when compiled
with Thumb-2. Provide a smaller version that does not support printf()-style
arguments and use it in two commonly compiled places.

Signed-off-by: Simon Glass <sjg@chromium.org>
9 years agodm: core: Remove unbind operations when not required
Simon Glass [Sat, 28 Feb 2015 05:06:31 +0000 (22:06 -0700)]
dm: core: Remove unbind operations when not required

The CONFIG_DM_DEVICE_REMOVE option takes out code related to removing
devices. It should also remove the 'unbind' code since if we cannot
remove we probably don't need to unbind.

Signed-off-by: Simon Glass <sjg@chromium.org>
9 years agodm: core: Allow sequence alias support to be removed for SPL
Simon Glass [Sat, 28 Feb 2015 05:06:30 +0000 (22:06 -0700)]
dm: core: Allow sequence alias support to be removed for SPL

In many cases SPL only uses a single serial port and there is no need for
alias sequence support. We will just use the serial port pointed to by
stdout-path in the /chosen node.

Signed-off-by: Simon Glass <sjg@chromium.org>
9 years agoserial: ns16550: Remove unnecessary init on UART setup
Simon Glass [Sat, 28 Feb 2015 05:06:28 +0000 (22:06 -0700)]
serial: ns16550: Remove unnecessary init on UART setup

It is not necessary to write a zero baud rate to the device, and for some
chips this will cause problems. Drop this code.

Signed-off-by: Simon Glass <sjg@chromium.org>
9 years agodm: ns16550: Support non-byte register spacing with driver model
Simon Glass [Sat, 28 Feb 2015 05:06:26 +0000 (22:06 -0700)]
dm: ns16550: Support non-byte register spacing with driver model

Allow this driver to support boards where the register shift is not 0.
This fixes some compiler warnings which appear in that case.

Signed-off-by: Simon Glass <sjg@chromium.org>
9 years agoserial: ns16550: Add an option to specify the debug UART register shift
Simon Glass [Sat, 28 Feb 2015 05:06:25 +0000 (22:06 -0700)]
serial: ns16550: Add an option to specify the debug UART register shift

This UART permits different register spacing. To support the debug UART on
devices which have a spacing other than 1 byte, allow the shift value to
be specified.

Signed-off-by: Simon Glass <sjg@chromium.org>
9 years agoexynos: sandbox: ti: Add SPDX license identifiers and notes
Simon Glass [Mon, 13 Apr 2015 17:19:07 +0000 (11:19 -0600)]
exynos: sandbox: ti: Add SPDX license identifiers and notes

For some files I neglected to add a license. Rectify this:

arch/arm/dts/exynos4210-pinctrl-uboot.dtsi
arch/arm/dts/exynos4x12-pinctrl-uboot.dtsi
arch/arm/dts/exynos5250-pinctrl-uboot.dtsi
arch/arm/dts/exynos54xx-pinctrl-uboot.dtsi
arch/arm/dts/s5pc100-pinctrl.dtsi
arch/arm/dts/s5pc110-pinctrl.dtsi

This file came from Linux and has no license information there, so add a
comment to that effect:

arch/sandbox/include/asm/bitops.h

This file also came from Linux - presumably someone from TI could add the
license:

include/dt-bindings/pinctrl/omap.h

Signed-off-by: Simon Glass <sjg@chromium.org>
Reported-by: Ingrid Viitanen <ingrid.viitanen@nokia.com>
9 years agodm: test: Don't clear global_data in dm_test_uclass_before_ready()
Simon Glass [Sun, 19 Apr 2015 13:21:02 +0000 (07:21 -0600)]
dm: test: Don't clear global_data in dm_test_uclass_before_ready()

We must not clear global_data even in tests, since the ram_buffer (which
is used by malloc()) will also be lost, and subsequent tests will fail.

Zero only the global_data fields that are required for the test to function.

Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Joe Hershberger <joe.hershberger@ni.com>
Tested-by: Joe Hershberger <joe.hershberger@ni.com>
9 years agodm: Update the README to reflect the current test output
Simon Glass [Sun, 19 Apr 2015 13:21:01 +0000 (07:21 -0600)]
dm: Update the README to reflect the current test output

There are a lot more tests now. To avoid confusion add the updated test
output to the driver model README.

Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Joe Hershberger <joe.hershberger@ni.com>
9 years agodm: usb: Add a terminator to the string destructor list
Simon Glass [Sun, 19 Apr 2015 13:20:59 +0000 (07:20 -0600)]
dm: usb: Add a terminator to the string destructor list

The terminator is missing. Add it for completeness.

Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Joe Hershberger <joe.hershberger@ni.com>
Tested-by: Joe Hershberger <joe.hershberger@ni.com>
9 years agodm: core: Handle recursive unbinding of uclass devices
Simon Glass [Sun, 19 Apr 2015 13:20:58 +0000 (07:20 -0600)]
dm: core: Handle recursive unbinding of uclass devices

Since a device can have children in the same uclass as itself, we need
to handle unbinding carefully: we must allow that unbinding a device in a
uclass may cause another device in the same uclass to be unbound.

Adjust the code to cope.

Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Joe Hershberger <joe.hershberger@ni.com>
Tested-by: Joe Hershberger <joe.hershberger@ni.com>
9 years agodm: core: precise comments for get/find device by name
Przemyslaw Marczak [Mon, 20 Apr 2015 11:32:34 +0000 (13:32 +0200)]
dm: core: precise comments for get/find device by name

The functions:
- uclass_find_device_by_name()
- uclass_get_device_by_name()
searches the required device for the exactly given name.
This patch, presice this fact for both function's comments.

