ARM: uniphier: set {kernel, ramdisk, fdt}_addr_r at boot-time
The base of DRAM will be changed for the next generation SoC.
The addresses needed for booting the kernel should be shifted
according to the DRAM base.
The base of DRAM will be changed for the next generation SoC.
To support it along with existing SoCs in the single defconfig,
set 'loadaddr' at boot-time by adding the offset to the DRAM base.
CONFIG_SYS_LOAD_ADDR is still hard-coded for compilation, but the
value from environment variable 'loadaddr' should be used.
The base address of DRAM was 0x80000000 for all the ARM SoCs of this
family in the past. It will be changed to 0x20000000 for a planned new
SoC. To support multiple SoCs by the single uniphier_v8_defconfig, the
base must be run-time determined.
Currently, mem_map is hard-coded, and it worked well until the last
SoC. For a planned new SoC, the addresses of peripherals and DRAM
will be changed. Set it up run-time.
ARM: uniphier: detect register base addresses run-time
Until the last SoC, the register addresses have been hard-coded because
they are always constant. For a planned new SoC, the register bases
will be completely changed. I insist on supporting multiple SoCs/boards
by a single defconfig (uniphier_v8_defconfig) since duplicating similar
defconfig files is a maintenance burden. The base addresses must be
fixed-up at run-time somehow.
Previously, the board init code identified the SoC by reading out the
SG_REVISION register. This is much easier than parsing DT.
You cannot do it any more because the base address of SG will be
changed. The SG_REVISION register exists to read out the SoC ID, but
you never know its address before identifying the SoC. Oh well.
So, the possible solution is to parse the DT, and find out the node
with "*-soc-glue" compatible string. Then, sg_base is set to the value
of the "reg" property. The sc_base is set up likewise.
It is worth noting a pit-fall. Having sc_base and sg_base in the global
scope will make the life easier, but the global variables are poorly
supported before the relocation. In fact, the .bss section overwraps
with DT. Allocating them in the .bss section would break DT. So, I gave
dummy initializers to assign them in the .data section.
The memory regions specified by /memreserve/ are passed to
early_init_dt_reserve_memory_arch() with nomap=false, so it is
not suitable for reserving memory for Trusted Firmware-A etc.
Use the more robust /reserved-memory node with the no-map property
to prevent the kernel from mapping it.
Masahiro Yamada [Wed, 26 Jun 2019 04:51:46 +0000 (13:51 +0900)]
arm64: add an option to switch visibility of CONFIG_SYS_INIT_SP_BSS_OFFSET
By default, CONFIG_SYS_INIT_SP_BSS_OFFSET was made invisible by not
giving a prompt to it.
The only way to define it is to hard-code an extra entry in SoC/board
Kconfig, like arch/arm/mach-tegra/tegra{186,210}/Kconfig.
Add a prompt to it in order to allow defconfig files to specify the
value of CONFIG_SYS_INIT_SP_BSS_OFFSET.
With this, CONFIG_SYS_INIT_SP_BSS_OFFSET would become always visible.
So, we need a new bool option to turn it off by default.
I move the 'default 524288' to the common place. This value is not too
big, but is big enough to avoid the overwrap of DT in most platforms.
If 512KB is not a suitable choice for your platform, you can change it
from your defconfig or menuconfig etc.
Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com> Acked-by: Stephen Warren <swarren@nvidia.com>
Cyrille Pitchen [Tue, 18 Jun 2019 08:52:01 +0000 (08:52 +0000)]
ARM: at91: sama5: add common environment for QSPI
Use the same memory layout as we use for the NAND boot on the other boards.
QSPI flashes are present on the following boards:
sama5d2_xplained RevB: 32 Mbyte flash (mx25l3273fm2i-08g)
sama5d2_xplained RevC: 8 Mbyte flash (sst26vf064b-104i/sn)
sama5d27_som1_ek: 8 Mbyte flash (sst26vf064b-104i/sn)
sama5d2_ptc_ek: 8 Mbyte flash (sst26vf064b-104i/sn)
The 8 Mbyte limit is enough to cope with the memory layout used in the NAND
boot. rootfs exceeds the 8 Mbyte limit and will stay in eMMC in the
sama5d2_xplained case. The final scope is to use a single memory layout for
all boot medias.
