oweals/u-boot.git
7 years agobootstage: Change CONFIG_BOOTSTAGE_USER_COUNT to an int
Simon Glass [Mon, 22 May 2017 11:05:24 +0000 (05:05 -0600)]
bootstage: Change CONFIG_BOOTSTAGE_USER_COUNT to an int

There is no good read to make this hex, and integer is more natural for
this type of setting. Update it.

Signed-off-by: Simon Glass <sjg@chromium.org>
7 years agobootstage: Require timer_get_boot_us() to be defined
Simon Glass [Mon, 22 May 2017 11:05:23 +0000 (05:05 -0600)]
bootstage: Require timer_get_boot_us() to be defined

At present we provide a default version of this function for use by
bootstage. However it uses the system timer and therefore likely requires
driver model. This makes it impossible to time driver-model init.

Drop the function and require boards to provide their own. Add a sandbox
version also. There is a default implememtation in lib/time.c for boards
which use CONFIG_SYS_TIMER_COUNTER.

Signed-off-by: Simon Glass <sjg@chromium.org>
7 years agobootstage: Provide a default timer function
Simon Glass [Mon, 22 May 2017 11:05:22 +0000 (05:05 -0600)]
bootstage: Provide a default timer function

If CONFIG_SYS_TIMER_COUNTER is used we can provide a default microsecond
timer implementation.

Signed-off-by: Simon Glass <sjg@chromium.org>
7 years agoARM: k2g: Fix passing main pll info for higher speeds
Lokesh Vutla [Sat, 20 May 2017 00:19:27 +0000 (05:49 +0530)]
ARM: k2g: Fix passing main pll info for higher speeds

Main pll is marked as arm plls for higher speeds. Fix this.

Signed-off-by: Lokesh Vutla <lokeshvutla@ti.com>
7 years agocmd/elf.c: Support passing arguments with bootelf
Tom Rini [Thu, 18 May 2017 21:03:07 +0000 (17:03 -0400)]
cmd/elf.c: Support passing arguments with bootelf

The bootelf command could, but does not, pass additional arguments along
on the command line.  Make do_bootelf consume bootelf/flags/address as
needed and then pass along anything else to the ELF application we've
launched.

Reported-by: Thomas Doerfler <thomas.doerfler@embedded-brains.de>
Signed-off-by: Tom Rini <trini@konsulko.com>
7 years agoreset: sti: add deassert counter in reset channel descriptor
Patrice Chotard [Thu, 18 May 2017 07:58:00 +0000 (09:58 +0200)]
reset: sti: add deassert counter in reset channel descriptor

This deassert counter allow to manage "shared" reset lines
encountered in some specific case. On STiH410 SoC, DWC3,
EHCI and OHCI are all using a respective PHY, but all of
these PHYs shared a "global" reset.

Currently, during command "usb stop", all host controller are
stopped (XHCI, EHCI and OHCI). XHCI is first shutdowned, which
means that PHY global reset is asserted. Then EHCI is shutdowned,
but its PHY reset has already been asserted which make handshake()
call failed in ehci_shutdown().

This counter allows to really assert a reset lines only when the
"last" user is asserting it.

Signed-off-by: Patrice Chotard <patrice.chotard@st.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
7 years agotest: py: hush: Add echo dependency
Michal Simek [Thu, 18 May 2017 07:23:15 +0000 (09:23 +0200)]
test: py: hush: Add echo dependency

Some tests depends on echo command to be present.

Reported-by: Stephen Warren <swarren@nvidia.com>
Signed-off-by: Michal Simek <michal.simek@xilinx.com>
7 years agotest: py: Use global pytestmark for hush tests
Michal Simek [Thu, 18 May 2017 07:23:14 +0000 (09:23 +0200)]
test: py: Use global pytestmark for hush tests

All tests in test_hush_if_test depends on hush parser to be
present. This patch simplify test dependencies by using global
pytestmark.

Reported-by: Stephen Warren <swarren@nvidia.com>
Signed-off-by: Michal Simek <michal.simek@xilinx.com>
Acked-by: Stephen Warren <swarren@nvidia.com>
7 years agoarm: am57xx: cl-som-am57x: adjust default env to the installation system
Uri Mashiach [Wed, 17 May 2017 14:29:28 +0000 (17:29 +0300)]
arm: am57xx: cl-som-am57x: adjust default env to the installation system

The SD card automatic installation system depends on the default
environment of the previous U-Boot.

Add the missing environment variables.

Signed-off-by: Uri Mashiach <uri.mashiach@compulab.co.il>
Reviewed-by: Tom Rini <trini@konsulko.com>
7 years agoARM: am57xx: cl-som-am57x: support for AM5718
Uri Mashiach [Wed, 17 May 2017 14:29:27 +0000 (17:29 +0300)]
ARM: am57xx: cl-som-am57x: support for AM5718

Disable SDRAM controller EMIF2 for single core SOC
Set SDRAM size size to 1GB

Signed-off-by: Uri Mashiach <uri.mashiach@compulab.co.il>
Reviewed-by: Tom Rini <trini@konsulko.com>
7 years agoarm: am57xx: cl-som-am57x: change the shell prompt
Uri Mashiach [Wed, 17 May 2017 14:29:26 +0000 (17:29 +0300)]
arm: am57xx: cl-som-am57x: change the shell prompt

Change the shell prompt to "U-Boot# ".

Signed-off-by: Uri Mashiach <uri.mashiach@compulab.co.il>
Reviewed-by: Tom Rini <trini@konsulko.com>
7 years agoarm: am57xx: cl-som-am57x: support for FS boot
Uri Mashiach [Wed, 17 May 2017 14:29:25 +0000 (17:29 +0300)]
arm: am57xx: cl-som-am57x: support for FS boot

Supported boot devices are raw QSPI and raw SD card.
Add support for a FAT16/32 file system for SD card.

The SOC's boot ROM only supports FAT file system.
Therefore remove the SPL support for the EXT file system.

