Kever Yang [Fri, 23 Jun 2017 08:11:06 +0000 (16:11 +0800)]
rockchip: use common sdram function
Replace the sdram_init() in board init and rockchip_sdram_size() in
sdram driver for all the Rockchip SoCs which enable CONFIG_RAM.
Signed-off-by: Kever Yang <kever.yang@rock-chips.com> Reviewed-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
Make dram_init() in rk3036-board.c conditional on CONFIG_RAM: Signed-off-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
Kever Yang [Fri, 23 Jun 2017 08:11:05 +0000 (16:11 +0800)]
rockchip: add sdram_common for common functions
There are some functions like sdram_size_mb can be re-used for
different rockchip SoCs, just put them into common file.
Add board_get_usable_ram_top() for ram_top init base on
SDRAM_MAX_SIZE.
Signed-off-by: Kever Yang <kever.yang@rock-chips.com> Reviewed-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
Added SDRAM_MAX_SIZE definition for RK3036: Signed-off-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
fixup: 3036 fix for sdram_common
Kever Yang [Tue, 13 Jun 2017 13:00:12 +0000 (21:00 +0800)]
rockchip: rk3328: correct mem_region
According to rk3328 TRM:
0~0xff000000 is ddr space;
0xff000000~0xffffffff is device space.
Signed-off-by: Kever Yang <kever.yang@rock-chips.com> Reviewed-by: Simon Glass <sjg@chromium.org> Reviewed-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
Philipp Tomsich [Wed, 7 Jun 2017 16:46:03 +0000 (18:46 +0200)]
rockchip: dm: convert fdt_get to dev_read
With the new dev_read functions available, we can convert the rockchip
architecture-specific drivers and common drivers used by these devices
over to the dev_read family of calls.
This covers the DRAM controller initialisation for the RK3188, RK3288
and RK3399... all of these read some of the tuning/setup/timing
parameters from the device-tree.
Signed-off-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com> Reviewed-by: Simon Glass <sjg@chromium.org>
Philipp Tomsich [Wed, 7 Jun 2017 16:46:02 +0000 (18:46 +0200)]
rockchip: ns16550: dm: convert fdt_get to dev_read
With the new dev_read functions available, we can convert the rockchip
architecture-specific drivers and common drivers used by these devices
over to the dev_read family of calls.
This covers the serial driver (ns16550 and compatible) used for the
Rockchip devices.
Signed-off-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com> Reviewed-by: Simon Glass <sjg@chromium.org>
Philipp Tomsich [Wed, 7 Jun 2017 16:46:01 +0000 (18:46 +0200)]
rockchip: net: dm: convert fdt_get to dev_read
With the new dev_read functions available, we can convert the rockchip
architecture-specific drivers and common drivers used by these devices
over to the dev_read family of calls.
This covers the Gigabit Ethernet MAC (i.e. common designware driver and
rockchip-specific wrapper).
Signed-off-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com> Reviewed-by: Simon Glass <sjg@chromium.org>
Philipp Tomsich [Wed, 7 Jun 2017 16:46:00 +0000 (18:46 +0200)]
rockchip: mmc: dm: convert fdt_get to dev_read
With the new dev_read functions available, we can convert the rockchip
architecture-specific drivers and common drivers used by these devices
over to the dev_read family of calls.
This covers the dw_mmc and sdhci wrapper drivers for Rockchip.
Signed-off-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com> Reviewed-by: Simon Glass <sjg@chromium.org>
Philipp Tomsich [Wed, 7 Jun 2017 16:45:59 +0000 (18:45 +0200)]
rockchip: xhci: dm: convert fdt_get to dev_read
With the new dev_read functions available, we can convert the rockchip
architecture-specific drivers and common drivers used by these devices
over to the dev_read family of calls.
This change covers the USB3 (xhci) driver for the Rockchip devices.
Signed-off-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com> Reviewed-by: Simon Glass <sjg@chromium.org>
Philipp Tomsich [Wed, 7 Jun 2017 16:45:58 +0000 (18:45 +0200)]
rockchip: spi: dm: convert fdt_get to dev_read
With the new dev_read functions available, we can convert the rockchip
architecture-specific drivers and common drivers used by these devices
over to the dev_read family of calls.
This change covers the rk_spi.c (SPI driver) used in Rockchip devices.