Signed-off-by: Przemyslaw Marczak <p.marczak@samsung.com>
Cc: Simon Glass <sjg@chromium.org>
Acked-by: Simon Glass <sjg@chromium.org>
9 years agodm: test: Add tests for get/find uclass's device by name
Przemyslaw Marczak [Mon, 20 Apr 2015 11:32:33 +0000 (13:32 +0200)]
dm: test: Add tests for get/find uclass's device by name

This commit introduces simple tests for functions:
- uclass_find_device_by_name()
- uclass_get_device_by_name()

Tests added by this commit:
- Test: dm_test_uclass_devices_find_by_name: for uclass id: UCLASS_TEST_FDT
 * get uclass's devices by uclass_find_first/next_device() each as 'testdev',
 * for each returned device, call: uclass_find_device_by_name(),
   with previously returned device's name as an argument ('testdev->name').
 * for the found device ('founddev') check if:
   * founddev != NULL
   * testdev == founddev
   * testdev->name == founddev->name (by strcmp)

- Test: dm_test_uclass_devices_get_by_name: for uclass id: UCLASS_TEST_FDT
 * get uclass's devices by uclass_get_first/next_device() each as 'testdev',
 * for each returned device, call: uclass_get_device_by_name(),
   with previously returned device's name as an argument ('testdev->name').
 * for the found device ('founddev') check if:
   * founddev != NULL
   * founddev is active
   * testdev == founddev
   * testdev->name == founddev->name (by strcmp)

Signed-off-by: Przemyslaw Marczak <p.marczak@samsung.com>
Cc: Simon Glass <sjg@chromium.org>
Acked-by: Simon Glass <sjg@chromium.org>
9 years agodm: core: remove type 'static' of function uclass_get_device_tail()
Przemyslaw Marczak [Mon, 20 Apr 2015 11:32:32 +0000 (13:32 +0200)]
dm: core: remove type 'static' of function uclass_get_device_tail()

Uclass API provides a few functions for get/find the device.
To provide a complete function set of uclass-internal functions,
for use by the drivers, the function uclass_get_device_tail()
should be non-static.

Signed-off-by: Przemyslaw Marczak <p.marczak@samsung.com>
Cc: Simon Glass <sjg@chromium.org>
Acked-by: Simon Glass <sjg@chromium.org>
9 years agodm: core: device: add function: dev_get_uclass_name()
Przemyslaw Marczak [Wed, 15 Apr 2015 11:07:25 +0000 (13:07 +0200)]
dm: core: device: add function: dev_get_uclass_name()

This commit extends the driver model device's API by function:
- dev_get_uclass_name()

And this function returns the device's uclass driver name if:
- given dev pointer, is non_NULL
otherwise, the NULL pointer is returned.

Signed-off-by: Przemyslaw Marczak <p.marczak@samsung.com>
Cc: Simon Glass <sjg@chromium.org>
Acked-by: Simon Glass <sjg@chromium.org>
9 years agodm: core: device: add function: dev_get_driver_ops()
Przemyslaw Marczak [Wed, 15 Apr 2015 11:07:24 +0000 (13:07 +0200)]
dm: core: device: add function: dev_get_driver_ops()

This commit extends the driver model device's API by function:
- dev_get_driver_ops()

And this function returns the device's driver's operations if given:
- dev pointer, is non-NULL
- dev->driver->ops pointer, is non-NULL
in other case the, the NULL pointer is returned.

Signed-off-by: Przemyslaw Marczak <p.marczak@samsung.com>
Cc: Simon Glass <sjg@chromium.org>
Acked-by: Simon Glass <sjg@chromium.org>
9 years agodm: core: uclass: add function: uclass_get_device_by_name()
Przemyslaw Marczak [Wed, 15 Apr 2015 11:07:23 +0000 (13:07 +0200)]
dm: core: uclass: add function: uclass_get_device_by_name()

This commit extends the driver model uclass's API by function:
- uclass_get_device_by_name()

And this function returns the device if:
- uclass with given ID, exists,
- device with exactly given name(dev->name), exists,
- device probe, doesn't return an error.

The returned device is activated and ready to use.

Note:
This function returns the first device, which name is equal
to the given one. This means, that using this function you must
assume, that the device name is unique in the given uclass's ID
device list.

Signed-off-by: Przemyslaw Marczak <p.marczak@samsung.com>
Cc: Simon Glass <sjg@chromium.org>
Acked-by: Simon Glass <sjg@chromium.org>
9 years agodm: core: uclass: add function: uclass_find_device_by_name()
Przemyslaw Marczak [Wed, 15 Apr 2015 11:07:22 +0000 (13:07 +0200)]
dm: core: uclass: add function: uclass_find_device_by_name()

This commit extends the driver model uclass's API by function:
- uclass_find_device_by_name()

And this function returns the device if:
- uclass with given ID, exists,
- device with exactly given name(dev->name), exists.

The returned device is not activated - need to be probed before use.

Note:
This function returns the first device, which name is equal
to the given one. This means, that using this function you must
assume, that the device name is unique in the given uclass's ID
device list.

uclass-internal.h: cleanup - move the uclass_find_device_by_seq()
declaration and description, near the other uclass_find*() functions.

Signed-off-by: Przemyslaw Marczak <p.marczak@samsung.com>
Cc: Simon Glass <sjg@chromium.org>
Acked-by: Simon Glass <sjg@chromium.org>
9 years agodm: test: Add tests for get/find uclass devices
Przemyslaw Marczak [Wed, 15 Apr 2015 11:07:20 +0000 (13:07 +0200)]
dm: test: Add tests for get/find uclass devices

This commit introduces simple tests for functions:
- uclass_find_first_device()
- uclass_find_next_device()
- uclass_first_device()
- uclass_next_device()

Tests added by this commit:
- Test: dm_test_uclass_devices_find:
  * call uclass_find_first_device(), then check if: (dev != NULL), (ret == 0)
  * for the rest devices, call uclass_find_next_device() and do the same check

- Test: dm_test_uclass_devices_get:
  * call uclass_first_device(), then check if:
    -- (dev != NULL), (ret == 0), device_active()
  * for the rest devices, call uclass_next_device() and do the same check

Signed-off-by: Przemyslaw Marczak <p.marczak@samsung.com>
Cc: Simon Glass <sjg@chromium.org>
Acked-by: Simon Glass <sjg@chromium.org>
9 years agodm: test: Add tests for device's uclass platform data
Przemyslaw Marczak [Wed, 15 Apr 2015 11:07:19 +0000 (13:07 +0200)]
dm: test: Add tests for device's uclass platform data

This test introduces new test structure type:dm_test_perdev_uc_pdata.
The structure consists of three int values only. For the test purposes,
three pattern values are defined by enum, starting with TEST_UC_PDATA_INTVAL1.

This commit adds two test cases for uclass platform data:
- Test: dm_test_autobind_uclass_pdata_alloc - this tests if:
  * uclass driver sets: .per_device_platdata_auto_alloc_size field
  * the devices's: dev->uclass_platdata is non-NULL

- Test: dm_test_autobind_uclass_pdata_valid - this tests:
  * if the devices's: dev->uclass_platdata is non-NULL
  * the structure of type 'dm_test_perdev_uc_pdata' allocated at address
    pointed by dev->uclass_platdata. Each structure field, should be equal
    to proper pattern data, starting from .intval1 == TEST_UC_PDATA_INTVAL1.

