oweals/u-boot.git
6 years agophy: add support for STM32 usb phy controller
Patrice Chotard [Fri, 27 Apr 2018 09:01:55 +0000 (11:01 +0200)]
phy: add support for STM32 usb phy controller

This patch adds phy tranceiver driver for STM32 USB PHY
Controller (usbphyc) that provides dual port High-Speed
phy for OTG (single port) and EHCI/OHCI host controller
(two ports).
One port of the phy is shared between the two USB controllers
through a UTMI+ switch.

Signed-off-by: Christophe Kerello <christophe.kerello@st.com>
Signed-off-by: Amelie Delaunay <amelie.delaunay@st.com>
Signed-off-by: Patrice Chotard <patrice.chotard@st.com>
6 years agogadget: f_thor: update to support more than 4GB file as thor 5.0
Seung-Woo Kim [Thu, 10 May 2018 01:52:15 +0000 (10:52 +0900)]
gadget: f_thor: update to support more than 4GB file as thor 5.0

During file download, it only uses 32bit variable for file size and
it limits maximum file size less than 4GB. Update to support more
than 4GB file with using two 32bit variables for file size as thor
protocol 5.0.

Signed-off-by: Seung-Woo Kim <sw0312.kim@samsung.com>
6 years agogadget: f_thor: fix filename overflow
Seung-Woo Kim [Thu, 10 May 2018 01:52:14 +0000 (10:52 +0900)]
gadget: f_thor: fix filename overflow

The thor sender can send filename without null character and it is
used without consideration of overflow. Actually, character array
for filename is assigned with DEFINE_CACHE_ALIGN_BUFFER() and it
is bigger than size of memcpy, so there was no real overflow.
Fix filename overflow for code level integrity.

Signed-off-by: Seung-Woo Kim <sw0312.kim@samsung.com>
6 years agoMerge git://www.denx.de/git/u-boot-marvell
Tom Rini [Thu, 17 May 2018 16:38:30 +0000 (12:38 -0400)]
Merge git://www.denx.de/git/u-boot-marvell

6 years agonet: MVGBE don't automatically select PHYLIB
Chris Packham [Thu, 17 May 2018 12:12:04 +0000 (00:12 +1200)]
net: MVGBE don't automatically select PHYLIB

When Kconfig support was added for MVGBE it included automatically
selected PHYLIB support. But MVGBE does not need PHYLIB it will build
fine without it. Commit ed52ea507f12 ("net: add Kconfig for MVGBE")
should have been a no-op in terms of build size but because of the
selecting PHYLIB the openrd configs increased in size.

Remove the automatic selection of PHYLIB, boards that need it will have
already enabled it in their config header file.

Fixes: commit ed52ea507f12 ("net: add Kconfig for MVGBE")
Signed-off-by: Chris Packham <judge.packham@gmail.com>
Reviewed-by: Tom Rini <trini@konsulko.com>
Acked-by: Joe Hershberger <joe.hershberger@ni.com>
Signed-off-by: Stefan Roese <sr@denx.de>
6 years agoARM: kirkwood: Add device-tree for sheevaplug
Chris Packham [Tue, 8 May 2018 10:34:21 +0000 (22:34 +1200)]
ARM: kirkwood: Add device-tree for sheevaplug

Import the dts files from Linux 4.17 and enable device tree control in
u-boot.

Signed-off-by: Chris Packham <judge.packham@gmail.com>
Signed-off-by: Stefan Roese <sr@denx.de>
6 years agoARM: kirkwood: Add device-tree for pogo_e02
Chris Packham [Tue, 8 May 2018 10:34:20 +0000 (22:34 +1200)]
ARM: kirkwood: Add device-tree for pogo_e02

Import the dts files from Linux 4.17 and enable CONFIG_OF_CONTROL.

Signed-off-by: Chris Packham <judge.packham@gmail.com>
Signed-off-by: Stefan Roese <sr@denx.de>
6 years agoARM: kirkwood: Add device-tree for openrd
Chris Packham [Tue, 8 May 2018 10:34:19 +0000 (22:34 +1200)]
ARM: kirkwood: Add device-tree for openrd

Import the dts files from Linux 4.17 and enable CONFIG_OF_CONTROL.

Signed-off-by: Chris Packham <judge.packham@gmail.com>
Signed-off-by: Stefan Roese <sr@denx.de>
6 years agoARM: kirkwood: Add device-tree for nas220
Chris Packham [Tue, 8 May 2018 10:34:17 +0000 (22:34 +1200)]
ARM: kirkwood: Add device-tree for nas220

Import the dts file from Linux 4.17 and enable CONFIG_OF_CONTROL.

Signed-off-by: Chris Packham <judge.packham@gmail.com>
Signed-off-by: Stefan Roese <sr@denx.de>
6 years agoARM: kirkwood: Add device-tree for iconnect
Chris Packham [Tue, 8 May 2018 10:34:16 +0000 (22:34 +1200)]
ARM: kirkwood: Add device-tree for iconnect

Import the dts files from Linux 4.17 and enable CONFIG_OF_CONTROL.

Signed-off-by: Chris Packham <judge.packham@gmail.com>
Signed-off-by: Stefan Roese <sr@denx.de>
6 years agoARM: kirkwood: Add device-tree for ib62x0
Chris Packham [Tue, 8 May 2018 10:34:15 +0000 (22:34 +1200)]
ARM: kirkwood: Add device-tree for ib62x0

Import the dts files from Linux 4.17 and enable CONFIG_OF_CONTROL.

Signed-off-by: Chris Packham <judge.packham@gmail.com>
Signed-off-by: Stefan Roese <sr@denx.de>
6 years agoARM: kirkwood: Add device-tree for guruplug
Chris Packham [Tue, 8 May 2018 10:34:14 +0000 (22:34 +1200)]
ARM: kirkwood: Add device-tree for guruplug

Import the dts files from Linux 4.17 and enable CONFIG_OF_CONTROL.

Signed-off-by: Chris Packham <judge.packham@gmail.com>
Signed-off-by: Stefan Roese <sr@denx.de>
6 years agoARM: kirkwood: Add device-tree for goflexhome
Chris Packham [Tue, 8 May 2018 10:34:13 +0000 (22:34 +1200)]
ARM: kirkwood: Add device-tree for goflexhome

Import the dts files from Linux 4.17 and enable CONFIG_OF_CONTROL.

Signed-off-by: Chris Packham <judge.packham@gmail.com>
Signed-off-by: Stefan Roese <sr@denx.de>
6 years agoARM: kirkwood: Add device-tree for dockstar
Chris Packham [Tue, 8 May 2018 10:34:12 +0000 (22:34 +1200)]
ARM: kirkwood: Add device-tree for dockstar

Import the dts files from Linux 4.17 and enable CONFIG_OF_CONTROL.

Signed-off-by: Chris Packham <judge.packham@gmail.com>
Signed-off-by: Stefan Roese <sr@denx.de>
6 years agoARM: kirkwood: Add device-tree for dns325
Chris Packham [Tue, 8 May 2018 10:34:11 +0000 (22:34 +1200)]
ARM: kirkwood: Add device-tree for dns325

Import the dts files from Linux 4.17 and enable CONFIG_OF_CONTROL.