Signed-off-by: Uri Mashiach <uri.mashiach@compulab.co.il>
Reviewed-by: Tom Rini <trini@konsulko.com>
7 years agocommon: microblaze: Drop arch-specific declarations
Simon Glass [Wed, 17 May 2017 14:23:11 +0000 (08:23 -0600)]
common: microblaze: Drop arch-specific declarations

These are not needed and should not be in common.h. Drop them.

Signed-off-by: Simon Glass <sjg@chromium.org>
7 years agocommon: arm: freescale: layerscape: Move header files out of common.h
Simon Glass [Wed, 17 May 2017 14:23:10 +0000 (08:23 -0600)]
common: arm: freescale: layerscape: Move header files out of common.h

We should not have an arch-specific header file in common.h. Adjust the
board files a little so it is not needed, and drop it.

Signed-off-by: Simon Glass <sjg@chromium.org>
7 years agocommon: arm: davinci: Move header file out of common
Simon Glass [Wed, 17 May 2017 14:23:09 +0000 (08:23 -0600)]
common: arm: davinci: Move header file out of common

We should not have an arch-specific header file in common.h. Instead, use
the asm/hardware.h header to provide the required declarations, and drop
the common.h changes.

Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Tom Rini <trini@konsulko.com>
7 years agocommon: ep93xx: Move arch-specific declarations out of common
Simon Glass [Wed, 17 May 2017 14:23:08 +0000 (08:23 -0600)]
common: ep93xx: Move arch-specific declarations out of common

These declarations should not be in common. Remove those that are not
needed and move the others to an arch-specific location.

Signed-off-by: Simon Glass <sjg@chromium.org>
7 years agocommon: freescale: Move arch-specific imx code to arch-imx
Simon Glass [Wed, 17 May 2017 14:23:07 +0000 (08:23 -0600)]
common: freescale: Move arch-specific imx code to arch-imx

These declarations should not be in common.h. Move them to an
arch-specific header.

Signed-off-by: Simon Glass <sjg@chromium.org>
7 years agocommon: freescale: Move arch-specific declarations
Simon Glass [Wed, 17 May 2017 14:23:06 +0000 (08:23 -0600)]
common: freescale: Move arch-specific declarations

The declarations should not be in common.h. Move them to the arch-specific
headers.

Signed-off-by: Simon Glass <sjg@chromium.org>
[trini: Fixup thinko defined(FSL_LSCH3) -> defined(CONFIG_FSL_LSCH3)]
Signed-off-by: Tom Rini <trini@konsulko.com>
7 years agocommon: powerpc: Move arch-specific headers
Simon Glass [Wed, 17 May 2017 14:23:05 +0000 (08:23 -0600)]
common: powerpc: Move arch-specific headers

Set up a new asm/ppc.h header file to hold this arch-specific stuff. It
should not be in common.h. It probably should be refactored to use
asm/arch instead, but that is a job for the maintainer.

Signed-off-by: Simon Glass <sjg@chromium.org>
7 years agocommon: Move PPC4xx_SYS_INFO() et al to arch-specific header
Simon Glass [Wed, 17 May 2017 14:23:04 +0000 (08:23 -0600)]
common: Move PPC4xx_SYS_INFO() et al to arch-specific header

These definitions should not be in common.h. Move them to an arch-specific
header file.

Signed-off-by: Simon Glass <sjg@chromium.org>
7 years agocommon: Drop determine_sysper() and determine_pci_clock_per()
Simon Glass [Wed, 17 May 2017 14:23:03 +0000 (08:23 -0600)]
common: Drop determine_sysper() and determine_pci_clock_per()

These arch-specific declarations should not be in common.h. Drop them.

Signed-off-by: Simon Glass <sjg@chromium.org>
7 years agocommon: Drop cpu_init_f() declarations
Simon Glass [Wed, 17 May 2017 14:23:02 +0000 (08:23 -0600)]
common: Drop cpu_init_f() declarations

These arch-specific functions are not needed here.

Signed-off-by: Simon Glass <sjg@chromium.org>
7 years agoarm: Remove include files from common.h
Simon Glass [Wed, 17 May 2017 14:23:01 +0000 (08:23 -0600)]
arm: Remove include files from common.h

With a small tweak we can avoid including these files for all boards.

Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Tom Rini <trini@konsulko.com>
7 years agoarm: Include asm/setup.h explictly
Simon Glass [Wed, 17 May 2017 14:23:00 +0000 (08:23 -0600)]
arm: Include asm/setup.h explictly

Include this header where needed so we do not need to rely on common.h.

Signed-off-by: Simon Glass <sjg@chromium.org>
7 years agoarm: Add declarations to avoid needing to include headers
Simon Glass [Wed, 17 May 2017 14:22:59 +0000 (08:22 -0600)]
arm: Add declarations to avoid needing to include headers

At present common.h includes various ARM-specific headers. In preparation
for dropping this, add a few explicit declarations.

Signed-off-by: Simon Glass <sjg@chromium.org>
7 years agoarm: Add explicit include of <asm/mach-types.h>
Simon Glass [Thu, 1 Jun 2017 01:47:48 +0000 (19:47 -0600)]
arm: Add explicit include of <asm/mach-types.h>

Rather than relying on common.h to provide this include, which is going
away at some point, include it explicitly in each file.

Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Tom Rini <trini@konsulko.com>
7 years agonds32: Remove include files from common.h
Simon Glass [Wed, 17 May 2017 14:22:57 +0000 (08:22 -0600)]
nds32: Remove include files from common.h

With a few tweaks we can avoid including these files, which are only
needed by two C files.

Signed-off-by: Simon Glass <sjg@chromium.org>
7 years agox86: Don't include asm/u-boot.h in common
Simon Glass [Wed, 17 May 2017 14:22:56 +0000 (08:22 -0600)]
x86: Don't include asm/u-boot.h in common

With a small fixup to u-boot-x86.h, this is not actually needed anywhere,
so drop it.