Signed-off-by: Przemyslaw Marczak <p.marczak@samsung.com>
Cc: Simon Glass <sjg@chromium.org>
Acked-by: Simon Glass <sjg@chromium.org>
9 years agodm: core: Extend struct udevice by '.uclass_platdata' field.
Przemyslaw Marczak [Wed, 15 Apr 2015 11:07:18 +0000 (13:07 +0200)]
dm: core: Extend struct udevice by '.uclass_platdata' field.

This commit adds 'uclass_platdata' field to 'struct udevice', which
can be automatically allocated at bind. The allocation size is defined
in 'struct uclass_driver' as 'per_device_platdata_auto_alloc_size'.

New device's flag is added: DM_FLAG_ALLOC_UCLASS_PDATA, which is used
for memory freeing at device unbind method.

As for other udevice's fields, a complementary function is added:
- dev_get_uclass_platdata()

Signed-off-by: Przemyslaw Marczak <p.marczak@samsung.com>
Cc: Simon Glass <sjg@chromium.org>
Acked-by: Simon Glass <sjg@chromium.org>
9 years agodm: core: add internal functions for getting the device without probe
Przemyslaw Marczak [Wed, 15 Apr 2015 11:07:17 +0000 (13:07 +0200)]
dm: core: add internal functions for getting the device without probe

This commit extends the uclass-internal functions by:
- uclass_find_first_device()
- uclass_find_next_device()
For both functions, the returned device is not probed.

After some cleanup, the above functions are called by:
- uclass_first_device()
- uclass_next_device()
for which, the returned device is probed.

Signed-off-by: Przemyslaw Marczak <p.marczak@samsung.com>
Cc: Simon Glass <sjg@chromium.org>
Acked-by: Simon Glass <sjg@chromium.org>
9 years agoserial: pl01x: fix PL010 regression
Linus Walleij [Tue, 21 Apr 2015 13:10:06 +0000 (15:10 +0200)]
serial: pl01x: fix PL010 regression

commit aed2fbef5e9a0ab5a7cd01e742039a962f0b24ef
"dm: serial: Tidy up the pl01x driver"
caused a regression on (real hardware) PL010 by omitting
to update the line control register when switching baudrate.

Fix this by inlining the missing write to the baud control
register.

Also renaming the set_line_control() function to
pl011_set_line_control() since this function is clearly
PL011-specific, and it won't suffice to call that to
set up line control.

Tested on the Integrator/AP hardware.

Cc: Simon Glass <sjg@chromium.org>
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
9 years agokconfig: remove duplicated CMD_DNS option
Andrey Skvortsov [Sun, 19 Apr 2015 11:58:43 +0000 (14:58 +0300)]
kconfig: remove duplicated CMD_DNS option

two CMD_DNS options were added by commit 60296a835cb17 ("commands: add more
command entries in Kconfig")

Signed-off-by: Andrey Skvortsov <andrej.skvortzov@gmail.com>
Acked-by: Masahiro Yamada <yamada.masahiro@socionext.com>
9 years agoMerge branch 'master' of git://git.denx.de/u-boot-net
Tom Rini [Tue, 21 Apr 2015 00:16:21 +0000 (20:16 -0400)]
Merge branch 'master' of git://git.denx.de/u-boot-net

9 years agonet: pch_gbe: Fix pch_gbe device name
Bin Meng [Wed, 15 Apr 2015 03:18:20 +0000 (11:18 +0800)]
net: pch_gbe: Fix pch_gbe device name

The name "pch_gbe.%x" exceeds the limit of the name in the
'struct eth_device'. Rename it as just "pch_gbe".

Signed-off-by: Bin Meng <bmeng.cn@gmail.com>
9 years agonet: gem: Use correct type for casting
Michal Simek [Wed, 15 Apr 2015 11:31:28 +0000 (13:31 +0200)]
net: gem: Use correct type for casting

Use phys_addr_t which is used in function prototype
in system.h.

Signed-off-by: Michal Simek <michal.simek@xilinx.com>
9 years agonet/phy: fixup for get_phy_id
Shengzhou Liu [Tue, 7 Apr 2015 10:46:32 +0000 (18:46 +0800)]
net/phy: fixup for get_phy_id

commit 3c6928fd7b0f84 "net: phy: fix warnings with W=1" caused
some PHYs(e.g. CS4315/CS4340) not working. This patch fixes the
warning and make those special PHYs working as well.

Signed-off-by: Shengzhou Liu <Shengzhou.Liu@freescale.com>
9 years agonet: phy: micrel: add support for KSZ8081MNX
Luca Ellero [Tue, 24 Mar 2015 10:32:24 +0000 (11:32 +0100)]
net: phy: micrel: add support for KSZ8081MNX

This patch adds a support for KSZ8081MNX in MII mode.

Signed-off-by: Luca Ellero <luca.ellero@brickedbrain.com>
Acked-by: Pavel Machek <pavel@denx.de>
9 years agomii: add read-modify-write option to mii command
Tim James [Wed, 25 Mar 2015 11:55:15 +0000 (11:55 +0000)]
mii: add read-modify-write option to mii command

When accessing PHY registers it is often desirable to only update
selected bits, so it is necessary to first read the current value
before writing back an modified value with the relevant bits
updated.

To simplify this and to allow such operations to be incorporated
into simple shell scripts propose adding a 'modify' option to the
existing mii command, which takes a mask indicating the bits to
be updated in addition to a data value containing the new bits,
ie, <updated> = (<data> & <mask>) | (<current> & ~<mask>).

Signed-off-by: Tim <tim.james@macltd.com>
Cc: Nobuhiro Iwamatsu <iwamatsu@nigauri.org>
Cc: Joe Hershberger <joe.hershberger@gmail.com>
Cc: Jeroen Hofstee <jeroen@myspectrum.nl>
Cc: Tom Rini <trini@konsulko.com>
Cc: Tim <tim.james@macltd.com>
9 years agoUpdate MAINTAINERS and git-mailrc for net
Joe Hershberger [Fri, 20 Mar 2015 18:25:57 +0000 (13:25 -0500)]
Update MAINTAINERS and git-mailrc for net

Update to my corporate email and make the supported filter and aliases
more accurate.