Signed-off-by: Chris Packham <judge.packham@gmail.com>
Signed-off-by: Stefan Roese <sr@denx.de>
6 years agoARM: add devicetree files for kirkwood SoC
Chris Packham [Tue, 8 May 2018 10:34:10 +0000 (22:34 +1200)]
ARM: add devicetree files for kirkwood SoC

These files are taken verbatim from the Linux kernel 4.17

Signed-off-by: Chris Packham <judge.packham@gmail.com>
Signed-off-by: Stefan Roese <sr@denx.de>
6 years agoMerge git://git.denx.de/u-boot-dm
Tom Rini [Wed, 16 May 2018 21:32:59 +0000 (17:32 -0400)]
Merge git://git.denx.de/u-boot-dm

6 years agoARM: re-enable MVGBE for edminiv2
Chris Packham [Wed, 16 May 2018 08:31:43 +0000 (20:31 +1200)]
ARM: re-enable MVGBE for edminiv2

This was unintentionally disabled when moving MVGBE to Kconfig.

Fixes: commit ed52ea507f12 ("net: add Kconfig for MVGBE")
Signed-off-by: Chris Packham <judge.packham@gmail.com>
6 years agopinctrl: do not set_state for device without valid ofnode
Kever Yang [Wed, 18 Apr 2018 09:54:04 +0000 (17:54 +0800)]
pinctrl: do not set_state for device without valid ofnode

Not all the udevice have a available DT node, eg. rksdmmc@ff500000.blk
which add by mmc_bind(), these device do not have/need set pinctrl
state.

Signed-off-by: Kever Yang <kever.yang@rock-chips.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
6 years agotools: buildman: Don't use the working dir as build dir
Lothar Waßmann [Sun, 8 Apr 2018 11:14:11 +0000 (05:14 -0600)]
tools: buildman: Don't use the working dir as build dir

When the U-Boot base directory happens to have the same name as the branch
that buildman is directed to use via the '-b' option and no output
directory is specified with '-o', buildman happily starts removing the
whole U-Boot sources eventually only stopped with the error message:

OSError: [Errno 20] Not a directory: '../<branch-name>/boards.cfg

Add a check to avoid this and also deal with the case where '-o' points
to the source directory, or any subdirectory of it.

Finally, tidy up the confusing logic for removing the old tree when using
-b. This is only done when building a branch.

Signed-off-by: Lothar Waßmann <LW@KARO-electronics.de>
Signed-off-by: Simon Glass <sjg@chromium.org>
Tested-by: Lothar Waßmann <LW@KARO-electronics.de>
6 years agousb: composite convert __set_bit to generic_set_bit
Bryan O'Donoghue [Mon, 30 Apr 2018 14:56:10 +0000 (15:56 +0100)]
usb: composite convert __set_bit to generic_set_bit

Compiling the f_mass_storage driver for an x86 target results in a
compilation error as set_bit and clear_bit are provided by bitops.h

To address that situation we discussed on the list moving to
genetic_set_bit() instead.

Doing a quick grep for similar situations in drivers/usb shows that the
composite device is using __set_bit().

This patch switches over to generic_set_bit to maintain consistency between
the two gadget drivers.

Signed-off-by: Bryan O'Donoghue <pure.logic@nexus-software.ie>
Cc: Lukasz Majewski <lukma@denx.de>
Cc: Marek Vasut <marex@denx.de>
6 years agousb: f_mass_storage: Fix set_bit and clear_bit usage
Bryan O'Donoghue [Mon, 30 Apr 2018 14:56:09 +0000 (15:56 +0100)]
usb: f_mass_storage: Fix set_bit and clear_bit usage

Compiling the f_mass_storage driver for an x86 target results in a
compilation error as set_bit and clear_bit are provided by bitops.h

Looking at the provenance of the current u-boot code and the git change
history in the kernel, it looks like we have a local copy of set_bit and
clear_bit as a hold-over from porting the Linux driver into u-boot.

These days __set_bit and __clear_bit are optionally provided by an arch and
can be used as inputs to generic_bit_set and generic_bit_clear.

This patch switches over to generic_set_bit and generic_clear_bit to
accommodate.

Tested on i.MX WaRP7 and Intel Edison

Signed-off-by: Bryan O'Donoghue <pure.logic@nexus-software.ie>
Cc: Lukasz Majewski <lukma@denx.de>
Cc: Marek Vasut <marex@denx.de>
6 years agonds32: Define PLATFORM__CLEAR_BIT for generic_clear_bit()
Bryan O'Donoghue [Mon, 30 Apr 2018 14:56:08 +0000 (15:56 +0100)]
nds32: Define PLATFORM__CLEAR_BIT for generic_clear_bit()

nds2 bitops.h provides a __clear_bit() but does not define
PLATFORM__CLEAR_BIT as a result generic_clear_bit() is used instead of the
architecturally provided __clear_bit().

This patch defines PLATFORM__CLEAR_BIT which means that __clear_bit() in
nds32 bitops.h will be called whenever generic_clear_bit() is called - as
opposed to the default cross-platform generic_clear_bit().

Signed-off-by: Bryan O'Donoghue <pure.logic@nexus-software.ie>
Cc: Macpaul Lin <macpaul@andestech.com>
6 years agonds32: Define PLATFORM__SET_BIT for generic_set_bit()
Bryan O'Donoghue [Mon, 30 Apr 2018 14:56:07 +0000 (15:56 +0100)]
nds32: Define PLATFORM__SET_BIT for generic_set_bit()

nds32 bitops.h provides a __set_bit() but does not define PLATFORM__SET_BIT
as a result generic_set_bit() is used instead of the architecturally
provided __set_bit().

This patch defines PLATFORM__SET_BIT which means that __set_bit() in nds32
bitops.h will be called whenever generic_set_bit() is called - as opposed
to the default cross-platform generic_set_bit().

Signed-off-by: Bryan O'Donoghue <pure.logic@nexus-software.ie>
Cc: Macpaul Lin <macpaul@andestech.com>
6 years agonios2: Define PLATFORM__CLEAR_BIT for generic_clear_bit()
Bryan O'Donoghue [Mon, 30 Apr 2018 14:56:06 +0000 (15:56 +0100)]
nios2: Define PLATFORM__CLEAR_BIT for generic_clear_bit()

nios2 bitops.h provides a __clear_bit() but does not define
PLATFORM__CLEAR_BIT as a result generic_clear_bit() is used instead of the
architecturally provided __clear_bit().

This patch defines PLATFORM__CLEAR_BIT which means that __clear_bit() in
nios2 bitops.h will be called whenever generic_clear_bit() is called - as
opposed to the default cross-platform generic_clear_bit().

Signed-off-by: Bryan O'Donoghue <pure.logic@nexus-software.ie>
Cc: Thomas Chou <thomas@wytron.com.tw>
6 years agonios2: Define PLATFORM__SET_BIT for generic_set_bit()
Bryan O'Donoghue [Mon, 30 Apr 2018 14:56:05 +0000 (15:56 +0100)]
nios2: Define PLATFORM__SET_BIT for generic_set_bit()

nios2 bitops.h provides a __set_bit() but does not define PLATFORM__SET_BIT
as a result generic_set_bit() is used instead of the architecturally
provided __set_bit().

This patch defines PLATFORM__SET_BIT which means that __set_bit() in nios2
bitops.h will be called whenever generic_set_bit() is called - as opposed
to the default cross-platform generic_set_bit().

Signed-off-by: Bryan O'Donoghue <pure.logic@nexus-software.ie>
Cc: Thomas Chou <thomas@wytron.com.tw>
6 years agoriscv: Define PLATFORM__CLEAR_BIT for generic_clear_bit()
Bryan O'Donoghue [Mon, 30 Apr 2018 14:56:04 +0000 (15:56 +0100)]
riscv: Define PLATFORM__CLEAR_BIT for generic_clear_bit()

riscv bitops.h provides a __clear_bit() but does not define
PLATFORM__CLEAR_BIT as a result generic_clear_bit() is used instead of the
architecturally provided __clear_bit().