Signed-off-by: Simon Glass <sjg@chromium.org>
7 years agosandbox: Don't include asm/u-boot.h in common
Simon Glass [Wed, 17 May 2017 14:22:55 +0000 (08:22 -0600)]
sandbox: Don't include asm/u-boot.h in common

This is not actually needed anywhere, so drop it.

Signed-off-by: Simon Glass <sjg@chromium.org>
7 years agomips: Don't include asm/u-boot.h in common
Simon Glass [Wed, 17 May 2017 14:22:54 +0000 (08:22 -0600)]
mips: Don't include asm/u-boot.h in common

This is not actually needed anywhere, so drop it.

Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Daniel Schwierzeck <daniel.schwierzeck@gmail.com>
7 years agoarc: Don't include asm/u-boot.h in common
Simon Glass [Wed, 17 May 2017 14:22:53 +0000 (08:22 -0600)]
arc: Don't include asm/u-boot.h in common

This is not actually needed anywhere, so drop it.

Signed-off-by: Simon Glass <sjg@chromium.org>
7 years agonds32: Make u-boot-nds32.h a private header
Simon Glass [Wed, 17 May 2017 14:22:52 +0000 (08:22 -0600)]
nds32: Make u-boot-nds32.h a private header

Rather than including this arch-specific header file in common.h, include
it from within nds32's u-boot.h header.

Signed-off-by: Simon Glass <sjg@chromium.org>
7 years agomips: Make u-boot-mips.h a private header
Simon Glass [Wed, 17 May 2017 14:22:51 +0000 (08:22 -0600)]
mips: Make u-boot-mips.h a private header

Rather than including this arch-specific header file in common.h, include
it from within mips's u-boot.h header.

Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Daniel Schwierzeck <daniel.schwierzeck@gmail.com>
7 years agoarm: Make u-boot-arm.h a private header
Simon Glass [Wed, 17 May 2017 14:22:50 +0000 (08:22 -0600)]
arm: Make u-boot-arm.h a private header

Rather than including this arch-specific header file in common.h, include
it from within arm's u-boot.h header.

Also drop the comment about something to be fixed. It has been there
forever and it is not clear what it means.

Signed-off-by: Simon Glass <sjg@chromium.org>
7 years agox86: Make u-boot-x86.h a private header
Simon Glass [Wed, 17 May 2017 14:22:49 +0000 (08:22 -0600)]
x86: Make u-boot-x86.h a private header

Rather than including this arch-specific header file in common.h, include
it from within x86's u-boot.h header.

Also drop the comment about something to be fixed. It is not clear what
needs fixing.

Signed-off-by: Simon Glass <sjg@chromium.org>
7 years agosandbox: Make u-boot-sandbox.h a private header
Simon Glass [Wed, 17 May 2017 14:22:48 +0000 (08:22 -0600)]
sandbox: Make u-boot-sandbox.h a private header

Rather than including this arch-specific header file in common.h, include
it from within sandbox's u-boot.h header.

Also drop the comment about something to be fixed.

Signed-off-by: Simon Glass <sjg@chromium.org>
7 years agoarc: Make u-boot-arc.h a private header
Simon Glass [Wed, 17 May 2017 14:22:47 +0000 (08:22 -0600)]
arc: Make u-boot-arc.h a private header

Rather than including this arch-specific header file in common.h, include
it from within arc's u-boot.h header.

Signed-off-by: Simon Glass <sjg@chromium.org>
7 years agosamsung: Drop more references fo s3c24x0
Simon Glass [Wed, 17 May 2017 14:22:46 +0000 (08:22 -0600)]
samsung: Drop more references fo s3c24x0

This is dead code now. Drop it.

Signed-off-by: Simon Glass <sjg@chromium.org>
7 years agosamsung: nand: Drop s3c2410_nand driver
Simon Glass [Wed, 17 May 2017 14:22:45 +0000 (08:22 -0600)]
samsung: nand: Drop s3c2410_nand driver

This is not used anymore. Drop it.

Signed-off-by: Simon Glass <sjg@chromium.org>
7 years agosamsung: usb: Drop ohci-s3c24xx driver
Simon Glass [Wed, 17 May 2017 14:22:44 +0000 (08:22 -0600)]
samsung: usb: Drop ohci-s3c24xx driver

This is not used anymore. Drop it.

Signed-off-by: Simon Glass <sjg@chromium.org>
7 years agosamsung: mmc: Drop s3c_sdi driver
Simon Glass [Wed, 17 May 2017 14:22:43 +0000 (08:22 -0600)]
samsung: mmc: Drop s3c_sdi driver

This is no-longer used in U-Boot. Drop it.

Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Jaehoon Chung <jh80.chung@samsung.com>
7 years agosamsung: Drop s3c24x0 arch-specific code
Simon Glass [Wed, 17 May 2017 14:22:42 +0000 (08:22 -0600)]
samsung: Drop s3c24x0 arch-specific code

This is not used anymore. Drop it.

Signed-off-by: Simon Glass <sjg@chromium.org>
7 years agogpio: samsung: Drop s3c2440_gpio driver
Simon Glass [Wed, 17 May 2017 14:22:41 +0000 (08:22 -0600)]
gpio: samsung: Drop s3c2440_gpio driver

This is no longer used in U-Boot. Drop it.

Signed-off-by: Simon Glass <sjg@chromium.org>
7 years agoapi: Add a header for api_init()
Simon Glass [Wed, 17 May 2017 14:22:40 +0000 (08:22 -0600)]
api: Add a header for api_init()

Put this in its own header instead of using common.h.

Signed-off-by: Simon Glass <sjg@chromium.org>
7 years agocommon: Move get_OPB_freq() and get_PCI_freq() to PPC header
Simon Glass [Wed, 17 May 2017 14:22:39 +0000 (08:22 -0600)]
common: Move get_OPB_freq() and get_PCI_freq() to PPC header

These should not be in common.h. Move the to an arch-specific header.