Signed-off-by: Joe Hershberger <joe.hershberger@ni.com>
9 years agonet: rtl8169: Build warning fixes for 64-bit
Thierry Reding [Fri, 20 Mar 2015 11:41:21 +0000 (12:41 +0100)]
net: rtl8169: Build warning fixes for 64-bit

Turn ioaddr into an unsigned long rather than a sized 32-bit variable.
While at it, fix a couple of pointer to integer cast size mismatch
warnings by casting through unsigned long going from pointers to
integers and vice versa.

Cc: Joe Hershberger <joe.hershberger@gmail.com>
Signed-off-by: Thierry Reding <treding@nvidia.com>
Acked-by: Joe Hershberger <joe.hershberger@ni.com>
9 years agonet: phy: realtek: Disable interrupt on Realtek Ethernet PHY drivers
Codrin Ciubotariu [Fri, 13 Feb 2015 12:47:58 +0000 (14:47 +0200)]
net: phy: realtek: Disable interrupt on Realtek Ethernet PHY drivers

Some Realtek Ethernet PHYs, like RTL8211D(G/N) and RTL8211E(G), have
interrupts enabled by default. If the interrupt is not treated later by
the OS and the PHY's interrupt line is enabled and shared with other
interrupts, the system will get an interrupt storm. This patch disables
the interrupt for PHY devices that use one of the current Realtek
Ethernet PHY drivers. Some of Realtek Ethernet PHYs, such as RTL8211B(L)
have the interrupt masked. In this case, the functionality of the PHY
should not be afected since this patch brings INER and INSR registers to
their default values.

Signed-off-by: Codrin Ciubotariu <codrin.ciubotariu@freescale.com>
Acked-by: Joe Hershberger <joe.hershberger@ni.com>
9 years agoMerge branch 'master' of git://git.denx.de/u-boot-mpc85xx
Tom Rini [Mon, 20 Apr 2015 21:12:45 +0000 (17:12 -0400)]
Merge branch 'master' of git://git.denx.de/u-boot-mpc85xx

9 years agopowerpc/mpc8641hpcn: Move environment to avoid conflict
Scott Wood [Wed, 15 Apr 2015 21:13:48 +0000 (16:13 -0500)]
powerpc/mpc8641hpcn: Move environment to avoid conflict

U-Boot on this board grew a long time ago past the 384 KiB that
it reserves for the U-Boot image, before the environment.  Thus,
saveenv overwrites the U-Boot image and bricks the board.

I tried to find out when U-Boot grew beyond this point, but there is a
long stretch in the history where this board did not build -- and
AFAICT when it did fit in 384 KiB, it was missing vital features such
as fdt support.  Turning off CONFIG_VIDEO was not enough to make it
fit.  Thus, I don't think we have any choice other than to move the
environment.

Signed-off-by: Scott Wood <scottwood@freescale.com>
Reviewed-by: York Sun <yorksun@freescale.com>
9 years agoboard/t2080rdb: enable CONFIG_PHY_AQUANTIA
Shengzhou Liu [Wed, 8 Apr 2015 03:12:15 +0000 (11:12 +0800)]
board/t2080rdb: enable CONFIG_PHY_AQUANTIA

CONFIG_PHY_AQ1202 is no longer needed, use CONFIG_PHY_AQUANTIA.

Signed-off-by: Shengzhou Liu <Shengzhou.Liu@freescale.com>
Reviewed-by: York Sun <yorksun@freescale.com>
9 years agopowerpc/mpc85xx: Remove some dead code
Scott Wood [Wed, 8 Apr 2015 01:20:01 +0000 (20:20 -0500)]
powerpc/mpc85xx: Remove some dead code

U-Boot does not have system calls (the services it exposes to
standalone commands use a different mechanism), so the syscall handler
is dead code.  It's also broken code, as it assumes it is located at
0xc00 -- while even before the patch to stop relocating exception
vectors to 0, U-Boot had the syscall at 0x900.

The critical and machine check return paths are never called -- the
regular exception return path is used instead, which works because
xSRR0/1 have already been saved and can be restored via the regular
SRR0/1 (we don't care too much in U-Boot about taking a critical/mcheck
inside another exception prolog/epilog).

Also remove a few other small unused functions.

Signed-off-by: Scott Wood <scottwood@freescale.com>
Reviewed-by: York Sun <yorksun@freescale.com>
9 years agopowerpc/mpc85xx: Don't relocate exception vectors
Scott Wood [Wed, 8 Apr 2015 01:20:00 +0000 (20:20 -0500)]
powerpc/mpc85xx: Don't relocate exception vectors

Booke does not require exception vectors to be located at address zero.
U-Boot was doing so anyway, simply because that's how it had been done
on other PPC.  The downside of this is that once the OS is loaded to
address zero, the exception vectors have been overwritten -- which
makes it difficult to diagnose a crash that happens after that point.

The IVOR setup and trap entry code is simplified somewhat as a result.

Also, there is no longer a need to align individual exceptions on 0x100
byte boundaries.

Signed-off-by: Scott Wood <scottwood@freescale.com>
Reviewed-by: York Sun <yorksun@freescale.com>
9 years agopowerpc/t2080rdb: update ddr to support 1866MT/s
Shengzhou Liu [Fri, 27 Mar 2015 07:53:14 +0000 (15:53 +0800)]
powerpc/t2080rdb: update ddr to support 1866MT/s

Support SODIMM D3XP12081XL10AA 1866MT/s on T2080RDB.
Enable CONFIG_CMD_MEMTEST as well.

Signed-off-by: Shengzhou Liu <Shengzhou.Liu@freescale.com>
Reviewed-by: York Sun <yorksun@freescale.com>
9 years agoT4240RDB: Enable CONFIG_SYS_CORTINA_FW_IN_NOR config
Chunhe Lan [Tue, 24 Mar 2015 07:10:41 +0000 (15:10 +0800)]
T4240RDB: Enable CONFIG_SYS_CORTINA_FW_IN_NOR config

Now cortina driver uses macro CONFIG_SYS_CORTINA_FW_IN_NOR
to define that firmware of cortina driver is stored in the
nor flash.

Signed-off-by: Chunhe Lan <Chunhe.Lan@freescale.com>
Reviewed-by: York Sun <yorksun@freescale.com>
9 years agoboard/t208xrdb: VID support
Ying Zhang [Tue, 10 Mar 2015 06:21:36 +0000 (14:21 +0800)]
board/t208xrdb: VID support

The fuse status register provides the values from on-chip
voltage ID efuses programmed at the factory.
These values define the voltage requirements for
the chip. u-boot reads FUSESR and translates the values
into the appropriate commands to set the voltage output
value of an external voltage regulator.