Cyrille Pitchen [Tue, 18 Jun 2019 08:51:58 +0000 (08:51 +0000)]
ARM: dts: at91: sama5d2_xplained: fix QSPI0 node
Fix the following:
- use "jedec,spi-nor" binding, we use jedec compatible flashes
- set bus width to 4, we use quad capable flashes
- differentiate bewteen data and clk and cs pins
- drop partions as we don't use them in u-boot.
Tom Rini [Mon, 8 Jul 2019 20:02:51 +0000 (16:02 -0400)]
Merge tag 'efi-2019-10-rc1' of https://gitlab.denx.de/u-boot/custodians/u-boot-efi
Pull request for UEFI sub-system for v2019.10-rc1
Fix a possible overflow for GUID partition tables.
For some runtime services we only have implementations valid at boottime.
So we replace them when leaving boottime. Move this from
SetVirtualAddressMap() to ExitBootServices() as SetVirtualAddressMap() is
not called by all operating systems. Adjust the Python tests accordingly.
Bump the supported UEFI specification version to 2.8.
Weijie Gao [Thu, 16 May 2019 09:19:50 +0000 (17:19 +0800)]
arm: mediatek: remove arch_misc_init
The watchdog of mediatek chips is enabled by bootrom before u-boot is
running. Previously we choose to enable the wdt driver only to disable the
watchdog hardware.
Now wdt service is enabled by default. The function arch_misc_init which is
only used to disable wdt is no longer needed.
Reviewed-by: Stefan Roese <sr@denx.de> Reviewed-by: Ryder Lee <ryder.lee@mediatek.com> Signed-off-by: Weijie Gao <weijie.gao@mediatek.com>
Weijie Gao [Thu, 16 May 2019 09:19:13 +0000 (17:19 +0800)]
board_r: move initr_watchdog to be called after initr_serial
The initr_watchdog is currently placed before initr_serial. The
initr_watchdog calls printf and printf finally calls ops->putc of a serial
driver.
However, gd->cur_serial_dev points to a udevice allocated in board_f. The
gd->cur_serial_dev->driver->ops->putc points the the code region before
relocation.
Some serial drivers call WATCHDOG_RESET() in ops->putc. When DM is enabled
for watchdog, watchdog_reset() is called. watchdog_reset() calls get_timer
to get current timer.
On some platforms the timer driver is also a DM driver. initr_watchdog is
placed right after initr_dm, which means the timer driver hasn't been
initialized. So dm_timer_init() is called. To create a new udevice, calloc
is called.
However start from ops->putc, u-boot execution flow is redirected into the
memory region before relocation (board_f). In board_f, dlmalloc hasn't
been initialized. The call to calloc will fail, and this will cause DM to
print out an error message, and it will call printf again, causing
recursive error outputs.
This patch places initr_watchdog after initr_serial to solve this issue.
Cc: Stefan Roese <sr@denx.de> Reviewed-by: Ryder Lee <ryder.lee@mediatek.com> Signed-off-by: Weijie Gao <weijie.gao@mediatek.com> Reviewed-by: Stefan Roese <sr@denx.de> Tested-by: Frank Wunderlich <frank-w@public-files.de> Tested-by: Suniel Mahesh <sunil.m@techveda.org>
Marc Dietrich [Tue, 2 Jul 2019 20:08:33 +0000 (22:08 +0200)]
video: backlight: fix pwm inversion
set_pwm() will always fail with -ENOSYS if pwm_ops set_invert() is
not implemented, leaving the backlight dark. Fix this by returning
no error if set_invert() is not implemented and no polarity change
is requested.
Fixes: 57e7775413 ("video: backlight: Parse PWM polarity cell") Signed-off-by: Marc Dietrich <marvin24@gmx.de>
Linux can be called with a command line parameter efi=novamap, cf.
commit 4e46c2a95621 ("efi/arm/arm64: Allow SetVirtualAddressMap() to be
omitted"). In this case SetVirtualAddressMap() is not called after
ExitBootServices().
OpenBSD 32bit does not call SetVirtualAddressMap() either.
Runtime services must be set to an implementation supported at runtime
in ExitBootServices().
Reported-by: Ard Biesheuvel <ard.biesheuvel@linaro.org> Suggested-by: Alexander Graf <agraf@csgraf.de> Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
We are implementing UEFI variable RuntimeServicesSupported and set the
unimplemented runtime functions return EFI_UNSUPPORTED as described in UEFI
specification 2.8. So let's also advertise this specification version in
our system table.
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de> Reviewed-by: Alexander Graf <agraf@csgraf.de>
efi_loader: split off detaching SetVirtualAddress()
The runtime services SetVirtualAddress() and ConvertPointer() become
unavailable after SetVirtualAddress(). Other runtime services become
unavailable after ExitBootServices.
