Michal Simek [Wed, 2 Oct 2019 13:55:57 +0000 (15:55 +0200)]
arm64: zynqmp: Add support for u-boot.itb generation with ATF
Follow i.MX, Sunxi, RISC-V and Rockchip to generate u-boot.itb which
includes U-Boot proper, ATF and DTBs in FIT format. ZynqMP supports FIT for
quite a long time but with using out of tree solution. The patch is filling
this gap.
Michal Simek [Thu, 10 Oct 2019 11:00:38 +0000 (13:00 +0200)]
test/py: hush_if_test: Add tests to cover octal/hex values
Extend test suite to cover also automatic octal/hex converstions which
haven't been implemented in past.
Signed-off-by: Michal Simek <michal.simek@xilinx.com> Acked-by: Stephen Warren <swarren@nvidia.com> Reviewed-by: Simon Goldschmidt <simon.k.r.goldschmidt@gmail.com>
Remove hardcoded base addresses of smc controller and nand controller.
Get those addresses from dt and replace wherever they are used.
Remove smc and nand base address from header file too.
Signed-off-by: Ashok Reddy Soma <ashok.reddy.soma@xilinx.com> Signed-off-by: Michal Simek <michal.simek@xilinx.com>
handoff_setup() was used to generate fixed handoff structure for ATF on
ZynqMP platform.
Switching to bl2_plat_get_bl31_params() platform brings more flexibility
because information can be taken from fit image where /fit-images node is
created at run time.
Signed-off-by: Michal Simek <michal.simek@xilinx.com>
Michal Simek [Thu, 19 Dec 2019 17:13:31 +0000 (18:13 +0100)]
spl: atf: Enable option to rewrite bl2_plat_get_bl31_params()
Xilinx ZynqMP platform is passing information to ATF in private format and
ATF bl31 parameters are not used. That's why enable option to rewrite this
function by platform specific implementation.
The patch also move and update kernel-doc format with missing parameters.
Signed-off-by: Michal Simek <michal.simek@xilinx.com>
When two instances of AXI QSPI with flash are added and tested
simultaneously the spi driver operations are relocated twice.
As a result code is accessing addresses outside of RAM when
relocated second time which is causing a crash.
Tested on Microblaze.
Similar change was done in past by:
commit f238b3f0fbc9 ("watchdog: dm: Support manual relocation for watchdogs")
commit 2588f2ddfd60 ("dm: sf: Add support for all targets which requires MANUAL_RELOC")
commit 1b4c2aa25bdf ("gpio: dm: Support manual relocation for gpio")
Signed-off-by: Ashok Reddy Soma <ashok.reddy.soma@xilinx.com> Signed-off-by: Michal Simek <michal.simek@xilinx.com>
Michal Simek [Fri, 6 Dec 2019 11:05:34 +0000 (12:05 +0100)]
arm: xilinx: Use device tree which can be passed on cmd line
DEVICE_TREE variable has been added by commit 6ab6b2afa091
("dts: re-write dts/Makefile more simply with Kbuild")
where DT can be specified on command line. If it is not then
CONFIG_DEFAULT_DEVICE_TREE is used to ensure backward compatibility.
Signed-off-by: Michal Simek <michal.simek@xilinx.com>
Michal Simek [Mon, 9 Dec 2019 12:00:57 +0000 (13:00 +0100)]
arm64: zynqmp: Add board_boot_order for MMC boot extension
In past SPL_ZYNQMP_TWO_SDHCI symbol was introduced to handle boards with
two sdhci controllers. The problem was that U-Boot is registering
controllers based on aliases in DT but bootmode targets specific controller
ID. That's why on boards with one "second" sdhci controller bootmode was
pointing to second controller(MMC2) but alias was setup to mmc0 (the first
controller). And SPL requires to point to mmc0 in this case.
Long time ago commit f101e4bd3703
("spl: add support for alternative boot device") added support for handling
multiple bootmodes in SPL. Use this functionality and setup second sdhci
controller as backup boot device.
Below is table with behavior:
HW/bootmode bootorder
sd0/sd0 mmc0/mmc1 (mmc1 never called)
sd1/sd1 mmc1/mmc0 (mmc0 fails and mmc1 is called)
sd0+sd1/sd0 mmc0/mmc1 (mmc1 never called)
sd0+sd1/sd1 mmc1/mmc0 (mmc0 never called)
All other bootmodes are not affected but order can be extended to cover
advance boot flows.
Signed-off-by: Michal Simek <michal.simek@xilinx.com>
Michal Simek [Tue, 3 Dec 2019 14:02:50 +0000 (15:02 +0100)]
arm64: zynqmp: Do not assing MIO34 that early on zcu100
MIO34 is connected to POWER_KILL signal. When MIO configuration is done in
psu_init() and this pin is assigned to PMU but PMU configuration is not
loaded yet. PMU gpio output is high that means board is powered off
immediately.