This patch defines PLATFORM__CLEAR_BIT which means that __clear_bit() in
riscv bitops.h will be called whenever generic_clear_bit() is called - as
opposed to the default cross-platform generic_clear_bit().

Signed-off-by: Bryan O'Donoghue <pure.logic@nexus-software.ie>
Cc: Rick Chen <rick@andestech.com>
Cc: Greentime Hu <green.hu@gmail.com>
6 years agoriscv: Define PLATFORM__SET_BIT for generic_set_bit()
Bryan O'Donoghue [Mon, 30 Apr 2018 14:56:03 +0000 (15:56 +0100)]
riscv: Define PLATFORM__SET_BIT for generic_set_bit()

riscv bitops.h provides a __set_bit() but does not define PLATFORM__SET_BIT
as a result generic_set_bit() is used instead of the architecturally
provided __set_bit().

This patch defines PLATFORM__SET_BIT which means that __set_bit() in x86
bitops.h will be called whenever generic_set_bit() is called - as opposed
to the default cross-platform generic_set_bit().

Signed-off-by: Bryan O'Donoghue <pure.logic@nexus-software.ie>
Cc: Rick Chen <rick@andestech.com>
Cc: Greentime Hu <green.hu@gmail.com>
6 years agox86: Define PLATFORM__SET_BIT for generic_set_bit()
Bryan O'Donoghue [Mon, 30 Apr 2018 14:56:02 +0000 (15:56 +0100)]
x86: Define PLATFORM__SET_BIT for generic_set_bit()

x86 bitops.h provides a __set_bit() but does not define PLATFORM__SET_BIT
as a result generic_set_bit() is used instead of the architecturally
provided __set_bit().

This patch defines PLATFORM__SET_BIT which means that __set_bit() in x86
bitops.h will be called whenever generic_set_bit() is called - as opposed
to the default cross-platform generic_set_bit().

Signed-off-by: Bryan O'Donoghue <pure.logic@nexus-software.ie>
Cc: Simon Glass <sjg@chromium.org>
Cc: Bin Meng <bmeng.cn@gmail.com>
6 years agoARM: mvebu: a38x: Add missing SPDX license identfier
Chris Packham [Tue, 15 May 2018 01:31:11 +0000 (13:31 +1200)]
ARM: mvebu: a38x: Add missing SPDX license identfier

mv_ddr_build_message.c is generated in Marvell's standalone mv_ddr code.
When imported into u-boot we need to add the appropriate SPDX tag and
re-format it slightly.

Signed-off-by: Chris Packham <judge.packham@gmail.com>
6 years agoMerge branch 'master' of git://git.denx.de/u-boot-video
Tom Rini [Tue, 15 May 2018 12:29:24 +0000 (08:29 -0400)]
Merge branch 'master' of git://git.denx.de/u-boot-video

6 years agodts: sunxi: add PWM node for sun50i
Vasily Khoruzhick [Mon, 14 May 2018 15:16:21 +0000 (08:16 -0700)]
dts: sunxi: add PWM node for sun50i

Add PWM definition to sun50i-a64.dtsi

Signed-off-by: Vasily Khoruzhick <anarsoul@gmail.com>
6 years agopwm: sunxi: add support for PWM found on Allwinner A64
Vasily Khoruzhick [Mon, 14 May 2018 15:16:20 +0000 (08:16 -0700)]
pwm: sunxi: add support for PWM found on Allwinner A64

This commit adds basic support for PWM found on Allwinner A64.
It can be used for pwm_backlight driver (e.g. for Pinebook)

Signed-off-by: Vasily Khoruzhick <anarsoul@gmail.com>
6 years agovideo: dw_hdmi: fix HSYNC and VSYNC polarity settings
Vasily Khoruzhick [Mon, 14 May 2018 20:49:53 +0000 (13:49 -0700)]
video: dw_hdmi: fix HSYNC and VSYNC polarity settings

Currently dw_hdmi configures HSYNC polarity using VSYNC setting from
EDID and vice versa. Fix it, since it breaks displays where HSYNC
and VSYNC polarity differs

Signed-off-by: Vasily Khoruzhick <anarsoul@gmail.com>
Reviewed-by: Jernej Skrabec <jernej.skrabec@siol.net>
6 years agosunxi: video: HDMI: use correct bits for HSYNC and VSYNC polarity.
Vasily Khoruzhick [Mon, 14 May 2018 20:49:52 +0000 (13:49 -0700)]
sunxi: video: HDMI: use correct bits for HSYNC and VSYNC polarity.

HSYNC is bit 8, and VSYNC is bit 9.

Signed-off-by: Vasily Khoruzhick <anarsoul@gmail.com>
Reviewed-by: Jernej Skrabec <jernej.skrabec@siol.net>
6 years agommc: sdhci: Check that ops are defined
Ramon Fried [Mon, 14 May 2018 12:02:30 +0000 (15:02 +0300)]
mmc: sdhci: Check that ops are defined

The check is necessary to avoid NULL pointer dereference.

Signed-off-by: Ramon Fried <ramon.fried@gmail.com>
Reviewed-by: Michal Simek <michal.simek@xilinx.com>
6 years agonet: add Kconfig for MVGBE
Chris Packham [Thu, 3 May 2018 11:00:35 +0000 (23:00 +1200)]
net: add Kconfig for MVGBE

Add Kconfig for MVGBE and update boards to select this.

Signed-off-by: Chris Packham <judge.packham@gmail.com>
Reviewed-by: Stefan Roese <sr@denx.de>
6 years agonet: mvgbe: remove CONFIG_DOVE
Chris Packham [Thu, 3 May 2018 11:00:34 +0000 (23:00 +1200)]
net: mvgbe: remove CONFIG_DOVE

Nothing defines CONFIG_DOVE so remove the code that uses it.

Signed-off-by: Chris Packham <judge.packham@gmail.com>
Reviewed-by: Stefan Roese <sr@denx.de>
Acked-by: Joe Hershberger <joe.hershberger@ni.com>
6 years agonet: bootp: Fix compile error processing ntpserver option
Chris Packham [Thu, 3 May 2018 08:19:03 +0000 (20:19 +1200)]
net: bootp: Fix compile error processing ntpserver option

When the following configuration is set

  # CONFIG_CMD_DHCP is not set
  CONFIG_CMD_BOOTP=y
  CONFIG_BOOTP_NTPSERVER=y

The following compile error is observed

  error: used struct type value where scalar is required
    if (net_ntp_server)
        ^~~~~~~~~~~~~~

Resolve this by checking net_ntp_server.s_addr instead.

Signed-off-by: Chris Packham <judge.packham@gmail.com>
Acked-by: Joe Hershberger <joe.hershberger@ni.com>
6 years agonet: Add Kconfig option for BOOTP_NTPSERVER
Chris Packham [Thu, 3 May 2018 08:19:02 +0000 (20:19 +1200)]
net: Add Kconfig option for BOOTP_NTPSERVER

Add a Kconfig option for BOOTP_NTPSERVER to enable the DHCP/BOOTP option
to configure the sntp server address.