Move the update of SetVirtualAddress() and ConvertPointer() to
efi_relocate_runtime_table().
Use functions with the correct signature when detaching.
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
efi_loader: let the variable driver patch out the runtime
Our variable services are only provided at boottime. Therefore when
leaving boottime the variable function are replaced by dummy functions
returning EFI_UNSUPPORTED. Move this patching of the runtime table to the
variable services implementation. Executed it in ExitBootServices().
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
The detaching of the runtime will have to move to ExitBootServices() to
encompass operating system that do not call SetVirtualAddressMap().
This patch changes the logic for the relocation of the pointers in the
runtime table such that the relocation becomes independent of the entries
in the detach list.
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
efi_loader: remove NULL entries from runtime detach list
Some entries in the system table are set to NULL in ExitBootServices(). We
had them in the runtime detach list to avoid relocation of NULL. Let's
instead assign the pointers dynamically in efi_initialize_system_table() to
avoid the relocation entry.
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
In part_get_info_efi() we use the output of print_efiname() to set
info->name[]. The size of info->name is PART_NAME_LEN = 32 but
print_efiname() returns a string with a maximum length of
PARTNAME_SZ + 1 = 37.
Use snprintf() instead of sprintf() to avoid buffer overflow.
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Stefan Roese [Tue, 28 May 2019 06:11:37 +0000 (08:11 +0200)]
mips: mt76xx: Implement new d-cache fix in last_stage_init()
With commit 06985289d452 ("watchdog: Implement generic watchdog_reset()
version") the init sequence has changed in arch_misc_init(), resulting
in a re-appearance of the d-cache issue on MT7688 boards (e.g. gardena).
When this happens, the first (or sometimes later ones as well) TFTP
command hangs and does not complete correctly. This leads to the
assumption that the d-cache is not in a clean state once the ethernet
driver is called (d-cache is used here for the buffers). The old work-
around with the cache flush somehow does not work any more now with
the new code change.
Unfortunately adding CONFIG_SYS_MALLOC_CLEAR_ON_INIT also did not fix
this issue. With v2019.07-rc3 it shows again. The time of accessing
the data seems to be very important here. It needs to be "very late"
in the boot process.
Testing has shown, that copying a 64KiB area in DDR at a very late
bootup time, directly before calling into the prompt, fixes this issue.
Flushing of the complete d-cache does not seem to necessary, as this
copy alone seems to fix this problem.
Signed-off-by: Stefan Roese <sr@denx.de> Cc: Daniel Schwierzeck <daniel.schwierzeck@gmail.com>
This uclass is intended for devices that do not need any features from the
uclass, including binding children.
This will typically be used by devices that are used to bind child devices
but do not use dm_scan_fdt_dev() to do it. That is for example the case of
several USB wrappers that have 2 child devices (1 for device and 1 for
host) but bind only one at a any given time.
Signed-off-by: Jean-Jacques Hiblot <jjhiblot@ti.com> Reviewed-by: Simon Glass <sjg@chromium.org>
Marek Vasut [Mon, 24 Jun 2019 17:05:47 +0000 (19:05 +0200)]
usb: ehci-mx6: Fix bus enumeration for DM case
The EHCI iMX6 driver is only partly converted to DT probing and
still uses a tremendous amount of hard-coded addresses. Worse,
the driver uses hard-coded SoC-model-specific base addresses, which
are derived from values protected by SoC-specific macros, hence the
driver is also compiled for a specific SoC model. Even worse, the
driver depends on specific sequential indexing of the controllers,
from which it derives offsets in the PHY and ANATOP register sets.
However, when the driver is probed from DT, the indexing is not
correct. In fact, each controller has index 0. This patch derives
the index for DT probing case from the controller base addresses,
which is not the way this should be done, however it is the least
intrusive approach, favorable this close to release.
The necessary steps to convert this driver fully to DT probing are
described inside the patch, however this should be done in the next
release and depends on iMX clock driver patches.
Signed-off-by: Marek Vasut <marex@denx.de> Cc: Abel Vesa <abel.vesa@nxp.com> Cc: Adam Ford <aford173@gmail.com> Cc: Fabio Estevam <festevam@gmail.com> Cc: Ludwig Zenz <lzenz@dh-electronics.com> Cc: Lukasz Majewski <lukma@denx.de> Cc: Peng Fan <peng.fan@nxp.com> Cc: Stefano Babic <sbabic@denx.de> Cc: Vagrant Cascadian <vagrant@debian.org>
Neil Armstrong [Wed, 12 Jun 2019 09:49:07 +0000 (11:49 +0200)]
ARM: meson: add unique MAC address generation
Add support for generating an unique MAC address using the SoC internal
serial number from the Secure Monitor interface.