Signed-off-by: Simon Glass <sjg@chromium.org>
7 years agocommon: Move pcie_setup_hoses() to PPC header
Simon Glass [Wed, 17 May 2017 14:22:38 +0000 (08:22 -0600)]
common: Move pcie_setup_hoses() to PPC header

Only one board needs this definition. Move it to an arch-specific header.

Signed-off-by: Simon Glass <sjg@chromium.org>
7 years agocommon: Drop pci_master_init()
Simon Glass [Wed, 17 May 2017 14:22:37 +0000 (08:22 -0600)]
common: Drop pci_master_init()

This should not be in common.h. Drop it.

Signed-off-by: Simon Glass <sjg@chromium.org>
7 years agocommon: Move pci_target_init() to PPC header
Simon Glass [Wed, 17 May 2017 14:22:36 +0000 (08:22 -0600)]
common: Move pci_target_init() to PPC header

Only one boards needs this definition. Move it to an arch-specific header.

Signed-off-by: Simon Glass <sjg@chromium.org>
7 years agocommon: Drop pci_pre_init() and is_pci_host()
Simon Glass [Wed, 17 May 2017 14:22:35 +0000 (08:22 -0600)]
common: Drop pci_pre_init() and is_pci_host()

These should not be in common.h. They are used in some legacy PowerPC
code. Just drop them.

Signed-off-by: Simon Glass <sjg@chromium.org>
7 years agocommon: Drop inclusion of pci.h
Simon Glass [Wed, 17 May 2017 14:22:34 +0000 (08:22 -0600)]
common: Drop inclusion of pci.h

This should not be in common.h - remove it and update the only file that
needs it.

Signed-off-by: Simon Glass <sjg@chromium.org>
7 years agot81xx: Migrate TI81XX/TI816X/TI814X symbols to Kconfig
Tom Rini [Tue, 16 May 2017 18:46:40 +0000 (14:46 -0400)]
t81xx: Migrate TI81XX/TI816X/TI814X symbols to Kconfig

The symbol CONFIG_TI81XX is used for the parts that are common to the
TI816x and TI814x SoCs and are not part of CONFIG_ARCH_OMAP2PLUS nor
CONFIG_AM33XX.  It however has so few uses that we can just modify the
code to check for both and drop the symbol. The symbols CONFIG_TI816X
and CONFIG_TI814X are for the repective SoCs.

Signed-off-by: Tom Rini <trini@konsulko.com>
7 years agoti816x: Modernize the defconfig
Tom Rini [Tue, 16 May 2017 18:46:39 +0000 (14:46 -0400)]
ti816x: Modernize the defconfig

- Switch to using <configs/ti_armv7_omap.h> and family.  This lets us
  drop lots of custom defines.
- Ensure that our default environment uses DEFAULT_LINUX_BOOT_ENV so
  that Linux will boot correctly.
- Enable CONFIG_DISTRO_DEFAULTS
- Switch to using CONFIG_OF_CONTROL
- Various other cleanups to match other SoCs in the family line.

Signed-off-by: Tom Rini <trini@konsulko.com>
7 years agoti816x: Import dts files from Linux Kernel v4.11
Tom Rini [Tue, 16 May 2017 18:46:38 +0000 (14:46 -0400)]
ti816x: Import dts files from Linux Kernel v4.11

This brings in the required dts/dtsi files for the TI8168-EVM from the
Linux Kernel v4.11 release.

Signed-off-by: Tom Rini <trini@konsulko.com>
7 years agoti816x: Enable NAND
Tom Rini [Tue, 16 May 2017 18:46:37 +0000 (14:46 -0400)]
ti816x: Enable NAND

The TI8168-EVM comes with NAND on board.  Enable it and move environment
over there.

Signed-off-by: Tom Rini <trini@konsulko.com>
7 years agoti816x_evm: Disable CONFIG_USE_PRIVATE_LIBGCC
Tom Rini [Tue, 16 May 2017 18:46:36 +0000 (14:46 -0400)]
ti816x_evm: Disable CONFIG_USE_PRIVATE_LIBGCC

On this platform, we can trace a general failure to boot to enabling /
disabling this option.  When this is enabled, we go off into the
weeds during SPL and are unable to talk with the SD card and
mmc_initialize() fails.

Signed-off-by: Tom Rini <trini@konsulko.com>
7 years agoti816x: Rework DDR initialization sequence
Tom Rini [Tue, 16 May 2017 18:46:35 +0000 (14:46 -0400)]
ti816x: Rework DDR initialization sequence

The ti816x/am389x SoC is the first generation in what U-Boot calls the
"am33xx" family.  In the first generation of this family the DDR
initialization sequence is quite different from all of the subsequent
generations.  Whereas with ti814x (second generation) we can easily work
the minor differenced between that and am33xx (third generation), our
attempts to do this for ti816x weren't sufficient.  Rather than add a
large amount of #ifdef logic to make this different sequence work we add
a new file, ti816x_emif4.c to handle the various required undocumented
register writes and sequence and leverage what we can from
arch/arm/mach-omap2/am33xx/ddr.c still.  As DDR2 has similar problems
today but I am unable to test it, we drop the DDR2 defines from the code
rather than imply that it works by leaving it.  We also remove a bunch
of other untested code about changing the speed the DDR runs at.

Signed-off-by: Tom Rini <trini@konsulko.com>
7 years agoarmv7: Mark the default lowlevel_init function as weak
Tom Rini [Tue, 16 May 2017 18:46:34 +0000 (14:46 -0400)]
armv7: Mark the default lowlevel_init function as weak

Rather than have a long and if check in the Makefile, mark the default
lowlevel_init function as weak (as we do on armv8) so that SoCs can
override it if needed, and it will still be discarded if unused.
Provide a weak s_init as well to allow for this to link and be
discarded.