Signed-off-by: Chris Packham <judge.packham@gmail.com>
Acked-by: Joe Hershberger <joe.hershberger@ni.com>
6 years agotreewide: Move CONFIG_PHY_MARVELL to Kconfig
Mario Six [Fri, 27 Apr 2018 12:52:21 +0000 (14:52 +0200)]
treewide: Move CONFIG_PHY_MARVELL to Kconfig

The CONFIG_PHY_MARVELL has already been migrated to Kconfig (some boards
already had it in their Kconfig), but had not been moved for older
boards.

Move it to the defconfigs for all boards.

Signed-off-by: Mario Six <mario.six@gdsys.cc>
Acked-by: Joe Hershberger <joe.hershberger@ni.com>
6 years agoMerge branch 'master' of git://git.denx.de/u-boot-rockchip
Tom Rini [Mon, 14 May 2018 22:13:59 +0000 (18:13 -0400)]
Merge branch 'master' of git://git.denx.de/u-boot-rockchip

6 years agorockchip: clk: rk3288: handle clk_enable requests for GMAC
Jonathan Gray [Tue, 8 May 2018 09:49:05 +0000 (19:49 +1000)]
rockchip: clk: rk3288: handle clk_enable requests for GMAC

Since b0ba1e7e9d9b9441a18048ec67a3b3100c096975
(rockchip: clk: rk3288: add clk_enable function and support USB HOST0/HSIC)
Ethernet no longer probes on RK3288.

Add no-ops for GMAC clocks observed to be requested which match the
clk_enable cases in RK3368 and RK3399.

Signed-off-by: Jonathan Gray <jsg@jsg.id.au>
Cc: Wadim Egorov <w.egorov@phytec.de>
Cc: Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
Acked-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
Reviewed-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
6 years agorockchip: set SYS_NS16550_MEM32 for all SoCs
Jonathan Gray [Tue, 8 May 2018 10:43:01 +0000 (20:43 +1000)]
rockchip: set SYS_NS16550_MEM32 for all SoCs

Add back part of patch send out as
'rockchip: enable SYS_NS16550 for all SoCs by default' that seems to have
gotten lost when it got merged to set SYS_NS16550_MEM32.

Allows serial output to work on tinker-rk3288 again after
c3c0331db1fb7b1f4ff41e144fc04353b37c785c.

Signed-off-by: Kever Yang <kever.yang@rock-chips.com>
Signed-off-by: Jonathan Gray <jsg@jsg.id.au>
Cc: Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
Reviewed-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
Acked-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
Tested-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
6 years agoMerge git://git.denx.de/u-boot-marvell
Tom Rini [Mon, 14 May 2018 12:52:48 +0000 (08:52 -0400)]
Merge git://git.denx.de/u-boot-marvell

6 years agophy: marvell: a3700: Fix compatible string for ehci
Marek Behún [Fri, 11 May 2018 08:03:39 +0000 (10:03 +0200)]
phy: marvell: a3700: Fix compatible string for ehci

The DTS file for armada-37xx uses the string "marvell,armada3700-ehci",
but the code searched for "marvell,armada-3700-ehci".

Signed-off-by: Marek Behun <marek.behun@nic.cz>
Signed-off-by: Stefan Roese <sr@denx.de>
6 years agoARM: mvebu: a38x: use non-zero size for ddr scrubbing
Chris Packham [Thu, 10 May 2018 01:28:31 +0000 (13:28 +1200)]
ARM: mvebu: a38x: use non-zero size for ddr scrubbing

Make ddr3_calc_mem_cs_size() global scope and use it in
ddr3_new_tip_ecc_scrub to correctly initialize all of DDR memory.

Signed-off-by: Chris Packham <judge.packham@gmail.com>
Signed-off-by: Stefan Roese <sr@denx.de>
6 years agoARM: mvebu: a38x: restore support for setting timing
Chris Packham [Thu, 10 May 2018 01:28:30 +0000 (13:28 +1200)]
ARM: mvebu: a38x: restore support for setting timing

This restores support for configuring the timing mode based on the
ddr_topology. This was originally implemented in commit 90bcc3d38d2b
("driver/ddr: Add support for setting timing in hws_topology_map") but
was removed as part of the upstream sync.

Signed-off-by: Chris Packham <judge.packham@gmail.com>
Signed-off-by: Stefan Roese <sr@denx.de>
6 years agoARM: mvebu: a38x: sync ddr training code with upstream
Chris Packham [Thu, 10 May 2018 01:28:29 +0000 (13:28 +1200)]
ARM: mvebu: a38x: sync ddr training code with upstream

This syncs drivers/ddr/marvell/a38x/ with the mv_ddr-armada-17.10 branch
of https://github.com/MarvellEmbeddedProcessors/mv-ddr-marvell.git.

The upstream code is incorporated omitting the ddr4 and apn806 and
folding the nested a38x directory up one level. After that a
semi-automated step is used to drop unused features with unifdef

  find drivers/ddr/marvell/a38x/ -name '*.[ch]' | \
    xargs unifdef -m -UMV_DDR -UMV_DDR_ATF -UCONFIG_DDR4 \
-UCONFIG_APN806 -UCONFIG_MC_STATIC \
-UCONFIG_MC_STATIC_PRINT -UCONFIG_PHY_STATIC \
-UCONFIG_64BIT

INTER_REGS_BASE is updated to be defined as SOC_REGS_PHY_BASE.

Some now empty files are removed and the ternary license is replaced
with a SPDX GPL-2.0+ identifier.

Signed-off-by: Chris Packham <judge.packham@gmail.com>
Signed-off-by: Stefan Roese <sr@denx.de>
6 years agoARM: mvebu: a38x: remove some unused code
Chris Packham [Thu, 10 May 2018 01:28:28 +0000 (13:28 +1200)]
ARM: mvebu: a38x: remove some unused code

No in-tree code defines SUPPORT_STATIC_DUNIT_CONFIG or
STATIC_ALGO_SUPPORT. Remove ddr3_a38x_mc_static.h and use unifdef to
remove unused sections in the rest of the ddr/marvell/a38x code.

Signed-off-by: Chris Packham <judge.packham@gmail.com>
Signed-off-by: Stefan Roese <sr@denx.de>
6 years agoARM: mvebu: a38x: move sys_env_device_rev_get
Chris Packham [Thu, 10 May 2018 01:28:27 +0000 (13:28 +1200)]
ARM: mvebu: a38x: move sys_env_device_rev_get

Move sys_env_device_rev_get() from the ddr training code to
sys_env_lib.c (which currently resides with the serdes code). This
brings sys_env_device_rev_get() into line with sys_env_device_id_get()
and sys_env_model_get().

Signed-off-by: Chris Packham <judge.packham@gmail.com>
Signed-off-by: Stefan Roese <sr@denx.de>
6 years agoARM: mvebu: a38x: move definition of PEX_CFG_DIRECT_ACCESS
Chris Packham [Thu, 10 May 2018 01:28:26 +0000 (13:28 +1200)]
ARM: mvebu: a38x: move definition of PEX_CFG_DIRECT_ACCESS

PEX_CFG_DIRECT_ACCESS was defined in ddr3_hws_hw_training_def.h despite
only being used in the serdes code. Move this definition to ctrl_pex.h
where all the other PEX defines are. Also remove the duplicate
definition of PEX_DEVICE_AND_VENDOR_ID which is already defined in
ctrl_pex.h.