Signed-off-by: Ying Zhang <b40530@freescale.com>
Reviewed-by: York Sun <yorksun@freescale.com>
9 years agopowerpc/t2080: enable erratum_a007186 for t2080 rev1.1
Shengzhou Liu [Mon, 9 Mar 2015 09:12:22 +0000 (17:12 +0800)]
powerpc/t2080: enable erratum_a007186 for t2080 rev1.1

T2080 rev1.1 also needs erratum a007186.

Signed-off-by: Shengzhou Liu <Shengzhou.Liu@freescale.com>
Reviewed-by: York Sun <yorksun@freescale.com>
9 years agoqemu-ppce500: Add support for 64bit CCSR map
Alexander Graf [Sat, 7 Mar 2015 01:10:09 +0000 (02:10 +0100)]
qemu-ppce500: Add support for 64bit CCSR map

QEMU 2.3 changes the address layout of the CCSR map in the PV ppce500 machine
to reside in higher address space.

Unfortunately, this exposed a glitch in u-boot for ppce500: While providing
a function to dynamically evaluate the CCSR region's position in physical
address space, we never used it. Plus we forgot to support 64bit physical
addresses.

This patch fixes that mishap, making u-boot work fine with latest QEMU again.

Signed-off-by: Alexander Graf <agraf@suse.de>
Reviewed-by: Scott Wood <scottwood@freescale.com>
Reviewed-by: York Sun <yorksun@freescale.com>
9 years agoMPC8541/MPC8555: Enable SS_EN in DDR_SDRAM_CLK_CNLT register
Curt Brune [Fri, 13 Feb 2015 18:57:11 +0000 (10:57 -0800)]
MPC8541/MPC8555: Enable SS_EN in DDR_SDRAM_CLK_CNLT register

According to the MPC8555/MPC8541 reference manual the SS_EN (source
synchronous enable) bit in the DDR_SDRAM_CLK_CNLT register must be set
during initialization.

>From section 9.4.1.8 of that manual:

   Source synchronous enable. This bit field must be set during
   initialization. See Section 9.6.1, "DDR SDRAM Initialization
   Sequence," details.

   0 - Reserved
   1 - The address and command are sent to the DDR SDRAMs source
       synchronously.

In addition, Freescale application note AN2805 is also very clear that
this bit must be set.

This patch reverts a change introduced by commit
457caecdbca3df21a93abff19eab12dbc61b7897.

Testing Done:

Compiled targets CONFIG_TARGET_MPC8555CDS and CONFIG_TARGET_MPC8541CDS
and inspected the generated assembly code to verify the SS_EN bit was being
set.  There is one extra instruction emitted:

  fff9b774: 65 29 80 00  oris    r9,r9,32768

Compiled the CONFIG_TARGET_MPC8548CDS target and verified that no
additional instructions were emitted related to this patch.

Booted an image on a MPC8541 based board successfully.

Signed-off-by: Curt Brune <curt@cumulusnetworks.com>
Reviewed-by: York Sun <yorksun@freescale.com>
9 years agoMerge branch 'master' of git://git.denx.de/u-boot-video
Tom Rini [Mon, 20 Apr 2015 13:13:52 +0000 (09:13 -0400)]
Merge branch 'master' of git://git.denx.de/u-boot-video

9 years agovideo, lg4573: add support for the lg4573 display
Heiko Schocher [Sun, 12 Apr 2015 08:20:19 +0000 (10:20 +0200)]
video, lg4573: add support for the lg4573 display

Signed-off-by: Heiko Schocher <hs@denx.de>
9 years agovideo, ipu: make ldb clock frequency overwritable through board code
Heiko Schocher [Mon, 20 Apr 2015 05:53:48 +0000 (07:53 +0200)]
video, ipu: make ldb clock frequency overwritable through board code

the ldb clock can be setup in board code (for example set through PLL5).
Update the ldb_clock rate also through board code.

This should be removed, if a clock framework is availiable.

Signed-off-by: Heiko Schocher <hs@denx.de>
Tested-by: Eric Nelson <eric.nelson@boundarydevices.com>
9 years agovideo, ipu: make ldb_clock configurable
Heiko Schocher [Mon, 20 Apr 2015 05:52:21 +0000 (07:52 +0200)]
video, ipu: make ldb_clock configurable

make the ldb_clock configurable through the new define
CONFIG_SYS_LDB_CLOCK. This is needed as the ldb clock is not
always 650000000, for example on the aristainetos2 board,
where the ldb clock derives from PLL5 clock.

Signed-off-by: Heiko Schocher <hs@denx.de>
Tested-by: Eric Nelson <eric.nelson@boundarydevices.com>
9 years agosandbox: add config_distro_defaults and config_distro_bootcmd
Sjoerd Simons [Mon, 13 Apr 2015 20:54:27 +0000 (22:54 +0200)]
sandbox: add config_distro_defaults and config_distro_bootcmd

Make the sandbox setup more generic/examplary by including
config_distro_defaults.h and config_distro_bootcmd.h.

Among other things this makes it easy to test whether images will boot
though with the standard distro bootcmds by running e.g:
  u-boot -c 'host bind 0 myimage.img ; boot'

By default there are 2 target host devices to emulate device with
multiple storage devices (e.g. internal ("host 0") and external
("host 1") and verify that the prioritization and fallbacks do work
correctly.

Signed-off-by: Sjoerd Simons <sjoerd.simons@collabora.co.uk>
Reviewed by: Simon Glass <sjg@chromium.org>
Acked-by: Simon Glass <sjg@chromium.org>
9 years agoconfig: Add default client arch defines for intel architectures
Sjoerd Simons [Mon, 13 Apr 2015 20:54:26 +0000 (22:54 +0200)]
config: Add default client arch defines for intel architectures

Define default PXE client architecture identifiers for IA32 (0x0 aka
Intel x86PC) and Intel x86-64 (0x9 aka EFI x86-64).

This prepares for usage for config_distro_defaults in the sandbox
architecture

Signed-off-by: Sjoerd Simons <sjoerd.simons@collabora.co.uk>
9 years agopxe: Ensure all memory access is to mapped memory
Sjoerd Simons [Mon, 13 Apr 2015 20:54:25 +0000 (22:54 +0200)]
pxe: Ensure all memory access is to mapped memory

Properly map memory through map_sysmem so that pxe can be used from the
sandbox.