The patch is fixing this sequence that MIO34 stays assing to ps gpio IP.
PMU config is loaded in SPL and then pin assigned to PMU through
psu_post_config_data().
Signed-off-by: Michal Simek <michal.simek@xilinx.com>
Michal Simek [Fri, 20 Dec 2019 10:35:52 +0000 (11:35 +0100)]
arm64: Enable INIT_SP_RELATIVE by default when POC is enabled
When position-independent pre-relocation code is enable there is also
necessary to enable relative early stack pointer not to use origin location
pointed by CONFIG_SYS_INIT_SP_ADDR macro.
Signed-off-by: Michal Simek <michal.simek@xilinx.com>
Michal Simek [Thu, 19 Dec 2019 12:56:33 +0000 (13:56 +0100)]
arm64: zynqmp: Do not call bss init and board_init_r from board_init_f
There is no reason to clear bss and call board_init_r() from board_init_f()
beca it can be called directly from crt0_64.S with also support for SPL
stack relocation to SDRAM.
For more information please take a look at arch/arm/lib/crt0_64.S
Signed-off-by: Michal Simek <michal.simek@xilinx.com>
Michal Simek [Thu, 19 Dec 2019 14:42:13 +0000 (15:42 +0100)]
spl: atf: Add missing ndepth initialization
ndepth needs to be initialized before it is used in fdt_next_node().
Uninitialized value is causing that node is found and depth increase but
won't pass condition below because initial state wasn't setup.
Signed-off-by: Michal Simek <michal.simek@xilinx.com>
T Karthik Reddy [Wed, 18 Dec 2019 10:34:41 +0000 (03:34 -0700)]
board: xilinx: Add support for user configurable boot script offset
Currently "script_offset_f" env variable is hardcoded, this variable
specifies from which offset of the flash boot.scr should be read/write.
As flashes are of different sizes having a fixed offset makes it
difficult to load other images into the flash which may overwrite the
boot script or cannot utilize the full memory. This current fix
creates a new config "CONFIG_BOOT_SCRIPT_OFFSET" which holds the
offset address, overwrites the "script_offset_f" variable.
Also removed existing variable with default values, as the default
values are held by CONFIG_BOOT_SCRIPT_OFFSET
Signed-off-by: T Karthik Reddy <t.karthik.reddy@xilinx.com> Signed-off-by: Michal Simek <michal.simek@xilinx.com>
Emmc boot devcie is mounted to mmc@ff160000 controller on zynqmp.
Detect mmc alias at run time for setting up proper boot_targets
sequence instead of setting boot_targets manually.
Removed setting "modeboot" variable as it not longer need as we
switched to distro boot.
Signed-off-by: T Karthik Reddy <t.karthik.reddy@xilinx.com> Signed-off-by: Michal Simek <michal.simek@xilinx.com>
Emmc boot devcie is mounted to sdhci@f1050000 controller on versal, while
in emmc boot mode, boot tragets are set to mmc0. So removed mmc0 and using
the dev seq number from the sdhci@f1050000 controller.
Dev seq number is setup based on DT aliases that's why needs to be setup at
run time.
Signed-off-by: T Karthik Reddy <t.karthik.reddy@xilinx.com> Signed-off-by: Michal Simek <michal.simek@xilinx.com>
Eugeniu Rosca [Tue, 24 Dec 2019 16:51:08 +0000 (17:51 +0100)]
cmd: adtimg: Refactor usage style
Trying to extend 'adtimg' functionality [1], we've been severely hit
by a major limitation in the command's usage scheme. Specifically, the
command's user interface appears to be too centric to getting the
DTB/DTBO entry [3] based on the index of the desired DT in the image,
which makes it really difficult retrieving the DT entry based on
alternative criteria (e.g. filtering by id/rev fields), the
latter being demanded by real life customer use-cases [1].
This went to the point of receiving below feedback from Sam [2]:
-- snip --
As for 'dtimg' command: after giving it some thought, I think not much
people using it yet. So in this particular case I don't have some
strong preference, and if you think the 'dtimg' interface is ugly, and
it overcomes "don't break interfaces" rule, maybe now is a good time
to rework it (before it gets widely used).
-- snip --
Given the above, rework the usage pattern from [4] to [5], in order to
allow an intuitive enablement of "by id|rev" DT search [6].