Signed-off-by: Chris Packham <judge.packham@gmail.com>
Signed-off-by: Stefan Roese <sr@denx.de>
6 years agoarm64: mvebu: Add basic support for the Turris Mox board
Marek Behún [Tue, 24 Apr 2018 15:21:31 +0000 (17:21 +0200)]
arm64: mvebu: Add basic support for the Turris Mox board

This adds basic support for the Turris Mox board from CZ.NIC, which is
currently being crowdfunded on Indiegogo.

Turris Mox is as modular router based on the Armada 3720 SOC (same as
EspressoBin).

The basic module can be extended by different modules. The device tree
binary for the kernel can be dependent on which modules are connected,
and in what order. Because of this, the board specific code creates
in U-Boot a variable called module_topology, which carries this
information.

Signed-off-by: Marek Behun <marek.behun@nic.cz>
Signed-off-by: Stefan Roese <sr@denx.de>
6 years agowatchdog: Add support for Armada 37xx CPU watchdog
Marek Behún [Tue, 24 Apr 2018 15:21:30 +0000 (17:21 +0200)]
watchdog: Add support for Armada 37xx CPU watchdog

This adds support for the CPU watchdog found on Marvell Armada 37xx
SoCs.

There are 4 counters which can be set as CPU watchdog counters.
This driver uses the second counter (ID 1, counting from 0)
(Marvell's Linux also uses second counter by default).
In the future it could be adapted to use other counters, with
definition in the device tree.

Signed-off-by: Marek Behun <marek.behun@nic.cz>
Signed-off-by: Stefan Roese <sr@denx.de>
6 years agonet: mvneta: Fix fault when wrong device tree
Marek Behún [Tue, 24 Apr 2018 15:21:29 +0000 (17:21 +0200)]
net: mvneta: Fix fault when wrong device tree

The driver does not check id phy_connect failed (for example on wrong
property name in device tree). In such a case a fault occurs and the
CPU is restarted.

Signed-off-by: Marek Behun <marek.behun@nic.cz>
Reviewed-by: Stefan Roese <sr@denx.de>
Signed-off-by: Stefan Roese <sr@denx.de>
6 years agophy: marvell: core: Cosmetic fixes
Marek Behún [Tue, 24 Apr 2018 15:21:28 +0000 (17:21 +0200)]
phy: marvell: core: Cosmetic fixes

Move the reg_set* functions into comphy.h as static inline functions.
Change return type of get_*_string to const char *.

Signed-off-by: Marek Behun <marek.behun@nic.cz>
Reviewed-by: Stefan Roese <sr@denx.de>
Signed-off-by: Stefan Roese <sr@denx.de>
6 years agoclk: armada-37xx: Support soc_clk_dump
Marek Behún [Tue, 24 Apr 2018 15:21:27 +0000 (17:21 +0200)]
clk: armada-37xx: Support soc_clk_dump

Add support for the clk dump command on Armada 37xx.

Signed-off-by: Marek Behun <marek.behun@nic.cz>
Reviewed-by: Stefan Roese <sr@denx.de>
Signed-off-by: Stefan Roese <sr@denx.de>
6 years agospi: mvebu_a3700_spi: Use Armada 37xx clk driver for SPI clock frequency
Marek Behún [Tue, 24 Apr 2018 15:21:26 +0000 (17:21 +0200)]
spi: mvebu_a3700_spi: Use Armada 37xx clk driver for SPI clock frequency

Since now we have driver for clocks on Armada 37xx, use it to determine
SQF clock frequency for the SPI driver.

Also change the default config files for Armada 37xx devices so that
the clock driver is enabled by default, otherwise the SPI driver cannot
be enabled.

Signed-off-by: Marek Behun <marek.behun@nic.cz>
Signed-off-by: Stefan Roese <sr@denx.de>
6 years agodriver: clk: Add support for clocks on Armada 37xx
Marek Behún [Tue, 24 Apr 2018 15:21:25 +0000 (17:21 +0200)]
driver: clk: Add support for clocks on Armada 37xx

The drivers are based on Linux driver by Gregory Clement.

The TBG clocks support only the .get_rate method.
  - since setting rate is not supported, the driver computes the rates
    when probing and so subsequent calls to the .get_rate method do not
    read the corresponding registers again

The peripheral clocks support methods .get_rate, .enable and .disable.

  - the .set_parent method theoretically could be supported on some clocks
    (the parent would have to be one of the TBG clocks)

  - the .set_rate method would have to try all the divider values to find
    the best approximation of a given rate, and it doesn't seem like
    this should be needed in U-Boot, therefore not implemented

Signed-off-by: Marek Behun <marek.behun@nic.cz>
Reviewed-by: Stefan Roese <sr@denx.de>
Signed-off-by: Stefan Roese <sr@denx.de>
6 years agophy: marvell: a3700: Save/restore selector reg in SGMII init
Marek Behún [Tue, 24 Apr 2018 15:21:24 +0000 (17:21 +0200)]
phy: marvell: a3700: Save/restore selector reg in SGMII init

In SGMII initialization PIN_PIPE_SEL has to be zero when resetting
the PHY. Since comphy_mux already set the selector register to
correct values, we have to store it's value before setting it to 0
and restore it after SGMII init.

Signed-off-by: Marek Behun <marek.behun@nic.cz>
Reviewed-by: Stefan Roese <sr@denx.de>
Signed-off-by: Stefan Roese <sr@denx.de>
6 years agophy: marvell: a3700: Use comphy_mux on Armada 37xx.
Marek Behún [Tue, 24 Apr 2018 15:21:23 +0000 (17:21 +0200)]
phy: marvell: a3700: Use comphy_mux on Armada 37xx.

Lane 0 supports SGMII1 and USB3.
Lane 1 supports SGMII0 and PEX0.
Lane 2 supports SATA0 and USB3.

This is needed for Armada 37xx.

This introduces new device tree bindings. AFAIK there is currently no
driver for Armada 37xx comphy in Linux. When such a driver will be
pushed into Linux, this will need to be rewritten accordingly.

Signed-off-by: Marek Behun <marek.behun@nic.cz>
Signed-off-by: Stefan Roese <sr@denx.de>
6 years agophy: marvell: a3700: Fix SGMII cfg and stat register addresses
Marek Behún [Tue, 24 Apr 2018 15:21:22 +0000 (17:21 +0200)]
phy: marvell: a3700: Fix SGMII cfg and stat register addresses

The register addresses on lanes 0 and 1 are switched, first comes 1 and
then 0.

Signed-off-by: Marek Behun <marek.behun@nic.cz>
Signed-off-by: Stefan Roese <sr@denx.de>
6 years agophy: marvell: mux: Support nontrivial node order in selector register
Marek Behún [Tue, 24 Apr 2018 15:21:21 +0000 (17:21 +0200)]
phy: marvell: mux: Support nontrivial node order in selector register

Currently comphy_mux supports only trivial order of nodes in pin
selector register, that is lane N on position N*bitcount.

Add support for nontrivial order, with map stored in device tree
property mux-lane-order.

This is needed for Armada 37xx.

As far as I know, there is no driver for Armada 37xx comphy in the
kernel. When such a driver comes, this will need to be rewritten to
support the device tree bindings from the kernel.