Signed-off-by: Tom Rini <trini@konsulko.com>
7 years agoconfigs: davinci: omapl138_lcdk: add random eth address support
Anna, Suman [Tue, 16 May 2017 17:50:18 +0000 (12:50 -0500)]
configs: davinci: omapl138_lcdk: add random eth address support

Any TFTP or DHCP boot on the Davinci OMAP-L138 LCDK board requires
that the 'ethaddr' variable be defined. There are no e-fuses to store
the ethernet mac address for this platform, and neither is a MAC
address reserved in any format. So enable random MAC address support
so that networking boot can be supported.

Signed-off-by: Suman Anna <s-anna@ti.com>
Reviewed-by: Tom Rini <trini@konsulko.com>
Reviewed-by: Lokesh Vutla <lokeshvutla@ti.com>
7 years agoEnable PXE boot on meson-gxbb.
Vagrant Cascadian [Fri, 5 May 2017 21:11:26 +0000 (14:11 -0700)]
Enable PXE boot on meson-gxbb.

Enable distro_bootcmd PXE functions on meson-gxbb systems.

Signed-off-by: Vagrant Cascadian <vagrant@debian.org>
Reviewed-by: Simon Glass <sjg@chromium.org>
7 years agoMerge git://git.denx.de/u-boot-x86
Tom Rini [Mon, 5 Jun 2017 02:29:33 +0000 (22:29 -0400)]
Merge git://git.denx.de/u-boot-x86

7 years agox86: fsp: Remove the call to set up internal uart in fsp_init()
Bin Meng [Thu, 1 Jun 2017 10:41:14 +0000 (03:41 -0700)]
x86: fsp: Remove the call to set up internal uart in fsp_init()

First of all, it's inappropriate to call setup_internal_uart() in a
generic API fsp_init(), as CONFIG_INTERNAL_UART is an option that
is only available on BayTrail platform. Secondly even for BayTrail,
there is no need to call setup_internal_uart() at all, as Intel FSP
will do this for us.

Signed-off-by: Bin Meng <bmeng.cn@gmail.com>
Reviewed-by: Stefan Roese <sr@denx.de>
Reviewed-by: Simon Glass <sjg@chromium.org>
7 years agox86: baytrail: Fix boot hang with a debug build
Bin Meng [Thu, 1 Jun 2017 10:41:13 +0000 (03:41 -0700)]
x86: baytrail: Fix boot hang with a debug build

It was observed that when -DDEBUG is used to generate a debug build,
U-Boot does not boot on MinnowMax board. A workaround is to disable
CONFIG_DEBUG_UART. The real issue is that in order to have the debug
uart to work, BayTrail SoC needs to be configured so that its internal
uart is available to be used as the debug uart.

Signed-off-by: Bin Meng <bmeng.cn@gmail.com>
Reviewed-by: Stefan Roese <sr@denx.de>
Reviewed-by: Simon Glass <sjg@chromium.org>
7 years agox86: baytrail: Change lpe/lpss-sio/scc FSP properties to integer
Bin Meng [Wed, 31 May 2017 08:04:15 +0000 (01:04 -0700)]
x86: baytrail: Change lpe/lpss-sio/scc FSP properties to integer

At present lpe/lpss-sio/scc FSP properties are all boolean, but in
fact for "enable-lpe" it has 3 possible options. This adds macros
for these options and change the property from a boolean type to
an integer type, and change their names to explicitly indicate what
the property is really for.

Signed-off-by: Bin Meng <bmeng.cn@gmail.com>
Reviewed-by: Stefan Roese <sr@denx.de>
Reviewed-by: Simon Glass <sjg@chromium.org>
7 years agox86: baytrail: Use macros instead of magic numbers for FSP settings
Bin Meng [Wed, 31 May 2017 08:04:14 +0000 (01:04 -0700)]
x86: baytrail: Use macros instead of magic numbers for FSP settings

Introduce various meaningful macros for FSP settings and switch over
to use them instead of magic numbers.

Signed-off-by: Bin Meng <bmeng.cn@gmail.com>
Reviewed-by: Stefan Roese <sr@denx.de>
Reviewed-by: Simon Glass <sjg@chromium.org>
7 years agox86: baytrail: Remove "serial-debug-port-*" settings
Bin Meng [Wed, 31 May 2017 08:04:13 +0000 (01:04 -0700)]
x86: baytrail: Remove "serial-debug-port-*" settings

"serial-debug-port-address" and "serial-debug-port-type" settings
are actually reserved in the FSP UPD data structure. Remove them.

Signed-off-by: Bin Meng <bmeng.cn@gmail.com>
Reviewed-by: Stefan Roese <sr@denx.de>
Reviewed-by: Simon Glass <sjg@chromium.org>
7 years agox86: baytrail: Change "fsp, mrc-init-tseg-size" default value to 1
Bin Meng [Wed, 31 May 2017 08:04:12 +0000 (01:04 -0700)]
x86: baytrail: Change "fsp, mrc-init-tseg-size" default value to 1

The default value of "fsp,mrc-init-tseg-size" should be 1 (1MB) per
FSP default settings. 0 is not valid.

Signed-off-by: Bin Meng <bmeng.cn@gmail.com>
Reviewed-by: Stefan Roese <sr@denx.de>
Reviewed-by: Simon Glass <sjg@chromium.org>
7 years agoMerge git://git.denx.de/u-boot-fdt
Tom Rini [Sun, 4 Jun 2017 17:13:29 +0000 (13:13 -0400)]
Merge git://git.denx.de/u-boot-fdt

7 years agoMerge git://git.denx.de/u-boot-net
Tom Rini [Sat, 3 Jun 2017 22:05:28 +0000 (18:05 -0400)]
Merge git://git.denx.de/u-boot-net

7 years agoMerge git://git.denx.de/u-boot-fsl-qoriq
Tom Rini [Sat, 3 Jun 2017 22:05:04 +0000 (18:05 -0400)]
Merge git://git.denx.de/u-boot-fsl-qoriq

7 years agoMerge git://git.denx.de/u-boot-sunxi
Tom Rini [Sat, 3 Jun 2017 22:04:54 +0000 (18:04 -0400)]
Merge git://git.denx.de/u-boot-sunxi

7 years agoKconfig: Migrate FS_FAT / FAT_WRITE
Tom Rini [Fri, 2 Jun 2017 15:03:50 +0000 (11:03 -0400)]
Kconfig: Migrate FS_FAT / FAT_WRITE

Now that these symbols are in Kconfig, migrate all users.  Use imply on
a number of platforms that default to having this enabled.  As part of
this we must migrate some straglers for CMD_FAT and DOS_PARTITION.