[1] https://patchwork.ozlabs.org/cover/1202575/
("cmd: dtimg: Enhance with --id and --rev options (take #1)")
[2] https://patchwork.ozlabs.org/patch/1182207/#2317020
[3] https://source.android.com/devices/architecture/dto/partitions
[4] Old usage
adtimg dump <addr> - Print image contents
adtimg start <addr> <index> <varname> - Get DT address by index
adtimg size <addr> <index> <varname> - Get DT size by index
[5] New usage
adtimg addr <addr> - Set image location to <addr>
adtimg dump - Print out image contents
adtimg get dt --index=<i> [avar [svar]] - Get DT address and size by index
[6] Soon-to-be-provided "by id|rev" add-on functionality
adtimg get dt --id=<id> --rev=<rev> [avar [svar [ivar]]]
- Get DT address/size/index by id|rev fields
Signed-off-by: Eugeniu Rosca <erosca@de.adit-jv.com> Reviewed-by: Sam Protsenko <semen.protsenko@linaro.org>
Eugeniu Rosca [Tue, 24 Dec 2019 16:51:06 +0000 (17:51 +0100)]
dtimg/am57xx_evm_defconfig: Rename dtimg to adtimg
Rename the existing 'dtimg' command to 'adtimg', in order to:
- Suggest the Android origins and scope
- Be consistent with the upcoming 'abootimg' command (naming
suggested by Simon [*])
The change in _not_ backward compatible, but its benefits outweigh its
downsides, given that we don't expect active users of 'dtimg' today.
Perform the rename in several steps:
1. Rename *.c file and Kconfig symbol. This should allow
'git log --follow' to properly track the history of 'adtimg.c'
2. 's/dtimg/adtimg/g' in the internal namespace of 'adtimg.c'
ELF comparison [**] before and after shows no functional change.
Cc: Tom Rini <trini@konsulko.com> Signed-off-by: Eugeniu Rosca <erosca@de.adit-jv.com> Reviewed-by: Simon Glass<sjg@chromium.org> Reviewed-by: Sam Protsenko <semen.protsenko@linaro.org>
Rajesh Ravi [Fri, 22 Nov 2019 22:50:01 +0000 (14:50 -0800)]
arm: cpu: armv8: add support for arm psci reset2.
Current U-Boot has only support for psci reset.
Adding support for arm psci reset2 allows passing of reset level
and other platform sepcific parameters like strap settings
to lowlevel psci implementation.
Signed-off-by: Rajesh Ravi <rajesh.ravi@broadcom.com> Signed-off-by: Vladimir Olovyannikov <vladimir.olovyannikov@broadcom.com>
Programs the following:
1. Redistributor PROCBASER configuration table (which
is common for all redistributors)
2. Redistributor pending table (PENDBASER), for all the
available redistributors.
Signed-off-by: Bharat Kumar Reddy Gooty <bharat.gooty@broadcom.com> Signed-off-by: Vladimir Olovyannikov <vladimir.olovyannikov@broadcom.com>
Philippe Reynes [Thu, 9 Jan 2020 16:34:02 +0000 (17:34 +0100)]
test/cmd_ut.c: fix cmd_ut_category
In the function cmd_ut_category, the prefix is used with
the function strncmp to know if the prefix should be
removed from the test name, even if the prefix is NULL.
To avoid this issue, we consider that a prefix NULL
mean no prefix. So we only try to remove the prefix
from the test_name if the prefix is not NULL, then
we avoid to call the function strncmp with a NULL
prefix.
Reported-by: Coverity CID 281110 Fixes: 4ad4edfe ("cmd_ut: add a parameter prefix to the function cmd_ut_category") Signed-off-by: Philippe Reynes <philippe.reynes@softathome.com>
Andy Shevchenko [Fri, 29 Nov 2019 17:47:59 +0000 (19:47 +0200)]
gcc-9: silence 'address-of-packed-member' warning
GCC 9.x starts complaining about potential misalignment of the pointer to
the array (in this case alignment=2) in the packed (alignment=1) structures.
Repeating Linus' Torvalds commit 6f303d60534c in the Linux kernel.
Original commit message:
We already did this for clang, but now gcc has that warning too.
Yes, yes, the address may be unaligned. And that's kind of the point.
This in particular hides the warnings like
drivers/usb/gadget/composite.c:545:23: warning: taking address of packed member of ‘struct usb_string_descriptor’ may result in an unaligned pointer value [-Waddress-of-packed-member]
545 | collect_langs(sp, s->wData);
drivers/usb/gadget/composite.c:550:24: warning: taking address of packed member of ‘struct usb_string_descriptor’ may result in an unaligned pointer value [-Waddress-of-packed-member]
550 | collect_langs(sp, s->wData);
drivers/usb/gadget/composite.c:555:25: warning: taking address of packed member of ‘struct usb_string_descriptor’ may result in an unaligned pointer value [-Waddress-of-packed-member]
555 | collect_langs(sp, s->wData);
Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Holger Brunck [Mon, 25 Nov 2019 16:24:14 +0000 (17:24 +0100)]
km/boards: add MAC address offset as argument to ivm_populate_env
For upcoming SECU board we have different MAC address offsets depending
on which functional unit we are running. In this case we need to pass
this value to the ivm_populate_env function instead of using the hard
coded config option there.