Signed-off-by: Marek Behun <marek.behun@nic.cz>
Signed-off-by: Stefan Roese <sr@denx.de>
6 years agophy: marvell: a3700: revise the USB3 comphy setting during power on
zachary [Tue, 24 Apr 2018 15:21:20 +0000 (17:21 +0200)]
phy: marvell: a3700: revise the USB3 comphy setting during power on

This commit is based on commit d9899826 by
  zachary <zhangzg@marvell.com>
from u-boot-marvell, see
github.com/MarvellEmbeddedProcessors/u-boot-marvell/commit/d9899826

- According to design specification, the transmitter should be set to high
  impedence mode during electrical idle. Thus transmitter should detect RX
  at high impedence mode also, and delay is needed to accommodate high
  impedence off latency. Otherwise the USB3 will have detection issue that
  most of the time the USB3 device can not be detected at all, or be
  detected as USB2 device sometimes.
  Modified registers: RD005C302h (R181h) (0051h) Lane Configuration 1
  Bit 6: set to 1 to let Tx detect Rx at HiZ mode
  Bit [3:4]: set to 2 to be delayed by 2 clock cycles
  Bit 0: set to 1 to set transmitter to high impedance mode during idle.
- USB3 De-emphasize level of -3.5dB is mandatory, but USB3 MAC selects 0x2
  (emphasize disabled) in the MAC_PHY_TXDEEMPH [1:0], while it is supposed
  to select 0x1(3.5dB emphasize). Thus need to override what comes from
  the MAC(by setting register 0x1c2 bit2 to 0x1) and to configure the
  overridded values of MAC_PHY_TXDEEMPH [1:0] to 0x1(bit15 of register
  0x181 and bit0 of register 0x180).
- According to USB3 application note, need to update below comphy
  registers:
  Set max speed generation to USB3.0 5Gbps(set RD005C04Ah bit[11:10] to 1)
  Set capacitor value to 0xF(set RF005C224 bit[3:0] to 0xF)

Signed-off-by: Marek Behun <marek.behun@nic.cz>
Signed-off-by: Stefan Roese <sr@denx.de>
6 years agophy: marvell: a3700: Set USB3 RX wait depending on ref clock
Marek Behún [Tue, 24 Apr 2018 15:21:19 +0000 (17:21 +0200)]
phy: marvell: a3700: Set USB3 RX wait depending on ref clock

According to specification, CFG_PM_RXDLOZ_WAIT should be set to 0x7
when reference clock is at 25 MHz. The specification (at least the
version I have) does not mentoin the setting for 40 MHz reference
clock, but Marvell's U-Boot sets 0xC in that case.

Signed-off-by: Marek Behun <marek.behun@nic.cz>
Reviewed-by: Stefan Roese <sr@denx.de>
Signed-off-by: Stefan Roese <sr@denx.de>
6 years agophy: marvell: a3700: Access USB3 register indirectly on lane 2
Marek Behún [Tue, 24 Apr 2018 15:21:18 +0000 (17:21 +0200)]
phy: marvell: a3700: Access USB3 register indirectly on lane 2

When USB3 is on comphy lane 2 on the Armada 37xx, the registers
have to be accessed indirectly via SATA indirect access.

This is the case of the Turris Mox board from CZ.NIC.

Signed-off-by: Marek Behun <marek.behun@nic.cz>
Signed-off-by: Stefan Roese <sr@denx.de>
6 years agophy: marvell: a3700: Use reg_set_indirect istead of 2 reg_sets
Marek Behún [Tue, 24 Apr 2018 15:21:17 +0000 (17:21 +0200)]
phy: marvell: a3700: Use reg_set_indirect istead of 2 reg_sets

Create a special function for indirect register setting,
reg_set_indirect, and use it instead of the two calls to reg_set.

Signed-off-by: Marek Behun <marek.behun@nic.cz>
Reviewed-by: Stefan Roese <sr@denx.de>
Signed-off-by: Stefan Roese <sr@denx.de>
6 years agophy: marvell: a3700: Use (!ret) instead of (ret == 0)
Marek Behún [Tue, 24 Apr 2018 15:21:16 +0000 (17:21 +0200)]
phy: marvell: a3700: Use (!ret) instead of (ret == 0)

In U-Boot it is usually written this way.

Signed-off-by: Marek Behun <marek.behun@nic.cz>
Reviewed-by: Stefan Roese <sr@denx.de>
Signed-off-by: Stefan Roese <sr@denx.de>
6 years agophy: marvell: a3700: Use same timeout for all register polling
Marek Behún [Tue, 24 Apr 2018 15:21:15 +0000 (17:21 +0200)]
phy: marvell: a3700: Use same timeout for all register polling

The timeout is set to PLL_LOCK_TIMEOUT in every call to
comphy_poll_reg. Remove this parameter from the function.

Signed-off-by: Marek Behun <marek.behun@nic.cz>
Reviewed-by: Stefan Roese <sr@denx.de>
Signed-off-by: Stefan Roese <sr@denx.de>
6 years agophy: marvell: a3700: Don't create functional macro for each register
Marek Behún [Tue, 24 Apr 2018 15:21:14 +0000 (17:21 +0200)]
phy: marvell: a3700: Don't create functional macro for each register

Currently there is for each register special functional macro, ie:
  LANE_CFG1_ADDR(u)
  GLOB_CLK_SRC_LO_ADDR(u)
  ...
where can be either PCIE or USB3.

Change this to one function PHY_ADDR(unit, addr). The code becomes:
  phy_addr(PCIE, LANE_CFG1)
  phy_addr(PCIE, GLOB_CLK_SRC_LO)
  ...

Signed-off-by: Marek Behun <marek.behun@nic.cz>
Signed-off-by: Stefan Roese <sr@denx.de>
6 years agophy: marvell: a3700: Use reg_set16 instead of phy_write16
Marek Behún [Tue, 24 Apr 2018 15:21:13 +0000 (17:21 +0200)]
phy: marvell: a3700: Use reg_set16 instead of phy_write16

The macro phy_write16 is not used by the rest of the code,
phy_read16 is not used at all.
We also change the macro SGMIIPHY_ADDR to a static inline function.

Signed-off-by: Marek Behun <marek.behun@nic.cz>
Signed-off-by: Stefan Roese <sr@denx.de>
6 years agophy: marvell: a3700: Change return type of macro MVEBU_REG
Marek Behún [Tue, 24 Apr 2018 15:21:12 +0000 (17:21 +0200)]
phy: marvell: a3700: Change return type of macro MVEBU_REG

All the calls to reg_set and friends have to cast the first argument
to void __iomem *. Lets change the return type of the MVEBU_REG macro
instead.

Signed-off-by: Marek Behun <marek.behun@nic.cz>
Signed-off-by: Stefan Roese <sr@denx.de>
6 years agommc: Improve tinification
Marek Vasut [Sat, 14 Apr 2018 22:37:11 +0000 (00:37 +0200)]
mmc: Improve tinification

Drop all the extra content from the MMC core, so that tiny MMC support
is really tiny, no fancy anything. That means the tiny MMC support does
only 1-bit transfers at default speed settings. Moreover, this patch
drops duplicate instance of struct mmc mmc_static, which wasted about
360 bytes. Furthermore, since MMC tiny supports only one controller
at all times, get rid of mallocating the ext csd backup and replace
it with static array. All in all, this patch saves ~4 kiB of bloat
from the MMC core, which on platforms with severe limitations can be
beneficial.

Signed-off-by: Marek Vasut <marek.vasut@gmail.com>
Cc: Jaehoon Chung <jh80.chung@samsung.com>
[trini: Fixup checkpatch.pl style warnings]
Signed-off-by: Tom Rini <trini@konsulko.com>
6 years agoMerge git://git.denx.de/u-boot-tegra
Tom Rini [Fri, 11 May 2018 19:22:36 +0000 (15:22 -0400)]
Merge git://git.denx.de/u-boot-tegra

6 years agoSPDX: Correct SPDX tags from recent xilinx merge
Tom Rini [Fri, 11 May 2018 18:54:57 +0000 (14:54 -0400)]
SPDX: Correct SPDX tags from recent xilinx merge

Correct the SPDX tag format.