The algorithm generates an unicast locally administered 6bytes minus 2bits
address using an crc16 of the serial for the top 16bits with the lower 2 bits
masked to setup the unicast locally administered property and a crc24 for
the lower 24bits.
Signed-off-by: Neil Armstrong <narmstrong@baylibre.com>
Kever Yang [Mon, 1 Jul 2019 03:49:14 +0000 (11:49 +0800)]
rockchip: config: tiner-rk3288: extend CONFIG_SYS_MONITOR_LEN to 600KB
The raw u-boot.bin for tinker board has been about 450KB without
debug option, and 550KB with all debug on, and the default value is 200KB,
which is not enough for run raw u-boot.bin.
Signed-off-by: Kever Yang <kever.yang@rock-chips.com>
Kever Yang [Mon, 1 Jul 2019 03:49:13 +0000 (11:49 +0800)]
rockchip: config: update config for TPL support on tinker-rk3288
We need to update TEXT BASE for TPL/SPL/U-Boot;
SPL no need relocate STACK after enable TPL, so remove it;
Don't enable pinctrl names so that SPL can get pinctrl dts;
Signed-off-by: Kever Yang <kever.yang@rock-chips.com>
Kever Yang [Mon, 1 Jul 2019 03:49:11 +0000 (11:49 +0800)]
rockchip: dts: tinker: migrate the dm-pre-reloc tag into -u-boot dts
Migrate all the "u-boot,dm-pre-reloc" tag from rk3288-tinker.dts
into rk3288-tinker-u-boot.dtsi.
When both board level and soc level '-u-boot.dtsi' files exist,
we need to include the soc level 'rk3288-u-boot.dtsi' manually.
Signed-off-by: Kever Yang <kever.yang@rock-chips.com>
Mark Kettenis [Fri, 21 Jun 2019 21:16:45 +0000 (23:16 +0200)]
rockchip: dts: rk3399: rockpro64: Provide init voltage
Add missing regulator-init-microvolt property to vdd_log regulator.
Signed-off-by: Mark Kettenis <kettenis@openbsd.org> Reviewed-by: Kever Yang <kever.yang@rock-chips.com>
(Rebase on latest u-boot-rockchip master) Signed-off-by: Kever Yang <kever.yang@rock-chips.com>
Change-Id: I13b24fb81e8ad269d7dbb0c7b67f5f4795d2e775
Frieder Schrempf [Thu, 27 Jun 2019 07:03:16 +0000 (07:03 +0000)]
spl: atf: Fix uninitialized pointer to bl31_image_info
The pointer to struct atf_image_info in
bl31_params_mem.bl31_params.bl31_image_info is not initialized before
being dereferenced. This can cause U-Boot to crash right before jumping
to the BL31 ATF binary.
Signed-off-by: Frieder Schrempf <frieder.schrempf@kontron.de> Fixes: bcc1726a7bdd ("spl: add support to booting with ATF") Reviewed-by: Kever Yang <kever.yang@rock-chips.com>
Shawn Guo [Mon, 17 Jun 2019 03:08:34 +0000 (11:08 +0800)]
poplar: increase SYS_MALLOC_F_LEN for DM_FLAG_PRE_RELOC support
There is a regression seen on Poplar platform, which doesn't even show
a U-Boot version banner on booting. It turns out that due to landing
of commit 3a7c45f6a772 ("simple-bus: add DM_FLAG_PRE_RELOC flag to
simple-bus driver"), we need to increase SYS_MALLOC_F_LEN from its
default size 0x400, as pre-relocation requires more memory there. Let's
increase SYS_MALLOC_F_LEN to 0x4000 to fix the regression.
Thanks to Andreas Färber <afaerber@suse.de> for reporting, and Bin Meng
<bmeng.cn@gmail.com> for trouble shooting.
Reported-by: Andreas Färber <afaerber@suse.de> Signed-off-by: Shawn Guo <shawn.guo@linaro.org> Reviewed-by: Bin Meng <bmeng.cn@gmail.com> Tested-by: Andreas Färber <afaerber@suse.de>
projects / u-boot.git / log