Signed-off-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com> Reviewed-by: Simon Glass <sjg@chromium.org>
Philipp Tomsich [Wed, 7 Jun 2017 16:45:57 +0000 (18:45 +0200)]
rockchip: pinctrl: dm: convert fdt_get to dev_read
With the new dev_read functions available, we can convert the rockchip
architecture-specific drivers and common drivers used by these devices
over to the dev_read family of calls.
This change covers the pinctrl drivers for the Rockchip devices.
Signed-off-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com> Reviewed-by: Simon Glass <sjg@chromium.org>
If the value of either "nand options" or "bbt options" has a zero in
the most significant nibble, the '0x' prefix will be isolated from the
value like shown below:
|Device 0: nand0, sector size 128 KiB
| Page size 2048 b
| OOB size 64 b
| Erase size 131072 b
| subpagesize 2048 b
| options 0x40000200
| bbt options 0x 60000
Change the format string to produce leading zeroes filling the gap.
Simon Glass [Mon, 10 Jul 2017 03:30:17 +0000 (21:30 -0600)]
libfdt: Drop -FDT_ERR_TOODEEP
This error code has not been upstreamed and is not really needed since it
is unlikely to be triggered. Drop it to maintain compatability with
upstream.
Reported-by: Peter Robinson <pbrobinson@gmail.com> Signed-off-by: Simon Glass <sjg@chromium.org> Tested-by: Peter Robinson <pbrobinson@gmail.com>
powerpc, 8xx: Properly set CPM frequency in the device tree
For processors whose core runs at twice the bus frequency,
the fallback frequency calculation in Linux provides a wrong
result. Therefore, U-boot needs to pass the correct value.
CS Systemes d'Information (CSSI) manufactures 8xx boards for
critical communication systems. Those boards have been
running U-Boot since 2010 and will have to be maintained
until at least 2027.
This commit partially restores support for the 8xx, with the
following limitations:
- Restores support for MPC866 and MPC885 only
- Does not restore IDE, PCMCIA, I2C, USB
- Does not restore examples
- Does not restore POST
- Does not restore Ethernet on SCC
- Does not restore console on SCC
- Does not restore bedbug and kgdb support
As the 866 and 885 do not support the following features,
they are not restored either:
- VIDEO / LCD
- RTC clock
The CPM uCODE patch is not restored either, because:
- 866 and 885 already have support for I2C and SPI relocation
without a uCODE patch
- relocation of SMC, I2C or SPI is only needed for using SCCs
for Ethernet or QMC
The dynamic setup/calculation of clocks is removed, we
expect the target being use with the clock and PLPRCR register
defined in the configuration.
All the clock settings for 8xx prior to 866 is removed as
well as we now only support 866 and 885.
This code is mature and addresses mature boards. Therefore
all code enclosed in '#if 0/#endif' and '#if XX_DEBUG/#endif'
is unneeded.
The following files are not restored by this patch:
Some of the restored files are not located in a proper location.
In order to keep traceability of the changes, they will be
moved to their correct location and moved to Kconfig in a
followup patch.
This patch also declares CSSI as point of contact for the update
of the 8xx platform, as those boards are the only ones still
being maintained on the 8xx area. A later patch will add
those boards to the tree.
On startup the SL50 board halt at U-Boot prompt. Use CONFIG_AUTOBOOT_KEYED
to enable autoboot for this board and define the <SPACE> key to get the
U-Boot prompt.
Signed-off-by: Enric Balletbo i Serra <enric.balletbo@collabora.com>
Andy Shevchenko [Wed, 5 Jul 2017 13:25:22 +0000 (16:25 +0300)]
avr32: Retire AVR32 for good
AVR32 is gone. It's already more than two years for no support in Buildroot,
even longer there is no support in GCC (last version is heavily patched 4.2.4).
Linux kernel v4.12 got rid of it (and v4.11 didn't build successfully).
There is no good point to keep this support in U-Boot either.
Reviewed-by: Simon Glass <sjg@chromium.org> Reviewed-by: Tom Rini <trini@konsulko.com> Reviewed-by: Heiko Schocher <hs@denx.de> Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Simon Glass [Sat, 1 Jul 2017 17:06:09 +0000 (11:06 -0600)]
stm32: Correct positioning of declaration
The current code gives a warning:
arch/arm/mach-stm32/stm32f7/soc.c: In function 'arch_cpu_init':
arch/arm/mach-stm32/stm32f7/soc.c:38:2: error: 'for' loop initial
declarations are only allowed in C99 or C11 mode
for (int i = 0; i < ARRAY_SIZE(stm32_region_config); i++)
^
arch/arm/mach-stm32/stm32f7/soc.c:38:2: note: use option -std=c99,
-std=gnu99, -std=c11 or -std=gnu11 to compile your code
Fix it by moving the declaration to the top of the function.