Signed-off-by: Tom Rini <trini@konsulko.com>
7 years agoboard: ti: enable support for writing to fat partition
Sekhar Nori [Fri, 2 Jun 2017 12:24:04 +0000 (17:54 +0530)]
board: ti: enable support for writing to fat partition

Enable support for writing to FAT partitions on
TI's boards.

Signed-off-by: Sekhar Nori <nsekhar@ti.com>
7 years agoconfig_fallbacks: add additional fallbacks for fat filesystem
Sekhar Nori [Fri, 2 Jun 2017 12:24:02 +0000 (17:54 +0530)]
config_fallbacks: add additional fallbacks for fat filesystem

Add fallbacks needed to keep all boards building
while they are migrated to use Kconfig symbols
instead of defines in <board>_config.h files for
FAT filesystem.

These should eventually go away once Kconfig select
or imply statements are put in place and duplicated
defines in <board>_config.h removed.

Signed-off-by: Sekhar Nori <nsekhar@ti.com>
[trini: Update logic since CMD_FAT / CONFIG_SPL_FAT_SUPPORT are
        selecting FS_FAT]
Signed-off-by: Tom Rini <trini@konsulko.com>
7 years agoconfigs: k2g_evm: make sure config fallbacks take effect
Sekhar Nori [Fri, 2 Jun 2017 12:24:01 +0000 (17:54 +0530)]
configs: k2g_evm: make sure config fallbacks take effect

Since config fallbacks contained in include/config_fallbacks.h
come into k2g_evm.h file through ti_armv7_keystone2.h, it should
be the last file included.

Without this, #define of FAT_WRITE when environment is in FAT
does not happen as the environment location is decided later
in the file.

Similar issues can come with other config fallbacks implemented.

Signed-off-by: Sekhar Nori <nsekhar@ti.com>
7 years agoconfigs: k2*_evm: let each board decide env location
Sekhar Nori [Fri, 2 Jun 2017 12:24:00 +0000 (17:54 +0530)]
configs: k2*_evm: let each board decide env location

Not all TI Keystone2 EVMs want environment in NAND flash.
K2G EVM which has an MMC/SD slot, keep environment in a
FAT partition on SD card.

Since ti_armv7_keystone2.h defines environment is in NAND,
boards which do not follow that have to #undef'ine that
configuration. This leads to ugly ordering issues around
where exactly the include of ti_armv7_keystone2.h can come
in within the k2*_evm.h files.

Move environment location to config file of each board.
This should make it easy to change it for any one board
without affecting all other boards.

Signed-off-by: Sekhar Nori <nsekhar@ti.com>
7 years agofs: fat: add kbuild configuration support
Sekhar Nori [Fri, 2 Jun 2017 12:23:59 +0000 (17:53 +0530)]
fs: fat: add kbuild configuration support

Add Kconfig symbols for various configurations
supported by FAT filesystem support code.

CONFIG_SUPPORT_VFAT has been left out since its
force enabled in include/fat.h and probably
should get removed at some point.

Signed-off-by: Sekhar Nori <nsekhar@ti.com>
[trini: add select FS_FAT for CMD_FAT and SPL_FAT_SUPPORT]
Signed-off-by: Tom Rini <trini@konsulko.com>
7 years agonet: pch_gbe: Add cache maintenance
Paul Burton [Sun, 30 Apr 2017 19:57:08 +0000 (21:57 +0200)]
net: pch_gbe: Add cache maintenance

On MIPS systems DMA isn't coherent with the CPU caches unless an IOCU is
present. When there is no IOCU we need to writeback or invalidate the
data caches at appropriate points. Perform this cache maintenance in
the pch_gbe driver which is used on the MIPS Boston development board.

Signed-off-by: Paul Burton <paul.burton@imgtec.com>
Reviewed-by: Bin Meng <bmeng.cn@gmail.com>
Tested-by: Bin Meng <bmeng.cn@gmail.com>
Signed-off-by: Daniel Schwierzeck <daniel.schwierzeck@gmail.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
Acked-by: Joe Hershberger <joe.hershberger@ni.com>
7 years agonet: pch_gbe: CPU accessible addresses are virtual
Paul Burton [Sun, 30 Apr 2017 19:57:07 +0000 (21:57 +0200)]
net: pch_gbe: CPU accessible addresses are virtual

Use the virt_to_bus & bus_to_virt functions rather than phys_to_bus &
bus_to_phys, since the addresses accessed by the CPU will be virtual
rather than physical. On MIPS physical & virtual addresses differ as we
use virtual addresses in kseg0, and attempting to use physical addresses
directly caused problems as they're in the user segment which would be
mapped via the uninitialised TLB.

Signed-off-by: Paul Burton <paul.burton@imgtec.com>
Signed-off-by: Daniel Schwierzeck <daniel.schwierzeck@gmail.com>
Reviewed-by: Bin Meng <bmeng.cn@gmail.com>
Tested-by: Bin Meng <bmeng.cn@gmail.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
Acked-by: Joe Hershberger <joe.hershberger@ni.com>
7 years agonet: pch_gbe: Fix rx descriptor buffer addresses
Paul Burton [Sun, 30 Apr 2017 19:57:06 +0000 (21:57 +0200)]
net: pch_gbe: Fix rx descriptor buffer addresses

The loop to set up buffer addresses in rx descriptors always operated on
descriptor 0, rather than on each descriptor sequentially. Fix this in
order to setup correct buffer addresses for each descriptor.