Tested in sandbox as well as on jetson-tk1, odroid-xu3, snow as peach-pi
boards

Signed-off-by: Sjoerd Simons <sjoerd.simons@collabora.co.uk>
Acked-by: Simon Glass <sjg@chromium.org>
9 years agoconfig_distro_bootcmd.h: Add shared block definition for the host interface
Sjoerd Simons [Mon, 13 Apr 2015 20:54:24 +0000 (22:54 +0200)]
config_distro_bootcmd.h: Add shared block definition for the host interface

Define the common shared block environment for the host interface in
preperation for the sandbox build to use config_distro_bootcmd.

Signed-off-by: Sjoerd Simons <sjoerd.simons@collabora.co.uk>
Acked-by: Simon Glass <sjg@chromium.org>
Acked-by: Stephen Warren <swarren@nvidia.com>
9 years agosandbox: Implement host dev [device]
Sjoerd Simons [Mon, 13 Apr 2015 20:54:23 +0000 (22:54 +0200)]
sandbox: Implement host dev [device]

A common pattern to check if a certain device exists (e.g. in
config_distro_bootcmd) is to use: <interface> dev [device]

Implement host dev [device] so this pattern can be used for sandbox host
devices.

Signed-off-by: Sjoerd Simons <sjoerd.simons@collabora.co.uk>
Acked-by: Simon Glass <sjg@chromium.org>
9 years agosandbox: Renamed sb command to host
Sjoerd Simons [Mon, 13 Apr 2015 20:54:22 +0000 (22:54 +0200)]
sandbox: Renamed sb command to host

As suggested by Simon Glass, rename the sb command to host but keep the
old sb command as an alias

Signed-off-by: Sjoerd Simons <sjoerd.simons@collabora.co.uk>
Acked-by: Simon Glass <sjg@chromium.org>
9 years agosandbox: Add support for bootz
Sjoerd Simons [Mon, 13 Apr 2015 20:54:21 +0000 (22:54 +0200)]
sandbox: Add support for bootz

Add dummy bootz_setup implementation allowing the u-boot sandbox to
run bootz. This recognizes both ARM and x86 zImages to validate a
valid zImage was loaded.

Signed-off-by: Sjoerd Simons <sjoerd.simons@collabora.co.uk>
Acked-by: Simon Glass <sjg@chromium.org>
9 years agosandbox: Split bootm code out into lib/bootm
Sjoerd Simons [Mon, 13 Apr 2015 20:54:20 +0000 (22:54 +0200)]
sandbox: Split bootm code out into lib/bootm

Follow the convention of other architectures and move the platform
specific linux bootm code into sandbox/lib/bootm.c.

Signed-off-by: Sjoerd Simons <sjoerd.simons@collabora.co.uk>
Acked-by: Simon Glass <sjg@chromium.org>
9 years agosandbox: only do sandboxfs for hostfs interface
Sjoerd Simons [Mon, 13 Apr 2015 20:54:19 +0000 (22:54 +0200)]
sandbox: only do sandboxfs for hostfs interface

Only do sandbox filesystem access when using the hostfs device
interface, rather then falling back to it in all cases. This prevents
confusion situations due to the fallback being taken rather then an
unsupported error being raised.

Signed-off-by: Sjoerd Simons <sjoerd.simons@collabora.co.uk>
Reviewed-by: Simon Glass <sjg@chromium.org>
Acked-by: Simon Glass <sjg@chromium.org>
9 years agoMerge branch 'buildman' of git://git.denx.de/u-boot-x86
Tom Rini [Sat, 18 Apr 2015 23:24:13 +0000 (19:24 -0400)]
Merge branch 'buildman' of git://git.denx.de/u-boot-x86

9 years agopatman: cover letter shows like 00/xx if more than 10 patches
Wu, Josh [Fri, 3 Apr 2015 02:51:17 +0000 (10:51 +0800)]
patman: cover letter shows like 00/xx if more than 10 patches

Make cover letter shows like 0/x, 00/xx and 000/xxx etc.

Signed-off-by: Josh Wu <josh.wu@atmel.com>
Acked-by: Simon Glass <sjg@chromium.org>
9 years agobuildman: Make -V (verbose_build) really be verbose
Tom Rini [Wed, 1 Apr 2015 11:47:41 +0000 (07:47 -0400)]
buildman: Make -V (verbose_build) really be verbose

The help text for -V says we will pass V=1 but all it really did was not
pass in -s.  Change the logic to pass make V=1 with given to buildman -V or
-s to make otherwise.

Cc: Simon Glass <sjg@chromium.org>
Signed-off-by: Tom Rini <trini@konsulko.com>
Acked-by: Simon Glass <sjg@chromium.org>
9 years agobuildman: Keep more outputs with the --keep-outputs flag
Tom Rini [Fri, 20 Mar 2015 14:50:38 +0000 (10:50 -0400)]
buildman: Keep more outputs with the --keep-outputs flag

When told to keep outputs, be much more liberal in what files we keep.
In addition to adding 'MLO', keep anything that matches u-boot-spl.* (so
that we keep the map file as well) and anything we generate about
'u-boot itself.  A large number of bootable formats now match this and
thus it's easier to build many targets and then boot them afterwards
using buildman.

Cc: Simon Glass <sjg@chromium.org>
Signed-off-by: Tom Rini <trini@konsulko.com>
Acked-by: Simon Glass <sjg@chromium.org>
9 years agobuildman: Allow comparison of build configuration
Simon Glass [Fri, 6 Feb 2015 05:06:15 +0000 (22:06 -0700)]
buildman: Allow comparison of build configuration

It is useful to be able to see CONFIG changes made by commits. Add this
feature to buildman using the -K flag so that all CONFIG changes are
reported.

The CONFIG options exist in a number of files. Each is reported
individually as well as a summary that covers all files. The output
shows three parts: green for additions, red for removals and yellow for
changes.

Signed-off-by: Simon Glass <sjg@chromium.org>
9 years agobuildman: Store build config files
Simon Glass [Fri, 6 Feb 2015 05:06:14 +0000 (22:06 -0700)]
buildman: Store build config files

Store all config file output so that we can compare changes if requested.

Signed-off-by: Simon Glass <sjg@chromium.org>
9 years agobuildman: Adjust the 'aborted' heuristic for writing output
Simon Glass [Fri, 6 Feb 2015 05:06:13 +0000 (22:06 -0700)]
buildman: Adjust the 'aborted' heuristic for writing output

At present buildman tries to detect an aborted build and doesn't record a
result in that case. This is to make sure that an abort (e.g. with Ctrl-C)
does not mark the build as done. Without this option, buildman would never
retry the build unless -f/-F are provided. The effect is that aborting the
build creates 'fake errors' on whatever builds buildman happens to be
working on at the time.