Signed-off-by: Simon Glass <sjg@chromium.org>
Series-cc trini
The Odroid C2 comes with several expansion options such as an RTC module.
As these expansions are not considered in the Linux device tree they have
to be enabled via device tree overlays.
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Baruch Siach [Thu, 29 Jun 2017 17:37:08 +0000 (20:37 +0300)]
mkimage: fix display of image types list
Since commit 5b9d44df2307f (mkimage: Display a better list of available image
types) mkimage usage text suggest to "use -T to see a list of available image
types". Unfortunately, commit 02221f29deb8 (mkimage: Convert to use getopt())
broke that feature, because getopt() fails when -T has no option argument.
Add a pseudo image type name 'list' that lists all image types. Update the
usage text accordingly.
Cc: Simon Glass <sjg@chromium.org> Signed-off-by: Baruch Siach <baruch@tkos.co.il>
Peter Robinson [Mon, 26 Jun 2017 13:25:06 +0000 (14:25 +0100)]
CHIP: add device tree overlay support
CHIP and CHIP Pro devices have options of DIP addon boards some of which need
overlays from the beginning so it makes sense to enable device tree overlays.
Signed-off-by: Peter Robinson <pbrobinson@gmail.com>
Peter Robinson [Mon, 26 Jun 2017 13:25:05 +0000 (14:25 +0100)]
96boards: dragonboard/hikey: add device tree overlay support
96boards CE devices such as the DragonBoard and Hikey devices have numerous
mezzanine options some of which need overlays from the beginning so it makes
sense to enable device tree overlays.
Signed-off-by: Peter Robinson <pbrobinson@gmail.com>
Peter Robinson [Mon, 26 Jun 2017 13:25:04 +0000 (14:25 +0100)]
am335x: beagle/evm: add device tree overlay support
TI am33xx devices such as the BeagleBone devices have numerous cape options such
as screens some of which need overlays from the beginning so it makes sense to
enable evice tree overlays.
Signed-off-by: Peter Robinson <pbrobinson@gmail.com>
Peter Robinson [Mon, 26 Jun 2017 13:25:03 +0000 (14:25 +0100)]
Raspberry Pi: add device tree overlay support
Raspberry Pi has numerous HAT and other HW expansion options such as screens and
cameras some of which need overlays from the beginning so it makes sense to
enable evice tree overlays.
Signed-off-by: Peter Robinson <pbrobinson@gmail.com>
UEFI spec 2.7 indicates that HandleProtocol can be implemented
by calling OpenProtocol with
attributes = EFI_OPEN_PROTOCOL_BY_HANDLE_PROTOCOL.
Currently we pass attributes = 0 to efi_open_protocol. 0 is not a
valid value when calling OpenProtocol. This does not cause any errors
yet because our implementation of OpenProtocol is incomplete.
We should pass the correct value to enable a fully compliant
implementation of OpenProtocol in the future.
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de> Signed-off-by: Alexander Graf <agraf@suse.de>
Heiko Schocher [Tue, 27 Jun 2017 14:49:14 +0000 (16:49 +0200)]
powerpc: remove 4xx support
There was for long time no activity in the 4xx area.
We need to go further and convert to Kconfig, but it
turned out, nobody is interested anymore in 4xx,
so remove it.
Jonathan Gray [Sun, 12 Mar 2017 08:26:07 +0000 (19:26 +1100)]
efi_loader: check CreateEvent() parameters
Add some of the invalid parameter checks described in the UEFI
specification for CreateEvent(). This does not include checking
the validity of the type and tpl parameters.
Signed-off-by: Jonathan Gray <jsg@jsg.id.au> Acked-By: Heinrich Schuchardt <xypron.glpk@gmx.de>
[agraf: fix checkpatch.pl indent warning] Signed-off-by: Alexander Graf <agraf@suse.de>
Jonathan Gray [Sun, 12 Mar 2017 08:26:06 +0000 (19:26 +1100)]
efi_loader: run CreateEvent() notify function based on flags
The UEFI specification states that the tpl, function and context
arguments are to be ignored if neither EVT_NOTIFY_WAIT or
EVT_NOTIFY_SIGNAL are specified. This matches observed behaviour with
an AMI EDK2 based UEFI implementation.