Fixes: 3b52847a451a ("Merge tag 'xilinx-for-v2018.07' of git://www.denx.de/git/u-boot-microblaze")
Signed-off-by: Tom Rini <trini@konsulko.com>
6 years agoMerge tag 'xilinx-for-v2018.07' of git://www.denx.de/git/u-boot-microblaze
Tom Rini [Fri, 11 May 2018 15:45:28 +0000 (11:45 -0400)]
Merge tag 'xilinx-for-v2018.07' of git://denx.de/git/u-boot-microblaze

Xilinx changes for v2018.07

microblaze:
- Align defconfig

zynq:
- Rework fpga initialization and cpuinfo handling

zynqmp:
- Add ZynqMP R5 support
- Wire and enable watchdog on zcu100-revC
- Setup MMU map for DDR at run time
- Show board info based on DT and cleanup IDENT_STRING

zynqmp tools:
- Add read partition support
- Add initial support for Xilinx bif format for boot.bin generation

mmc:
- Fix get_timer usage on 64bit cpus
- Add support for SD3.0 UHS mode

nand-zynq:
- Add support for 16bit buswidth
- Use address cycles from onfi params

scsi:
- convert ceva sata to UCLASS_AHCI

timer:
- Add Cadence TTC for ZynqMP r5

watchdog:
- Minor cadence driver cleanup

6 years agoMerge git://git.denx.de/u-boot-fsl-qoriq
Tom Rini [Fri, 11 May 2018 11:09:21 +0000 (07:09 -0400)]
Merge git://git.denx.de/u-boot-fsl-qoriq

6 years agoarm64: zynqmp: Enable UHS support for ZCU102 Rev1.0 board
Siva Durga Prasad Paladugu [Thu, 19 Apr 2018 07:07:10 +0000 (12:37 +0530)]
arm64: zynqmp: Enable UHS support for ZCU102 Rev1.0 board

This patch enables UHS support for ZynqMP zcu102 rev 1.0
board.

Signed-off-by: Siva Durga Prasad Paladugu <siva.durga.paladugu@xilinx.com>
Signed-off-by: Michal Simek <michal.simek@xilinx.com>
6 years agommc: zynq_sdhci: Add support for SD3.0
Siva Durga Prasad Paladugu [Thu, 19 Apr 2018 07:07:09 +0000 (12:37 +0530)]
mmc: zynq_sdhci: Add support for SD3.0

This patch adds support of SD3.0 for ZynqMP.

Signed-off-by: Siva Durga Prasad Paladugu <siva.durga.paladugu@xilinx.com>
Signed-off-by: Michal Simek <michal.simek@xilinx.com>
6 years agommc: sdhci: Read capabilities register1 and update host caps
Siva Durga Prasad Paladugu [Thu, 19 Apr 2018 07:07:08 +0000 (12:37 +0530)]
mmc: sdhci: Read capabilities register1 and update host caps

This patch reads the capabilities register1 and update the host
caps accordingly for mmc layer usage. This patch mainly reads
for UHS capabilities inorder to support SD3.0.

Signed-off-by: Siva Durga Prasad Paladugu <sivadur@xilinx.com>
Signed-off-by: Michal Simek <michal.simek@xilinx.com>
6 years agommc: sdhci: Invoke platform specific tuning and delay routines
Siva Durga Prasad Paladugu [Thu, 19 Apr 2018 07:07:07 +0000 (12:37 +0530)]
mmc: sdhci: Invoke platform specific tuning and delay routines

This patch adds support to invoke any platform specific tuning
and delay routines if available.

Signed-off-by: Siva Durga Prasad Paladugu <sivadur@xilinx.com>
Signed-off-by: Michal Simek <michal.simek@xilinx.com>
6 years agosdhci: Add new sdhci ops for platform specific tuning and delays
Siva Durga Prasad Paladugu [Thu, 19 Apr 2018 07:07:06 +0000 (12:37 +0530)]
sdhci: Add new sdhci ops for platform specific tuning and delays

This patch adds new hooks for any platform specific tuning and
tap delays programing. These are needed for supporting
SD3.0.

Signed-off-by: Siva Durga Prasad Paladugu <sivadur@xilinx.com>
Signed-off-by: Michal Simek <michal.simek@xilinx.com>
6 years agommc: sdhci: Handle execute tuning command in sdhci_send_command
Siva Durga Prasad Paladugu [Thu, 19 Apr 2018 07:07:05 +0000 (12:37 +0530)]
mmc: sdhci: Handle execute tuning command in sdhci_send_command

This patch upadted sdhci_send_command to handle execute tuning
command.

Signed-off-by: Siva Durga Prasad Paladugu <sivadur@xilinx.com>
Signed-off-by: Michal Simek <michal.simek@xilinx.com>
6 years agommc: sdhci: Add support for disabling clock
Siva Durga Prasad Paladugu [Thu, 19 Apr 2018 07:07:04 +0000 (12:37 +0530)]
mmc: sdhci: Add support for disabling clock

This patch adds support to disable clock if clk_disable
was set and then enable or set clock if the clock was changed
or clock was disabled when clock needs to be enabled.

Signed-off-by: Siva Durga Prasad Paladugu <sivadur@xilinx.com>
Signed-off-by: Michal Simek <michal.simek@xilinx.com>
6 years agommc: Changed the datatype of the variable to handle 64-bit arch
Vipul Kumar [Thu, 3 May 2018 06:50:54 +0000 (12:20 +0530)]
mmc: Changed the datatype of the variable to handle 64-bit arch

This patch changed the datatype of variable "start" from uint to ulong
to work properly on 64-bit machines as well. Also the return type of
get_timer() function is ulong.

Signed-off-by: Vipul Kumar <vipul.kumar@xilinx.com>
Signed-off-by: Michal Simek <michal.simek@xilinx.com>
6 years agoarm: zynqmp: Add ZynqMP minimal R5 support
Michal Simek [Thu, 12 Apr 2018 15:39:46 +0000 (17:39 +0200)]
arm: zynqmp: Add ZynqMP minimal R5 support

Xilinx ZynqMP also contains dual Cortex R5 which can run U-Boot.
This patch is adding minimal support to get U-Boot boot.
U-Boot on R5 runs out of DDR with default configuration that's why
DDR needs to be partitioned if there is something else running on arm64.
Console is done via Cadence uart driver and the first Cadence Triple
Timer Counter is used for time.

This configuration with uart1 was tested on zcu100-revC.

U-Boot 2018.05-rc2-00021-gd058a08d907d (Apr 18 2018 - 14:11:27 +0200)

Model: Xilinx ZynqMP R5
DRAM:  512 MiB
WARNING: Caches not enabled
MMC:
In:    serial@ff010000
Out:   serial@ff010000
Err:   serial@ff010000
Net:   Net Initialization Skipped
No ethernet found.
ZynqMP r5>

There are two ways how to run this on ZynqMP.
1. Run from ZynqMP arm64
tftpb 20000000 u-boot-r5.elf
setenv autostart no && bootelf -p 20000000
cpu 4 disable && cpu 4 release 10000000 lockstep
or
cpu 4 disable && cpu 4 release 10000000 split

2. Load via jtag when directly to R5

Signed-off-by: Michal Simek <michal.simek@xilinx.com>
6 years agotools: zynqmpimage: Add bif support
Alexander Graf [Fri, 13 Apr 2018 12:18:52 +0000 (14:18 +0200)]
tools: zynqmpimage: Add bif support

The officially described way to generate boot.bin files for ZynqMP is to
describe the contents of the target binary using a file of the "bif"
format.  This file then links to other files that all get packed into a
bootable image.