Unfortunately the current test is not reliable and this detection can
trigger if a required toolchain tool is missing. In this case the toolchain
problem is never reported.

Adjust the logic to continue processing the build result, mark the build as
done (and failed), but with a return code which indicates that it should be
retried.

The correct fix is to fully and correctly detect an aborted build, quit
buildman immediately and not write any partial build results in this case.
Unfortunately this is currently beyond my powers and is left as an exercise
for the reader (and patches are welcome).

Signed-off-by: Simon Glass <sjg@chromium.org>
9 years agobuildman: Show 'make' command line when -V is used
Simon Glass [Fri, 6 Feb 2015 05:06:12 +0000 (22:06 -0700)]
buildman: Show 'make' command line when -V is used

When a verbose build it selected, show the make command before the output of
that command.

Signed-off-by: Simon Glass <sjg@chromium.org>
9 years agoCreate a .cfg file containing the CONFIG options used to build
Simon Glass [Fri, 6 Feb 2015 05:06:10 +0000 (22:06 -0700)]
Create a .cfg file containing the CONFIG options used to build

At present CONFIG options are split across Kconfig and board config headers
files. Also we have multiple files containing these CONFIG options.

In order to see exactly what is being used for building, create a .cfg
file which holds these options as reported by the C preprocessor.

Signed-off-by: Simon Glass <sjg@chromium.org>
9 years agoahci: mmio_base is a virtual address
Scott Wood [Fri, 17 Apr 2015 14:19:01 +0000 (09:19 -0500)]
ahci: mmio_base is a virtual address

Don't store it in a u32.

Don't dereference the bus address as if it were a virtual address
(fixes 284231e49a2b4 ("ahci: Support splitting of read transactions
into multiple chunks")).

Fixes crash on boot in MPC8641HPCN_36BIT target.

Signed-off-by: Scott Wood <scottwood@freescale.com>
Cc: Vadim Bendebury <vbendeb@chromium.org>
Acked-by: York Sun <yorksun@freescale.com>
9 years agosandbox: exynos: Move CONFIG_SOUND_SANDBOX to Kconfig
Simon Glass [Fri, 6 Mar 2015 20:19:14 +0000 (13:19 -0700)]
sandbox: exynos: Move CONFIG_SOUND_SANDBOX to Kconfig

Move this over to Kconfig and tidy up.

Signed-off-by: Simon Glass <sjg@chromium.org>
9 years agosandbox: exynos: Move CONFIG_SOUND_WM8994 to Kconfig
Simon Glass [Fri, 6 Mar 2015 20:19:13 +0000 (13:19 -0700)]
sandbox: exynos: Move CONFIG_SOUND_WM8994 to Kconfig

Move this over to Kconfig and tidy up.

Signed-off-by: Simon Glass <sjg@chromium.org>
9 years agosandbox: exynos: Move CONFIG_SOUND_MAX98095 to Kconfig
Simon Glass [Fri, 6 Mar 2015 20:19:12 +0000 (13:19 -0700)]
sandbox: exynos: Move CONFIG_SOUND_MAX98095 to Kconfig

Move this over to Kconfig and tidy up.

Signed-off-by: Simon Glass <sjg@chromium.org>
9 years agosandbox: exynos: Move CONFIG_I2S_SAMSUNG to Kconfig
Simon Glass [Fri, 6 Mar 2015 20:19:11 +0000 (13:19 -0700)]
sandbox: exynos: Move CONFIG_I2S_SAMSUNG to Kconfig

Move this over to Kconfig and tidy up.

Signed-off-by: Simon Glass <sjg@chromium.org>
9 years agosandbox: exynos: Move CONFIG_I2S to Kconfig
Simon Glass [Fri, 6 Mar 2015 20:19:10 +0000 (13:19 -0700)]
sandbox: exynos: Move CONFIG_I2S to Kconfig

Move this over to Kconfig and tidy up.

Signed-off-by: Simon Glass <sjg@chromium.org>
9 years agosandbox: exynos: Move CONFIG_CMD_SOUND to Kconfig
Simon Glass [Fri, 6 Mar 2015 20:19:09 +0000 (13:19 -0700)]
sandbox: exynos: Move CONFIG_CMD_SOUND to Kconfig

Move this over to Kconfig and tidy up.

Signed-off-by: Simon Glass <sjg@chromium.org>
9 years agosandbox: exynos: Move CONFIG_SOUND to Kconfig
Simon Glass [Fri, 6 Mar 2015 20:19:08 +0000 (13:19 -0700)]
sandbox: exynos: Move CONFIG_SOUND to Kconfig

Move this over to Kconfig and tidy up.

Signed-off-by: Simon Glass <sjg@chromium.org>
9 years agosandbox: Move CONFIG_TPM_TIS_SANDBOX to Kconfig
Simon Glass [Fri, 6 Mar 2015 20:19:07 +0000 (13:19 -0700)]
sandbox: Move CONFIG_TPM_TIS_SANDBOX to Kconfig

Move this over to Kconfig and tidy up.

Signed-off-by: Simon Glass <sjg@chromium.org>
9 years agosandbox: Move CONFIG_SPI_FLASH_SANDBOX to Kconfig
Simon Glass [Fri, 6 Mar 2015 20:19:06 +0000 (13:19 -0700)]
sandbox: Move CONFIG_SPI_FLASH_SANDBOX to Kconfig

Move this over to Kconfig and tidy up.

Signed-off-by: Simon Glass <sjg@chromium.org>
9 years agosandbox: Move CONFIG_SANDBOX_SPI to Kconfig
Simon Glass [Fri, 6 Mar 2015 20:19:05 +0000 (13:19 -0700)]
sandbox: Move CONFIG_SANDBOX_SPI to Kconfig

Move this over to Kconfig and tidy up.

Signed-off-by: Simon Glass <sjg@chromium.org>
9 years agosandbox: Move CONFIG_SYS_I2C_SANDBOX to Kconfig
Simon Glass [Fri, 6 Mar 2015 20:19:04 +0000 (13:19 -0700)]
sandbox: Move CONFIG_SYS_I2C_SANDBOX to Kconfig

Move this over to Kconfig and tidy up.