Signed-off-by: Holger Brunck <holger.brunck@ch.abb.com> CC: Stefan Roese <sr@denx.de> CC: Valentin Longchamp <valentin.longchamp@ch.abb.com> CC: Marek Vasut <marex@denx.de> Reviewed-by: Stefan Roese <sr@denx.de>
Han Nandor [Tue, 12 Nov 2019 08:39:38 +0000 (08:39 +0000)]
rtc: add support for DS3232 device
DS3232 is an i2c RTC with 236 bytes of battery-backed SRAM.
Add an RTC driver for DS3232 device, which provides time and
date support. Also read and write functions are provided,
which can be used to access the SRAM memory.
Signed-off-by: Nandor Han <nandor.han@vaisala.com>
spl: fix stack usage check if gd is not initialized
Most platforms do not set up gd->start_addr_sp in SPL. Since this is
required for CONFIG_SPL_SYS_REPORT_SACK_F_USAGE to work correctly, set
up gd->start_addr_sp in SPL to the value passed to
board_init_f_init_reserve if it is not set yet.
Tom Rini [Wed, 8 Jan 2020 23:57:11 +0000 (18:57 -0500)]
Merge tag 'efi-2020-04-rc1' of https://gitlab.denx.de/u-boot/custodians/u-boot-efi
Pull request for UEFI sub-system for efi-2020-04-rc1
This pull request provides:
* support for FIT images for UEFI binaries
* drivers for hardware random number generators
* an implementation of the EFI_RNG_PROTOCOL
* a sub-command for efidebug to display configuration tables
Peng Ma [Wed, 4 Dec 2019 10:36:45 +0000 (10:36 +0000)]
ata: fsl_sata: Continue probing other sata port when failed current port.
In the initialization of sata driver, we want to initialize all port
probes, Therefore, any detection failure between of them should continue
initialization by skipping the current port instead of exit.
Peng Ma [Wed, 4 Dec 2019 10:36:42 +0000 (10:36 +0000)]
ata: sata_sil: Continue probing other sata port when failed current port.
In the initialization of sata driver, we want to initialize all port
probes, Therefore, any detection failure between of them should continue
initialization by skipping the current port instead of exit.
Patrice Chotard [Fri, 6 Dec 2019 14:01:49 +0000 (15:01 +0100)]
treewide: Remove CONFIG_SYS_UBOOT_START from configs board files
As previous CONFIG_SYS_UBOOT_START is now set by default to
CONFIG_SYS_TEXT_BASE when not defined, CONFIG_SYS_UBOOT_START
can be removed from include/configs board files.
Peng Fan [Mon, 6 Jan 2020 08:16:32 +0000 (16:16 +0800)]
imx: imx8mn_evk: add board_mmc_get_env_dev
Add board_mmc_get_env_dev, otherwise,
Loading Environment from MMC...
MMC Device 0 not found
*** Warning - No MMC card found, using default environment
Reviewed-by: Fabio Estevam <festevam@gmail.com> Signed-off-by: Peng Fan <peng.fan@nxp.com>
CPU: Freescale i.MX8MP rev1.0 at 1000 MHz
Reset cause: POR
Model: NXP i.MX8MPlus EVK board
DRAM: 6 GiB
MMC: FSL_SDHC: 1, FSL_SDHC: 2
Loading Environment from MMC... OK
In: serial
Out: serial
Err: serial
Net: No ethernet found.
Hit any key to stop autoboot: 0
u-boot=> mmc list
FSL_SDHC: 1 (SD)
FSL_SDHC: 2
Peng Fan [Mon, 30 Dec 2019 08:52:30 +0000 (16:52 +0800)]
imx: imx8m: only support non-dm code in clock_imx8mm.c
The drivers/clk/imx/*.c are used for CLK dm case, the
clock_imx8mm.c is used for non CLK dm case, let's split
it. Sometimes it is hard to enable CLK dm in SPL stage,
considering code size, malloc size requirement, the splittion
will make it easy to use non CLK dm in SPL stage.
Peng Fan [Mon, 30 Dec 2019 02:03:44 +0000 (10:03 +0800)]
arm: dts: freescale: Add i.MX8MP dtsi support
The i.MX8M Plus Media Applications Processor is part of the growing
mScale family targeting the consumer and industrial market. It brings
an effective Machine Learning and AI accelerator that enables a new
class of applications. It is built in Samsung 14LPP to achieve both
high performance and low power consumption and relies on a powerful
fully coherent core complex based on a quad core ARM Cortex-A53 cluster
and Cortex-M7 low-power coprocessor, audio digital signal processor,
machine learning and graphics accelerators.
Add the basic dtsi support for i.MX8MP.
Patch from Anson Huang for Kernel
https://patchwork.kernel.org/patch/11310915/