Skip calling the notify function if neither flag is present.
Signed-off-by: Jonathan Gray <jsg@jsg.id.au> Acked-By: Heinrich Schuchardt <xypron.glpk@gmx.de> Signed-off-by: Alexander Graf <agraf@suse.de>
Alexander Graf [Mon, 3 Jul 2017 11:32:35 +0000 (13:32 +0200)]
efi_loader: Add check for fallback fdt memory reservation
When running bootefi, we allocate new space but never check whether
the allocation succeeded. This patch adds a check so that in case
things go wrong, we at least know they did.
Christophe Leroy [Tue, 27 Jun 2017 14:00:24 +0000 (16:00 +0200)]
doc: restore doc/README.fsl-clk
doc/README.fsl-clk was removed in commit 5b8e76c35ec31 ("powerpc, 8xx: remove support for 8xx")
allthought CONFIG_SYS_FSL_CLK is defined in
arch/arm/cpu/armv8/fsl-layerscape/Kconfig and still in use
in the following configs:
Marek Vasut [Tue, 2 May 2017 18:27:48 +0000 (20:27 +0200)]
ARM: at91: ma5d4: Switch environment start to eMMC
The redesigned version of the SoM which was released onto the market
does no longer contain SPI flash, but boots from the eMMC. Move the
environment storage to the eMMC.
Signed-off-by: Marek Vasut <marex@denx.de> Cc: Andreas Bießmann <andreas.devel@googlemail.com>
Marek Vasut [Tue, 2 May 2017 18:27:47 +0000 (20:27 +0200)]
ARM: at91: ma5d4: Boot from MMC2 when using SAM-BA
Continue loading U-Boot from MMC2 when the SPL was loaded using SAM-BA
loader. This allows the board to boot system from the removable media
instead of the eMMC, which is useful for commissioning purposes. When
booting from the eMMC, always boot from it as it is not possible to
boot from the SD interface directly.
Signed-off-by: Marek Vasut <marex@denx.de> Cc: Andreas Bießmann <andreas.devel@googlemail.com>
Marek Vasut [Tue, 2 May 2017 18:27:45 +0000 (20:27 +0200)]
ARM: at91: ma5d4: Swap SD/MMC controller order
The SDHCI1 is the primary boot controller on rev. 2.1 SoM, which
is the version available on the market. Swap the controller order
to match this and future versions of the SoM.
Signed-off-by: Marek Vasut <marex@denx.de> Cc: Andreas Bießmann <andreas.devel@googlemail.com>
Marek Vasut [Tue, 2 May 2017 18:27:42 +0000 (20:27 +0200)]
ARM: at91: ma5d4: Reset CAN controllers late
The CAN controllers need slight delay between toggling of their reset
line. Move this action into board_init(), otherwise timer will not be
initialized and the board might hang.
Signed-off-by: Marek Vasut <marex@denx.de> Cc: Andreas Bießmann <andreas.devel@googlemail.com>
Marek Vasut [Tue, 2 May 2017 18:27:41 +0000 (20:27 +0200)]
ARM: at91: ma5d4: Switch DDR2 controller to sequencial address decoding
According to the datasheet, sequential mapping is used for DDR
SDRAM, while interleaved mapping is used for regular SDRAM.
Incorrect configuration of this bit does indeed cause sporadic
memory instability.
Signed-off-by: Marek Vasut <marex@denx.de> Cc: Andreas Bießmann <andreas.devel@googlemail.com> Cc: Wenyou Yang <wenyou.yang@atmel.com>
Alexey Brodkin [Fri, 25 Nov 2016 13:23:43 +0000 (16:23 +0300)]
arc: Add support for HS Development Kit board
ARC HS Development Kit board is a new low-cost
development platform sporting ARC HS38 in real silicon
with nice set of features such as:
* Quad-core ARC HS38 with 512 kB L2 cache and running @1GHz
* 4Gb of DDR (we use only lowest 1Gb out of it now)
* Lots of DesigWare peripherals
* Different connectivity modules:
- Synopsys HAPS HT3
- Arduino-compatible connector
- MikroBUS
This initial commit supports the following peripherals:
* UART (DW 8250)
* Ethernet (DW GMAC)
* SD/MMC (DW Mobile Storage)
* USB 1.1 & 2.0
Alexey Brodkin [Mon, 26 Jun 2017 08:46:47 +0000 (11:46 +0300)]
arcv2: Set IOC aperture so it covers available DDR
We used to use the same memory layout and size for a couple of
boards and thus we just hardcoding IOC aperture start and size.