This patch adds support to read such a .bif file and generate a respective
ZynqMP boot.bin file that can include the normal image and pmu files, but
also supports image partitions now. This makes it a handy replacement for
the proprietary "bootgen" utility that is currently used to generate
boot.bin files with FSBL.

Signed-off-by: Alexander Graf <agraf@suse.de>
Signed-off-by: Michal Simek <michal.simek@xilinx.com>
6 years agoMAINTAINERS: Declare tools/zynqmp* as Xilinx maintained
Alexander Graf [Fri, 13 Apr 2018 12:18:51 +0000 (14:18 +0200)]
MAINTAINERS: Declare tools/zynqmp* as Xilinx maintained

The zynqmpimage.c and the new zynqmpbif.c files are all maintained by
Xilinx for the Zynq platforms. Let's match them accordingly
in the MAINTAINERS file.

Signed-off-by: Alexander Graf <agraf@suse.de>
Signed-off-by: Michal Simek <michal.simek@xilinx.com>
6 years agotools: zynqmpimage: Move defines to header
Alexander Graf [Fri, 13 Apr 2018 12:18:50 +0000 (14:18 +0200)]
tools: zynqmpimage: Move defines to header

We will add support for ZynqMP bif input files later, so let's move
all structure definitions into a header file that can be used by that
one as well.

Signed-off-by: Alexander Graf <agraf@suse.de>
Signed-off-by: Michal Simek <michal.simek@xilinx.com>
6 years agotools: zynqmpimage: Add partition read support
Alexander Graf [Fri, 13 Apr 2018 12:18:49 +0000 (14:18 +0200)]
tools: zynqmpimage: Add partition read support

The zynqmp image format has support for inline partitions which are
used by FSBL to describe payloads that are loaded by FSBL itself.

While we can't create images that contain partitions (yet), we should
still at least be able to examine them and show the user what's inside
when we analyze an image created by bootgen.

Signed-off-by: Alexander Graf <agraf@suse.de>
Signed-off-by: Michal Simek <michal.simek@xilinx.com>
6 years agoarm64: zynqmp: Show model information instead of custom IDENT_STRING
Michal Simek [Mon, 30 Apr 2018 07:26:47 +0000 (09:26 +0200)]
arm64: zynqmp: Show model information instead of custom IDENT_STRING

DISPLAY_BOARDINFO in OF case show model identification string from DT.
Enable this feature instead of custom IDENT_STRING which does the same
thing.

Signed-off-by: Michal Simek <michal.simek@xilinx.com>
Reviewed-by: Alexander Graf <agraf@suse.de>
6 years agoarm64: zynqmp: Simplify boot_target variable composition
Michal Simek [Wed, 25 Apr 2018 09:10:34 +0000 (11:10 +0200)]
arm64: zynqmp: Simplify boot_target variable composition

Call calloc for space allocation only at one location and include if/else
to sprintf. This will simplify run time device adding based on id aliases.

Signed-off-by: Michal Simek <michal.simek@xilinx.com>
Reviewed-by: Alexander Graf <agraf@suse.de>
6 years agoarm64: zynqmp: Setup MMU map for DDR at run time
Nitin Jain [Fri, 20 Apr 2018 07:00:40 +0000 (12:30 +0530)]
arm64: zynqmp: Setup MMU map for DDR at run time

This patch fills the MMU map for DDR at run time based on information read
from Device Tree or automatically detected from static configuration.

The patch is needed because for systems which has for example 1GB of memory
but MMU map is 2GB there could be spurious accesses which was seen in past
when mapping is not fitting with actual memory installed.

Signed-off-by: Nitin Jain <nitin.jain@xilinx.com>
Signed-off-by: Siva Durga Prasad Paladugu <sivadur@xilinx.com>
Signed-off-by: Michal Simek <michal.simek@xilinx.com>
6 years agoarm64: zynqmp: Enable cadence WDT for zcu100
Michal Simek [Thu, 19 Apr 2018 13:43:50 +0000 (15:43 +0200)]
arm64: zynqmp: Enable cadence WDT for zcu100

Enable watchdog on zcu100 to make sure if there is a bug in the u-boot
there is proper reset.
Watchdog expires and PMU fw is informed and based on setting proper
action is taken.

The patch is enabling reset-on-timeout feature and also fixing fixed
clock rate for watchdog where 100MHz is max (and also default) clock value.

Signed-off-by: Michal Simek <michal.simek@xilinx.com>
6 years agoarm64: zynqmp: Wire watchdog internals
Michal Simek [Thu, 19 Apr 2018 13:43:38 +0000 (15:43 +0200)]
arm64: zynqmp: Wire watchdog internals

Enable watchdog in full U-Boot.

Similar changes were done by:
"arm: zynq: Wire watchdog internals"
(sha1: e6cc3b25d721c3001019f8b44bfaae2a57255162)

Signed-off-by: Michal Simek <michal.simek@xilinx.com>
6 years agowatchdog: cadence: Show used timeout value
Michal Simek [Thu, 19 Apr 2018 13:22:04 +0000 (15:22 +0200)]
watchdog: cadence: Show used timeout value

Debug message was showing timeout value which was passed to start
function but there is a checking if this value can be setup.
The patch is moving this debug printf function below checking.

Signed-off-by: Michal Simek <michal.simek@xilinx.com>
6 years agoarm64: zynqmp: Reset FPD Watchdog on zcu100
Michal Simek [Thu, 19 Apr 2018 12:41:11 +0000 (14:41 +0200)]
arm64: zynqmp: Reset FPD Watchdog on zcu100

Low level configuration didn't reset FPD Watchdog that's why accessing
it caused u-boot hang.

Signed-off-by: Michal Simek <michal.simek@xilinx.com>
6 years agonand: zynq: Send address cycles as per onfi parameter page
Siva Durga Prasad Paladugu [Thu, 25 May 2017 06:45:24 +0000 (12:15 +0530)]
nand: zynq: Send address cycles as per onfi parameter page

Send address cycles as per value read from onfi parameter
page for Read and write commands instead of using a
hard coded value. This may vary for different parts and
hence use it from onfi parameter page value.

Signed-off-by: Siva Durga Prasad Paladugu <siva.durga.paladugu@xilinx.com>
Signed-off-by: Michal Simek <michal.simek@xilinx.com>
6 years agonand: zynq: Add support for 16-bit buswidth
Siva Durga Prasad Paladugu [Thu, 25 May 2017 08:55:55 +0000 (14:25 +0530)]
nand: zynq: Add support for 16-bit buswidth

This patch adds support for 16-bit buswidth by determining
the bus width based on mio configuration.

Signed-off-by: Siva Durga Prasad Paladugu <siva.durga.paladugu@xilinx.com>
Signed-off-by: Michal Simek <michal.simek@xilinx.com>
6 years agoMAINTAINERS: zynqmp: Point to proper zynqmp folder
Michal Simek [Wed, 18 Apr 2018 13:00:43 +0000 (15:00 +0200)]
MAINTAINERS: zynqmp: Point to proper zynqmp folder

Point to Zynqmp arm64 cpu folder not to Zynq arm32.

Signed-off-by: Michal Simek <michal.simek@xilinx.com>