Signed-off-by: Paul Burton <paul.burton@imgtec.com>
Signed-off-by: Daniel Schwierzeck <daniel.schwierzeck@gmail.com>
Reviewed-by: Bin Meng <bmeng.cn@gmail.com>
Tested-by: Bin Meng <bmeng.cn@gmail.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
Acked-by: Joe Hershberger <joe.hershberger@ni.com>
7 years agonet: pch_gbe: Reset during probe
Paul Burton [Sun, 30 Apr 2017 19:57:05 +0000 (21:57 +0200)]
net: pch_gbe: Reset during probe

Using the EG20T gigabit ethernet controller on the MIPS Boston board, we
find that we have to reset the controller in order for the RGMII link to
the PHY to become functional. Without doing so we constantly time out in
pch_gbe_mdio_ready.

Signed-off-by: Paul Burton <paul.burton@imgtec.com>
Reviewed-by: Bin Meng <bmeng.cn@gmail.com>
Tested-by: Bin Meng <bmeng.cn@gmail.com>
Signed-off-by: Daniel Schwierzeck <daniel.schwierzeck@gmail.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
Acked-by: Joe Hershberger <joe.hershberger@ni.com>
7 years agonet: zynq_gem: Dont flush dummy descriptors
Siva Durga Prasad Paladugu [Tue, 30 May 2017 12:28:40 +0000 (14:28 +0200)]
net: zynq_gem: Dont flush dummy descriptors

Dont flush dummy descriptors as they are already
allocated from a region with dcache off. Tested
this on Zynq(zc702) and ZynqMP(zcu102) boards.

Signed-off-by: Siva Durga Prasad Paladugu <sivadur@xilinx.com>
Signed-off-by: Michal Simek <michal.simek@xilinx.com>
Acked-by: Joe Hershberger <joe.hershberger@ni.com>
7 years agonet: zynq_gem: Use wait_for_bit with non breakable
Siva Durga Prasad Paladugu [Tue, 30 May 2017 12:28:39 +0000 (14:28 +0200)]
net: zynq_gem: Use wait_for_bit with non breakable

Use wait_for_bit to be non breakable as using it with
breakable causes issue of un interruptible auto negotiation.
This is due to the ctrlc pressed will taken for wait_for_bit()
abort during phy_read() and hence not coming out of
auto negotiation.

Signed-off-by: Siva Durga Prasad Paladugu <sivadur@xilinx.com>
Signed-off-by: Michal Simek <michal.simek@xilinx.com>
Acked-by: Joe Hershberger <joe.hershberger@ni.com>
7 years agonet: phy: marvell 88e151x: Fix handling of RGMII interface types
Phil Edworthy [Wed, 24 May 2017 13:43:06 +0000 (14:43 +0100)]
net: phy: marvell 88e151x: Fix handling of RGMII interface types

The 88E1518 code is programming the wrong registers for rgmii-id,
rgmii-txid and rgmii-rxid interfaces.

Since the PHY defaults to rgmii-id, it would appear that the code
was previously only used with sgmii and rgmii-id interfaces.

Tested on 88E1512 PHY in rgmii-id mode which is from the same family
as 88E1518.

Signed-off-by: Phil Edworthy <phil.edworthy@renesas.com>
Acked-by: Joe Hershberger <joe.hershberger@ni.com>
7 years agonet: core: avoid possible NULL pointer dereference
xypron.glpk@gmx.de [Tue, 16 May 2017 03:07:01 +0000 (05:07 +0200)]
net: core: avoid possible NULL pointer dereference

Checking if dev is NULL after dereferencing it does not make sense.

Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Acked-by: Joe Hershberger <joe.hershberger@ni.com>
7 years agodrivers: net: cpsw: abort init() on aneg timeout
Sekhar Nori [Mon, 8 May 2017 15:19:56 +0000 (20:49 +0530)]
drivers: net: cpsw: abort init() on aneg timeout

Abort CPSW driver init when auto-negotiation of link
times out. Currently, the code ignores return status
of phy_startup(), and goes ahead with network operation
(like DHCP) even though the link may be down.

Instead, abort init process if link is down or if there
is another error, so phy_startup() can easily be retried
again. This also helps quick fallback to next network interface
(like USB RNDIS) without inordinate delay.

Tested on AM571x IDK and AM335x BeagleBone black.

Signed-off-by: Sekhar Nori <nsekhar@ti.com>
Reviewed-by: Tom Rini <trini@konsulko.com>
Acked-by: Joe Hershberger <joe.hershberger@ni.com>
7 years agonet: macb: Fix GMAC not work when enable DM_ETH
Wenyou Yang [Thu, 20 Apr 2017 03:13:13 +0000 (11:13 +0800)]
net: macb: Fix GMAC not work when enable DM_ETH

Always search the PHY to determine the macb->phy_addr before using
the PHY to fix "No PHY present" error.

Fix the wrong test of the GMAC's phy interface mode, it should be
PHY_INTERFACE_MODE_RGMII.

Signed-off-by: Wenyou Yang <wenyou.yang@atmel.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
Acked-by: Joe Hershberger <joe.hershberger@ni.com>
7 years agonet: mvpp2.c: Enable 10G support for port 0 (SFI)
Stefan Chulski [Thu, 6 Apr 2017 13:39:08 +0000 (15:39 +0200)]
net: mvpp2.c: Enable 10G support for port 0 (SFI)

This patch fixes some remaining issues in the mvpp2 driver for the 10GB
support on port 0. These changes are:

- Incorrect PCS configuration
- Skip PHY configuration when no PHY is connected
- Skip GMAC configurations if 10G SFI mode set