Signed-off-by: Simon Glass <sjg@chromium.org>
Acked-by: Heiko Schocher<hs@denx.de>
9 years agosandbox: Move CONFIG_SYS_VSNPRINTF to Kconfig
Simon Glass [Fri, 6 Mar 2015 20:19:02 +0000 (13:19 -0700)]
sandbox: Move CONFIG_SYS_VSNPRINTF to Kconfig

Move this over to Kconfig and tidy up.

Signed-off-by: Simon Glass <sjg@chromium.org>
9 years agosandbox: Move GPIO CONFIGs to Kconfig
Simon Glass [Fri, 6 Mar 2015 20:19:01 +0000 (13:19 -0700)]
sandbox: Move GPIO CONFIGs to Kconfig

Move these over to Kconfig and tidy up.

Signed-off-by: Simon Glass <sjg@chromium.org>
9 years agoKconfig: Move CONFIG_BOOTSTAGE to Kconfig
Simon Glass [Tue, 3 Mar 2015 00:04:37 +0000 (17:04 -0700)]
Kconfig: Move CONFIG_BOOTSTAGE to Kconfig

Move CONFIG_BOOT_STAGE and its associated options to Kconfig. Adjust
existing users and code.

Signed-off-by: Simon Glass <sjg@chromium.org>
9 years agodm: sunxi: Use driver model for Ethernet on Linksprite pcDuino3
Simon Glass [Sun, 5 Apr 2015 22:07:43 +0000 (16:07 -0600)]
dm: sunxi: Use driver model for Ethernet on Linksprite pcDuino3

Switch this board over to use driver model for Ethernet.

Signed-off-by: Simon Glass <sjg@chromium.org>
9 years agodm: sunxi: Support driver model for Ethernet
Simon Glass [Sun, 5 Apr 2015 22:07:42 +0000 (16:07 -0600)]
dm: sunxi: Support driver model for Ethernet

Adjust the Ethernet initialisation code to support driver model.
It is no-longer necessary to call designware_initialize(). The device will
be probed when it is used. The PHY type and GMAC base will come from the
device tree.

Signed-off-by: Simon Glass <sjg@chromium.org>
9 years agodm: net: Adjust designware driver to support driver model
Simon Glass [Sun, 5 Apr 2015 22:07:41 +0000 (16:07 -0600)]
dm: net: Adjust designware driver to support driver model

Add driver model support to the designware driver. This reuses most of the
existing code except for some duplication in the probe() method.

Signed-off-by: Simon Glass <sjg@chromium.org>
Acked-by: Joe Hershberger <joe.hershberger@ni.com>
9 years agodm: net: Tidy up designware driver ready for driver model
Simon Glass [Sun, 5 Apr 2015 22:07:40 +0000 (16:07 -0600)]
dm: net: Tidy up designware driver ready for driver model

Adjust the error handling to use errno.h instead of returning -1. Change
leaf functions to pass in the arguments they require rather than struct
eth_device. Apart from simplifying the code it makes is easier to reuse
these functions for driver model, since mostly they actually only use
struct dw_eth_priv (which we can keep).

Create a stub for each Ethernet operation function. This will allow use to
share code with the driver model versions.

Signed-off-by: Simon Glass <sjg@chromium.org>
Acked-by: Joe Hershberger <joe.hershberger@ni.com>
9 years agodm: net: Adjust PHY interface to work with CONFIG_DM_ETH
Simon Glass [Sun, 5 Apr 2015 22:07:39 +0000 (16:07 -0600)]
dm: net: Adjust PHY interface to work with CONFIG_DM_ETH

When driver model is used for Ethernet a few functions are passed a udevice
instead of an eth_device. Also add a function to find a PHY type given its
name. This will be used to decode the device tree node.

Finally, put a phy_interface field in struct eth_pdata since this is an
important part of the platform data for Ethernet.

Signed-off-by: Simon Glass <sjg@chromium.org>
Acked-by: Joe Hershberger <joe.hershberger@ni.com>
9 years agoAvoid calling print_eths() with driver model
Simon Glass [Sun, 5 Apr 2015 22:07:38 +0000 (16:07 -0600)]
Avoid calling print_eths() with driver model

This function is not supported with driver model.

Signed-off-by: Simon Glass <sjg@chromium.org>
Acked-by: Joe Hershberger <joe.hershberger@ni.com>
9 years agodm: net: Use existing Ethernet init for driver model
Simon Glass [Sun, 5 Apr 2015 22:07:37 +0000 (16:07 -0600)]
dm: net: Use existing Ethernet init for driver model

At present even with driver model is used there is still much manual init
of related devices: PHY, environment and board init. Until these requirements
are dealt with in another way we need to keep them around.

Break out the init portion of the legacy eth_initialize() into a separate
function and call it from both the legacy and driver model eth_initialize()
functions.

Signed-off-by: Simon Glass <sjg@chromium.org>
Acked-by: Joe Hershberger <joe.hershberger@ni.com>
9 years agodts: sunxi: Bring in Ethernet device tree bindings
Simon Glass [Sun, 5 Apr 2015 22:07:35 +0000 (16:07 -0600)]
dts: sunxi: Bring in Ethernet device tree bindings

Since we will use these bindings on sunxi, bring them in from Linux
4.0-rc1.

Signed-off-by: Simon Glass <sjg@chromium.org>
Acked-by: Ian Campbell <ijc@hellion.org.uk>
9 years agoKconfig: Move CONFIG_DESIGNWARE_ETH to Kconfig
Simon Glass [Sun, 5 Apr 2015 22:07:34 +0000 (16:07 -0600)]
Kconfig: Move CONFIG_DESIGNWARE_ETH to Kconfig

Move this to Kconfig and clean up board config files that use it. Also
rename it to CONFIG_ETH_DESIGNWARE to fit with the naming that exists
in drivers/net/Kconfig.

Signed-off-by: Simon Glass <sjg@chromium.org>
Version 1:
Acked-by: Joe Hershberger <joe.hershberger@ni.com>
9 years agosunxi: Replace the pcDuino3 config with FDT version
Simon Glass [Sun, 5 Apr 2015 22:07:33 +0000 (16:07 -0600)]
sunxi: Replace the pcDuino3 config with FDT version

We currently have Linksprite_pcDuino3 and Linksprite_pcDuino3_fdt. Drop the
former in favour of the latter.

Signed-off-by: Simon Glass <sjg@chromium.org>
Acked-by: Hans de Goede <hdegoede@redhat.com>
Acked-by: Ian Campbell <ijc@hellion.org.uk>