Now when we're getting more boards with more memory on board we
need to have an ability to set IOC so it matches real DDR layout
and size.
Even though it is not really a must but for simplicity we assume
IOC covers all the DDR we have, that gives us a chance to not
bother where DMA buffers are allocated - any part of DDR is OK.
Alexey Brodkin [Thu, 1 Jun 2017 14:09:38 +0000 (17:09 +0300)]
axs10x: Move environment from I2C EEPROM to SD-card
With deprecation of I2C EEPROM we we left without a permamnent
storage for U-Boot environment, but luckily we may simply use SD-card
with FAT partition for that.
Having environment on SD-card is much more convenient as it
allows us to preserve all the settings when moving from one board to
another. Moreover instead of 256 bytes of EEPROM we're now virtually
unlimited in stuff being placed in environment like complicated scripts
etc which are usually required in case of full-scale distros.
Alexey Brodkin [Thu, 1 Jun 2017 13:27:27 +0000 (16:27 +0300)]
axs10x: Get rid of both I2C and EEPROM who used to use I2C
With eb5ba3aefdf0 "i2c: Drop use of CONFIG_I2C_HARD" in place we
cannot use I2C EEPROM any longer so we're dropping all references to
both EEPROM and I2C which was only used for EEPROM.
Signed-off-by: Alexey Brodkin <abrodkin@synopsys.com> Cc: Simon Glass <sjg@chromium.org>
ARCompact cores are not supposed to be used in SMP designs
(this doesn't stop people from creation of heterogeneous chips,
for an example keep reading) so there's no point in
checking ARCNUM and halting somebody if we build for ARC700.
Moreover on AXS101 board we have ARC770 in the ASIC together with
other ARC cores and ARC770 happens to be the last node in JTAG chain
with ARCNUM = 4. And existing check halts the one and only core we
want keep running.
Heiko Schocher [Fri, 23 Jun 2017 18:13:58 +0000 (20:13 +0200)]
drivers, usb, gadget: fix compiler warnings for at91_udc.c
fix warnings:
drivers/usb/gadget/at91_udc.c:1344:12: warning: 'at91rm9200_udc_init' defined but not used [-Wunused-function]
drivers/usb/gadget/at91_udc.c:1379:13: warning: 'at91rm9200_udc_pullup' defined but not used [-Wunused-function]
drivers/usb/gadget/at91_udc.c:1476:12: warning: 'at91sam9263_udc_init' defined but not used [-Wunused-function]
Bin Meng [Thu, 29 Jun 2017 03:36:25 +0000 (11:36 +0800)]
scripts: config_whitelist: Handle lines with leading spaces/tabs
Some Kconfig options are defined in a line with leading spaces/tabs.
Update build-whitelist/check-config scripts to handle such cases.
Generate a new config_whitelist.txt with new scripts.
Andy Yan [Wed, 28 Jun 2017 08:27:37 +0000 (16:27 +0800)]
ARM: make memset and memcpy prompt message more clearly
The origin SPL_USE_ARCH_MEMSET/MEMCPY use same prompt message
as USE_ARCH_MEMSET/MEMCPY, which makes it's hard to distinguish
them in menuconfig interface. This patch gives them different
prompt messages for spl and none-spl config.
Lukasz Majewski [Mon, 26 Jun 2017 11:04:58 +0000 (13:04 +0200)]
usb: gadget: Call g_dnl_bind_fixup() before testing g_dnl_serial length
After the commit SHA1: 842778a091 - the serial number descriptor is only
visible when we have non zero length of g_dnl_serial.
However, on some platforms (e.g. Siemens) the serial number is set at
g_dnl_bind_fixup(), so with the current code we will always omit the
serial (since it is not set).
This commit moves the g_dnl_bind_fixup() call before the g_dnl_serial
length test.
Tom Rini [Wed, 10 May 2017 16:01:02 +0000 (12:01 -0400)]
ti816x: Enable ethernet support
The ti816x SoC revision of the ethernet IP block is handled by the
"davinci_emac" driver, rather than the "cpsw" driver as done by later
members of the family. Enable the relevant plumbing.