Signed-off-by: Stefan Chulski <stefanc@marvell.com>
Signed-off-by: Stefan Roese <sr@denx.de>
Cc: Kostya Porotchkin <kostap@marvell.com>
Cc: Nadav Haklai <nadavh@marvell.com>
Cc: Joe Hershberger <joe.hershberger@ni.com>
Acked-by: Joe Hershberger <joe.hershberger@ni.com>
7 years agoarmv8/ls1046a: RGMII PHY requires internal delay on Tx
Madalin Bucur [Mon, 3 Apr 2017 14:43:56 +0000 (17:43 +0300)]
armv8/ls1046a: RGMII PHY requires internal delay on Tx

Signed-off-by: Madalin Bucur <madalin.bucur@nxp.com>
Acked-by: Joe Hershberger <joe.hershberger@ni.com>
7 years agoarmv8/ls1043a: RGMII PHY requires internal delay on Tx
Madalin Bucur [Mon, 3 Apr 2017 14:43:55 +0000 (17:43 +0300)]
armv8/ls1043a: RGMII PHY requires internal delay on Tx

Signed-off-by: Madalin Bucur <madalin.bucur@nxp.com>
Acked-by: Joe Hershberger <joe.hershberger@ni.com>
7 years agonet: zynq_gem: Do not return -ENOSYS on success
Olliver Schinagl [Mon, 3 Apr 2017 14:18:53 +0000 (16:18 +0200)]
net: zynq_gem: Do not return -ENOSYS on success

The .read_rom_hwaddr net_ops hook does not check the return value, which
is why it was never caught that we are currently returning 0 if the
read_rom_hwaddr function return -ENOSYS and -ENOSYS otherwise.

In this case we can simplify this by just returning the result of the
function.

Signed-off-by: Olliver Schinagl <oliver@schinagl.nl>
Acked-by: Joe Hershberger <joe.hershberger@ni.com>
Acked-by: Michal Simek <michal.simek@xilinx.com>
7 years agonet: designware: Add phy supply support
Jacob Chen [Mon, 27 Mar 2017 08:54:17 +0000 (16:54 +0800)]
net: designware: Add phy supply support

Some board need a regulator for gmac phy, so add this code to handle it.
Signed-off-by: Jacob Chen <jacob2.chen@rock-chips.com>
Acked-by: Joe Hershberger <joe.hershberger@ni.com>
7 years agonet: Kconfig:make PHY_GIGE and individual Micrel PHYs selectable
Philipp Tomsich [Sun, 26 Mar 2017 16:50:23 +0000 (18:50 +0200)]
net: Kconfig:make PHY_GIGE and individual Micrel PHYs selectable

This change migrate the following configuration options for Kconfig:
 * PHY_GIGE, indicates that a controller (with an appropriate PHY) is
   Gigabit capable and enables extra support in the miiutil for
   parsing the status of Gigabit PHYs
 * adds configuration options for Micrel KSZ9021 and KSZ9031 GbE PHYs,
   which previously had to enabled through a board-specific config file

Signed-off-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
Acked-by: Maxime Ripard <maxime.ripard@free-electrons.com>
Acked-by: Joe Hershberger <joe.hershberger@ni.com>
7 years agofdt: Drop fdt_select.py
Simon Glass [Sat, 27 May 2017 13:38:30 +0000 (07:38 -0600)]
fdt: Drop fdt_select.py

This file was used to select between the normal and fallback libfdt
implementations. Now that we only have one, it is not needed.

Drop it and fix up all users.

Signed-off-by: Simon Glass <sjg@chromium.org>
7 years agobinman: Rename fdt variable to dtb
Simon Glass [Sat, 27 May 2017 13:38:29 +0000 (07:38 -0600)]
binman: Rename fdt variable to dtb

Since fdt is the name of a module, use a different name for variables to
avoid a conflict.

Signed-off-by: Simon Glass <sjg@chromium.org>
7 years agofdt: Merge fdt_normal with its base class
Simon Glass [Sat, 27 May 2017 13:38:28 +0000 (07:38 -0600)]
fdt: Merge fdt_normal with its base class

Since we only have one Fdt implementation now we don't need to have a base
class. Merge the implementation and the base class together.

Signed-off-by: Simon Glass <sjg@chromium.org>
7 years agobinman: Drop a special case related to fdt_fallback
Simon Glass [Sat, 27 May 2017 13:38:27 +0000 (07:38 -0600)]
binman: Drop a special case related to fdt_fallback

Previously we were sometimes forced to collate x86 microcode due to not
having access to the offset of each individual piece. Now that we never
use fdt_fallback, we don't have this problem. Drop this special case from
the code.

Signed-off-by: Simon Glass <sjg@chromium.org>
7 years agofdt: Drop fdt_fallback library
Simon Glass [Sat, 27 May 2017 13:38:26 +0000 (07:38 -0600)]
fdt: Drop fdt_fallback library

Drop this now-unused library and associated tests.

Signed-off-by: Simon Glass <sjg@chromium.org>
7 years agofdt: Drop use of the legacy libfdt python module
Simon Glass [Sat, 27 May 2017 13:38:25 +0000 (07:38 -0600)]
fdt: Drop use of the legacy libfdt python module

Now that this is no-longer available, stop looking for it. The new module
will be used if available.

Signed-off-by: Simon Glass <sjg@chromium.org>
7 years agofdt: Stop building the old python libfdt module
Simon Glass [Sat, 27 May 2017 13:38:24 +0000 (07:38 -0600)]
fdt: Stop building the old python libfdt module

This is no-longer needed, so stop building it.

Signed-off-by: Simon Glass <sjg@chromium.org>
7 years agofdt: Makefile: Build python libfdt library if needed
Simon Glass [Sat, 27 May 2017 13:38:23 +0000 (07:38 -0600)]
fdt: Makefile: Build python libfdt library if needed

This is needed by binman and dtoc, so if those are being used, check that
the library is present and complain if not. Make sure that any error
appears on stderr so that buildman notices it.

This means that the fallback library (which uses fdtget) will not be used
anymore and swig will need to be installed to use binman / dtoc.

This affects any board which uses binman (currently sunxi and x86) or dtoc
(anything that uses CONFIG_SPL_OF_PLATDATA, currently some rockchip
boards).

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